bs_bs_banner Benchmarking recombinant Pichia pastoris for 3-hydroxypropionic acid production from glycerol Albert Fina,1 Gabriela Coelho Bre ^da,2 modifications aimed at increasing cofactor and meta- Mıriam Pe rez-Trujillo, 3 Denise Maria Guimara ~es bolic precursors availability allowed a 14-fold Freire,2 Rodrigo Volcan Almeida,2 increase in the production of 3-HP compared to the Joan Albiol1 and Pau Ferrer1,* initial strain. The best strain (PpHP6) was tested in a 1 Department of Chemical, Biological and Environmental fed-batch culture, achieving a final concentration of Engineering, Universitat Auto  noma de Barcelona, 3-HP of 24.75 g l 1, a product yield of 0.13 g g 1 and Bellaterra, Catalonia, Spain. a volumetric productivity of 0.54 g l 1 h 1, which, to 2 Departamento de Bioquımica, Instituto de Quımica, our knowledge, is the highest volumetric productivity Universidade Federal do Rio de Janeiro, Rio de Janeiro, reported in yeast. These results benchmark P. pas- Brazil. toris as a promising platform to produce bulk chemi- 3 Servei de Ressona ncia Magne  tica Nuclear, Facultat de cals for the revalorization of crude glycerol and, in Cie ncies i Biocie  ncies, Universitat Auto  noma de particular, to produce 3-HP. Barcelona, Bellaterra, Catalonia, Spain. Introduction Summary In 2004, the US Department of Energy (DOE) published The use of the methylotrophic yeast Pichia pastoris the list of the top value-added bio-based chemicals to be (Komagataella phaffi) to produce heterologous pro- produced from biomass (Werpy and Petersen, 2004). teins has been largely reported. However, investiga- The production of such bio-based products in a biorefin- tions addressing the potential of this yeast to ery is necessary to complement biofuel production, produce bulk chemicals are still scarce. In this which is a low value-added product with a high price study, we have studied the use of P. pastoris as a volatility. Co-producing biofuels and chemicals in an inte- cell factory to produce the commodity chemical 3- grated biorefinery would allow the design of an economi- hydroxypropionic acid (3-HP) from glycerol. 3-HP is cally robust and sustainable process, making it an a chemical platform which can be converted into attractive investment option. The DOE report ranked 3- acrylic acid and to other alternatives to petroleum- hydroxypropionic acid (3-HP) among one of these top based products. To this end, the mcr gene from value-added chemicals to be produced in a biorefinery. Chloroflexus aurantiacus was introduced into P. pas- The largest application of 3-HP is its conversion into toris. This single modification allowed the production acrylic acid and other products typically derived from of 3-HP from glycerol through the malonyl-CoA path- petroleum (Della Pina et al., 2011; Kumar et al., 2013). way. Further enzyme and metabolic engineering The global market size of acrylic acid is estimated to reach 22 550 M$ in 2022 (Grand View Research, 2016). Due to such promising market forecast, Novozymes and Received 6 February, 2021; revised 30 April, 2021; accepted 1 May, 2021. Cargill announced a joint agreement to develop a plat- For correspondence. E-mail Pau.Ferrer@uab.cat; Tel. +34 935812692; form to produce acrylic acid from biologically produced Fax +34 935812013. 3-HP in 2008 (Novozymes, 2008). Nevertheless, up to Microbial Biotechnology (2021) 14(4), 1671–1682 doi:10.1111/1751-7915.13833 date, the process has not been implemented at a large Funding Information scale, as further development is still required. This work was supported by projects PHBP 14/00087 and While the revalorization of lignocellulosic products con- CTQ2016-74959-R of the Spanish Ministry of Economy and Com- petitiveness; 2017-SGR-1462 from the Age ncia de Gestio  d’Ajuts sisting of glucose and xylose has been largely investi- Universitaris i de Recerca (AGAUR) of the Catalan Government; gated, there are other waste products that can be used PETROBRAS cooperation term number 0050.0081285.13.9; and as substrates to produce bulk chemicals. This is the CAPES-DGPU programme (1694/2015) from Coordenacßa ~o de Aper- feicßoamento de Pessoal de Nıvel Superior (CAPES). AF was sup- case of crude glycerol, a side-product obtained during ported with a FPU scholarship (FPU17/05434) from the Spanish the enzymatic production of biodiesel. It is a mixture Ministry of Science, and a FI scholarship (2018FI_B00713) from made of 60–80% glycerol, 10–20% methanol and 10– AGAUR. GCB was supported by Conselho Nacional de Desenvolvi- mento Cientıfico e Tecnologico (CNPq, Brazil) scholarship (158956/ 20% of soap or other undefined organic matter com- 2015-0). pounds (Luo et al., 2016). The use of this mixture as ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 1672 A. Fina et al. substrate is limited by the fact that methanol is toxic to the route which uses malonyl-CoA as a precursor has many microorganisms. The methylotrophic yeast Pichia been implemented in several microorganisms, including pastoris is a promising microorganism for the revaloriza- E. coli (Rathnasingh et al., 2012; Liu et al., 2013, 2016), tion of crude glycerol as it can efficiently grow using both S. cerevisiae (Chen et al., 2014; Kildegaard et al., 2016) glycerol and methanol as carbon sources. Moreover, and Shizosaccharomyces pombe (Suyama et al., 2017). P. pastoris can grow at a low pH, and it is reported that Using glucose as substrate, the maximum theoretical the cost of the downstream process is reduced if the fer- yield of the b-alanine to 3-HP pathway is higher than the mentation is performed at a pH below the pKa value of route starting from malonyl-CoA, as more ATP is required the acidic product (van Maris et al., 2004). For all these for the latter. However, when glycerol is used as a sub- reasons, in this study, we investigate the production of strate, the maximum theoretical yield for both pathways 3-HP from glycerol in the yeast P. pastoris. equals 1, as there is net ATP production in both cases Pichia pastoris is widely used in industrial biotechnol- (see Supplementary materials Section Data S1). While ogy as an efficient host for recombinant protein produc- the b-alanine pathway would require the expression of 3 tion, and it has received increasing interest as a platform heterologous genes (Borodina et al., 2015) to achieve 3- to produce fine and bulk chemicals (Schwarzhans et al., HP production in P. pastoris, it is reported that the single 2017; Pen ~a et al., 2018). Moreover, the efforts of the expression of the bifunctional enzyme malonyl-CoA P. pastoris community have allowed the implementation reductase from Chloroflexus aurantiacus (MCRCa) triggers of the necessary tools allowing its use in metabolic engi- 3-HP production in yeast through the malonyl-CoA route neering research, including up-to-date genetic engineer- (Chen et al., 2014). This enzyme performs two NADPH- ing tools like the GoldenMOCS (Prielhofer et al., 2017) consuming consecutive reactions sequentially converting and CRISPR-Cas9 (Weninger et al., 2016), metabolic malonyl-CoA to malonate semialdehyde (MSA), and then, genome-scale models (GSM; Toma s-Gamisans et al., MSA is converted to 3-HP (Fig. 1). 2016), metabolomics and fluxomics protocols (Carnicer In this study, the production of 3-HP using glycerol as et al., 2012; Ferrer and Albiol, 2014), and a wide knowl- substrate has been implemented in P. pastoris through edge of its behaviour at the bioreactor scale (Looser the malonyl-CoA pathway. The expression of mcrCa et al., 2014; Yang and Zhang, 2018). Another interesting leads to 3-HP production in P. pastoris. This base strain trait of P. pastoris is that it is a Crabtree-negative yeast. has been further modified using two different strategies: This is an interesting feature in metabolic engineering protein engineering and metabolic engineering. The because overflow metabolism to undesired by-products, independent expression of the two subunits of the such as ethanol or glycerol, can be minimized. The full malonyl-CoA reductase has shown higher 3-HP produc- oxidation of the carbon source leads to a higher energetic tion in other microorganisms (Liu et al., 2013). This strat- yield than in Crabtree-positive yeasts, like Saccha- egy has been further tested in P. pastoris, yielding a romyces cerevisiae, thus leading to potentially higher pro- substantial improvement compared to the initial strain. duct yields (Dai et al., 2018; Pen ~a et al., 2018). Moreover, Further modifications have been implemented to the extracellular concentration of some intermediate increase the fluxes producing the substrates of the metabolites of the TCA cycle (i.e. malate or citrate) is malonyl-CoA pathway to 3-HP (NADPH and malonyl- lower in P. pastoris than in S. cerevisiae under similar CoA). The strain producing the highest 3-HP titre has conditions (Carnicer et al., 2012). This trait simplifies the been characterized in a fed-batch culture using glycerol downstream processing of the products of interest. as a substrate. Overall, the potential of P. pastoris as a The biological production of 3-HP has been largely promising host for 3-HP production from this renewable investigated (de Fouche cour et al., 2018). Several path- feedstock was demonstrated for the first time. ways have been tested in a number of microorganisms, including the industrial workhorses Escherichia coli and Results and discussion S. cerevisiae. Each pathway to produce 3-HP is named Expression of mcrCa in P. pastoris leads to 3-HP according to its precursor. The route that has obtained the production highest yields and productivities is the coenzyme B12- dependant glycerol pathway (Raj et al., 2008; Rathnas- In the present work, we expressed mcrCa gene in ingh et al., 2009). However, the implementation of this P. pastoris under the control of the constitutive and route in P. pastoris is currently unfeasible at an industrial strong GAP promoter (Waterham et al., 1997). Compar- scale due to the high production costs caused by the ison of the resulting strain, PpHP1, to the reference requirement of coenzyme B12 addition, which is an strain (X-33) in triplicate shake flasks on buffered mini- expensive compound (Chen et al., 2014). The 3-HP route mal glycerol (BMG) medium showed that the sole through b-alanine has been investigated in S. cerevisiae expression of mcrCa resulted in 3-HP production, while it (Borodina et al., 2015) and E. coli (Song et al., 2016) and did not affect cell growth (the µmax of the control and ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 3-HP production in Pichia pastoris 1673 Fig. 1. Simplified representation of the conversion of glycerol to 3-HP through the malonyl-CoA route. The metabolic engineering targets to increase the availability of the precursors of the malonyl-CoA to 3-HP pathway are included. Acc, acetyl-CoA carboxylase; cPos5, cytosolic NADH kinase; MCR (C-ter), C-terminal domain of malonyl-CoA reductase; MCR (N-ter), N-terminal domain of malonyl-CoA reductase. PpHP1 strains was the same within the precision range, S. cerevisiae (0.008 U mg 1; Chen et al., 2014). The com- 0.24  0.01 h 1 and 0.24  0.01 h 1 respectively). bined effect of the strength of the P. pastoris GAP promoter Moreover, no by-products were detected in any of the and the fact that glycerol is more reduced than glucose, two strains (Fig. S2). which leads to net production of ATP from the substrate, Notably, the PpHP1 strain produced 0.19  0.03 g l 1 instead of net ATP consumption (Fig. S1), points at of 3-HP after 24 h of cultivation. Such 3-HP titre is consid- P. pastoris as a promising cell factory for the bioproduction erably higher than those achieved in S. pombe (Takayama of 3-HP from glycerol. et al., 2018) and S. cerevisiae (Chen et al., 2014) harbour- ing a similar genetic construction and using glucose as a Setting up the screening conditions for 3-HP producing substrate (0.016 g l 1 and 0.093 g l 1 respectively). Fur- P. pastoris clones thermore, the C-yield (Cmol of 3-HP per Cmol of substrate, i.e. glycerol or glucose) for the PpHP1 strain We established two independent analytical methods (0.015  0.002 Cmol Cmol 1) was remarkably higher based on NMR and HPLC-MS for 3-HP quantification, than the one observed in other yeasts (0.0003 Cmol which were cross-validated. The results using the HPLC- Cmol 1 in S. pombe, and 0.0048 Cmol Cmol 1 in MS method diverge from those obtained using NMR S. cerevisiae). Remarkably, the specific activity of MCR in when glycerol is still present in the medium (Fig. S3). PpHP1 was 0.30  0.06 U mg 1 of protein, which is This is explained by the matrix effect caused by the glyc- remarkably higher than the specific activity reported in erol co-eluting with 3-HP from the column, which affects ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 1674 A. Fina et al. the ionization efficiency at the ionization source. How- of the two reactions converting malonyl-CoA into 3-HP ever, once the glycerol is fully consumed, the HPLC-MS (Liu et al., 2013). The presence of three point mutations and the NMR methods yielded statistically identical causing three amino acidic changes (N940V, K1106R results when measuring the final 3-HP concentrations in and S1114R) had an additional positive impact on 3-HP the PpHP1 shake flask cultures described in Section 3.1 production in E. coli (Liu et al., 2016). The same out- (i.e. 3-HP final concentrations of 0.18  0.02 g l 1 of 3- come has been demonstrated in S. pombe, where the HP and 0.19  0.02 g l 1 of 3-HP respectively). expression of the N-terminal and the improved version of Consumption of 3-HP has been reported in several the C-terminal domain of MCR using two independent microorganisms (Zhou et al., 2014; Yang et al., 2017). In expression cassettes led to a 30-fold improvement in the order to evaluate whether the PpHP1 strain was able to final 3-HP titre, compared to the starting strain express- assimilate this compound, this strain was grown in tripli- ing the original sequence of MCRCa (Takayama et al., cate shake flasks using BMG, BMG medium supple- 2018). Therefore, we introduced two expression cas- mented with 2.5 g l 1 of 3-HP (BMG3HP) or buffered settes into P. pastoris allowing the independent expres- minimal medium supplemented with 2.5 g l 1 of 3-HP as sion of the coding DNA sequences for the N-terminal a sole carbon source (BM3HP). Samples were collected and the C-terminal domains of MCR (including the wild- after 24, 48 and 72 h. The 3-HP concentration remained type and the mutated version of the latter) under the constant throughout the 72 h in the cultures grown on control of the GAP promoter. BM3HP (no growth was observed). After 24 h, glycerol As shown in Fig. 2, the dissection of MCR (strain was exhausted from the BMG and BMG3HP, while the PpHP2) had a positive impact on 3-HP production, trig- concentration of 3-HP remained unaltered between the gering a 12.5-fold increase in 3-HP production 24 and the 72 h (data not shown), thereby indicating that (1.63  0.09 g l 1 3-HP). Nevertheless, the introduction P. pastoris does not assimilate 3-HP as C-source within of the 3 point mutations in the C-terminal domain (strain the experimental time frame tested. PpHP3) resulted in a non-producing strain. Similarly, the Considering these results, screening of the 3-HP- introduction of these 3 point mutations to a non- producing strains was performed taking end-point sam- dissected version of MCR resulted in a non-producing P. ples after 48 h of incubation to ensure full consumption pastoris strain (data not shown). These point mutations of glycerol, followed by 3-HP quantification using the are far from the reactive site and the NADPH binding HPLC-MS method. site of the enzyme. It remains poorly understood how distant mutations may affect catalytic properties. More- over, as the structure of the C-terminal domain of MCR Improvement of 3-HP production by MCR enzyme from C. aurantiacus is not available, the exact effect of engineering the point mutations on the protein conformational stabil- Recent studies have shown the positive effect of dissect- ity/quality is hard to predict, particularly when the protein ing the MCR enzyme in the two subunits catalysing each is synthetized at high rates (i.e. using a strong Fig. 2. Production of 3-HP for each strain in the screening experiments. The genes heterologously expressed in each strain are depicted in the left side of the graph. The grey bars show the average 3-HP concentration at the end of the culture, the discontinuous line shows the standard deviation, the solid line indicates the SE and the circles show the average result for each clone. The solid circle of a PpHP5 clone shows the result of a clone which was discarded for the calculations, as it had a different behaviour from the rest of the clones of that strain. For PpHP6, only 3 clones could be screened. Despite that more than 50 transformants of PpHP6 from 2 independent transformations were checked using colony PCR, only 3 clones resulted positive. ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 3-HP production in Pichia pastoris 1675 promoter). Notably, the 12.5-fold improvement obtained positive effect in 3-HP production in both S. cerevisiae by dissecting MCR is in the same range as those (Kildegaard et al., 2016) and S. pombe (Takayama reported in S. pombe and E. coli. et al., 2018). Overall, the product yield obtained by the strain PpHP6 during the screening phase (0.146  0.003 Cmol Cmol 1) Metabolic engineering of P. pastoris to improve 3-HP is significantly higher (around 1.8-fold) to the highest yield production observed in yeast under comparable conditions (deep-well Metabolic engineering for 3-HP production through the plate culture using defined medium), i.e. using the S. cere- malonyl-CoA pathway was aimed at increasing the avail- visiae strain 3HP-M11 producing 3-HP through the malonyl- ability of the two precursors of this route, namely CoA pathway, with a yield on glucose of 0.080  0.0 NADPH and malonyl-CoA. To do so, the ACCYl and 08 Cmol Cmol 1 (calculated from data given by Kildegaard cPOS5Sc genes have been heterologously expressed in et al., 2016). the PpHP2 strain. The reactions catalysed by the enzymes encoded by these two genes are depicted in Production of 3-HP in a fed-batch culture Fig. 1. Acc produces malonyl-CoA from acetyl-CoA, which is the precursor of the central carbon metabolism The strain PpHP6 was further cultivated in a controlled for 3-HP production. The overexpression of such fed-batch culture. First, a batch experiment was per- enzyme has already been performed in E. coli (Rathnas- formed in order to determine the µmax of PpHP6, and ingh et al., 2012), S. cerevisiae (Chen et al., 2014; Kilde- also the initial biomass concentration (X0) and the bio- gaard et al., 2016) and S. pombe (Takayama et al., mass to substrate yield (YX/S), which are the parameters 2018) to increase the production of 3-HP. The enzyme required to set the exponential feeding rate (see Sec- cPos5 produces NADPH by means of NADH and ATP tion 2.4). The µmax was 0.19  0.1 h 1, the X0 was consumption, and the overexpression of its gene leads 18.6  0.2 g l 1 and the YX/S 0.47  0.01 g g 1. to an increase in the NADPH/NADP ratio (Toma s- After the initial batch phase (22 h), the exponential Gamisans et al., 2020). feeding rate was set to maintain a growth rate equal to As shown in Fig. 2, the overexpression of Acc or 0.1 h 1 (approximately 50% of the µmax; Fig. 3A). After cPos5 in the strain PpHP2 – resulting in the strains 44 h of cultivation (22 h of feeding phase), glycerol PpHP4 and PpHP5, respectively – led to a small accumulation was observed, and the feeding pump was increase in 3-HP production. Still, such increases were stopped. Thereafter, the fermentation was terminated not statistically significant compared to the parental after 45.5 h of cultivation, when the pO2 increased, indi- strain PpHP2 (P-values of 0.27 and 0.15 respectively). cating that all the glycerol had been consumed. Overall, However, when both genes were overexpressed at the 195 g l 1 of glycerol was added into the reactor, and same time (PpHP6), a significant increase in the final 3- 24.75  0.54 g l 1 of 3-HP was produced. The only by- HP concentration was observed (P = 0.015). product detected by NMR was arabitol in the late stages The PpHP6 strain produced 1.81  0.04 g l 1 of 3- of the batch and the fed-batch phases. HP, which represents a 14-fold increase compared to As already inferred from the evolution of physiological the starting strain (PpHP1), and a 12% increase com- growth parameters over time, growth rate was not main- pared to PpHP2. The C-yield for PpHP6 was tained during the whole feeding phase. Furthermore, the 0.146  0.003 Cmol Cmol 1. decrease in the growth rate was accompanied with the Similar strategies in other microorganisms resulted in accumulation of glycerol and the production of arabitol similar outcomes. For example, in E. coli, the overex- towards the end of the cultivation. These results differ from pression of either ACC or PntAB (a transhydrogenase previous observations of P. pastoris producing heterolo- encoding gene) led to a twofold increase, while the co- gous proteins in fed-batch cultures operated under analo- expression of both genes led to a threefold increase gous conditions (i.e. using a pre-programmed exponential (Rathnasingh et al., 2012). It is also worth noticing that substrate feeding strategy for controlled specific growth in S. pombe, enzyme engineering and expression level rate at 0.1 h 1), where the µ remained constant through- adjustment of MCR led to a 30-fold increase in 3-HP out the feeding phase and no arabitol accumulation was titres, while increasing the availability of acetyl-CoA and detected, achieving up to 100 g l 1 of dry cell biomass CoA in this strain resulted in a twofold increase. Alto- (Garcia-Ortega et al., 2013). The spline curves fitted to gether, these results point at the flux from malonyl-CoA the evolution of the growth rate (µ), product yield (YP/S), to 3-HP as the main limiting factor for 3-HP production. qP3-HP and qPArabitol throughout the cultivation further Further metabolic studies should corroborate this hypoth- reveal that the decrease in the growth rate coincides with esis. Indeed, the increase of the number of copies of the the onset of glycerol accumulation and arabitol by-product mcr gene has already been demonstrated to have a excretion (Fig. 3B). Conversely, product yield and qP3-HP ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 1676 A. Fina et al. Fig. 3. Fed-batch culture of PpHP6. A. Biomass and metabolites concentration during cultivation and total glycerol added (per litre) into the reactor is shown. It was calculated con- sidering the volume of feeding added to the reactor and the actual culture volume. Error bars denote SE. B. Growth rate (µ), product yield (Yield3-HP/Glyc), q-rate of 3-HP (qP 3-HP), and q-rate of arabitol (qP Arabitol) of the strain PpHP6 throughout the feeding phase of the fed-batch culture at a pre-set µ of 0.1 h 1. Error bars show the SE. followed the opposite trend, i.e. the YP/S is reduced from the ATP expenditure for maintenance at high extracellu- 0.134  0.010 Cmol Cmol 1 at the first part of the feed- lar concentrations of weak organic acid, i.e. 3-HP. Proto- ing phase to yields close to 0.1 Cmol Cmol 1 coinciding nated acid molecules may enter the cytoplasm by with the highest qPArabitol values. These results show how diffusion and later on dissociate, leading to an ATP cost during the beginning of the feeding phase the C-yield was for the re-secretion of the organic acid and the restora- close to the one obtained in the screening phase tion of the intracellular pH. The expression of a 3-HP (0.146  0.003 Cmol Cmol 1), even though the average exporter in E. coli has proven to be beneficial, as it yield of the overall fed-batch culture was lower reduces stress caused by the intracellular dissociation of (0.130  0.003 Cmol Cmol 1 and 0.127  0.003 g g 1). 3-HP (Nguyen-Vo et al., 2020). Moreover, higher extra- A possible explanation for the observed trends is the cellular concentrations of 3-HP have a direct impact on increase in the 3-HP concentration. Accumulation of 3-HP the thermodynamics of its secretion, leading to an in the fermentation broth has been reported to be toxic for increase in ATP expenses for 3-HP export (van Maris other microorganisms like E. coli and S. cerevisiae. Such et al., 2004). All these phenomena would ultimately toxicity may be triggered by the conversion of 3-HP into impact on the YX/S in P. pastoris, explaining the 3-hydroxypropanaldehyde (reuterin), which causes oxida- decrease in the growth rate and the accumulation of tive stress to the cells through its interaction with reduced glycerol at the later stage of the fed-batch culture. glutathione (Schaefer et al., 2010; Kildegaard et al., Moreover, we observe the production of arabitol at the 2014). The sequestration of glutathione leads to a end of the batch and the fed-batch phases (Fig. 3A). decrease in the maximal growth rate of S. cerevisiae at Arabitol production in P. pastoris has been related to high 3-HP concentrations (Kildegaard et al., 2014). stress conditions caused by the unfolded protein Another plausible explanation of the decrease in the µ response (UPR; Tredwell et al., 2017), high osmolarity at the final stage of the culture may be the increase in (Dragosits et al., 2010) or hypoxic conditions, where ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 3-HP production in Pichia pastoris 1677 arabitol production has been proposed to be used as a under aerobic conditions, leading to higher biomass and redox sink (Baumann et al., 2010). Additionally, arabitol product yields compared to the use of glucose; and (iii) by-product formation has been observed in lactic acid- the use of glycerol as substrate allows for higher growth producing recombinant P. pastoris growing in batch cul- rates during the feeding phase compared to typical fed- tures using glycerol as carbon source (Melo et al., batch cultivations of Crabtree-positive yeasts growing on 2020). Therefore, the presence of arabitol points to a glucose as substrate, thereby supporting higher volumet- redox imbalance, either caused by 3-HP-derived toxicity ric productivities, while minimizing by-product formation. effects, or by overexpression of cPos5. Further metabo- lomic and fluoxmic studies will help elucidating the Conclusions cause. Altogether, these results suggest that the pre- In this study, we successfully introduced the malonyl- established bioreactor fed-batch cultivation protocols for CoA to 3-HP pathway in P. pastoris for 3-HP production. heterologous protein production in P. pastoris are not The use of pGAP, a strong constitutive promoter, to necessarily optimal for the production of weak acids drive expression of the biosynthetic mcr gene, combined such as 3-HP, as its accumulation in the extracellular with the use of glycerol as carbon source, proved to be space and/or the toxic effect at high concentrations may key to obtain higher space-time yields than in other lead to physiological changes negatively impacting on yeasts. The subsequent combination of protein and the bioprocess parameters such as the µmax or YX/S. metabolic engineering strategies performed in this study Moreover, the current cultivation protocol led to the pro- have led to a 14-fold increase in the 3-HP yield. More- duction of arabitol as a by-product at the end of the fed- over, a controlled fed-batch strategy has shown the abil- batch process, which caused a significant decrease in ity of P. pastoris to produce up to 24.75 g l 1 of 3-HP in the product yield. Deletion of the arabitol dehydrogenase 45.5 h, achieving an overall yield of 0.13 Cmol Cmol 1, (ArDH) gene in a lactic acid-producing strain reduced and a productivity of 0.54 g l 1 h 1. Overall, we bench- arabitol production, resulting in a 20% increase in pro- marked P. pastoris for 3-HP production, demonstrating duct titres (Melo et al., 2020). the potential of this cell factory for platform chemicals The tolerance to 3-HP should also be addressed in bioproduction. In addition, this study serves as a basis order to increase further the product yield. To this end, for further optimization of this platform through integrated both adaptive laboratory evolution (ALE) experiments systems metabolic engineering and bioprocess engineer- and rational engineering strategies have proven suc- ing strategies, which are paramount to reach economi- cessful to increase 3-HP tolerance in other yeast and cally attractive metrics for 3-HP production from crude could be therefore transferred to P. pastoris. Specifically, glycerol in P. pastoris. ALE experiments led to the conclusion that overexpres- sion of the S-(hydroxymethyl)glutathione dehydrogenase Experimental procedures gene (SFA1) in S. cerevisiae restored growth at 50 g l 1 Strain construction of 3-HP (Kildegaard et al., 2014). Expression of two mutated versions of SFA1 under the control of the native A series of strains derived from the parental strain P. SFA1 promoter was also able to restore cell growth at pastoris X-33 (Invitrogen-Thermo Fisher Scientific, Wal- 50 g l 1 of 3-HP. The SFA1 residues which were tham, MA, USA) was generated. Plasmids and strains mutated in S. cerevisiae are conserved in the homolo- used during this study are listed in Table 1. Three gous P. pastoris gene (PAS_chr3_1028). Therefore, a heterologous genes were expressed in P. pastoris, similar approach to avoid 3-HP toxicity could be tested encoding for: a bifunctional malonyl-CoA reductase in P. pastoris. from Chloroflexus aurantiacus (mcrCa, Uniprot: The overall productivity of the fed-batch culture was Q6QQP7_CHLAU), an acetyl-CoA carboxylase from Yar- 0.54  0.01 g l 1 h 1. To our knowledge, this is the rowia lipolytica (ACCYl, Uniprot: YALI0_C11407) and a highest 3-HP productivity reported in yeast (de cytosolic NADH kinase from S. cerevisiae (cPOSSc; Fouche cour et al., 2018; Lis et al., 2019), and it is Toma s-Gamisans et al., 2020). The expression of all the almost identical to the highest productivity reported using heterologous genes was controlled by the strong and the malonyl-CoA pathway, which was 0.56 g l 1 h 1 and constitutive GAP promoter. A detailed description of the it was reported in E. coli (Liu et al., 2016). Such promis- molecular cloning protocols used to generate the plas- ing results can be attributed to the combination of three mids for this study is available in the Supplementary main factors: (i) the high specific MCR activity due to the Materials Section (Data S1). strength of pGAP, extensively proven for heterologous Electrocompetent P. pastoris cells were prepared as protein production in P. pastoris; (ii) the use of glycerol described elsewhere (Sears et al., 1998). Plasmids as a substrate, which delivers more ATP than glucose derived from pBIZi_pGAP (Bioingenium SL, Barcelona, ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 1678 A. Fina et al. Table 1. List of plasmids and P. pastoris strains used during this study. Plasmid name Expression cassettes in the plasmid Source pBIZi_pGAP Bioingenium SL pBIZi_pGAP_MCR PGAP - mcrCa - AOX1tt This study pBIZi_pGAP_MCR_N PGAP-mcrCa(N-ter)-AOX1tt This study pBIZi_pGAP_MCR_C PGAP-mcrCa(C-ter)-AOX1tt This study pBIZi_pGAP_MCR_E PGAP-mcrCa(CterN940V/K1106W/S1114R)-AOX1tt This study pBIZi_pGAP_MCR_NC PGAP-mcrCa(N-ter)-AOX1tt This study PGAP-mcrCa(C-ter)-AOX1tt pBIZi_pGAP_MCR_NE PGAP-mcrCa(N-ter)-AOX1tt This study PGAP-mcrCa(CterN940V/K1106W/S1114R)-AOX1tt BB1_23 Prielhofer et al. (2017) BB1_12_pGAP Prielhofer et al. (2017) BB1_34_ScCYC1tt Prielhofer et al. (2017) BB1_34_RPS3tt Prielhofer et al. (2017) BB2_AB Prielhofer et al. (2017) BB2_BC Prielhofer et al. (2017) BB3eH_14 Prielhofer et al. (2017) BB3eH_AC Prielhofer et al. (2017) BB3eH_ACC PGAP-ACCYl(DBbsI)-ScCYC1tt This study BB3eH_cPOS5 PGAP-cPOS5Sc-ScCYC1tt This study BB3eH_ACC_cPOS5 PGAP-ACCYl(DBbsI)-ScCYC1tt This study PGAP-cPOS5Sc-ScCYC1tt Strain name Plasmid integration Source X-33 Invitrogen-Thermo Fisher Scientific PpHP1 pBIZi_pGAP_MCR This study PpHP2 pBIZi_pGAP_MCR_NC This study PpHP3 pBIZi_pGAP_MCR_NE This study PpHP4 pBIZi_pGAP_MCR_NC This study BB3eH_ACC PpHP5 pBIZi_pGAP_MCR_NC This study BB3eH_cPOS5 PpHP6 pBIZi_pGAP_MCR_NC This study BB3eH_ACC_cPOS5 Spain) were linearized using AvrII (New England Bio- (OD600) of 0.5 on 5 ml of YPG and grown for 8 h at the labs, Ipswich, MA, USA), while the ones derived from same cultivation conditions. Afterwards, these cultures BB3eH were linearized using PmeI (New England Bio- were used to inoculate a 500-ml baffled shake flask con- labs). Transformation was performed using 100 ng of taining 50 ml of buffered minimal glycerol medium purified linear DNA according to a previously described (BMG; 100 mM potassium phosphate buffer pH 6, protocol (Sears et al., 1998). 1.34% yeast nitrogen base (YNB), 0.4 mg l 1 biotin and Recombinant P. pastoris strains were selected on 1% v/v glycerol) at a starting OD600 of 0.05. The shake YPD agar plates (1% yeast extract, 2% peptone, 2% flask cultures were grown at 30°C and 180 rpm in an dextrose, 15 g l 1 agar) supplemented with the appropri- incubator shaker Multitron Standard (Infors HT, Bottmin- ate antibiotic (100 µg ml 1 zeocine or hygromycin gen, Switzerland) with a 2.5 cm orbit. 200 µg ml 1; InvivoGen, CA, USA). When the plates contained both zeocine and hygromycin, the concentra- Screening in deep-well plates tion of each antibiotic was reduced by half (50 µg ml 1 zeocine and hygromycin 100 µg ml 1). Pichia pastoris strains were inoculated from cryovials into a 24 deep-well plates containing YPG medium sup- plemented with the appropriate antibiotic. Cultures were Shake flask cultures placed on a platform with a slope of 20° in an incubator Cultures were first grown overnight on 5 ml of YPG med- shaker Multitron Standard with 25 mm orbit and grown ium (1% yeast extract, 2% peptone and 1% v/v glycerol) overnight at 28°C and 220 rpm. Afterwards, 50 µl of supplemented with the appropriate antibiotic in 50-ml overnight cultures was used to inoculate a 24 deep-well Falcon tubes, at 30°C, and 180 rpm. The overnight cul- plate containing 2 ml of BMG medium. If not stated tures were then diluted to an optical density at 600 nm otherwise, six clones of each strain were tested in ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 3-HP production in Pichia pastoris 1679 triplicates. The cultures were grown for 48 h at the same increasing the agitation was insufficient to maintain the conditions. pO2 above the set point, the agitation was set to 1200 rpm and pure oxygen was mixed with air at the air inlet of the bioreactor to maintain a pO2 above 25%, Enzymatic assay while maintaining an aeration rate of 2 l min 1. Pichia pastoris X-33 and P. pastoris PpHP1 cells were The batch phase was performed with a starting vol- grown overnight on YPG medium supplemented with ume equal to 2 l, and the reactor was inoculated at a zeocine when appropriate. The grown cultures were starting OD600 equal to 1. The feeding phase started used to inoculate a 100-ml shake flask containing 25 ml after a sudden increase in the pO2 indicating that all the of fresh BMG medium at an initial OD600 of 0.1. When glycerol from the batch phase had been consumed. A the cells reached an OD600 of about 2, 5 ml of each cul- pre-set exponential feeding rate was used when adding ture were centrifuged at 6000 g. The pellets were the feeding medium into the reactor. Such exponential washed twice with ice-cold PBS buffer and resuspended feeding rate was calculated according to previous litera- in 1 ml of breaking buffer (50 mM HEPES, 150 mM KCl, ture (Cos et al., 2005), and it aimed to maintain pseudo- 1 mM DTT, 1 mM EDTA, pH 7.5, Halt Protease Inhibitor steady state conditions at a constant growth rate of Cocktail (Thermo Fisher Scientific)). Resuspended cells 0.1 h 1 at C-limiting conditions. The working volume of were then lysed using glass beads applying 5 cycles of the reactor was calculated as described elsewhere 1 min of vortexing followed by 2 min of incubation on (Garcia-Ortega et al., 2013). To calculate the exponential ice. Cell lysates were centrifuged at 16 000 g and 4°C feeding rate, the biomass concentration at the end of the for 20 min. The supernatant (soluble fraction) was trans- batch phase (X0) and the biomass to substrate yield ferred to a new tube, and the amount of protein in the (YX/S) were used. cell extract was quantified using the Bradford method (PierceTM Coomassie Plus, Thermo Fisher Scientific). Analytical methods and data processing The MCR specific activity was quantified as previously reported (Chen et al., 2014). As MCR is a bifunctional The analytical and biological replicates were averaged, enzyme consuming 2 NAPDH molecules to convert and the standard error (SE) for each measurement was malonyl-CoA into 3-HP, one unit of enzyme activity was calculated considering the number of biological repli- defined as 2 µmol NADPH consumed per minute. cates (number of independent samples). The splines fitting the µ, the q-rates and the product yield during the feeding phase of the fed-batch culture Bioreactor cultivation were performed using Matlab 2019 (Mathworks, Natick, For the bioreactor cultivations, the batch medium con- MA, USA). The fit function was used together with the sisted of 40 g l 1 glycerol, 1.8 g l 1 citric acid, additional options: ‘smoothingspline’ and ‘Smooth- 0.02 g l 1 CaCl2  2 H2O, 12.6 g l 1 (NH4)2HPO4, ingParam’ = 0.5. 0.5 g l 1 MgSO4  7 H2O, 0.9 g l 1 KCl, 50 µl antifoam The OD600 measurements of the shake flask and the Glanapon 2000 kz (Bussetti and Co GmbH, Vienna, bioreactor cultures were performed using a Lange DR Austria), 4  10 5 g l 1 biotin and 4.6 ml l 1 of PTM1 3900 spectrophotometer (Hach, Loveland, CO, USA). trace salts (Maurer et al., 2006). The pH of the medium To quantify the metabolites, the culture samples were was set to 5 using 5 M HCl. The feeding medium com- first centrifuged at 12 000 g and 4°C and then filtered position was 400 g l 1 glycerol, 10 g l 1 KCl, 6.45 g l 1 using a 0.45-µm syringe filter of type HAWP (Millipore, MgSO4  7 H2O, 0.35 g l 1 CaCl2  2 H2O, 0.2 ml l 1 Temecula, CA, USA). Glycerol and arabitol were quanti- antifoam Glanapon 2000 kz, 0.3 mg l 1 biotin and fied using an HPLC Dionex Ultimate3000 (Dionex – 15 ml l 1 PTM1 trace salts. For the batch and the feed- Thermo Fisher Scientific) coupled to a UV detector at ing media, all components except the biotin and the 210 nm and an RI detector (Dionex – Thermo Fisher Sci- trace salts were mixed and autoclaved. Biotin and trace entific). The compounds were separated with an ionic salts were filter-sterilized and added to the mixture once exchange column ICSep ICE-COREGEL 87H3 (Transge- the medium had cooled down. nomic, Omaha, NE, USA) using 6 mM sulphuric acid as Fed-batch cultures were performed in duplicate in a 5- mobile phase at a flow rate of 0.6 ml min 1. As 3-HP and l Biostat B Bioreactor (Sartorius Stedim, Goettingen, glycerol elute the column at the same time, the 3-HP was Germany). The pH was maintained at 5 using ammonia quantified using two different methods: HPLC-MS and 15% (v/v). The temperature was set to 28°C, and air nuclear magnetic resonance (NMR) spectroscopy. Stan- was added into the reactor at an aeration rate of 1 vvm dard samples for calibration were prepared using 3- (2 l min 1). The agitation was gradually increased from hydroxypropionic acid sodium salt from Santa Cruz 600 to 1200 rpm to maintain a pO2 above 25%. When Biotechnology (Santa Cruz, CA, USA). ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 1680 A. Fina et al. For the HPLC-MS method, we used a Prominence (FPU17/05434) from the Spanish Ministry of Science, HPLC (Shimadzu, Kyoto, Japan) coupled to a single and a FI scholarship (2018FI_B00713) from AGAUR. quadrupole Shimadzu-2010A mass spectrometry detec- GCB was supported by Conselho Nacional de Desen- tor using electrospray ionization. The metabolites of the gico (CNPq, Brazil) volvimento Cientıfico e Tecnolo supernatant were separated using an ICSep 87H USP scholarship (158956/2015-0). L17 column (Transgenomic, NE, USA) and 16 mM for- mic acid as mobile phase at a flow rate of Conflict of interest 0.15 ml min 1. 2 µl samples were injected into the col- umn, and the MS analyser was set to 89 m/z for nega- The authors have no conflict of interest to declare. tively charged molecules, which corresponds to the m/z ratio for unprotonated 3-HP. The detector settings were References as follows: curved desolvation line (CDL) temperature at 200°C, heat block temperature at 200°C, voltage of the Baumann, K., Carnicer, M., Dragosits, M., Graf, A.B., Stadl- detector at 1.5 kV, nebulizing gas (nitrogen) flow at mann, J., Jouhten, P., et al. (2010) A multi-level study of 1.5 l min 1 and drying gas (nitrogen) flow at 10 l min 1. recombinant Pichia pastoris in different oxygen conditions. Each sample was analysed in duplicate. BMC Syst Biol 4: 141. Borodina, I., Kildegaard, K.R., Jensen, N.B., Blicher, T.H., For the NMR method, a Bruker AVANCE 600 spec- Maury, J., Sherstyk, S., et al. (2015) Establishing a syn- trometer (600.13 MHz frequency for 1H) equipped with a thetic pathway for high-level production of 3- 5 mm TBI probehead and an autosampler (Bruker BioS- hydroxypropionic acid in Saccharomyces cerevisiae via b- pin, Rheinstetten, Germany) was utilized. The probe alanine. Metab Eng 27: 57–64. temperature was maintained at 300.0 K for all experi- Carnicer, M., Canelas, A.B., ten Pierick, A., Zeng, Z., van ments. Once centrifuged and filtered, 300 µl of each cul- Dam, J., Albiol, J., et al. (2012) Development of quantitative ture aliquot was mixed with 300 µl of a D2O stock metabolomics for Pichia pastoris. Metabolomics 8: 284–298. Chen, Y., Bao, J., Kim, I.K., Siewers, V., and Nielsen, J. solution containing an internal standard (3-(trimethylsilyl)- (2014) Coupled incremental precursor and co-factor sup- [2,2,3,3-2H4]-propionic acid sodium salt (TSP), 11.1 mM) ply improves 3-hydroxypropionic acid production in Sac- and transferred to the NMR tube. All samples were anal- charomyces cerevisiae. Metab Eng 22: 104–109. ysed conducting standard quantitative 1D 1H NMR Cos, O., Resina, D., Ferrer, P., Montesinos, J.L., and experiments with presaturation of the residual water sig- Valero, F. (2005) Heterologous production of Rhizopus nal. Data were collected into 32k data points during an oryzae lipase in Pichia pastoris using the alcohol oxidase acquisition time of 1.7 s using a recycle delay of 15 s. and formaldehyde dehydrogenase promoters in batch and fed-batch cultures. Biochem Eng J 26: 86–94. Spectra were recorded in the time domain as interfero- Dai, Z., Huang, M., Chen, Y., Siewers, V., and Nielsen, J. grams (FID) across a spectral width of 9,615 Hz and as (2018) Global rewiring of cellular metabolism renders Sac- the sum of 128 transients. FIDs were automatically Four- charomyces cerevisiae Crabtree negative. Nat Commun ier transformed (FT) and the spectra were phased, and 9: 1–8. baseline corrected. TSP was used as internal reference Della Pina, C., Falletta, E., and Rossi, M. (2011) A green (d(1H) and d(13C) at 0.00 ppm). approach to chemical building blocks. the case of 3- The glycerol concentration obtained from the IR spec- hydroxypropanoic acid. Green Chem 13: 1624–1632. Dragosits, M., Stadlmann, J., Graf, A., Gasser, B., Maurer, tra of the HPLC analysis was corrected accounting for M., Sauer, M., et al. (2010) The response to unfolded pro- the 3-HP interference. The area at the IR detector corre- tein is involved in osmotolerance of Pichia pastoris. BMC sponding to the amount of 3-HP quantified using the Genom 11: 1–16. HPLC-MS or the NMR analyses was calculated and sub- Ferrer, P., and Albiol, J. (2014) 13C-based metabolic flux tracted from the total IR area. analysis of recombinant Pichia pastoris. Methods Mol Biol 1191: 291–313. de Fouche cour, F., Sa nchez-Castan ~eda, A.K., Saulou- Acknowledgements Be  (2018) Process engineering  rion, C., and Spinnler, H.E. for microbial production of 3-hydroxypropionic acid. This work was supported by projects PHBP 14/00087 Biotechnol Adv 36: 1207–1222. and CTQ2016-74959-R of the Spanish Ministry of Econ- Garcia-Ortega, X., Ferrer, P., Montesinos, J.L., and Valero, omy and Competitiveness; 2017-SGR-1462 from the F. (2013) Fed-batch operational strategies for recombi- Agencia de Gestio  d’Ajuts Universitaris i de Recerca nant Fab production with Pichia pastoris using the consti- (AGAUR) of the Catalan Government; PETROBRAS tutive GAP promoter. Biochem Eng J 79: 172–181. cooperation term number 0050.0081285.13.9; and Grand View Research (2016) Acrylic Acid Market Size Worth $22.55 Billion By 2022. [WWW document]. URL CAPES-DGPU programme (1694/2015) from Coor- https://www.grandviewresearch.com/press-release/global- denacßa~o de Aperfeicßoamento de Pessoal de Nıvel Supe- acrylic-acid-market. rior (CAPES). AF was supported with a FPU scholarship ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 3-HP production in Pichia pastoris 1681 Kildegaard, K.R., Hallstro € m, B.M., Blicher, T.H., Sonnen- Raj, S.M., Rathnasingh, C., Jo, J.E., and Park, S. (2008) schein, N., Jensen, N.B., Sherstyk, S., et al. (2014) Evolu- Production of 3-hydroxypropionic acid from glycerol by a tion reveals a glutathione-dependent mechanism of 3- novel recombinant Escherichia coli BL21 strain. Process hydroxypropionic acid tolerance. Metab Eng 26: 57–66. Biochem 43: 1440–1446. Kildegaard, K.R., Jensen, N.B., Schneider, K., Czarnotta, Rathnasingh, C., Raj, S.M., Jo, J.E., and Park, S. (2009) € E., Ozdemir, E., Klein, T., et al. (2016) Engineering and Development and evaluation of efficient recombinant systems-level analysis of Saccharomyces cerevisiae for Escherichia coli strains for the production of 3- production of 3-hydroxypropionic acid via malonyl-CoA hydroxypropionic acid from glycerol. Biotechnol Bioeng reductase-dependent pathway. Microb Cell Fact 15: 53. 104: 729–739. Kumar, V., Ashok, S., and Park, S. (2013) Recent advances Rathnasingh, C., Raj, S.M., Lee, Y., Catherine, C., Ashok, in biological production of 3-hydroxypropionic acid. S., and Park, S. (2012) Production of 3-hydroxypropionic Biotechnol Adv 31: 945–961. acid via malonyl-CoA pathway using recombinant Escheri- Lis, A.V., Schneider, K., Weber, J., Keasling, J.D., Jensen, chia coli strains. J Biotechnol 157: 633–640. M.K., and Klein, T. (2019) Exploring small-scale chemo- Schaefer, L., Auchtung, T.A., Hermans, K.E., Whitehead, stats to scale up microbial processes: 3-hydroxypropionic D., Borhan, B., and Britton, R.A. (2010) The antimicrobial acid production in S. cerevisiae. Microb Cell Fact 18: 50. compound reuterin (3-hydroxypropionaldehyde) induces Liu, C., Ding, Y., Zhang, R., Liu, H., Xian, M., and Zhao, G. oxidative stress via interaction with thiol groups. Microbiol- (2016) Functional balance between enzymes in malonyl- ogy 156: 1589–1599. CoA pathway for 3-hydroxypropionate biosynthesis. Schwarzhans, J.P., Luttermann, T., Geier, M., Kalinowski, J., Metab Eng 34: 104–111. and Friehs, K. (2017) Towards systems metabolic engi- Liu, C., Wang, Q., Xian, M., Ding, Y., and Zhao, G. (2013) neering in Pichia pastoris. Biotechnol Adv 35: 681–710. Dissection of malonyl-coenzyme A reductase of Chlo- Sears, I.B., O’Connor, J., Rossanese, O.W., and Glick, B.S. roflexus aurantiacus results in enzyme activity Improve- (1998) A versatile set of vectors for constitutive and regulated ment. PLoS One 8: 9. gene expression in Pichia pastoris. Yeast 14: 783–790. Looser, V., Bruhlmann, B., Bumbak, F., Stenger, C., Costa, Song, C.W., Kim, J.W., Cho, I.J., and Lee, S.Y. (2016) M., Camattari, A., et al. (2014) Cultivation strategies to Metabolic engineering of Escherichia coli for the produc- enhance productivity of Pichia pastoris: A review. Biotech- tion of 3-Hydroxypropionic acid and malonic acid through nol Adv 33: 1177–1193. b-Alanine route. ACS Synth Biol 5: 1256–1263. Luo, X., Ge, X., Cui, S., and Li, Y. (2016) Value-added pro- Suyama, A., Higuchi, Y., Urushihara, M., Maeda, Y., and cessing of crude glycerol into chemicals and polymers. Takegawa, K. (2017) Production of 3-hydroxypropionic Bioresour Technol 215: 144–154. acid via the malonyl-CoA pathway using recombinant fis- van Maris, A.J.A., Konings, W.N., van Dijken, J.P., and sion yeast strains. J Biosci Bioeng 124: 392–399. Pronk, J.T. (2004) Microbial export of lactic and 3- Takayama, S., Ozaki, A., Konishi, R., Otomo, C., Kishida, M., hydroxypropanoic acid: Implications for industrial fermen- Hirata, Y., et al. (2018) Enhancing 3-hydroxypropionic acid tation processes. Metab Eng 6: 245–255. production in combination with sugar supply engineering Maurer, M., Ku €hleitner, M., Gasser, B., and Mattanovich, D. by cell surface-display and metabolic engineering of (2006) Versatile modeling and optimization of fed batch Schizosaccharomyces pombe. Microb Cell Fact 17: 176. processes for the production of secreted heterologous Toma s-Gamisans, M., Andrade, C.C.P., Maresca, F., Mon- proteins with Pichia pastoris. Microb Cell Fact 5: 1–10. forte, S., Ferrer, P., and Albiol, J. (2020) Redox engineer- Melo, N.T.M., Pontes, G.C., Proco pio, D.P., de Cunha, ing by ectopic overexpression of NADH Kinase in G.C.G., Eliodo  rio, K.P., Paes, H.C., et al. (2020) Evalua- recombinant Pichia pastoris (Komagataella phaffii): Impact tion of product distribution in chemostat and batch fer- on cell physiology and recombinant production of mentation in lactic acid-producing Komagataella phaffii secreted proteins. Appl Environ Microbiol 86: 1–15. strains utilizing glycerol as substrate. Microorganisms 8: Toma s-Gamisans, M., Ferrer, P., and Albiol, J. (2016) Inte- 1–12. gration and validation of the genome-scale metabolic Nguyen-Vo, T.P., Ko, S., Ryu, H., Kim, J.R., Kim, D., and models of Pichia pastoris: A comprehensive update of Park, S. (2020) Systems evaluation reveals novel trans- protein glycosylation pathways, lipid and energy metabo- porter YohJK renders 3 - hydroxypropionate tolerance in lism. PLoS One 11: 1. Escherichia coli. Sci Rep 10: 1–12. Tredwell, G.D., Aw, R., Edwards-Jones, B., Leak, D.J., and Novozymes (2008) Cargill and Novozymes to enable pro- Bundy, J.G. (2017) Rapid screening of cellular stress duction of acrylic acid via 3HPA from renewable raw responses in recombinant Pichia pastoris strains using materials. [Press release]. URL https://www.novozymes.c metabolite profiling. J Ind Microbiol Biotechnol 44: 413–417. om/en/news/news-archive/2008/01/44469. Waterham, H.R., Digan, M.E., Koutz, P.J., Lair, S.V., and Pen~a, D.A., Gasser, B., Zanghellini, J., Steiger, M.G., and Cregg, J.M. (1997) Isolation of the Pichia pastoris Mattanovich, D. (2018) Metabolic engineering of Pichia glyceraldehyde-3-phosphate dehydrogenase gene and pastoris. Metab Eng 50: 2–15. regulation and use of its promoter. Gene 186: 37–44. Prielhofer, R., Barrero, J.J., Steuer, S., Gassler, T., Zahrl, Weninger, A., Hatzl, A.M., Schmid, C., Vogl, T., and Glieder, R., Baumann, K., et al. (2017) GoldenPiCS: A Golden A. (2016) Combinatorial optimization of CRISPR/Cas9 Gate-derived modular cloning system for applied synthetic expression enables precision genome engineering in the biology in the yeast Pichia pastoris. BMC Syst Biol 11: 1– methylotrophic yeast Pichia pastoris. J Biotechnol 235: 14. 139–149. ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682 1682 A. Fina et al. Werpy, T., and Petersen, G. (2004) Top value added chemi- Zhou, S., Ashok, S., Ko, Y., Kim, D.-M., and Park, S. (2014) cals from biomass Volume I - Results of screening for Development of a deletion mutant of Pseudomonas deni- potential candidates from sugars and synthesis gas. US trificans that does not degrade 3-hydroxypropionic acid. Department of Energy. https://www.osti.gov/biblio/15008859- Appl Microbiol Biotechnol 98: 4389–4398. top-value-added-chemicals-from-biomass-volume-results- screening-potential-candidates-from-sugars-synthesis-gas. Yang, Y.-M., Chen, W.-J., Yang, J., Zhou, Y.-M., Hu, B.o., Zhang, M., et al. (2017) Production of 3-hydroxypropionic Supporting information acid in engineered Methylobacterium extorquens AM1 and its reassimilation through a reductive route. Microb Additional supporting information may be found online in Cell Fact 16: 179. the Supporting Information section at the end of the arti- Yang, Z., and Zhang, Z. (2018) Engineering strategies for cle. enhanced production of protein and bio-products in Data. S1. Supplementary materials. Pichia pastoris: a review. Biotechnol Adv 36: 182–195. ª 2021 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology., Microbial Biotechnology, 14, 1671–1682