Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Microbial Cell Factories https://doi.org/10.1186/s12934-023-02241-9 RESEARCH Open Access Engineering the synthetic β-alanine pathway in Komagataella phaffii for conversion of methanol into 3-hydroxypropionic acid Sílvia Àvila‑Cabré1, Míriam Pérez‑Trujillo2, Joan Albiol1 and Pau Ferrer1* Abstract Background Methanol is increasingly gaining attraction as renewable carbon source to produce specialty and com‑ modity chemicals, as it can be generated from renewable sources such as carbon dioxide (­ CO2). In this context, native methylotrophs such as the yeast Komagataella phaffii (syn Pichia pastoris) are potentially attractive cell factories to pro‑ duce a wide range of products from this highly reduced substrate. However, studies addressing the potential of this yeast to produce bulk chemicals from methanol are still scarce. 3-Hydroxypropionic acid (3-HP) is a platform chemical which can be converted into acrylic acid and other commodity chemicals and biopolymers. 3-HP can be naturally produced by several bacteria through different metabolic pathways. Results In this study, production of 3-HP via the synthetic β-alanine pathway has been established in K. phaffii for the first time by expressing three heterologous genes, namely panD from Tribolium castaneum, yhxA from Bacil- lus cereus, and ydfG from Escherichia coli K-12. The expression of these key enzymes allowed a production of 1.0 g ­l−1 of 3-HP in small-scale cultivations using methanol as substrate. The addition of a second copy of the panD gene and selection of a weak promoter to drive expression of the ydfG gene in the PpCβ21 strain resulted in an additional increase in the final 3-HP titer (1.2 g ­l−1). The 3-HP-producing strains were further tested in fed-batch cultures. The best strain (PpCβ21) achieved a final 3-HP concentration of 21.4 g ­l−1 after 39 h of methanol feeding, a product yield ­ −1, and a volumetric productivity of 0.48 g ­l−1 ­h−1. Further engineering of this strain aiming at increasing of 0.15 g g NADPH availability led to a 16% increase in the methanol consumption rate and 10% higher specific productivity compared to the reference strain PpCβ21. Conclusions Our results show the potential of K. phaffii as platform cell factory to produce organic acids such as 3-HP from renewable one-carbon feedstocks, achieving the highest volumetric productivities reported so far for a 3-HP production process through the β-alanine pathway. Keywords 3-Hydroxypropionic acid, Pichia pastoris, Komagataella phaffii, Methanol, β-Alanine pathway, Metabolic engineering Background *Correspondence: Fossil fuels not only constitute the most important energy Pau Ferrer sources in our everyday life, but they are also critical to pau.ferrer@uab.cat 1 Department of Chemical, Biological and Environmental Engineering, produce plastics, synthetic materials, and other chemi- Universitat Autònoma de Barcelona, Bellaterra, Catalonia, Spain cal products. However, the accelerated depletion of fos- 2 Servei de Ressonància Magnètica Nuclear, Facultat de Ciències i sil resources and its undeniable relationship with global Biociències, Universitat Autònoma de Barcelona, Bellaterra, Catalonia, Spain warming are forcing us to identify alternative feedstocks. © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visithttp://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 2 of 16 Moreover, the use of plant-derived sugars as carbon 3-HP was identified as one of the top value-added sources comes into conflict with food commodities. chemicals from biomass by the US Department of Energy Methanol is increasingly gaining competitiveness as (DOE) in 2004, since it was considered a key building a microbial non-food feedstock for the chemical indus- block for both commodity and specialty chemicals pro- try and biomanufacturing. This one-carbon compound duction. 3-HP can be converted into 1,3-propanediol is broadly available, presently obtained from syn-gas. It (PDO), acrylamide, acrylic acid, and methyl acrylate [16]. can also be prepared through a catalytic conversion that There is a growing demand for acrylic acid-based poly- oxidizes a natural gas (methane). Another renewable mers used in the personal care industry, coatings, adhe- alternative to produce methanol includes the reduction sives, and others. The global market size of this 3-HP of atmospheric C ­ O2 with hydrogen, promoting a sustain- derivative is expected to reach $19.2 billion by 2030 [17]. able circular carbon economy [1]. In addition, this carbon The bioconversion of renewable methanol into this three- source has a greater degree of reduction per C-mol com- carbon acid building block will allow us to reduce the pared to most sugars, leading to potential higher product unsustainable use of fossil resources, meaning less envi- yields, and thus is considered a promising building block ronmental pollution [18]. Furthermore, since K. phaffii is for synthesizing biofuels, commodity chemicals and considered an acid-tolerant host, the fermentation pro- biopolymers [2]. cess could be performed at a pH below the pKa value of Komagataella phaffii (also known as Pichia pastoris) 3-HP, obtaining the undissociated form of the acid, there- is a methylotrophic yeast strain capable of assimilating fore reducing the costs of the downstream process [19]. methanol as a sole source of carbon and energy. This is Several microorganisms, both prokaryotes and eukary- possible due to the presence of strongly induced genes otes, are 3-HP natural producers, although none of them encoding for enzymes involved in methanol metabolism, produces a significant amount of this organic acid [20]. which are mostly located in the peroxisomes. After its Three predominant heterotrophic routes named accord- discovery in the late 1960s, K. phaffii was first used for ing to its precursor have been reconstructed into the biomass and single-cell protein production from metha- industrial workhorses Saccharomyces cerevisiae and nol [3, 4]. Later, K. phaffii was developed as a system to Escherichia coli for 3-HP synthesis. The highest 3-HP produce heterologous proteins for both academic and titer and overall process productivity were obtained with industrial purposes. The availability of the strong and the coenzyme ­ B12-dependent glycerol pathway imple- tightly regulated methanol inducible AOX1 promoter mented in E. coli [21]. Nevertheless, E. coli is not able to paved the ground for the exploitation of this non-con- produce vitamin B ­ 12 on its own. Therefore, this expensive ventional yeast for heterologous protein production. cofactor must be added to the culture medium, making Synthetic biology toolboxes such as GoldenPiCS [5] and the process economically unfeasible at an industrial scale CRISPi kit [6], together with the growing knowledge [22]. base of its metabolism, have led K. phaffii to stand out Alternatively, the gene encoding for a bifunctional as a specialized chassis not only to produce recombinant malonyl-CoA reductase from Chloroflexus aurantiacus proteins, but also to metabolically engineer this yeast ­(MCRCa) has been expressed to implement the malonyl- to produce added value metabolites [7]. In this context, CoA pathway in several model organisms, such as E. coli the deployment of technologies for renewable methanol [23–26], S. cerevisiae [27–29] and Schizosaccharomy- generation makes it an attractive feedstock for chemi- ces pombe [30, 31], allowing for the production of 3-HP. cals production using methylotrophic microorganisms. Recently, we have successfully introduced the malonyl- Methanol possesses more available electrons per car- CoA pathway in K. phaffii to produce 3-HP from glycerol bon atom (reduction degree) than most sugars, which [32]. After further metabolic engineering, the best strain implies that more reducing power is being generated, produced up to 37.1 g ­l−1 of 3-HP at 0.71 g ­l−1 ­h−1 with usually resulting in enhanced product yields of desired a final product yield of 0.19 g ­g−1 in fed-batch cultures, metabolites [2]. Over the past 10 years, several chemi- resulting in the highest volumetric productivity reported cals have been synthesized in K. phaffii using methanol as so far in yeast [33]. Notably, a free fatty acids (FFA) over- C-source, including carboxylic acids such as malic acid, producing S. cerevisiae strain engineered with the mal- d-lactic acid and 3-hydroxypropionic acid (3-HP) [8–10], onyl-CoA pathway has been recently optimized for an polyketides [11–13], and fatty acids derivatives [14]. enhanced supply of metabolic precursors (malonyl-CoA Recently, the production of amino acids (β-alanine) from and acetyl-CoA) and cofactor (NADPH), enabling the this renewable feedstock has also been reported in K. highest 3-HP production (56.5 g ­l−1) and product yield on phaffii for the first time [15]. This β-amino acid is a pre- glucose (0.31 g ­g−1) in fed-batch cultivations reported so cursor for the biosynthesis of several nitrogen-containing far in yeast [34]. Similarly, the malonyl-CoA pathway has chemicals, such as pantothenic acid (vitamin B ­ 5). also been implemented in a FFA-overproducing K. phaffii Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 3 of 16 strain to efficiently produce 3-HP from sole methanol, through the malonyl-CoA pathway was calculated to be achieving a final titer of 48.2 g l­−1, which is the highest 0.718 ­g3-HP ­gMetOH−1 (Fig. 1a), whereas a higher value was reported 3-HP production from one-carbon feedstocks, obtained when using the β-alanine route (0.795) (Fig. 1b), and a product yield of 0.23 g g­ −1 [10]. mainly owing to a lower ATP requirement to produce In addition, 3-HP has been produced through the syn- one molecule of 3-HP from 3 molecules of methanol thetic β-alanine pathway in E. coli [35, 36] and S. cer- through the β-alanine pathway (see Additional file 1), evisiae [28, 37]. This metabolic pathway consists of two resulting in slightly lower oxygen requirements compared steps, the first one converting β-alanine into malonyl to the malonyl-CoA pathway (Fig. 1a, b). The higher ATP semialdehyde (MSA) by either a β-alanine-pyruvate ami- requirement for 3-HP production through the malonyl- notransferase (BAPAT) or a γ-aminobutyrate transami- CoA pathway is mainly due to acetyl-CoA synthesis, nase (GABT), and the second one consisting in a which not only has an impact on the Υ3-HP max value, but reduction of MSA to 3-HP consuming either NADH or also results in higher oxygen dependence, as previously NADPH [38–40]. Several genes were tested for each one pointed out by Borodina et al. [37]. Overall, the β-alanine of these two steps to implement the pathway in S. cerevi- siae, finally determining the expression of a BAPAT from Bacillus cereus ­(BAPATBc), along with a 3-hydroxypropi- onate dehydrogenase from E. coli K-12 (­ YDFGEc), as the best option. Moreover, the endogenous biosynthesis of β-alanine is only active via spermine and has a very low flux. For this reason, an aspartate-1-decarboxylase from Tribolium castaneum ­(PANDTc), catalyzing the conver- sion of L-aspartate into β-alanine, was expressed in S. cerevisiae. After further metabolic engineering of this strain, a final titer of 13.7 g ­l−1 of 3-HP with a 0.14 g ­g−1 product yield was obtained in a glucose-limited con- trolled fed-batch cultivation [37]. Recently, 3-HP pro- duction through the β-alanine pathway has been further optimized by applying phosphate-limiting conditions to a glucose-based fed-batch system, achieving a final 3-HP concentration of approximately 27 g ­l−1 with a product yield of 0.26 g g­ −1 [41]. In the present study, the enzymes used by Boro- dina et al. in S. cerevisiae [37] were expressed for the first time in K. phaffii to enable 3-HP production from methanol via the β-alanine pathway in our base strains. These base strains were further modified through cofac- tor engineering by overexpressing a Pseudomonas sp. 101 ­NADP+-dependent formate dehydrogenase variant. Fed-batch cultivations using methanol as the only car- bon source in the feed were performed to evaluate the potential of this cell factory to generate an economically important product from this C1 feedstock. Fig. 1 Quantitative prediction of metabolic capabilities of K. phaffii Results and discussion for 3-HP production from methanol. a Through the malonyl-CoA In silico comparison of the malonyl‑CoA and the β‑alanine pathway. b Through the β-alanine pathway. A range of values pathways in K. phaffii for the specific growth rate was set, while limiting the system The K. phaffii’s iMT1026 v3.0 genome-scale metabolic to different oxygen availability conditions. The 3-HP production was defined as the objective function to be maximized, obtaining model [42] was used to evaluate the maximum 3-HP the above phase planes after calculating the corresponding product theoretical yield (Υ3-HP max) that could be achieved using yield for each predicted 3-HP production value. The color scale methanol as sole carbon source through the malonyl- code indicates the specific oxygen uptake rate ­(qO2), from 5 (yellow) CoA and the β-alanine pathways. Assuming that all the to maximal value set at 8 mmol O ­ 2 ­gDCW−1 ­h−1 (dark blue). Y-axis carbon source was used for product synthesis, that is, not represents the product yield (Υ3-HP) ­(g3-HP ­gMetOH−1), and the X-axis illustrates the growth rate (µ) ­(h−1) used for biomass growth (µ = 0 ­h−1), the maximum Υ3−HP Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 4 of 16 route stands out as a more attractive metabolic pathway ­YDFGEc (Fig. 2). Due to the NADPH-dependence of the towards 3-HP production at industrial scale. latter enzyme, we expected the potential generation of a redox imbalance by the last step of the pathway. For this The expression of panD, yhxA and ydfG genes in K. phaffii reason, instead of using a strong and methanol induc- generates a 3‑HP‑producing strain ible promoter to control the expression of the ydfG gene, To reconstruct the synthetic β-alanine pathway for 3-HP two different constitutive promoters showing a lower production, the three codon-optimized panD, yhxA and expression strength in comparison with the strong GAP ydfG genes encoding for a PAND from T. castaneum, a promoter under methanol feed conditions (µmax up to BAPAT from B. cereus and a YDFG from E. coli, respec- 0.1 ­h−1) were tested, namely the moderate strength mito- tively, were introduced into K. phaffii CBS7435, the chondrial porin (POR1) promoter, obtaining the PpCβ10 parental strain. The strong and methanol inducible alco- strain, and the weak pyruvate decarboxylase (PDC1) pro- hol oxidase 1 (AOX1) and formate dehydrogenase (FDH1) moter, obtaining the PpCβ20 strain [5]. promoters, were selected to conduct the expression of The PpCβ10 and PpCβ20 strains were grown up to an panD and yhxA genes, respectively. According to previ- optical density at 600 nm (­OD600) of 2.4 (see Additional ous studies [37], the utilization of strong promoters such file 2) on 24 deep-well plates with minimal medium con- as the translation elongation factor EF-1 alpha promoter taining 11.9 g ­l−1 (i.e., 1.5% v/v) of methanol, along with ­(PTEF1-α) for the expression of both panD and yhxA genes the reference strain (CBS7435). After 48 h of incubation, proved to be beneficial for 3-HP production with an methanol was completely depleted. Since K. phaffii does engineered S. cerevisiae strain growing on glucose, since not assimilate 3-HP as C-source [32], an endpoint sample the reaction catalyzed by the P ­ ANDTc was determined to was collected at the end of the culture to quantify the final be the main flux control step of the β-alanine pathway. In 3-HP titers. The heterologous expression of panD, yhxA these two first steps, an l-aspartate is transformed into and ydfG genes resulted in 0.93 ± 0.03 g ­l−1 of 3-HP after β-alanine by the ­PANDTc, which is in turn converted into 48 h of cultivation when the P ­ POR1 was used to control MSA by the ­BAPATBc enzyme. 3-HP is finally obtained the expression of the ydfG gene (PpCβ10 strain), whereas through the reduction of MSA by the action of the a slightly but significantly higher titer (p-value < 0.05), Fig. 2 Schematic representation of the biological β-alanine route towards 3-HP production using methanol as substrate. The enzyme abbreviations are as follows: PAND – aspartate-1-decarboxylase, BAPAT – β-alanine-pyruvate aminotransferase, YDFG – 3-hydroxypropionate dehydrogenase, FDH(V9) – mutated formate dehydrogenase. The three heterologous enzymes required for the expression of the synthetic β-alanine pathway in K. phaffii CBS7435 parental strain are highlighted in red. The NADPH regeneration by FDH(V9) is shaded in purple Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 5 of 16 1.04 ± 0.01 g ­l−1 3-HP, was obtained with the weak P ­ PDC1 multiple copies of B ­APATBc/YDFGEc, thus suggesting driving expression of this gene (PpCβ20 strain). Notably, that the major control point of the flux through the bio- the 3-HP titers obtained herein were slightly higher than synthetic pathway were the panD gene transcriptional those achieved with a base S. cerevisiae strain expressing levels [37]. Accordingly, we hypothesized the overexpres- the same key enzymes under the control of strong con- sion of this gene as the most straightforward strategy for stitutive promoters (TEF1 and PGK1) and growing on a 3-HP improvement. 20 g ­l−1 glucose medium (0.83 g ­l−1 3-HP) [37]. Moreover, The integration of a second copy of the ­PANDTc expres- the expression of these three heterologous genes in K. sion cassette into both PpCβ10 and PpCβ20 strains phaffii resulted in similar 3-HP titers to those previously resulted in the PpCβ11 and PpCβ21 strains, respectively, reported for a glucose-utilizing S. cerevisiae strain car- which were cultivated on buffered minimal methanol rying overexpression cassettes for 5 genes, namely pyru- medium (BMM) along with their parentals. After 48 h of vate carboxylase (PYC1 and PYC2), ­BAPATBc, ­PANDTc cultivation, all strains reached a similar ­OD600 between and ­YDFGEc. Only when xylose was used as substrate, a 2.3 and 2.5 (see Additional file 2). The 3-HP production higher production of 3-HP was achieved (1.84 g l­−1) with slightly increased by 3% (p-value = 0.04) in PpCβ11 strain the latter strain [28]. These initial screening experiments (0.92 ± 0.01 g ­l−1 3-HP), in comparison with its parental pointed at K. phaffii as a promising chassis cell for the PpCβ10 strain (0.89 ± 0.02 g l­−1), while the strain PpCβ21 bioproduction of 3-HP from methanol. reached the highest 3-HP concentration (1.18 ± 0.03 g ­l−1), i.e., a 12% increase respect to PpCβ20 (1.05 ± 0.01 g Improving 3‑HP production by optimizing the flux ­l−1). Notably, the 3-HP yield on methanol was increased through the β‑alanine synthetic pathway by 11% with the additional copy of the panD gene Once the functionality of the biosynthetic pathway (strain PpCβ21) while this effect was much lower in the was confirmed by the presence of 3-HP, we decided to PpCβ11 strain (3%) (Fig. 3). These results indicate that improve the flux towards the production of this carbox- lower expression levels of the ydfG gene (i.e., under the ylic acid by overexpressing the key biosynthetic enzymes. control of ­PPDC1) seem to be beneficial in terms of 3-HP Previous studies with a S. cerevisiae strain expressing production. This suggested the presence of a bottle- the β-alanine pathway, demonstrated that the effect of neck at the final step of the pathway, since the ­YDFGEc multiple integrations of ­PANDTc was larger than that of requires NADPH to catalyze the reduction of the malonic Fig. 3 Global product yields ­( YP/S) calculated for the base strains constructed in this study. The key heterologous enzymes expressed in each strain are depicted in the left side of the graph. One colored arrow corresponds to one copy of the related expression cassette (blue for ­PANDTc, purple for ­BAPATBc, and orange for ­YDFGEc). Promoters selected to drive expression of every transcriptional unit are mentioned above the arrows. On the right side, the grey bars show the average product yield (g ­g−1), and the black dots represent the averaged endpoint ­OD600. The error bars show the standard deviation, and the asterisk sign indicates a significant difference in global 3-HP yields between the strains indicated in brackets (p-values < 0.05) Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 6 of 16 semialdehyde to 3-HP, which is not being supplied. We cultivating a representative clone from each strain in hypothesized that this perturbation of the NADPH shake flask cultures using BMM medium. The introduc- homeostasis could be further exacerbated when the tion of the β-alanine route led to a reduction of the µmax β-alanine flux into the pathway is increased (i.e., with the from 0.14 ­h−1 (reference CBS7435 strain) to 0.11 ­h−1 and insertion of an additional copy of the panD gene under 0.09 ­h−1 for PpCβ20 and PpCβ10 strains, respectively the very strong AOX1 promoter), resulting in a poten- (see Additional file 2). The β-alanine pathway consumes tial limitation of NADPH equivalents for anabolism/cell 1 NADPH per 3-HP, thereby reducing NADPH avail- growth. ability for anabolic purposes, which has been previously correlated to decreased biomass yields on methanol Engineering the redox balance by overexpressing [44]. Coherently, lower expression levels of the NADPH- the PseFDH(V9) enzyme variant from Pseudomonas sp. consuming ­YDFGEc enzyme (i.e., under the control of 101, a ­NADP+‑dependent formate dehydrogenase ­PPDC1, strain PpCβ20) of the β-alanine pathway proved To increase NADPH supply, the gene coding for a to be beneficial, both in terms of 3-HP production and mutated formate dehydrogenase from Pseudomonas cell growth. Moreover, this effect was more pronounced sp. 101, namely PseFDH(V9), was expressed in both with the overexpression of the ­PANDTc expression cas- PpCβ20 and PpCβ21 strains, obtaining the PpCβ20-P sette, as inferred from the µmax comparison between and PpCβ21-P strains, respectively. The enzyme vari- strains PpCβ21 and PpCβ11 (see Additional file 2). Intro- ant PseFDH(V9) shows the highest catalytic efficiency ducing a second panD gene copy into the PpCβ10 strain ­(kcat/KM) for ­NADP+ that has ever been reported for substantially decreased the µmax value from 0.09 ­h−1 to any FDH towards this cofactor, mainly attributed to a 0.06 ­h−1 (PpCβ11), whereas this effect was minimal in the very low K ­ MNADP = 0.026 mM. This causes PseFDH(V9) PpCβ20 strain, leading to a slight (not statistically signifi- to reduce N ­ ADP+ with barely any competition from cant, p-value = 0.28) reduction from 0.11 ­h−1 to 0.10 ­h−1 + ­NAD , even though the high concentration of this com- (PpCβ21). According to previous studies [15], fine-tun- peting substrate in the cytosol [43]. The gene encod- ing the panD gene dosage aiming at increasing the flux ing for the PseFDH(V9) enzyme was expressed under towards the β-alanine intermediate through aspartate the control of the strong and methanol inducible FDH1 decarboxylation may cause an impaired cell growth of promoter. The PpCβ20-P and PpCβ21-P strains were cul- K. phaffii, since aspartate takes part in many biological tivated along with their corresponding parental strains processes. The µmax value obtained for the PpCβ11 strain on 24-deep well plates containing BMM. A comparable was consistent with this phenomenon, albeit we did not endpoint ­OD600 between 2.4 and 2.6 was reached after observe this effect in the PpCβ21 strain. Further meta- 48 h of cultivation by the strains PpCβ20, PpCβ21 and bolic flux analyses should help elucidating the cause of PpCβ20-P. Conversely, cell growth was slightly lower the differences between the strains PpCβ11 and PpCβ21 for the PpCβ21-P strain ­ (DO600 = 2.2) (see Additional growth. In any case, PpCβ11 was discarded for further file 2). The final 3-HP titer was significantly reduced by bioreactor-scale cultivation experiments. about 11% in strain PpCβ20-P (1.12 ± 0.01 g l­−1) com- The introduction of the mutated formate dehydroge- pared to its parental strain PpCβ20 (1.24 ± 0.03 g ­l−1) nase FDH(V9) from Pseudomonas in the PpCβ20 strain (p-value < 0.05). The introduction of the PseFDH(V9) did not affect cell growth on methanol, i.e., both PpCβ20 enzyme in strain PpCβ21 resulted in a minor but also and PpCβ20-P strains showed the same µmax (0.11 ­h−1). significant decrease (p-value < 0.0003) of about 6% on the In contrast, the PpCβ21-P strain showed a lower µmax final product titer, from 1.29 ± 0.01 g ­l−1 to 1.22 ± 0.01 g (0.08 ­h−1) compared to its parental strain PpCβ21. Since ­l−1 (PpCβ21-P). However, there were no significant dif- both strains showed similar 3-HP production titers, fur- ferences (p-value = 0.38) in the 3-HP produced per O ­ D600 ther characterization in bioreactor-scale experiments unit between these strains. In summary, the expression was required. of this ­ NADP+-dependent formate dehydrogenase did not improve neither 3-HP production nor product yields Production of 3‑HP in fed‑batch cultivations using methanol as sole carbon and energy source in Initial fed-batch cultivations of the 3-HP-producing small-scale experiments. strains using a growth defined mineral medium typically used for recombinant protein production in K. phaffii Evaluating the impact of the β‑alanine pathway on cell [45], which is only supplemented with biotin, revealed growth kinetics that such medium did not support neither growth nor To determine the maximum specific growth rate (µmax) 3-HP production during the feed phase on methanol. We of the different strains obtained in this study, series hypothesized this was probably due to the special co-fac- of growth kinetics experiments were performed by tor requirements described for some of the enzymes of Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 7 of 16 the β-alanine pathway, specifically for the ­PANDTc (Uni- catalyzed by an aspartate aminotransferase (AAT2), reac- ProtKB: A7U8C7) and B ­ APATBc (UniProtKB: C2VE79) tion that takes place upstream of the β-alanine pathway enzymes, which use at least one pyridoxal phosphate as (see Fig. 2). 3-Hydroxyisobutyrate, pyruvate and acetate cofactor for its synthesis according to the UniProtKB were also detected in much lower amounts as well as database [46]. This coenzyme is derived from pyridox- traces of methanol. In contrast to 3-HP-producing K. ine (vitamin B ­ 6) and plays an important role in amino phaffii strains engineered with the malonyl-CoA pathway acid metabolism. Moreover, β-alanine is an intermedi- growing in fed-batch cultures using glycerol as C-source ate in pantothenic acid (vitamin B ­ 5) biosynthesis, which, [32], no arabitol was accumulated during the fermenta- in turn, is an integral part of the coenzyme A and acyl tion process. carrier protein, co-factors required for several enzymes At the end of the culture, the PpCβ21 strain pro- [37]. The redirection of this β-amino acid flux towards duced up to 21.4 ± 0.7 g ­l−1 of 3-HP, whereas 20.3 ± 0.6 g the production of 3-HP instead of vitamin B ­ 5, could be ­l−1 were achieved by the strain PpCβ20. In both cases an explanation for the observed growth impairment in the 3-HP generated during the batch phase of the cul- the bioreactor-scale cultivations. Indeed, the presence tures was slightly above 2 g l­−1 (Fig. 4). The addition of of both vitamins ­B6 and ­B5 in the yeast nitrogen based a second copy of the panD gene in PpCβ21 resulted in (YNB)-containing medium used in the small-scale culti- a 12% improvement of the final 3-HP titer in the small- vations is likely to be the reason why the 3-HP-producing scale screenings. This increase was reduced to 6% when strains grow normally in those conditions. Subsequently, both strains were tested in bioreactor fed-batch cultiva- bioreactor cultivations were carried out using a growth tions. The 3-HP yield on biomass (­YP/X) calculated for medium supplemented with a vitamin solution contain- the fed-batch phase was increased by 18% in PpCβ21 ing, among others: calcium pantothenate (­ B5), pyridoxine with respect to PpCβ20. Consequently, this resulted hydrochloride ­(B6), and niacin ­(B3), which is an impor- in a 10% improvement of the specific productivity ­(qp) tant precursor for the essential redox cofactors ­NAD+ (Table 1), and thus demonstrating that an enhanced sup- and ­NADP+. This vitamin solution is usually added to ply of β-alanine intermediate in the metabolic pathway Delft medium, commonly employed for bioreactor culti- had a positive effect on 3-HP production, coherent with vations of S. cerevisiae [47], allowing for sustained growth previous studies [37]. The expression of the PseFDH(V9) on methanol and 3-HP production. enzyme in PpCβ21-P strain allowed for a final 3-HP titer The best performing strains PpCβ20 and PpCβ21, along of 19.0 ± 0.5 g ­l−1, along with lower biomass concentra- with the strain expressing the mutated FDH(V9) enzyme tion levels (38.8 g ­l−1 DCW) (Fig. 4), in comparison with from Pseudomonas, PpCβ21-P, were further evaluated in the above-mentioned strains. This yielded a further 14% aerobic fed-batch reactors. After the initial glycerol batch and 10% increase in Y ­ P/X and ­qp, respectively, while the and methanol adaptation (transition) phases, the fed- product yield on methanol ­(YP/S) remained unchanged batch phase was started, where methanol was fed follow- for both PpCβ21-P and PpCβ21 (Table 1). Furthermore, ing a pre-programmed exponential feeding strategy for the specific carbon dioxide evolution rate (­qCO2) was sig- controlled growth rate at µ = 0.03 ­h−1. The feeding phase nificantly increased by 29% (p-value = 0.02) when com- was finalized before reaching a dry cell weight (DCW) of paring the PpCβ21-P to the PpCβ21 strain, whereas there 50 g ­l−1 of biomass, accounting for 39.3 h of the total pro- were no significant differences (p-value = 0.10) in the cess duration, with a total methanol concentration added ­qCO2 between strains PpCβ20 and PpCβ21 (Table 1 and into the culture of almost 125 g ­l−1. Neither methanol nor Additional file 4). Consistently, the PpCβ21-P showed by-products were detected by HPLC analysis over the the highest C­ O2 yield on methanol (­YCO2/S) among the course of fermentation. Nonetheless, exometabolome three strains (Table 1). Therefore, it is plausible that the profiling analyses by NMR of the supernatant samples introduction of the ectopic NADPH-regenerating reac- from the early, mid, and late methanol-feeding phase tion from formate causes a carbon flux redistribution revealed the presence of small amounts (mM range, through the methanol dissimilation pathway, that is, below HPLC detection limit) of ethanol as well as some generating more C ­ O2 while providing a higher supply of metabolites potentially derived from leucine (Leu) and NADPH. This hypothesis is consistent with the higher isoleucine (Ile) biosynthesis and degradation pathways ­YP/X and q ­ P values observed for the strain PpCβ21-P, as (or pyruvate fermentation) such as 2-oxoisocaproate 3-HP synthesis requires a supply of reduction equiva- and 3-methyl-2-oxopentanoate (see Additional file 3 for lents. In addition, the biomass yield (­YX/S) obtained for NMR spectra of the culture supernatants). Both com- the PpCβ21-P strain was reduced compared to the other pounds are involved in transamination reactions with strains. Since the total amount of methanol consumed by 2-oxoglutarate and l-glutamate, which, in turn, are also all strains was the same, this was also reflected in a sig- present in the oxaloacetate conversion into l-aspartate nificant increase (p-value = 0.036, see Additional file 4) Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 8 of 16 Table 1 Averaged value of key process parameters obtained for the methanol fed-batch phase using a pre-programmed µ of 0.03 ­h−1 PpCβ20 PpCβ21 PpCβ21-P QP ­(g3-HP ­l−1 ­h−1) 0.46 ± 0.01 0.48 ± 0.02 0.46 ± 0.01 YX/S ­(gDCW ­gMetOH−1) 0.28 ± 0.01 0.25 ± 0.01 0.21 ± 0.01 YP/S ­(g3-HP ­gMetOH−1) 0.14 ± 0.01 0.15 ± 0.01 0.15 ± 0.01 YP/X ­(g3-HP ­gDCW−1) 0.52 ± 0.03 0.61 ± 0.03 0.69 ± 0.03 YCO2/S ­(gCO2 ­gMetOH−1) 0.70 ± 0.02 0.73 ± 0.02 0.82 ± 0.03 qS ­(gMetOH ­gDCW−1 ­h−1) 0.103 ± 0.002 0.107 ± 0.002 0.124 ± 0.002 −1 −1 qP ­(mmol3-HP ­gDCW  ­h ) 0.166 ± 0.001 0.183 ± 0.001 0.201 ± 0.001 qCO2 ­(mmolCO2 ­gDCW−1 ­h−1) 1.65 ± 0.01 1.79 ± 0.01 2.30 ± 0.01 µ ­(h−1) 0.029 ± 0.001 0.027 ± 0.001 0.026 ± 0.001 Volumetric productivity ­(QP), biomass yield on methanol (­ YX/S), 3-HP yield on methanol ­( YP/S), 3-HP yield on biomass ­( YP/X), ­CO2 yield on methanol ­( YCO2/S), specific substrate consumption rate ­(qS), specific 3-HP production rate ­(qP), specific carbon dioxide evolution rate ­(qCO2), and experimentally measured mean specific growth rate (µ). Cultivations were performed in duplicate and biomass concentration analyses were performed in triplicate. ± indicates SD of the biological replicates heterologous PseFDH(V9) gene caused a carbon flux redistribution towards ­CO2 and 3-HP production to the detriment of biomass generation. The iMT1026 v3.0 genome-scale metabolic model of K. phaffii [42] was used to validate the experimen- tal 3-HP yields resulting from fed-batch cultivations. For the simulations, the experimental values obtained for the main macroscopic variables (i.e., q ­ S, ­qO2, ­qCO2, and µ) of each methanol-grown strain were set as con- straints, and q ­ 3-HP was maximized. Y­ P/S and Y ­ P/X were calculated for each strain using the predicted ­q3-HP val- ues (Table 2). Notably, the model showed reasonably accurate predictions, with overall deviations around 6–15% in most cases, even though the mean elemental biomass composition and the estimated non-growth associated maintenance energy (NGAME) parameter that were used as constraints to perform these simu- lations were taken from chemostat cultivations of a Fig. 4 Fed-batch phase profiles from the bioreactor-scale experiments performed with the 3-HP-producer strains a PpCβ20, b Table 2 Comparison of theoretical performance of iMT1026 v3.0 PpCβ21 and, c PpCβ21-P. Concentration of dry cell weight and 3-HP with experimental data from the methanol fed-batch cultivations are represented in the left-side y-axis. The total amount of methanol added, normalized by the actual volume of the reactor at every at µ = 0.03 ­h−1 time, is represented using the right-side y-axis. The cultivation PpCβ20 PpCβ21 PpCβ21-P profile shown for each strain corresponds to the average result of two independent cultivations. The error bars denote the standard Predicted ­YP/S 0.16 0.17 0.16 deviation for the duplicate Experimental ­YP/S 0.14 ± 0.01 0.15 ± 0.01 0.15 ± 0.01 Relative deviation (%) 6–19 6–18 0-12.5 Predicted ­YP/X 0.57 0.68 0.77 Experimental ­YP/X 0.52 ± 0.03 0.61 ± 0.03 0.69 ± 0.03 of the specific methanol consumption rate (­qs) of the Relative deviation (%) 3.5–14 6–15 6.5–14 PpCβ21-P strain in comparison to the PpCβ21 strain (Table 1). Overall, it appears that overexpression of the 3-HP yield on methanol ­( YP/S) ­(g3-HP ­gMetOH−1), 3-HP yield on biomass ­( YP/X) ­(g3-HP ­gDCW−1). ± indicates SD of the biological replicates Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 9 of 16 non-producing (wild-type phenotype) K. phaffii strain Future perspectives growing on methanol [42]. The use of condition-spe- Although the 3-HP production yields and productivities cific biomass composition equations and the recalibra- we demonstrate herein are promising, further strain and tion of energetic parameters (i.e., NGAME and GAME) process engineering is clearly needed to reach economi- through experimental data derived from chemostat cally feasible metrics for carboxylic acid production at cultivations should allow for a greater accuracy in the industrial scale, that is, a productivity and product yield prediction of physiological macroscopic parameters. not lower than 2.5 g ­l−1 ­h−1 and 0.5 g ­g−1, respectively Interestingly, the highest C-yield (C-mol of 3-HP [16]. Moreover, downstream operations may account per C-mol of substrate) calculated for the methanol for 20 to 60% of the entire production process costs. To feeding phase of fed-batch cultivations (0.16 ± 0.01 reduce the overall separation and purification expenses, Cmol ­Cmol−1) is comparable to that achieved by a S. a final product concentration in the range of 50–100 g ­l−1 cerevisiae strain overexpressing AAT2, PYC1, PYC2, would be required for process scale up [49]. ­BAPATBc, ­YDFGEc, and multiple copies of ­ PANDTc Despite the great potential of K. phaffii to produce growing on glucose (0.15 ± 0.0 Cmol C ­ mol−1, calcu- metabolites such as 3-HP from methanol, a substantial lated from data provided in Borodina et al. [37]). The fraction of the methanol consumed by the cells (typi- conversion of 3 molecules of methanol to one mol- cally, up to 50% or even higher) is channeled through ecule of 3-HP through the β-alanine pathway is an the dissimilatory pathway to ­CO2, limiting the carbon ATP-consuming process, whereas it has no net ATP flux entering the assimilation pathway to pyruvate [50], consumption when glucose is used as substrate (see an important central metabolite as well as the criti- Additional file 1: Table S1). Therefore, the Υ3−HP max cal precursor for 3-HP production. Thereby, metabolic that could be achieved by a K. phaffii strain overex- engineering strategies aiming to shift the carbon flux pressing the β-alanine pathway is significantly higher split ratio between the assimilatory and dissimilatory for cells growing on glucose (> 0.9 ­ g3-HP ­gGlucose−1) pathways may push carbon flux into 3-HP production than for methanol-grown cells (0.795 ­g3-HP ­gMetOH−1), and reduce the carbon loss caused by the direct oxida- based on simulations performed with the K. phaffi’s tion of methanol into C ­ O2 in K. phaffii. In addition, iMT1026 v3.0 genome-scale metabolic model [42]. pulling pyruvate flux into the β-alanine pathway, e.g., by Since there are no fundamental differences between enhancing the generation of pathway precursors such as K. phaffii’s and S. cerevisiae’s respiratory glucose oxaloacetate and aspartate, can also be effective strategy. metabolisms, the maximum theoretical 3-HP yields Full exploitation of the improved supply of NADPH by on glucose are very similar between these yeasts [37]. overexpression of the ­NADP+-dependent formate dehy- Overall, we demonstrate the potential of K. phaffii to drogenase gene described in this study probably requires achieve 3-HP yields on methanol comparable to those further fine-tuning of the ydfG gene expression levels for obtained from glucose, albeit methanol assimilation enhanced 3-HP production. Besides, final product con- being less energy efficient. Furthermore, the volumet- centration demands for industrial scale production will ric productivities calculated for the fed-batch phase in probably require engineering K. phaffii’s tolerance for this study are more than twofold higher than the one high 3-HP concentrations, as demonstrated for other obtained for the above-mentioned S. cerevisiae strain yeasts. (0.21 g ­l−1 ­h−1, calculated from data given by Borodina et al. [37]). Production of 3-HP from methanol has also Conclusions been reported by an engineered Methylobacterium In this study, we successfully engineered the synthetic extorquens AM1 expressing the reductive malonyl-CoA β-alanine to 3-HP pathway in K. phaffii for the first time. pathway [48], although the yields obtained were an Minimal pathway optimization led to a final 3-HP titer order of magnitude smaller than the ones reported in of 21.4 g l­−1, a 0.15 g g­ −1 3-HP yield on methanol, and this study. Furthermore, the volumetric productivities a volumetric productivity of 0.48 g l­−1 ­h−1 in controlled achieved herein are remarkably higher (about 3.2-fold) fed-batch cultures. Notably, no major by-products were to the productivities observed in the recently published observed in the fermentation broth. Overexpression of K. phaffii strain producing 3-HP from sole methanol a mutated formate dehydrogenase from Pseudomonas (0.15 g ­l−1 ­h−1, calculated from data reported by Wu sp. 101 aiming at increasing NADPH availability led to a et al. [10]). To our knowledge, the values obtained in higher carbon uptake rate in K. phaffii. Exometabolome this study are the highest volumetric productivities profiling points at the pyruvate node as a potential target reported so far for a 3-HP production process based on for future metabolic engineering. Although further strain the β-alanine pathway. and process optimization is required to make this system Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 10 of 16 industrially attractive, we laid the basis for development NJ, USA) for its expression in K. phaffii (GenBank acces- of yeast strains for 3-HP production from renewable C1 sion numbers: OR360384, OR360385, OR360386, and feedstocks. OR360387, respectively). The modular cloning vectors used in this work are Materials and methods described in Prielhofer et al. [5], and are available at Computational methods Addgene (Watertown, MA, USA) as GoldenPiCS kit. The genome-scale metabolic model of K. phaffii used The CRISPR/Cas9-mediated homology-directed genome for all the simulations performed in this study was the editing protocol performed in this study is described in iMT1026 v3.0 [42]. To evaluate the Υ3−HP max of the mal- Gassler et al. [6], and a ready-to-use plasmid kit named onyl-CoA and the β-alanine pathways, a set of experi- CRISPi can also be found at Addgene. Further descrip- mental and calculated growth parameters based on data tion of the molecular cloning protocols followed in this obtained from chemostat methanol cultivations were study can be found in the Additional file 5. defined as constraints. When growing on this substrate, Different genomic regions were selected as targets the growth associated and the non-growth associated to efficiently design each sgRNA, focusing on inter- maintenance energy (GAME and NGAME parameters) genic regions such as upstream of the promoters, where were calculated to be 166.77 mmol ATP ­gDCW−1 and 0.44 the integration of exogenous genes usually leads to sta- mmol ATP ­gDCW−1 ­h−1, respectively. The maximum spe- ble expression, and cell fitness remains unaffected [52]. cific substrate uptake rate ­(qMetOH max) was determined to Accordingly, the ­ BAPATBc expression cassette was be 7.82 mmol MetOH ­gDCW−1 ­h−1 at µmax = 0.10 ­h−1 [42]. targeted within the 100-bp upstream of TEF1-α pro- The reactions for 3-HP production and excretion using moter (PP7435_Chr1-1535), while the ­YDFGEc expres- either the malonyl-CoA or β-alanine pathways were sion unit was located within the first 100-bp of P ­ FLD1 added to the original model. The 3-HP yields (Υ3-HP) (PP7435_Chr3-0140), hence confirming this target simulating different oxygen availability conditions ­(qO2) sequence as a non-critical region of the FLD1 promoter. and growth rates (µ) were calculated using the COBRA The ­PANDTc expression cassette was integrated into the Toolbox v3.0 [51], and the colored phase planes were RGI2 locus [5], almost 800-bp upstream of a hypotheti- generated using the ‘pcolor’ function from Matlab R2021a cal protein (PP7435_Chr1-0725). The second copy of the (Mathworks, Inc., Natick, MA, USA). ­PANDTc donor DNA template was targeted within 50-bp For the experimental data validation, the main physi- upstream of the GAP promoter (PP7435_Chr2-0858) for ological macroscopic parameters derived from the meth- homology-directed DNA repair. Finally, the PseFDH(V9) anol fed-batch cultivations were used as constraints, expression cassette was targeted within 50-bp upstream together with the calculated NGAME value for metha- of the AOX1 promoter (PP7435_Chr4-0130). The poten- nol growth, and 3-HP production was maximized by tial sgRNA candidates binding target regions within the performing a parsimonious enzyme usage Flux Balance selected loci were assessed with CHOPCHOP, a widely Analysis (pFBA) using the COBRA Toolbox v3.0 software used web tool for CRISPR-based genome editing [53]. [51] in Matlab R2020b (Mathworks, Inc., Natick, MA, CHOPCHOP v2 seeks and scores suitable 20 nucleotide USA). target sites, called protospacers, followed by a 5′-NGG- 3′ motif (PAM sequence). The existence of four con- Plasmid and strain construction secutive thymine bases (TTTT) within the protospacer The parental strain K. phaffii CBS7435 (CBS, Centraal- sequence might cause Pol III termination. Also, GC bureau voor Schimmelcultures, Utrecht, NL) was used to content without PAM sequence should be 20–80% [54]. generate a battery of 3-HP-producer strains by means of Considering that the choice of the sgRNA causes locus- CRISPR/Cas9-mediated homology-directed integrations specific variations on targeting efficiency, and even some of the heterologous genes. Plasmids and strains used in sgRNAs might not be functional, two high-scored sgRNA this study are listed in Table 3 from this section. The het- targets for each genome locus were selected for the erologous genes encoding for an aspartate-1-decarboxy- experiments (Table 4), eventually using one to perform lase from T. castaneum ­(PANDTc, GenBank: ABU25221), each CRISPR/Cas9-mediated gene insertion into the K. a β-alanine-pyruvate aminotransferase from B. cereus phaffii genome. ­(BAPATBc, GenBank: EEL86940), a 3-hydroxypropi- Electrocompetent K. phaffii cells were prepared as onate dehydrogenase from E. coli ­( YDFGEc, GenBank: described elsewhere [55]. Donor DNA plasmids car- 12932746), and a mutated formate dehydrogenase from rying each expression cassette flanked by the corre- Pseudomonas sp. (strain 101) (FDH(V9)Pse(A199G/D222Q/ sponding homology arms were linearized using the S381V/C256A/H380K), UniprotKB: P33160), were codon-opti- FastDigest BpiI (IIs class) (Thermo Fisher Scientific, mized and synthesized by GenScript Biotech (Piscataway, Waltham, MA, USA). Transformation was performed Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 11 of 16 Table 3 List of plasmids and strains used in this study Plasmids/strains Modules/genotype References Plasmids BB1_12_pAOX1 PAOX1, ­KanR+ [5] BB1_12_pFDH1 PFDH1, ­KanR+ BB1_12_pPOR1 PPOR1, ­KanR+ BB1_12_pPDC1 PPDC1, ­KanR+ BB1_23 Ø, ­KanR+ BB1_34_RPS3tt RPS3tt, ­KanR+ BB1_34_ScCYC1tt ScCYC1tt, ­KanR+ BB1_34_TDH3tt TDH3tt, ­KanR+ BB2_BC Ø, ­AmpR+ pCC1-4k_TcPAND panD+, T. castaneum gene encoding for ­PANDTc enzyme GenScript Biotech pCC1-4k_BcBAPAT yhxA+, B. cereus gene encoding for ­BAPATBc BB1_23_EcYDFG ydfG+, E. coli str. K-12 gene encoding for ­YDFGEc BB1_23_PseFDH(V9) fdh+, Pseudomonas sp. (strain 101) gene encoding for ­FDHPse(V9)(A199G/D222Q/S381V/C256A/H380K) BB2_BC_TcPAND PAOX1-PANDTc-RPS3tt This work BB2_BC_BcBAPAT PFDH1-BAPATBc-ScCYC1tt BB2_BC_EcYDFG PPOR1-YDFGEc-TDH3tt PPDC1-YDFGEc-TDH3tt BB2_BC_PseFDH(V9) PFDH1-FDH(V9)Pse-TDH3tt BB3nK_ext_AD Ø, ­KanR+ [6] BB3nK_ext_AD_TcPAND 5′-HR_PAOX1-PANDTc-RPS3tt_3′-HR(RGI2) This work 5′-HR_PAOX1-PANDTc-RPS3tt_3′-HR(pGAPUP) BB3nK_ext_AD_BcBAPAT 5′-HR_PFDH1-BAPATBc-ScCYC1tt_3′-HR(pTEF1-αUP) BB3nK_ext_AD_EcYDFG 5′-HR_PPOR1-YDFGEc-TDH3tt_3′-HR(pFLD1UP) 5′-HR_PPDC1-YDFGEc-TDH3tt_3′-HR(pFLD1UP) BB3nK_ext_AD_PseFDH(V9) 5′-HR_PFDH1-FDH(V9)Pse-TDH3tt_3′-HR(pAOX1UP) BB3cN_pGAP_23*_pLAT1_Cas9 PGAP_Ø_PLAT1_Cas9, ­NrsR+ [6] BB3cK_pGAP_23*_pLAT1_Cas9 PGAP_Ø_PLAT1_Cas9, ­KanR+ BB3cH_pGAP_23*_pLAT1_Cas9 PGAP_Ø_PLAT1_Cas9, ­HygR+ BB3cK_pGAP_gRNA1(RGI2)_pLAT1_Cas9 PGAP_gRNA1(RGI2)_PLAT1_Cas9, ­KanR+ This work UP BB3cH_pGAP_gRNA1(pGAP )_pLAT1_Cas9 PGAP_gRNA1(PGAPUP)_PLAT1_Cas9, ­HygR+ BB3cK_pGAP_gRNA2(pTEF1-αUP)_pLAT1_Cas9 PGAP_gRNA2(PTEF1-αUP)_PLAT1_Cas9, ­KanR+ UP BB3cN_pGAP_gRNA1(pFLD1 )_pLAT1_Cas9 PGAP_gRNA1(PFLD1UP)_PLAT1_Cas9, ­NrsR+ UP BB3cN_pGAP_gRNA1(pAOX1 )_pLAT1_Cas9 PGAP_gRNA1(PAOX1UP)_PLAT1_Cas9, ­NrsR+ K. phaffii strains CBS7435 Wild-type CBS, Centraalbureau voor Schimmelcul‑ tures PpCβ10 PAOX1_PANDTc + ­PFDH1_BAPATBc + ­PPOR1_YDFGEc This work PpCβ20 PAOX1_PANDTc + ­PFDH1_BAPATBc + ­PPDC1_YDFGEc PpCβ11 2x(PAOX1_PANDTc) + ­PFDH1_BAPATBc + ­PPOR1_YDFGEc PpCβ21 2x(PAOX1_PANDTc) + ­PFDH1_BAPATBc + ­PPDC1_YDFGEc PpCβ20-P PpCβ20 + ­PFDH1_FDH(V9)Pse PpCβ21-P PpCβ21 + ­PFDH1_FDH(V9)Pse according to the CRISPR/Cas9-mediated homology- sequencing using a set of primers covering the 5′ and directed genome editing protocol for K. phaffii [6]. the 3′ homology arms (see Additional file 5: Table S2). Correct genomic insertions were verified by Sanger The sequenced clones were confirmed to be genetically Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 12 of 16 Table 4 Variable sgRNA targets custom-designed in this study Target name Locus Sequence (5′→3′) Expression cassette pTEF1-αUP_sgRNA1 PP7435_Chr1-1535 AAC​AAC​ACT​AAA​C TA​CCT​TG PFDH1_BAPATBc_ScCYC1tt pTEF1-αUP_sgRNA2 TTA​AGG​ATG​TGT​AGT​GTC​AA pFLD1UP_sgRNA1 PP7435_Chr3-0140 TGC​TAA​TGG​TAG​TTA​TCC​AA PPOR1-YDFGEc-TDH3tt pFLD1UP_sgRNA2 CTA​TAG​GAT​AAA​AAC​AGG​AG PPDC1-YDFGEc-TDH3tt RGI2_sgRNA1 PP7435_Chr1-0725 TCT​CAA​CGT​ATT​TAT​ATG​GT PAOX1-PANDTc-RPS3tt RGI2_sgRNA2 ATG​AAG​CCA​C TT​CAA​C TA​CG pGAPUP_sgRNA1 PP7435_Chr2-0858 ATC​GAT​AAT​AGT​CGC​ATG​TG PAOX1-PANDTc-RPS3tt pGAPUP_sgRNA2 CGT​TAG​GTC​AGT​GAT​GAC​AA pAOX1UP_sgRNA1 PP7435_Chr4-0130 ATT​GTG​AAA​TAG​ACG​CAG​AT PFDH1-FDH(V9)Pse-TDH3tt pAOX1UP_sgRNA2 GCA​GTC​GAT​C TC​AAA​AGC​AA The sgRNA target used for each genomic insertion is highlighted in bold identical. Recombinant yeast strains were grown at Shake‑flask cultures for small‑scale growth kinetics 30 °C in YPD agar plates (1% yeast extract, 2% pep- experiments tone, 2% dextrose and 15 g ­l−1 agar). The medium was Komagataella phaffii strains were inoculated from cryo- supplemented with the appropriate antibiotic when vials into 5 ml of YPG medium in 50 ml Falcon tubes. required: Nourseothricin (200 µg m ­ l−1 working con- Cultures were grown overnight at 30 °C and 180 rpm. centration for K. phaffii) from Dismed S.A. (Asturias, The overnight cultures were then diluted to an O ­ D600 Spain), and Geneticin (500 µg ­ml−1) or Hygromycin of 0.2 in 250 ml shake flasks containing 25 ml of BMM (200 µg ­ml−1) from InvivoGen (San Diego, CA, USA). and grown in an incubator shaker Multitron Standard for 12 h at 25 °C, 180 rpm and a relative humidity of 80%, to adapt cells to methanol metabolism and setting the Screening in deep‑well plates screening conditions. Afterwards, these cultures were Komagataella phaffii strains were grown overnight in used to inoculate 250 ml shake flasks containing 25 ml 24 deep-well plates containing YPG medium (1% yeast of fresh BMM at a starting O ­ D600 of 0.2 and grown dur- extract, 2% peptone and 1% v/v glycerol). Cultures ing 24 h at the same cultivation conditions. Samples were were placed on a platform with a slope of 20° in an taken every 3 h to measure the ­OD600 in duplicate with incubator shaker Multitron Standard (Infors HT, Bott- a 96-well microtiter plate using the SPECTROstar Nano mingen, Switzerland) with a 2.5 cm orbit, and grown absorbance microplate reader (BMG Labtech, Ortenberg, at 30 °C and 220 rpm. Afterwards, a 24 deep-well Germany). plate containing 2 ml of buffered minimal methanol medium (BMM; 100 mM potassium phosphate buffer pH 6, 1.34% yeast nitrogen base (YNB), 0.4 mg l­−1 bio- Bioreactor cultivations tin and 0.5% v/v pure methanol) was inoculated with Fed-batch cultures were performed in duplicate using a the overnight cultures, at a starting O ­ D600 of 0.1. The DASGIP Parallel Bioreactor System (Eppendorf, Ger- cultures were grown for 48 h at 25 °C. Moreover, the many). The starting volume of each 1.3 l reactor vessel relative humidity (rh) in the incubation chamber was was 500 ml. The batch medium consisted of 40 g l­−1 glyc- fixed to 80%. Agitation was set to 220 rpm. After 24 h, erol, 1.8 g l­−1 citric acid, 0.02 g l­−1 ­CaCl2·2H2O, 12.6 g l­−1 1% v/v pure methanol pulse was added to the cultures. ­(NH4)2HPO4, 0.5 g ­l−1 ­MgSO4·7H2O, 0.9 g ­l−1 KCl, 50 µl Three independent transformants from each strain antifoam Glanapon 2000 kz (Bussetti and Co GmbH, were tested to discard any biological variability. The Vienna, Austria), 0.4 mg ­l−1 biotin, 2 ml ­l−1 of vitamin parental strain K. phaffii CBS7435 was used as a nega- stock solution [47], and 5 ml ­l−1 of PTM1 trace salts [45]. tive control. All clones were inoculated in triplicate. At The pH was adjusted to 5 using 5 M HCl. The vitamins, the end of the culture, the final biomass concentration the biotin and the trace salts were filter-sterilized and of each deep-well was determined in duplicate with a added to the autoclaved medium after cooling down. 96-well microtiter plate using a Multiskan FC Micro- Inocula were prepared as described elsewhere [56]. plate Photometer (Thermo Fisher Scientific, Waltham, Reactors were inoculated at a starting O ­ D600 of 1. The MA, USA) to ensure all the cultures were grown up to pH was controlled at 5 throughout the culture using 15% a similar endpoint ­OD600. ammonia (only base addition was used). The temperature Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 13 of 16 was set to 28 °C, and the inlet gas was fed into the reac- DCW and ­OD600 measures of the initial fed-batch cul- tors at an aeration rate of 1 vvm (0.5 l m­ in−1). The dis- tivations. To quantify the metabolites, 2 ml of culture solved oxygen (DO) was set to 30%, automatically samples were centrifuged 5 min at 13,400 rpm using a controlled using the following cascade: (1) Increasing MiniSpin (Eppendorf, Germany). The supernatant was the stirring rate from 400 to 1000 rpm; (2) Raising the then filtered with a 0.2 μm pore size single-use syringe percentage of oxygen in the inlet gas, while maintain- filter (SLLGX13NK, Merck Millipore, CA, USA). The ing an aeration rate of 0.5 l m ­ in−1. For the fed-batch filtered supernatant was stored at − 20 °C until HPLC phase, the temperature was set to 25 °C. Pure methanol analysis for methanol and 3-HP quantification, and for (ρ = 792 g ­l−1) and feeding salts were added separately to exometabolome profiling by NMR. avoid precipitation. The feeding salts medium composi- tion was 0.35 g l­−1 ­CaCl2·2H2O, 10 g l­−1 KCl, 6.45 g l­−1 Analytical methods ­MgSO4·7H2O, 200 µl antifoam Glanapon 2000 kz, 1.2 mg Methanol and 3-HP were quantified using an HPLC ­l−1 biotin, 6 ml l­−1 of vitamin stock solution, and 15 ml Dionex Ultimate3000 (Dionex – Thermo Fischer Sci- ­l−1 of PTM1 trace salts. This media was prepared at 2× entific). The compounds were separated with an concentration since pure methanol was used. Two sepa- ionic exchange column ICSep ICE-COREGEL 87H3 rated solutions were prepared and autoclaved, one con- (Transgenomic, Omaha, NE, USA) using 6 mM sulphuric taining ­CaCl2·2H2O, and the other containing the rest of acid as mobile phase at a flow rate of 0.6 ml ­min−1. Both the components. After that, both solutions were mixed metabolites were quantified from the Refractive Index under sterile conditions at room temperature. The vita- (RI) spectrum. mins, the biotin and the trace salts were filter-sterilized For the NMR method, a Bruker Avance 600 MHz NMR and added to this final feeding solution. spectrometer operating at a proton (1H) frequency of After reaching a peak in the % of ­CO2 from the exhaust 600.13 MHz equipped with a triple-resonance Bruker gas, indicating the end of the batch phase, two pulses of TXI 5 mm room-temperature probe and an autosampler pure methanol (1 and 2 g ­l−1, respectively) were sequen- (Bruker Biospin, Rheinstetten, Germany) was utilized. tially added into the reactors. Once the methanol of the The probe temperature was maintained at 298.0 K in all last pulse was completely depleted, the cultures were experiments. Once centrifuged and filtered, each super- fed with a constant feeding rate ­(Fo) for 6 h to complete natant aliquot (400 µl) was mixed with a D ­ 2O sodium the transition phase. After that, the feeding medium was phosphate buffer (200 µl, 0.2 M, pH 7.4) containing an added to the bioreactors using a pre-programmed expo- internal standard (3-(trimethylsilyl)-[2,2,3,3-2H4]-propi- nential feeding strategy for controlled specific growth onic acid sodium salt (TSP), 1 mM) and transferred to rate described in the following equation: the NMR tube. All samples were analysed conducting µ[X(to )V (to )] [µ(t−to )] 1D 1H NMR experiments with presaturation of the resid- F (t) = e ual water signal applying the pulse sequence commonly YX/S So termed 1D NOESY-presat [57]. Data were collected into where µ was 0.03 ­h−1, the ­X0 was set to 22 g ­l−1 and the 32K data points during an acquisition time of 2.3 s using ­YX/S to 0.273 g ­g−1. a recycle delay of 2 s. Spectra were recorded in the time The reactors were sampled every 5.5 h to measure domain as interferograms (FID) across a spectral width ­OD600, biomass dry cell weight and supernatant metabo- of 7211 Hz and as the sum of 1024 transients. Each FID lites. The O ­ D600 measurements were performed in trip- was multiplied by an exponential apodisation function licate using a Lange DR 3900 spectrophotometer (Hach, equivalent to a 0.2 Hz line broadening, prior to Fourier Loveland, CO, USA). The Relative Standard Deviation transform. The frequency-domain spectra were manually (RSD) was below 5%. For the biomass DCW determi- phased, baseline corrected, and referenced to the TSP nation, 10 ml of distilled water with 9 g l­−1 NaCl were resonance at δH 0.00 ppm. The identification of metabo- used to wet the pre-weighted glass microfiber filters lites was carried out using the BMRB spectral database (APFF04700, Merck Millipore) before filtering 2 ml of [58] and the software Chenomix NMR Suite 8.5 (Cheno- culture for each triplicate. After that, the filters were mix Inc., Edmonton, Canada). washed using the same volume of the NaCl solution and dried for 24 h at 105 °C. Filters containing the dry bio- Process parameters determination and data consistency mass were weighted to calculate the DCW. This param- checking eter was quantified for three samples throughout the Recorded offline and online state variables during the fed-batch, in which the RSD was below 2.5%. For the rest fed-batch phase of the bioreactor-scale experiments, of the samples, the DCW was calculated by interpola- namely biomass X, substrate S, product P, flow rate tion using an equation from a linear regression between F, and initial volume V ­ 0 (see Additional files 4 and 6) Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 14 of 16 were used to calculate all derived variables. Details of Additional file 5. Molecular cloning materials and methods. Table S2. all equations derived from the mass balances used to List of primers used for amplification and cloning of DNA parts, sgRNAs calculate yields and rates in the fed-batch experiments construction, and sequence verification of integrations. can be found elsewhere [59, 60]. In short, intermedi- Additional file 6. Raw data of the online monitored standard process ate values of the offline variables (X, P), correspond- parameters. ing to the time when online measured variables were taken, were calculated (interpolated) after applica- Acknowledgements We thank Albert Fina and Eric Antón for their assistance to perform MATLAB tion of the ‘smoothingspline’ function from Matlab simulations. R2021a (Mathworks, Inc., Natick, MA, USA). Variables measured as concentrations in the liquid phase were Author contributions SAC, JA, and PF conceived and designed the research project. SAC performed referred to the total volume. Recorded state vari- all experiments and analyzed the data. MPT performed the NMR analyses and ables were used to calculate the accumulated variables metabolite identification. PF and JA contributed to data interpretation. SAC (total amounts, XV, SV, PV, V) during the fed-batch wrote the first draft of the manuscript and PF contributed to manuscript final version. All authors read and approved the final manuscript. phase. Linear regression was applied on the time evo- lution of accumulated variables. Regression slopes and Funding its confidence intervals were taken as averaged rates This work was supported by project ‘innoVative bIo-based chains for ­CO2 VALorisation as aDded-value organic acids’—VIVALDI (ID: 101000441) from of total state variables. Averaged rates of state vari- the Horizon 2020 Program of the European Commission; 2017-SGR-1462 and ables during the fed-batch phase were used to calcu- 2021-SGR-00143 from the Agència de Gestió d’Ajuts Universitaris i de Recerca late the derived variables (growth rate µ, q-rates, and (AGAUR) of the Catalan Government. SAC was supported with a FI scholarship (2022FI_B1_00173) from AGAUR. yields). The carbon and electron balances were veri- fied using the averaged measures prior to apply com- Availability of data and materials mon reconciliation techniques [61]. For the elemental The data that supports the findings of this study are included in this published article/Additional files. Further datasets used and/or analyzed during the cur‑ mass balance verification, the mean elemental biomass rent study are available from the corresponding author on reasonable request. composition ­C H1.88O0.63N0.20S0.004 with an ash con- tent of 5.9%, previously reported for cells growing on Declarations methanol at µ = 0.035 ­h−1 [42], was assumed in this study. Loss of methanol by stripping was considered Ethics approval and consent to participate Not applicable. almost negligible since C-balances were satisfied with less than 6% of deviation. Statistical χ2 consistency test Consent for publication based on h-index, applied on measured data balances Not applicable. as described elsewhere [62, 63], was passed with a con- Competing interests fidence level of 95%. Consequently, there was no evi- The authors declare no competing interests. dence for gross measurement errors. The method used for this purpose is also described in detail elsewhere Received: 31 July 2023 Accepted: 1 November 2023 [59]. Supplementary Information The online version contains supplementary material available at https://​doi.​ References org/​10.​1186/​s12934-​023-​02241-9. 1. Olah GA. Beyond oil and gas: the methanol economy. Angew Chem Int Ed. 2005;44:2636–9. Additional file 1. Diagrams and stoichiometric balances of the malonyl- 2. Whitaker WB, Sandoval NR, Bennett RK, Fast AG, Papoutsakis ET. Synthetic CoA and β-alanine pathways. Figure S1. Glucose, glycerol, and methanol methylotrophy: engineering the production of biofuels and chemicals metabolization to 3-HP via the two main metabolic pathways reported in based on the biology of aerobic methanol utilization. Curr Opin Biotech‑ yeast. Table S1. Stoichiometric analysis of the main metabolic pathways nol. 2015;33:165–75. towards 3-HP using different carbon sources. 3. Sreekrishna K, Kropp KE. Pichia pastoris. In: Wolf K, editor. Nonconven‑ tional yeasts in biotechnology. A handbook. Berlin: Springer; 1996. p. Additional file 2. Raw data from the 24-deep well plate cultivations and 203–53. cell growth kinetics experiments. 4. Wegner GH. Biochemical conversions by yeast fermentation at high cell Additional file 3. NMR spectra for PpCβ20, PpCβ21 and PpCβ21-P strains densities. Patent US4414329, USA; 1983. exometabolome. Figure S2. 1H NMR spectra of PpCβ20 strain supernatant 5. Prielhofer R, Barrero JJ, Steuer S, Gassler T, Zahrl R, Baumann K, et al. samples at different times of the methanol-feeding phase. Figure S3.1H GoldenPiCS: a golden gate-derived modular cloning system for applied NMR spectra of PpCβ21 strain supernatant samples at different times of synthetic biology in the yeast Pichia pastoris. BMC Syst Biol. 2017;11:123. the methanol-feeding phase. Figure S4. 1H NMR spectra of PpCβ21-P 6. Gassler T, Heistinger L, Mattanovich D, Gasser B, Prielhofer R. CRISPR/ strain supernatant samples at different times of the methanol-feeding Cas9-mediated homology-directed genome editing in Pichia pastoris. phase. In: Gasser B, Mattanovich D, editors. Recombinant protein production in Additional file 4. Raw and processed data obtained from the bioreactor- yeast. New York: Springer; 2019. p. 211–25. scale experiments. Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 15 of 16 7. De S, Mattanovich D, Ferrer P, Gasser B. Established tools and emerging 30. Suyama A, Higuchi Y, Urushihara M, Maeda Y, Takegawa K. Production trends for the production of recombinant proteins and metabolites in of 3-hydroxypropionic acid via the malonyl-CoA pathway using recom‑ Pichia pastoris. Essays Biochem. 2021;65:293–307. binant fission yeast strains. J Biosci Bioeng. 2017;124:392–9. 8. Yamada R, Ogura K, Kimoto Y, Ogino H. Toward the construction of a 31. Takayama S, Ozaki A, Konishi R, Otomo C, Kishida M, Hirata Y, et al. technology platform for chemicals production from methanol: d-lactic Enhancing 3-hydroxypropionic acid production in combination with acid production from methanol by an engineered yeast Pichia pastoris. sugar supply engineering by cell surface-display and metabolic engi‑ World J Microbiol Biotechnol. 2019;35:37. neering of Schizosaccharomyces pombe. Microb Cell Fact. 2018;17:176. 9. Guo F, Dai Z, Peng W, Zhang S, Zhou J, Ma J, et al. Metabolic engineering 32. Fina A, Brêda GC, Pérez-Trujillo M, Freire DMG, Almeida RV, Albiol J, et al. of Pichia pastoris for malic acid production from methanol. Biotechnol Benchmarking recombinant Pichia pastoris for 3-hydroxypropionic acid Bioeng. 2021;118:357–71. production from glycerol. Microb Biotechnol. 2021;14:1671–82. 10. Wu X, Cai P, Gao L, Li Y, Yao L, Zhou YJ. Efficient bioproduction of 33. Fina A, Heux S, Albiol J, Ferrer P. Combining metabolic engineering and 3-hydroxypropionic acid from methanol by a synthetic yeast cell factory. multiplexed screening methods for 3-hydroxyprionic acid production ACS Sustain Chem Eng. 2023;11:6445–53. in Pichia pastoris. Front Bioeng Biotechnol. 2022;10:942304. 11. Gao L, Cai M, Shen W, Xiao S, Zhou X, Zhang Y. Engineered fungal 34. Yu W, Cao X, Gao J, Zhou YJ. Overproduction of 3-hydroxypropionate in polyketide biosynthesis in Pichia pastoris: a potential excellent host for a super yeast chassis. Bioresour Technol. 2022;361:127690. polyketide production. Microb Cell Fact. 2013;12:77. 35. Song CW, Kim JW, Cho IJ, Lee SY. Metabolic engineering of Escherichia 12. Xue Y, Kong C, Shen W, Bai C, Ren Y, Zhou X, et al. Methylotrophic yeast coli for the production of 3-hydroxypropionic acid and malonic acid Pichia pastoris as a chassis organism for polyketide synthesis via the full through β-alanine route. ACS Synth Biol. 2016;5:1256–63. citrinin biosynthetic pathway. J Biotechnol. 2017;242:64–72. 36. Lacmata ST, Kuiate JR, Ding Y, Xian M, Liu H, Boudjeko T, et al. Enhanced 13. Liu Y, Tu X, Xu Q, Bai C, Kong C, Liu Q, et al. Engineered monoculture and poly(3-hydroxypropionate) production via β-alanine pathway in co-culture of methylotrophic yeast for de novo production of monacolin recombinant Escherichia coli. PLoS ONE. 2017;12:0173150. J and lovastatin from methanol. Metab Eng. 2018;45:189–99. 37. Borodina I, Kildegaard KR, Jensen NB, Blicher TH, Maury J, Sherstyk S, 14. Cai P, Wu X, Deng J, Gao L, Shen Y, Yao L, et al. Methanol biotransforma‑ et al. Establishing a synthetic pathway for high-level production of tion toward high-level production of fatty acid derivatives by engineering 3-hydroxypropionic acid in Saccharomyces cerevisiae via β-alanine. the industrial yeast Pichia pastoris. Proc Natl Acad Sci USA. 2022;119:29. Metab Eng. 2015;27:57–64. 15. Miao L, Li Y, Zhu T. Metabolic engineering of methylotrophic Pichia pasto- 38. Liao HH, Gokarn RR, Gort SJ, Jessen HJ, Selifonova OV. Production of ris for the production of β-alanine. Bioresour Bioprocess. 2021;8:89. 3-hydroxypropionic acid using beta-alanine/pyruvate aminotrans‑ 16. Werpy T, Petersen G. Top value added chemicals from biomass: Volume ferase. Patent WO118719; 2005. I — Results of screening for potential candidates from sugars and synthe‑ 39. Jessen H, Liao HH, Gort SJ, Selifonova OV. Beta-alanine/alpha-ketoglut‑ sis gas. Springfield (VA): US Department of Energy; 2004. Report No. DOE/ arate aminotransferase for 3-hydroxypropionic acid production. Patent GO-102004-1992. WO027742; 2008. 17. Bhagyashri P, Pratik M, Eswara P. Acrylic acid market by derivative type, 40. Jessen H, Rush B, Huryta J, Mastel B, Berry A, Yaver D et al. Composi‑ and end-user: global opportunity analysis and industry forecast, 2021– tions and methods for 3-hydroxypropionic acid production. Patent 2030. Allied Market Research; 2022. Report No. A00133. https://​www.​allie​ WO074818; 2012. dmark​etres​earch.​com/​acryl​ic-​acid-​market. Accessed 8 Sept 2022. 41. Lis AV, Schneider K, Weber J, Keasling JD, Jensen MK, Klein T. Exploring 18. Jiang X, Meng X, Xian M. Biosynthetic pathways for 3-hydroxypropionic small-scale chemostats to scale up microbial processes: 3-hydroxypro‑ acid production. Appl Microbiol Biotechnol. 2009;82:995–1003. pionic acid production in S. cerevisiae. Microb Cell Fact. 2019;18:50. 19. Maris AJAV, Konings WN, Dijken JPV, Pronk JT. Microbial export of lactic 42. Tomàs-Gamisans M, Ferrer P, Albiol J. Fine-tuning the P. pastoris and 3-hydroxypropanoic acid: implications for industrial fermentation iMT1026 genome-scale metabolic model for improved prediction processes. Metab Eng. 2004;6:245–55. of growth on methanol or glycerol as sole carbon sources. Microb 20. Kumar V, Ashok S, Park S. Recent advances in biological production of Biotechnol. 2018;11:224–37. 3-hydroxypropionic acid. Biotechnol Adv. 2013;31:945–61. 43. Calzadiaz-Ramirez L, Calvó-Tusell C, Stoffel GMM, Lindner SN, Osuna 21. Chu HS, Kim YS, Lee CM, Lee JH, Jung WS, Ahn JH, et al. Metabolic S, Erb TJ, et al. In vivo selection for formate dehydrogenases with high engineering of 3-hydroxypropionic acid biosynthesis in Escherichia coli. efficiency and specificity toward NADP+. ACS Catal. 2020;10:7512–25. Biotechnol Bioeng. 2015;112:356–64. 44. Jordà J, Rojas H, Carnicer M, Wahl A, Ferrer P, Albiol J. Quantitative 22. de Fouchécour F, Sánchez-Castañeda AK, Saulou-Bérion C, Spinnler HÉ. metabolomics and instationary 13C-metabolic flux analysis reveals Process engineering for microbial production of 3-hydroxypropionic acid. impact of recombinant protein production on trehalose and energy Biotechnol Adv. 2018;36:1207–22. metabolism in Pichia pastoris. Metabolites. 2014;4:281–99. 23. Rathnasingh C, Raj SM, Lee Y, Catherine C, Ashok S, Park S. Production of 45. Maurer M, Kühleitner M, Gasser B, Mattanovich D. Versatile modeling 3-hydroxypropionic acid via malonyl-CoA pathway using recombinant and optimization of fed batch processes for the production of secreted Escherichia coli strains. J Biotechnol. 2012;157:633–40. heterologous proteins with Pichia pastoris. Microb Cell Fact. 2006;5:37. 24. Liu C, Wang Q, Xian M, Ding Y, Zhao G. Dissection of malonyl-coenzyme a 46. Consortium TU. UniProt: the Universal Protein knowledgebase in 2023. reductase of Chloroflexus aurantiacus results in enzyme activity improve‑ Nucleic Acids Res. 2023;51:523–31. ment. PLoS ONE. 2013;8(9):e75554. 47. Jensen NB, Strucko T, Kildegaard KR, David F, Maury J, Mortensen 25. Liu C, Ding Y, Zhang R, Liu H, Xian M, Zhao G. Functional balance UH, et al. EasyClone: method for iterative chromosomal integra‑ between enzymes in malonyl-CoA pathway for 3-hydroxypropionate tion of multiple genes in Saccharomyces cerevisiae. FEMS Yeast Res. biosynthesis. Metab Eng. 2016;34:104–11. 2014;14:238–48. 26. Liu B, Xiang S, Zhao G, Wang B, Ma Y, Liu W, et al. Efficient production of 48. Yang YM, Chen WJ, Yang J, Zhou YM, Hu B, Zhang M, et al. Production 3-hydroxypropionate from fatty acids feedstock in Escherichia coli. Metab of 3-hydroxypropionic acid in engineered Methylobacterium extorquens Eng. 2019;51:121–30. AM1 and its reassimilation through a reductive route. Microb Cell Fact. 27. Chen Y, Bao J, Kim IK, Siewers V, Nielsen J. Coupled incremental precursor 2017;16:179. and co-factor supply improves 3-hydroxypropionic acid production in 49. Mazzoli R. Current progress in production of building-block organic Saccharomyces cerevisiae. Metab Eng. 2014;22:104–9. acids by consolidated bioprocessing of lignocellulose. Fermentation. 28. Kildegaard KR, Wang Z, Chen Y, Nielsen J, Borodina I. Production of 2021;7:248. 3-hydroxypropionic acid from glucose and xylose by metabolically engi‑ 50. Guo F, Qiao Y, Xin F, Zhang W, Jiang M. Bioconversion of C1 feedstocks neered Saccharomyces cerevisiae. Metab Eng Commun. 2015;2:132–6. for chemical production using Pichia pastoris. Trends Biotechnol. 29. Tong T, Tao Z, Chen X, Gao C, Liu H, Wang X, et al. A biosynthesis pathway 2023;41:1066–79. for 3-hydroxypropionic acid production in genetically engineered Sac- 51. Becker SA, Feist AM, Mo ML, Hannum G, Palsson B, Herrgard MJ. charomyces cerevisiae. Green Chem. 2021;23:4502–9. Quantitative prediction of cellular metabolism with constraint-based models: the COBRA toolbox. Nat Protoc. 2007;2:727–38. Àvila‑Cabré et al. Microbial Cell Factories (2023) 22:237 Page 16 of 16 52. Liu Q, Shi X, Song L, Liu H, Zhou X, Wang Q, et al. CRISPR-Cas9-mediated genomic multiloci integration in Pichia pastoris. Microb Cell Fact. 2019;18:144. 53. Labun K, Montague TG, Gagnon JA, Thyme SB, Valen E. CHOPCHOP v2: a web tool for the next generation of CRISPR genome engineering. Nucleic Acids Res. 2016;44:W272–6. 54. Yang Y, Liu G, Chen X, Liu M, Zhan C, Liu X, et al. High efficiency CRISPR/ Cas9 genome editing system with an eliminable episomal sgRNA plas‑ mid in Pichia pastoris. Enzyme Microb Technol. 2020;138:109556. 55. Sears IB, O’connor J, Rossanese OW, Glick BS. A versatile set of vectors for constitutive and regulated gene expression in Pichia pastoris. Yeast. 1998;14:783–90. 56. Garcia-Ortega X, Ferrer P, Montesinos JL, Valero F. Fed-batch operational strategies for recombinant fab production with Pichia pastoris using the constitutive GAP promoter. Biochem Eng J. 2013;79:172–81. 57. Nicholson JK, Foxall PJD, Spraul M, Farrant RD, Lindon JC. 750 MHz 1H and 1H-13C NMR spectroscopy of human blood plasma. Anal Chem. 1995;67:793–811. 58. Hoch JC, Baskaran K, Burr H, Chin J, Eghbalnia HR, Fujiwara T, et al. Biologi‑ cal magnetic resonance data bank. Nucleic Acids Res. 2023;51:D368–76. 59. Ponte X, Montesinos-Seguí JL, Valero F. Bioprocess efficiency in Rhizopus oryzae lipase production by Pichia pastoris under the control of PAOX1 is oxygen tension dependent. Process Biochem. 2016;51:1954–63. 60. Barrigón JM, Montesinos JL, Valero F. Searching the best operational strategies for Rhizopus oryzae lipase production in Pichia pastoris mut+ phenotype: methanol limited or methanol non-limited fed-batch cul‑ tures? Biochem Eng J. 2013;75:47–54. 61. Verheijen PJT. Data reconciliation and error detection. In: Smolke CD, editor. The metabolic pathway engineering handbook. Boca Raton: CRC Press; 2010. p. 81–13. 62. Noorman HJ, Rornein B, Ch M, Luyben KA, Heijnen JJ. Classification, error detection, and reconciliation of process information in complex biochemical systems. Biotechnol Bioeng. 1996;49:364–76. 63. van der Heijden RT, Heijnen JJ, Hellinga C, Romein B, Luyben KC. Linear constraint relations in biochemical reaction systems: I. Classification of the calculability and the balanceability of conversion rates. Biotechnol Bioeng. 1994;43:3–10. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions