ORIGINAL RESEARCH published: 26 January 2022 doi: 10.3389/fbioe.2022.818434 Innovative Bioprocess Strategies Combining Physiological Control and Strain Engineering of Pichia pastoris to Improve Recombinant Protein Production Arnau Gasset 1, Xavier Garcia-Ortega 1,2, Javier Garrigós-Martínez 1, Francisco Valero 1* and José Luis Montesinos-Seguí 1 1 Department of Chemical, Biological and Environmental Engineering, School of Engineering, Universitat Autònoma de Barcelona, Bellaterra, Spain, 2QuBi Lab, Department of Biosciences, Faculty of Sciences and Technology, Universitat de Vic-Universitat Central de Catalunya, Vic, Spain Edited by: Priscilla Filomena Fonseca Amaral, The combination of strain and bioprocess engineering strategies should be considered to Federal University of Rio de Janeiro, Brazil obtain the highest levels of recombinant protein production (RPP) while assuring product Reviewed by: quality and process reproducibility of heterologous products. In this work, two Mehmet İnan, complementary approaches were investigated to improve bioprocess efficiency based Akdeniz University, Turkey on the yeast P. pastoris. Firstly, the performance of two Candida rugosa lipase 1 producer Rochelle Aw, Imperial College London, clones with different gene dosage under the regulation of the constitutive PGAP were United Kingdom compared in chemostat cultures with different oxygen-limiting conditions. Secondly, *Correspondence: hypoxic conditions in carbon-limited fed-batch cultures were applied by means of a Francisco Valero Francisco.Valero@uab.cat physiological control based on the respiratory quotient (RQ). Stirring rate was selected to maintain RQ between 1.4 and 1.6, since it was found to be the most favorable in Specialty section: chemostat. As the major outcome, between 2-fold and 4-fold higher specific This article was submitted to production rate (qP) values were observed when comparing multicopy clone (MCC) Industrial Biotechnology, a section of the journal and single-copy clone (SCC), both in chemostat and fed-batch. Additionally, when Frontiers in Bioengineering and applying oxygen limitation, between 1.5-fold and 3-fold higher qP values were obtained Biotechnology compared with normoxic conditions. Thus, notable increases of up to 9-fold in the Received: 19 November 2021 Accepted: 10 January 2022 production rates were reached. Furthermore, transcriptional analysis of certain key Published: 26 January 2022 genes related to RPP and central carbon metabolism were performed. Results seem Citation: to indicate the presence of a limitation in post-transcriptional protein processing steps and Gasset A, Garcia-Ortega X, a possible transcription attenuation of the target gene in the strains with high gene dosage. Garrigós-Martínez J, Valero F and Montesinos-Seguí JL (2022) The entire approach, including both strain and bioprocess engineering, represents a Innovative Bioprocess Strategies relevant novelty involving physiological control in Pichia cell factory and is of crucial interest Combining Physiological Control and Strain Engineering of Pichia pastoris to in bioprocess optimization, boosting RPP, allowing bioproducts to be economically Improve Recombinant competitive in the market, and helping develop the bioeconomy. Protein Production. Front. Bioeng. Biotechnol. 10:818434. Keywords: GAP promoter, Pichia pastoris (Komagataella phaffii), physiological control, transcriptional analysis, doi: 10.3389/fbioe.2022.818434 recombinant gene dosage, respiratory quotient, hypoxia Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 1 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control 1 INTRODUCTION improved (García-Ortega et al., 2019; Chen et al., 2021). Some examples are the application of substrate starving The need for recombinant proteins has increased exponentially periods (Garcia-Ortega et al., 2016; Çalık et al., 2018) and over the last years, generating a billionaire market in which the implementation of oxygen limiting conditions (Baumann biotechnology plays a crucial role (Dewan, 2017). Most et al., 2008), which have shown a positive effect on RPP under biopharmaceuticals, industrial enzymes, and metabolites are PGAP regulation. produced via biological systems, also named “cell factories.” In Previous studies have shown that oxygen limitation in P. this context, bioprocess development combining strain and pastoris cultures growing over glucose as the sole carbon bioprocess engineering has emerged as a significant target for source causes an increase in metabolic fluxes throughout the its optimization from an industrial point of view. glycolytic and ethanol generation pathways, as well as a reduction Among other workhorse cell factories, the yeast Pichia in TCA fluxes, which signifies a shift from respirative to respiro- pastoris (renamed as Komagataella phaffii) (Kurtzman, 2005) fermentative metabolism (Baumann et al., 2010). It was reported has received increasing attention in the field of recombinant that not only an increase in metabolic fluxes, but also an protein production (RPP). It has relevant advantages over overexpression of certain genes was observed, both at the other alternatives: the capacity to grow to high cell densities mRNA and protein level. One of the overexpressed genes was in defined medium, the ability to perform post-translational TDH3 (glyceraldehyde-3-phosphate dehydrogenase), which has modifications, and to secrete the product into the culture PGAP as its natural promoter. broth. Furthermore, there are many tools available for its In later studies, it was demonstrated that different degrees genetic modification, including Crispr/Cas9 (Weninger et al., of oxygen limitation have a positive effect on the production 2016). All these advantageous features have been thoroughly of an antibody fragment expressed under PGAP regulation, described elsewhere (Çalık et al., 2015; Yang and Zhang, increasing the specific production rate (q P) by up to 3-fold 2018). Due to these attractive characteristics, P. pastoris with respect to the non-limiting conditions (Garcia-Ortega has already been used to produce more than 5,000 et al., 2017). In that study, it was proposed that, in order to recombinant proteins, including biopharmaceuticals and implement the desired oxygen-limiting conditions in industrial enzymes. different equipment and operational modes, a Traditionally, most of these processes in P. pastoris have been physiological and system-independent parameter should be based on the use of the AOX1 promoter (PAOX1), the strong and used as a transferable operating criterion. Two physiological methanol-inducible promoter of Alcohol oxidase 1, which was parameters, which are system-independent, were proposed the first to be reported. Alternatively, the constitutive with this aim: specific ethanol production rate (q EtOH) and glyceraldehyde-3-phosphate dehydrogenase promoter PGAP has respiratory quotient (RQ). also been widely selected for RPP over the last two decades, since Regarding the heterologous product, lipases are considered it avoids the use of methanol, thus reducing the oxygen and heat promising biocatalysts for industrial applications (López- removal requirements compared to methanol-based processes. Fernández et al., 2020). Among them, Candida rugosa lipase Moreover, methanol transport and storage at an industrial scale is (Crl) has been widely used by flavor, oil, and pharma industries so also considered a relevant drawback for large-scale processes far (Ken Ugo et al., 2017; Vanleeuw et al., 2019). Specifically, (García-Ortega et al., 2019). seven isoenzymes of C. rugosa lipase (Crl1-7) have been Regarding strain engineering of this yeast, different identified, Crl1 being the most abundant in commercial lipase approaches have been applied with the objective of increasing preparations (Ferrer et al., 2001). the production and productivity of the new clone (Ahmad et al., In the present work, the combined effect of both strain 2014). The increase in heterologous gene dosage (Cámara et al., engineering and bioprocess optimization strategies has been 2016; Sallada et al., 2019) and the overexpression of studied. Two distinct P. pastoris producer clones, with transcriptional factors (Jiao et al., 2018; Liu et al., 2020) are different gene copy number, producing Crl1 under PGAP successful examples of how genetic tools can provide new strains regulation have been tested to determine the effect of gene with much improved RPP performance. Novel approaches, dosage on RPP. In addition, oxygen-limiting conditions have including both promoter engineering and discovery, have been applied to test its effect on lipase production. Initially, recently been reported to have a high impact on RPP both clones were characterized in chemostat cultures to (Prielhofer et al., 2013; Portela et al., 2018; Ergün et al., 2019; compare their performance and to assess the combined Garrigós-Martínez et al., 2021). effect of the two aforementioned strategies. Subsequently, Concerning bioprocess optimization, the operational optimal conditions regarding oxygen limitation obtained in mode is a key factor in bioprocess design, batch or fed- chemostat have been applied to fed-batch cultures with both batch being the most established modes for production at clones using RQ as the controlled variable. Additionally, an industrial scale, while chemostat is more commonly used transcriptional analysis of key genes involved in cell for strain characterization and bioprocess design (Looser metabolism and RPP were performed to identify the et al., 2014; Kumar et al., 2020; Nieto-Taype et al., 2020a). impact of oxygen limitation on gene expression regulation. Additionally, the application of different types of cell stress The combination of both gene dosage and physiological has been demonstrated to have a significant positive effect on control represents a relevant novelty to further improve RPP, so the established process strategies can be remarkably the performance of Pichia-based RPP. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 2 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control 2 MATERIALS AND METHODS oxygen transfer rate to the culture media and, therefore, decrease the oxygen uptake rate. To reduce RQ, the opposite 2.1 Clone Construction and Gene Dosage action was performed. Determination Two selected recombinant clones of P. pastoris with one and five copies of the CRL1 gene under the regulation of the constitutive 2.4 Biomass Analysis and Consistency GAP promoter (PGAP) were used for this study. The clones were Check named as follows: Single-Copy Clone (SCC) and Multi-Copy Biomass concentration in terms of dry cell weight (DCW) was Clone (MCC). By using the α-mating factor of Saccharomyces determined as described previously (Cos et al., 2005). The cerevisiae the recombinant protein is secreted into the culture Relative Standard Deviation (RSD) was below 5%. broth. Both clones had already been used in a previous study, thus Biomass elemental composition was determined as described the clone construction and CRL1 copy number determination previously (Carnicer et al., 2009). For all chemostat and fed-batch were described previously (Nieto-Taype et al., 2020b). fermentations, the carbon recovery data was above 90%. The experimental data was further verified using previously described 2.2 Chemostat Cultivation standard data consistency and reconciliation procedures (Wang and Biological duplicates of chemostat cultivations were performed Stephanopoulos, 1983; van der Heijden et al., 1994; Verheijen, 2009), in a 2 L Biostat B fermenter (Braun Biotech, Melsungen, applying the constraint that carbon and electron conservation Germany) with a working volume of 1 L at a dilution rate relations were satisfied. In all cases, a 95% confidence level was (D) of 0.10 h−1 (Garcia-Ortega et al., 2016). The batch medium reached in the statistical consistency test, so it was considered that no contained per liter of distilled water: citric acid 1.8 g, glycerol significant measurement errors were made. 40 g, (NH4)2HPO4 12.6 g, MgSO4·7H2O 0.5 g, KCl 0.9 g, CaCl2·2H2O 0.02 g, antifoam Glanapon 2000kz (Bussetti & Co., GmbH, Vienna, Austria) 0.05 ml, biotin 0.4 mg, and 2.5 Carbon Source and by-Product 4.6 ml PTM1 trace salts solution. The PTM1 solution Quantification contained per liter: CuSO4·5H2O 6.0 g, NaI 0.08 g, Glycerol, glucose, as well as the potential fermentation by- MnSO4·H2O 3.36 g, Na2MoO4·2H2O 0.2 g, H3BO3 0.02 g, products (ethanol, arabitol, and succinate) were determined CoCl2 0.82 g, ZnCl2 20.0 g, FeSO4·7H2O 65.0 g, and 5.0 ml via HPLC. The column and software used for this purpose are H2SO4 (95%–98%). The chemostat medium contained per described elsewhere (Jordà et al., 2014). RSD was below 1%. liter of distilled water: citric acid 0.9 g, glucose 50 g, (NH4)2HPO4 4.35 g, MgSO4·7H2O 0.65 g, KCl 1.7 g, 2.6 Inlet- and Off-Gas Analyses CaCl2·2H2O 0.01 g, antifoam Glanapon 2000kz 0.2 ml, A BlueInOne FERM gas analyzer (BlueSens, Herten, Germany) was biotin 1.2 mg, and 15 ml PTM1 trace salts solution. In order used with both chemostat and fed-batch cultures. CO2 and O2 molar to reduce the oxygen supply without altering the mixing fractions and absolute humidity were monitored and recorded online conditions of the process, different mixtures of air and N2, from exhaust gas, and intermittently measured from inlet gas. The gas maintaining a constant aeration flowrate of 0.8 vvm, were analyzer was re-calibrated every steady-state in chemostats and every implemented. Temperature was kept at 25°C and pH was fed-batch in order to assure an accurate measurement. The data maintained at 6.0 with NH4OH 15% v/v. For each condition obtained were used to calculate the respirometric parameters: oxygen tested, samples were taken and analyzed at 3, 4, and 5 residence uptake rate (OUR), carbon dioxide evolution rate (CER), their times (τ). corresponding specific rates (qO2 and qCO2) and respiratory quotient (RQ). RSD was below 5%. 2.3 Fed-Batch Cultivation Fed-batch cultures were performed in duplicates using a 5 L 2.7 Lipolytic Activity Determination Biostat B fermenter (Sartorius Stedim, Goettingen, Germany). Crl1 enzymatic activity was analyzed using a p-nitrophenyl butyrate The specific growth rate (μ) was also fixed at 0.10 h−1 by applying (pNPB)-based assay as previously described (Garrigós-Martínez et al., an exponential pre-programmed feeding profile. The batch 2019). One activity unit is defined as the amount of enzyme needed to medium composition was the same as for chemostat obtain 1 µmol of product per min under assay conditions. RSD was cultivation. The fed-batch medium contained per liter of below 1%. distilled water: glucose 400 g, MgSO4·7H2O 6.45 g, KCl 10 g, CaCl2·2H2O 0.35 g, antifoam Glanapon 2000kz 0.2 ml, biotin 2.8 Transcriptional Analysis 0.2 mg, and 1.6 ml PTM1 trace salts solution. Again, Samples of 1 ml were taken from both chemostat and fed-batch temperature was also controlled at 25°C and pH at 6.0 with cultures, mixed with 0.5 ml of a 95% ethanol:5% phenol mixture NH4OH 15% v/v. The procedure is described in detail elsewhere and centrifuged at maximum speed for 2 min at 4°C. Samples (Garcia-Ortega et al., 2013). were then stored at −80°C until RNA extraction. Target degree of oxygen limitation was reached by For RNA extraction, samples were first mixed with 300 μL of maintaining the RQ at the corresponding set-point value. To glass beads (425–600 μm, Sigma-Aldrich, St. Louis, MO, do so, the agitation was modified following heuristic rules: to United States) and vortexed with TissueLyser (Qiagen, Hilden, increase RQ, agitation was reduced in order to decrease the Germany). RNA extraction was then carried out with the SV Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 3 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control FIGURE 1 | Main physiological and production parameters for continuous cultures at different molar fractions of oxygen in the inlet gas for single-copy clone (SCC, filled symbols) and multicopy clone (MCC, void symbols). (A) Specific substrate consumption rate, qS (▼); overall biomass-to-substrate yield, YX/S (C); specific ethanol production rate, qEtOH (∎); and specific Crl1 production rate, qP (▲). (B) Respiratory quotient, RQ (◆); specific oxygen consumption rate, qO2 (▲); and specific carbon dioxide production rate, qCO2 (▼). Error bars represent the SD of biological replicates. Total RNA Isolation System kit (Promega, Madison, WI, 21% to 8% of oxygen in the inlet gas. A previous study established a United States) following manufacturer’s instructions. dilution rate of D = μ = 0.10 h−1 to maximize qP (Nieto-Taype et al., cDNA synthesis and determination of transcriptional levels 2020b), therefore this D = μ was selected for testing all conditions. were performed as reported previously (Garrigós-Martínez et al., The variation of different key parameters (specific substrate 2019; Nieto-Taype et al., 2020b), with the only exception that in consumption rate, qS; biomass-to-substrate yield, YX/S; specific this case the β-actin gene (ACT1) was selected as the house- ethanol production rate, qEtOH; and qP) at different oxygen supply keeping gene, as in earlier transcriptional analyses under hypoxic levels can be observed in Figure 1A for both SCC and MCC. The conditions (Adelantado et al., 2017). numeric values of these parameters are also shown in Supplementary Table S1. Generally, the behavior of both clones can be seen to be rather 3 RESULTS AND DISCUSSION similar, except for qP, so it can be stated that having a higher gene 3.1 Characterization of Hypoxia and Gene dosage does not significantly affect the yeast’s physiology except for Crl1 production. Dosage Effects on Recombinant Crl1 In addition, the results show that the increase in qP between Production and Cell Physiology Using both clones is not proportional to the increase in gene dosage, Chemostat Cultures since MCC has between 2- and 3-fold higher qP values than SCC 3.1.1 Effect of Hypoxia and Gene Dosage on Substrate while the gene copy number is 5 versus 1. Similar effects were also Consumption and Crl1 Production reported previously (Macauley-Patrick et al., 2005; Aw and To evaluate the combined effect of gene dosage and oxygen Polizzi, 2013; Cámara et al., 2016). limitation on P. pastoris physiology and Crl1 production, a set Regarding oxygen limitation, no physiological changes are of chemostat cultures was performed using SCC and MCC and observed within the range of 21%–14% of oxygen in the inlet gas. applying seven different oxygen supply conditions, ranging from As of 12% and lower, oxygen availability becomes a crucial factor. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 4 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control This implies that cells cannot completely oxidize the substrate, and so less energy is obtained from glucose, which causes a decrease in overall biomass-to-substrate yield (YX/S). Then, since the same amount of substrate is consumed and less biomass is generated, the specific substrate consumption rate (qS) increases. This fact is confirmed by the absence of glucose in the culture medium. At this point, around 12% of oxygen, there is a shift from respirative to respiro-fermentative metabolism (Baumann et al., 2010; Garcia-Ortega et al., 2017), confirmed by the accumulation of fermentative by-products such as ethanol, arabinitol, and succinic acid, which were already observed in similar glucose- based hypoxic cultures (Carnicer et al., 2009; Baumann et al., 2010; Adelantado et al., 2017; Garcia-Ortega et al., 2017). The specific ethanol production rate (qEtOH = FIGURE 2 | Relationship between specific Crl1 production rate (qP) and respiratory quotient (RQ) for single-copy clone (SCC, filled symbols) and 0.0210 gEtOH·gDCW−1·h−1) is comparable to that observed in multicopy clone (MCC, void symbols) obtained in chemostat cultures. Error similar hypoxic chemostat cultures (qEtOH = bars represent the SD of biological replicates. 0.0189 gEtOH·gDCW−1·h−1) with the same D = μ = 0.10 h−1, and the same hypoxic level (RQ = 1.4 ± 0.1) expressing a Fab antibody fragment (Adelantado et al., 2017). It has been described on RQ values, which present a linear increase from 1.1 under that high ethanol concentrations can result in yeast growth normoxic conditions to 1.6 with the most severe hypoxic inhibition (Ergün et al., 2019; Wehbe et al., 2020). However, conditions, indicating a direct correlation between RQ and the the concentrations reached in these fermentations did not attain degree of oxygen limitation. these inhibitory levels. Although the molar fraction of oxygen in the inlet gas is an As the oxygen molar fraction in the inlet gas decreases, easy-to-modify parameter, it cannot be generically correlated hypoxia increases, but no glucose accumulation is observed with hypoxia and the effect that it generates. The reason being until the oxygen molar fraction in the inlet gas drops below that oxygen availability also relies on other factors: gas flow 8%, where bioreactor washout could be observed. This indicates rate, stirring rate, temperature of the culture broth, and the that, under such conditions, critical dilution (DC) < 0.10 h−1 configuration of the agitation/aeration system. On the other because μmax decreases substantially. hand, RQ is a parameter that provides very valuable Regarding Crl1 production, under hypoxic conditions information regarding the degree of oxygen limitation that around a 5-fold and 2.5-fold increase in qP is observed for affects the biomass. Moreover, it does not depend on the SCC and MCC, respectively, compared with normoxic culture system, but solely and exclusively on the conditions. Strikingly, it is worth mentioning that qP physiological state of the cells. Thus, RQ can be used as a reaches rather similar values under all normoxic conditions transferable operating criterion to apply the same hypoxic level and under all hypoxic conditions, regardless of the oxygen to a different system. limitation level. The highest variability in terms of qP is qEtOH could be also considered a reporting parameter of the observed in the transition between normoxic and hypoxic hypoxic conditions, but its on-line determination would conditions (around 12% of oxygen molar fraction in the represent additional complexity because the on-line inlet gas) as previously described (Garcia-Ortega et al., 2017). determination of biomass and ethanol concentrations would be necessary. In contrast, RQ control is expected to be less 3.1.2 Effect of Hypoxia and Gene Dosage on complex, being a parameter commonly determined on-line Respirometric Parameters from off-gas analysis. In Figure 1B, the variation in respirometric parameters In Figure 2, the relationships between qP and RQ for SCC and (specific oxygen consumption rate, qO2; specific carbon MCC are presented. In both cases, a set of grouped dots with low dioxide evolution rate, qCO2; and RQ) can be observed. The qP and RQ values is observed, which correspond to normoxic value of these parameters under normoxic conditions is very conditions. When RQ reaches values around 1.2, a sudden rise in similar to those reported in previous glucose-based cultures qP can be observed, which becomes saturated when RQ > 1.3. (Nieto-Taype et al., 2020b; Zavec et al., 2020). As observed in From this point and above, qP remains rather constant regardless the previous section, both clones performed similarly, of the RQ value. regardless of the gene dosage. Considering both strategies applied, namely gene dosage From 21% to 14% of oxygen in the inlet gas, the respirometric increase and oxygen limitation, an improvement of about one behavior of both clones is constant, corresponding to the order of magnitude for qP is achieved comparing SCC in normoxic state. When the oxygen molar fraction is reduced normoxia (qP = 106 Crl1 UA·gDCW−1·h−1) with MCC in below 12%, the biomass has less oxygen available, so qO2 hypoxia (qP = 934 Crl1 UA·gDCW−1·h−1). decreases and dissolved oxygen is around 0% (air saturation). Once the combined effect of gene dosage and oxygen This, combined with a slight increase in qCO2, has a direct impact limitation was characterized, the next step was to transfer the Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 5 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control bioprocess to a fed-batch mode, which represents a novel reduction of more than 50% (data not shown). This ethanol approach, as yet untested. inhibitory effect was not observed within the RQ range of 1.2–1.6. Consequently, an RQ set-point of 1.4 was selected. Thus, possible minor deviations of RQ due to the non-automatization of the RQ 3.2 Fed-Batch Culture Strategy to Improve control are expected to have a smaller effect on qP, since it is quite Bioprocess Efficiency constant within this RQ range, as observed in Figure 2. 3.2.1 Agitation Rate as Manipulated Variable to Achieve Target RQ 3.2.2 Crl1 Production in Fed-Batch Mode A set of fed-batch cultures were performed with a feeding strategy The time evolution of biomass and ethanol concentrations, Crl1 consisting of a pre-programmed exponential profile with μ = titer, and RQ throughout the feeding phase, both under normoxic 0.10 h−1, mimicking the carbon-limited growth applied in [dissolved oxygen higher than 30% (air saturation), RQ ≈ 1.1] and chemostat. hypoxic conditions [dissolved oxygen ≈0% (air saturation), RQ Applying a constant level of oxygen limitation in a dynamic set-point = 1.4] is shown in Figure 3. Accordingly, in Table 1, the culture poses a challenge, since the control action over RQ will not averaged values of related key process parameters are presented. be constant. If RQ is to be maintained at a certain value and As observed in chemostat cultures, both clones (SCC and biomass grows exponentially, oxygen supply to the culture should MCC) exhibit similar behavior, except for Crl1 production. also be increased exponentially. Biomass production decreases and Crl1 titers increase when As observed in chemostat cultures, the variation of the oxygen applying hypoxia. Besides, YX/S decreases by about 20% and qS fraction in the inlet gas is an easy-to-modify parameter by means increases accordingly compared with normoxic conditions. of mixing nitrogen with air. However, at the industrial scale, the Ethanol accumulation up to 10 g/L is also detected under use of pure gases involves associated costs that should be avoided hypoxic conditions, although no growth inhibition is observed. if possible. Instead, agitation is also an easy-to-modify parameter Furthermore, arabitol (up to 7 g/L) and succinic acid (up to 1 g/L) and is directly related to oxygen supply, since it notably affects the are also detected in the culture broth. volumetric oxygen transfer coefficient (kLa). It can be stated that RQ has been satisfactorily controlled Thus, after having checked that the modification of both the within the desired range, being around 1.1 in normoxic oxygen molar fraction in the inlet gas and the stirring rate can be fermentations and between 1.2 and 1.4 in hypoxic ones. successfully used to maintain RQ within a particular range, it was However, small peaks in the RQ profile are detected, decided to implement an RQ control based on agitation rate corresponding to manual control actions on the agitation variation. This entails a new strategy, different from that rate carried out to regulate RQ (Supplementary Figure S1). implemented in chemostat cultures, which allows a saving in As seen in Table 1, qO2 decreases by about 15% under energy and pure gases, and, a priori, seems easier to implement in hypoxic conditions, while qCO 2 has the same value an industrial process, whenever proper mixing can be guaranteed regardless of the oxygenation conditions. No significant and no heterogeneities in the bioreactor are produced. differences were observed between both clones. The entire strategy applied in this work aims to be more robust With regards to Crl1 production, qP increases considerably and reproducible than that reported in previous studies, which with oxygen limitation. As observed in chemostat, the qP increase was based on a start-and-stop substrate feeding that only targeted for SCC (2-fold) is higher than for MCC (1.5-fold) when applying the avoidance of high ethanol accumulation, considered to be hypoxia. inhibitory, without taking into account other parameters than can From the results obtained in chemostat cultures, fed-batch affect qP and productivity, such as µ and degree of oxygen data seem to indicate that, in this operational mode the limitation (Baumann et al., 2008). application of oxygen limitation is less effective than in As seen in Figure 2, the maximum qP values were reached at chemostat, since higher Crl1 titers were expected when RQ > 1.2. Even so, RQ values that are too high also cause an implementing hypoxic conditions to fed-batch cultures, important ethanol production and a significant YX/S decrease. especially with MCC. This could be explained by the Consequently, in a discontinuous culture, ethanol accumulates in combination of some factors, among them: firstly, oxygen the culture broth and can reach inhibitory concentrations (Potvin limitation causes metabolic stress, forcing cell metabolism et al., 2016; Ergün et al., 2019; Wehbe et al., 2020). On the other towards the fermentative branch, generating a lower quantity hand, a significant YX/S decrease entails a remarkably lower of energetic molecules such as ATP (Carnicer et al., 2009; biomass production, since in a carbon-limited fed-batch, the Baumann et al., 2010); secondly, cell aging caused by the time- open-loop feeding profile is designed considering a constant effect related to fed-batch bioprocesses, which has also been YX/S. If YX/S decreases during the feeding phase due to a described to affect RPP (Curvers et al., 2001; Cos et al., 2006), highly fermentative metabolism, so will μ. This, in turn, might probably amplifies the metabolic stress, which does not occur in cause a decrease in the specific production rate and productivity chemostats. Finally, saturation of secretion capacity has been also of the bioprocess. reported to be a possible bottleneck for high productive systems A preliminary fed-batch, with RQ controlled above 1.8, was that could be limiting the expected Crl1 titers (Ahmad et al., 2014; performed to check the possible ethanol inhibition effect. Ethanol Garcia-Ortega et al., 2016). accumulation achieved inhibitory concentrations (>30 g/L) and Crl1 production is presented in Figure 4, where caused a progressive halt in biomass growth, as well as a YX/S total Crl1 activity time profiles are plotted. SCC has a 70% Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 6 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control FIGURE 3 | Time evolution of respiratory quotient (◆), biomass (C) and ethanol (∎) concentrations, and Crl1 activity (▲) in fed-batch cultures under different oxygen transfer conditions: normoxia (filled symbols) and hypoxia (void symbols). (A) Single-copy clone (SCC). (B) Multicopy clone (MCC). Error bars represent the SD of biological replicates. TABLE 1 | Averaged value of key process parameters obtained in fed-batch fermentations. Specific substrate consumption rate (qS), biomass-to-substrate yield (YX/S), specific ethanol production rate (qEtOH), specific oxygen consumption rate (qO2), specific carbon dioxide evolution rate (qCO2), respiratory quotient (RQ) and specific Crl1 production rate (qP). ± indicate SD of the biological replicates. Single-copy clone Multicopy clone O2 supply condition Normoxia Hypoxia Normoxia Hypoxia qS (g S g DCW−1 h−1) 0.18 ± 0.01 0.21 ± 0.01 0.18 ± 0.01 0.22 ± 0.01 YX/S (g DCW·g X−1) 0.59 ± 0.01 0.47 ± 0.04 0.58 ± 0.01 0.50 ± 0.01 qEtOH (g EtOH g DCW−1 h−1) n.d. 0.03 ± 0.01 n.d. 0.03 ± 0.01 qO2 (mmols O2 g DCW−1 h−1) 1.93 ± 0.07 1.67 ± 0.04 2.02 ± 0.04 1.78 ± 0.24 qCO2 (mmols CO2 g DCW−1 h−1) 2.13 ± 0.08 2.30 ± 0.10 2.23 ± 0.05 2.42 ± 0.19 RQ 1.11 ± 0.01 1.38 ± 0.09 1.10 ± 0.01 1.37 ± 0.07 qP (AU g DCW−1 h−1) 164 ± 8 329 ± 27 696 ± 7 1,029 ± 45 P·V (AU) (6.67 ± 0.45) × 105 (1.07 ± 0.17) × 106 (2.82 ± 0.10) × 106 (3.62 ± 0.12) × 106 higher Crl1 production under hypoxic rather than This represents an overall 6.3-fold increase in terms of q P, normoxic conditions, whereas the increase for MCC is considering both increased gene dosage and only 20%. On the other hand, MCC presents an oxygen limitation, lower than that achieved in chemostat approximately 4-fold higher Crl1 production than SCC. cultures. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 7 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control described as an UPR-related transcription factor and an UPR- related chaperone, respectively (Guerfal et al., 2010; Raschmanová et al., 2019). The transcription levels of the genes of interest have been studied in accordance with the protocol described in Section 2.8. The results of the relative transcription levels (RTL) for CRL1, TDH3, and PGK1, combined with the averaged qP values, are shown via bar graphs in Figure 5. In order to facilitate a comprehensive comparison of both chemostat and fed-batch results, these are grouped into normoxic (1.1 < RQ < 1.2) and hypoxic results (RQ > 1.2) as the RTL of all the genes plotted presented no significant differences between any condition for each of the two groups (normoxic and hypoxic), as observed in Supplementary Table S2. No clear trends could be observed for HAC1 and KAR2 RTLs with respect to neither RQ nor CRL1 gene dosage, so they are excluded from the graphs and the discussion FIGURE 4 | Total Crl1 activity for SCC (▲) and MCC (▼) obtained under normoxic (filled symbols) and hypoxic (void symbols) conditions. Error bars (results presented in the Supplementary Table S2). Previous represent the SD of biological replicates. studies, in which the same two Crl1-producer clones were studied, showed that these genes remained unregulated under different specific growth rate (µ) conditions (Nieto-Taype et al., 2020b). 3.3 Impact of Hypoxia and Gene Dosage at Regarding the glycolytic genes selected, TDH3 and PGK1, as the Transcriptional Level expected, there are no significant differences with respect to CRL1 3.3.1 Consistency Between Physiological Parameters gene dosage in terms of transcript levels. Interestingly, the and Gene Expression Levels transcription levels of the glycolytic genes analyzed are Transcriptional analysis of key genes was performed with the aim boosted under oxygen limitation conditions. Accordingly, it of identifying potential correlations between gene regulation and could be stated that oxygen-limiting conditions have a protein production under hypoxic conditions. significant effect on the upregulation of glycolysis-related gene The genes selected were the recombinant CRL1 gene, which expression. Additionally, in previous studies using the same allows the expression of the target product of the bioprocess, the clones but analyzing the effect of the specific growth rate, a glycolytic genes TDH3 and PGK1 that encode two crucial direct correlation between TDH3 RTL and µ was also observed. enzymes for central carbon metabolism, and the Unfolded However, this effect was not that evident with PGK1 RTL (Nieto- Protein Response (UPR)-related genes HAC1 and KAR2, Taype et al., 2020b). FIGURE 5 | Comparison of qP and relative transcript levels (RTL) of key genes under all conditions tested. Error bars represent the SD of all samples belonging to the same group of culture conditions (hypoxic and normoxic). Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 8 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control FIGURE 6 | Ratios of qP and relative transcript levels (RTL) of key genes under all conditions tested. (A) Ratio between MCC and SCC for each condition (hypoxic and normoxic). (B) Ratio between hypoxic conditions and normoxic conditions for each clone (SCC and MCC). Error bars calculated by error propagation of the corresponding quotients. When comparing qP and CRL1 RTL, a similar trend could be Very interestingly, the values of the MCC/SCC CRL1 RTL and observed: they increase in MCC with respect to SCC. In addition, MCC/SCC qP ratios in chemostat with D = µ = 0.10 h−1 and under oxygen-limiting conditions also increases qP and CRL1 RTL. normoxic conditions are very similar to those obtained in the Interestingly, it is clearly seen that the largest increase is due previous work, which demonstrates the robustness and accuracy to the higher gene copy number rather than the effect of oxygen of both the chemostat cultures and the transcriptional analysis limitation. (Nieto-Taype et al., 2020b). It is also important to note that no clear differences could be According to results shown in Figure 5, both parameters observed between chemostat and fed-batch transcription results, studied in this work enhance the CRL1 transcription rate. although the effects of gene dosage and oxygen limitation seem to However, the gene dosage effect is reduced under hypoxic impact more on chemostat than on fed-batch cultures, except for conditions, where both MCC/SCC CRL1 RTL and qP ratios the TDH3 gene, where the differences between the normoxic and are lower. This could indicate that, when combining both the hypoxic state are clearer in fed-batch cultures. heterologous gene copy number increase and hypoxic conditions, the transcriptional machinery is working closer to its upper rate 3.3.2 Effect of Gene Dosage on Gene Expression and limit, apart from the potential protein processing limitation Crl1 Production mentioned previously. A plausible explanation for such an To provide further insights regarding the transcriptional results, effect could be a finite availability of TDH3-related two new parameters were defined: MCC/SCC Ratio, which transcription factors, such as Gal4-like (CRA1), found to be corresponds to the RTL quotient of the aforementioned genes crucial in the regulation of respiro-fermentative metabolism between multicopy and single-copy clones; and HPX/NX Ratio, through the interaction with PGAP and possibly other which corresponds to the RTL quotient between hypoxic glycolytic promoters (Ata et al., 2017, 2018; Kalender and conditions and normoxic conditions. The former facilitates the Çalık, 2020). Dramatically increasing the number of comprehension of the effect of gene dosage on gene expression transcription factor binding sites, which occurs in MCC, with regulation throughout the different conditions tested, while the a limited pool of these transcription factors, could cause an latter highlights the effect of oxygen limitation. The values of attenuation of the transcription of all PGAP regulated genes, as these ratios are plotted in Figure 6. previously described for PAOX1 (Cámara et al., 2017; Garrigós- In relation to gene dosage (MCC/SCC Ratio), as expected, no Martínez et al., 2019). However, in this work, no significant effect on glycolytic reporting genes (TDH3, PGK1) is observed, attenuation of either TDH3 transcription or qS was observed. since their gene regulation mechanisms remained unchanged. However, these results identified a possible bottleneck regarding 3.3.3 Effect of Hypoxia on Gene Expression and Crl1 Crl1 production in MCC, considering that the MCC/SCC CRL1 Production RTL ratio is always slightly superior to the MCC/SCC qP ratio, Concerning the oxygen limitation effect (HPX/NX Ratio), irrespective of the oxygenation conditions. This fact suggests a Figure 6 shows that glycolytic genes are overexpressed by limitation in recombinant protein processing, including about 1.5-fold under hypoxic conditions, with small translation, folding, and/or secretion. Specifically, under deviations. This is in consonance with the generally observed normoxic conditions the MCC/SCC CRL1 RTL ratio values qS increase of about 1.25-fold. On the other hand, regarding Crl1 are around 5, in consonance with the increased number of production, and consistent with the above results, hypoxic CRL1 gene copies. On the other hand, MCC/SCC qP ratio conditions have less impact on MCC than on SCC, both at values are lower than 5. A similar effect at high µ with this the transcription (CRL1 RTL) and protein secretion (qP) levels. Crl1 producer MCC was also reported and thoroughly discussed To sum up, the combination of oxygen limitation and (Puxbaum et al., 2015; Nieto-Taype et al., 2020b). increased gene dosage has a synergic but not summatory effect Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 9 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control at the transcriptional level. Accordingly, the increase in Crl1 control. Accordingly, an adaptive proportional (P) or production achieved by the combination of these two strategies is proportional-integral (PI) feedback RQ-controller with stirring larger than that obtained using these two strategies separately. rate as the manipulate variable could be implemented. Moreover, the insights described shed light on the regulation of Finally, RQ has been proposed as a novel key physiological the expression of these genes, which could help identify parameter to be used as a criterion in the scale-up of P. pastoris bottlenecks in the RPP. bioprocesses. It allows working under comparable oxygen availability conditions in different fermenters with distinct oxygen transfer efficiencies, and even in different operating 4 CONCLUSIONS modes. Moreover, an enhanced RPP bioprocess for other proteins of interest, whose transcriptional regulation is also It can be concluded that both strain and bioprocess engineering can oxygen-modulated, could be achieved. boost RPP. In this work, a 5-fold increase in gene dosage generated a remarkable increase in qP. However, the gene dosage increase does not present a linear correlation with the qP increase, not being DATA AVAILABILITY STATEMENT proportional. On the other hand, the application of cell stress has been shown to increase Crl1 production without having a significative The original contributions presented in the study are included in impact on yeast growth. Overall, the application of oxygen limitation, the article/Supplementary Material, further inquiries can be specifically in the production of recombinant protein under PGAP directed to the corresponding author. regulation under hypoxic conditions, combined with the increase in CRL1 gene dosage produces a notable increase in qP—almost 9-fold higher—in chemostat cultures. When this novel strategy is transferred AUTHOR CONTRIBUTIONS to a different operational mode, such as fed-batch cultivation, a slightly lower but also important increase in qP is observed. This is the first AG performed experimental studies and data analysis, authoring time that the combinatorial effect of gene copy number and hypoxic and editing first draft. JG-M participated in transcriptional conditions has been studied. analyses. XG-O, FV, and JM conceived, designed, and Considering the results of the transcription analyses carried supervised work. All the authors contributed to editing and out, the presence of a bottleneck downstream of the transcription internal reviews of the work, interpretation of data, read and steps has been observed in those clones with higher levels of approved the final manuscript. recombinant production (MCC). This effect had already been observed with this MCC when growing at high μ, indicating a saturation of the protein production pathways. Moreover, it has FUNDING also been proposed that there might be a transcription factor limitation in MCC under those conditions where protein This work was funded by the Project PID 2019-104666GB-100 of production is maximized, since the CRL1 RLT increase under the Spanish Ministry of Science and Innovation. hypoxic conditions is higher for SCC than for MCC. Also, it has been demonstrated that both variable stirring rate and/or variable inlet gas composition can be used to implement ACKNOWLEDGMENTS oxygen limitation with the aim of carrying out a physiological control of the P. pastoris cell factory. Nevertheless, the first option AG acknowledges the award of a scholarship FI-DGR 2019 from newly applied could be a more feasible strategy in large scale the Generalitat de Catalunya. mainly because of the energy saving and reduction of used extra gases. In the present case, it can be stated that there is a wide RQ SUPPLEMENTARY MATERIAL range that corresponds with a rather similar increase in qP. Consequently, manual heuristic control of RQ has been The Supplementary Material for this article can be found online at: suitable for an exploratory study, but further work would be https://www.frontiersin.org/articles/10.3389/fbioe.2022.818434/ required to achieve an automated, accurate, and stable RQ- full#supplementary-material Protein Production. Appl. Microbiol. Biotechnol. 98, 5301–5317. doi:10.1007/ REFERENCES s00253-014-5732-5 Ata, Ö., Prielhofer, R., Gasser, B., Mattanovich, D., and Çalık, P. (2017). Adelantado, N., Tarazona, P., Grillitsch, K., García-Ortega, X., Monforte, S., Transcriptional Engineering of the Glyceraldehyde-3-Phosphate Valero, F., et al. (2017). The Effect of Hypoxia on the Lipidome of Dehydrogenase Promoter for Improved Heterologous Protein Production in Recombinant in Pichia pastoris. Microb. Cell Fact 16, 86. doi:10.1186/ Pichia Pastoris. Biotechnol. Bioeng. 114, 2319–2327. doi:10.1002/bit.26363 s12934-017-0699-4 Ata, Ö., Rebnegger, C., Tatto, N. E., Valli, M., Mairinger, T., Hann, S., et al. (2018). A Ahmad, M., Hirz, M., Pichler, H., and Schwab, H. (2014). Protein Expression in Single Gal4-like Transcription Factor Activates the Crabtree Effect in Komagataella Pichia pastoris: Recent Achievements and Perspectives for Heterologous phaffii. Nat. Commun. 9, 4911. doi:10.1038/s41467-018-07430-4 Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 10 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control Aw, R., and Polizzi, K. M. (2013). Can Too many Copies Spoil the Broth. Microb. Garcia-Ortega, X., Valero, F., and Montesinos-Seguí, J. L. (2017). Physiological Cell Fact 12, 128. doi:10.1186/1475-2859-12-128 State as Transferable Operating Criterion to Improve Recombinant Protein Baumann, K., Carnicer, M., Dragosits, M., Graf, A. B., Stadlmann, J., Jouhten, Production in Pichia pastoris Through Oxygen Limitation. J. Chem. Technol. P., et al. (2010). A Multi-Level Study of Recombinant Pichia pastoris in Biotechnol. 92, 2573–2582. doi:10.1002/jctb.5272 Different Oxygen Conditions. BMC Syst. Biol. 4, 141. doi:10.1186/1752- Garrigós-Martínez, J., Nieto-Taype, M. A., Gasset-Franch, A., Montesinos-Seguí, 0509-4-141 J. L., Garcia-Ortega, X., and Valero, F. (2019). Specific Growth Rate Governs Baumann, K., Maurer, M., Dragosits, M., Cos, O., Ferrer, P., and Mattanovich, D. AOX1 Gene Expression, Affecting the Production Kinetics of Pichia pastoris (2008). Hypoxic Fed-Batch Cultivation of Pichia pastoris Increases Specific and (Komagataella phaffii) PAOX1-Driven Recombinant Producer Strains with Volumetric Productivity of Recombinant Proteins. Biotechnol. Bioeng. 100, Different Target Gene Dosage. Microb. Cell Fact 18, 187. doi:10.1186/ 177–183. doi:10.1002/bit.21763 s12934-019-1240-8 Çalık, P., Ata, Ö., Güneş, H., Massahi, A., Boy, E., Keskin, A., et al. (2015). Garrigós-Martínez, J., Vuoristo, K., Nieto-Taype, M. A., Tähtiharju, J., Uusitalo, J., Recombinant Protein Production in Pichia pastoris under Glyceraldehyde-3- Tukiainen, P., et al. (2021). Bioprocess Performance Analysis of Novel Phosphate Dehydrogenase Promoter: From Carbon Source Metabolism to Methanol-Independent Promoters for Recombinant Protein Production with Bioreactor Operation Parameters. Biochem. Eng. J. 95, 20–36. doi:10.1016/j. Pichia pastoris. Microb. Cell Fact 20, 74. doi:10.1186/s12934-021-01564-9 bej.2014.12.003 Guerfal, M., Ryckaert, S., Jacobs, P. P., Ameloot, P., Van Craenenbroeck, K., Çalık, P., Hoxha, B., Çalık, G., and Özdamar, T. H. (2018). Hybrid Fed-Batch Derycke, R., et al. (2010). The HAC1 Gene from Pichia pastoris: Bioreactor Operation Design: Control of Substrate Uptake Enhances Characterization and Effect of its Overexpression on the Production of Recombinant Protein Production in High-Cell-Density Fermentations. Secreted, Surface Displayed and Membrane Proteins. Microb. Cel Fact. 9 (1), J. Chem. Technol. Biotechnol. 93, 3326–3335. doi:10.1002/jctb.5696 1–12. doi:10.1186/1475-2859-9-49 Cámara, E., Albiol, J., and Ferrer, P. (2016). Droplet Digital PCR-Aided Screening Jiao, L., Zhou, Q., Su, Z., and Yan, Y. (2018). Efficient Heterologous Production of and Characterization of Pichia pastoris Multiple Gene Copy Strains. Biotechnol. Rhizopus oryzae Lipase via Optimization of Multiple Expression-Related Bioeng. 113, 1542–1551. doi:10.1002/bit.25916 Helper Proteins. Ijms 19, 3372. doi:10.3390/ijms19113372 Cámara, E., Landes, N., Albiol, J., Gasser, B., Mattanovich, D., and Ferrer, P. (2017). Jordà, J., de Jesus, S. S., Peltier, S., Ferrer, P., and Albiol, J. (2014). Metabolic Flux Increased Dosage of AOX1 Promoter-Regulated Expression Cassettes Leads to Analysis of Recombinant Pichia pastoris Growing on Different Glycerol/ Transcription Attenuation of the Methanol Metabolism in Pichia pastoris. Sci. methanol Mixtures by Iterative Fitting of NMR-Derived 13C-Labelling Data Rep. 7, 44302. doi:10.1038/srep44302 from Proteinogenic Amino Acids. New Biotechnol. 31, 120–132. doi:10.1016/j. Carnicer, M., Baumann, K., Töplitz, I., Sánchez-Ferrando, F., Mattanovich, D., nbt.2013.06.007 Ferrer, P., et al. (2009). Macromolecular and Elemental Composition Analysis Kalender, Ö., and Çalık, P. (2020). Transcriptional Regulatory Proteins in central and Extracellular Metabolite Balances of Pichia pastoris Growing at Different Carbon Metabolism of Pichia pastoris and Saccharomyces cerevisiae. Appl. Oxygen Levels. Microb. Cell Fact 8, 65. doi:10.1186/1475-2859-8-65 Microbiol. Biotechnol. 104, 7273–7311. doi:10.1007/s00253-020-10680-2 Chen, X., Li, C., and Liu, H. (2021). Enhanced Recombinant Protein Production Ken Ugo, A., Vivian Amara, A., Cn, I., and Kenechuwku, U. (2017). Microbial under Special Environmental Stress. Front. Microbiol. 12, 1–11. doi:10.3389/ Lipases: A Prospect for Biotechnological Industrial Catalysis for Green fmicb.2021.630814 Products: A Review. Ferment. Technol. 06. doi:10.4172/2167-7972.1000144 Cos, O., Ramón, R., Montesinos, J. L., and Valero, F. (2006). Operational Strategies, Kumar, A., Udugama, I. A., Gargalo, C. L., and Gernaey, K. V. (2020). Why is Batch Monitoring and Control of Heterologous Protein Production in the Processing Still Dominating the Biologics Landscape? Towards an Integrated Methylotrophic Yeast Pichia pastoris under Different Promoters: A Review. Continuous Bioprocessing Alternative. Processes 8, 1641. doi:10.3390/ Microb. Cell Fact 5, 17. doi:10.1186/1475-2859-5-17 pr8121641 Cos, O., Serrano, A., Montesinos, J. L., Ferrer, P., Cregg, J. M., and Valero, F. (2005). Kurtzman, C. P. (2005). Description of Komagataella phaffii sp. nov. and the Combined Effect of the Methanol Utilization (Mut) Phenotype and Gene Transfer of Pichia pseudopastoris to the Methylotrophic Yeast Genus Dosage on Recombinant Protein Production in Pichia pastoris Fed-Batch Komagataella. Int. J. Syst. Evol. Microbiol. 55, 973–976. doi:10.1099/ijs.0. Cultures. J. Biotechnol. 116, 321–335. doi:10.1016/j.jbiotec.2004.12.010 63491-0 Curvers, S., Brixius, P., Klauser, T., Thommes, J., Weuster-Botz, D., Takors, R., et al. Liu, J., Han, Q., Cheng, Q., Chen, Y., Wang, R., Li, X., et al. (2020). Efficient (2001). Human Chymotrypsinogen B Production with Pichia pastoris by Expression of Human Lysozyme through the Increased Gene Dosage and Co- Integrated Development of Fermentation and Downstream Processing. Part expression of Transcription Factor Hac1p in Pichia pastoris. Curr. Microbiol. 1. Fermentation. Biotechnol. Prog. 17, 495–502. doi:10.1021/bp000164j 77, 846–854. doi:10.1007/s00284-019-01872-9 Dewan, S. S. (2017). Global Markets for Enzymes in Industrial Applications. [SI]: Looser, V., Bruhlmann, B., Bumbak, F., Stenger, C., Costa, M., Camattari, A., et al. BCC Research. Report ID: BIO030J. (2015). Cultivation Strategies to Enhance Productivity of Pichia pastoris: A Ergün, B. G., Gasser, B., Mattanovich, D., and Çalık, P. (2019). Engineering of Review. Biotechnol. Adv. 33, 1177–1193. doi:10.1016/j.biotechadv.2015.05.008 Alcohol Dehydrogenase 2 Hybrid-Promoter Architectures in Pichia pastoris to López-Fernández, J., Benaiges, M. D., and Valero, F. (2020). Rhizopus oryzae Enhance Recombinant Protein Expression on Ethanol. Biotechnol. Bioeng. 116, Lipase, A Promising Industrial Enzyme: Biochemical Characteristics, 2674–2686. doi:10.1002/bit.27095 Production and Biocatalytic Applications. Catalysts 10, 1277. doi:10.3390/ Ferrer, P., Montesinos, J. L., Valero, F., and Solà, C. (2001). Production of Native catal10111277 and Recombinant Lipases by Candida rugosa: A Review. Appl. Biochem. Macauley-Patrick, S., Fazenda, M. L., McNeil, B., and Harvey, L. M. (2005). Biotechnol. 95, 221–255. doi:10.1385/ABAB:95:3:221 Heterologous Protein Production Using the Pichia pastoris Expression Garcia-Ortega, X., Adelantado, N., Ferrer, P., Montesinos, J. L., and Valero, F. System. Yeast 22, 249–270. doi:10.1002/yea.1208 (2016). A Step Forward to Improve Recombinant Protein Production in Pichia Nieto-Taype, M. A., Garcia-Ortega, X., Albiol, J., Montesinos-Seguí, J. L., and pastoris: From Specific Growth Rate Effect on Protein Secretion to Carbon- Valero, F. (2020a). Continuous Cultivation as a Tool Toward the Rational Starving Conditions as Advanced Strategy. Process Biochem. 51, 681–691. Bioprocess Development with Pichia pastoris Cell Factory. Front. Bioeng. doi:10.1016/j.procbio.2016.02.018 Biotechnol. 8, 1–21. doi:10.3389/fbioe.2020.00632 García-Ortega, X., Cámara, E., Ferrer, P., Albiol, J., Montesinos-Seguí, J. L., and Nieto-Taype, M. A., Garrigós-Martínez, J., Sánchez-Farrando, M., Valero, F., Valero, F. (2019). Rational Development of Bioprocess Engineering Strategies Garcia-Ortega, X., and Montesinos-Seguí, J. L. (2020b). Rationale-Based for Recombinant Protein Production in Pichia pastoris (Komagataella phaffii) Selection of Optimal Operating Strategies and Gene Dosage Impact on Using the Methanol-free GAP Promoter. Where Do We Stand. New Biotechnol. Recombinant Protein Production in Komagataella phaffii (Pichia pastoris). 53, 24–34. doi:10.1016/j.nbt.2019.06.002 Microb. Biotechnol. 13, 315–327. doi:10.1111/1751-7915.13498 Garcia-Ortega, X., Ferrer, P., Montesinos, J. L., and Valero, F. (2013). Fed-batch Portela, R. M. C., Vogl, T., Ebner, K., Oliveira, R., and Glieder, A. (2018). Pichia Operational Strategies for Recombinant Fab Production with Pichia pastoris pastoris Alcohol Oxidase 1 (AOX1) Core Promoter Engineering by High Using the Constitutive GAP Promoter. Biochem. Eng. J. 79, 172–181. doi:10. Resolution Systematic Mutagenesis. Biotechnol. J. 13, 1700340. doi:10.1002/ 1016/j.bej.2013.07.013 biot.201700340 Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 11 January 2022 | Volume 10 | Article 818434 Gasset et al. Pichia RQ-Based Physiological Control Potvin, G., Zhang, Z., Defela, A., and Lam, H. (2016). Screening of Alternative Wehbe, O., Yaman, O. U., and Çalık, P. (2020). Ethanol Fed-Batch Bioreactor Carbon Sources for Recombinant Protein Production in Pichia pastoris. Int. Operation to Enhance Therapeutic Protein Production in Pichia pastoris under J. Chem. Reactor Eng. 14, 251–257. doi:10.1515/ijcre-2015-0092 Hybrid-Architectured ADH2 Promoter. Biochem. Eng. J. 164, 107782. doi:10. Prielhofer, R., Maurer, M., Klein, J., Wenger, J., Kiziak, C., Gasser, B., et al. (2013). 1016/j.bej.2020.107782 Induction Without Methanol: Novel Regulated Promoters Enable High-Level Weninger, A., Hatzl, A.-M., Schmid, C., Vogl, T., and Glieder, A. (2016). Expression in Pichia pastoris. Microb. Cell Fact 12, 5. doi:10.1186/1475-2859- Combinatorial Optimization of CRISPR/Cas9 Expression Enables Precision 12-5 Genome Engineering in the Methylotrophic Yeast Pichia pastoris. J. Biotechnol. Puxbaum, V., Mattanovich, D., and Gasser, B. (2015). Quo Vadis? the 235, 139–149. doi:10.1016/j.jbiotec.2016.03.027 Challenges of Recombinant Protein Folding and Secretion in Pichia Yang, Z., and Zhang, Z. (2018). Engineering Strategies for Enhanced Production of pastoris. Appl. Microbiol. Biotechnol. 99, 2925–2938. doi:10.1007/ Protein and Bio-Products in Pichia pastoris: A Review. Biotechnol. Adv. 36, s00253-015-6470-z 182–195. doi:10.1016/j.biotechadv.2017.11.002 Raschmanová, H., Zamora, I., Borčinová, M., Meier, P., Weninger, A., Mächler, D., Zavec, D., Gasser, B., and Mattanovich, D. (2020). Characterization of Methanol et al. (2019). Single-Cell Approach to Monitor the Unfolded Protein Response Utilization Negative Pichia pastoris for Secreted Protein Production: New during Biotechnological Processes with Pichia pastoris. Front. Microbiol. 10, Cultivation Strategies for Current and Future Applications. Biotechnol. 1–18. doi:10.3389/fmicb.2019.00335 Bioeng. 117, 1394–1405. doi:10.1002/bit.27303 Sallada, N. D., Harkins, L. E., and Berger, B. W. (2019). Effect of Gene Copy Number and Chaperone Coexpression on Recombinant Hydrophobin HFBI Conflict of Interest: The authors declare that the research was conducted in the Biosurfactant Production in Pichia pastoris. Biotechnol. Bioeng. 116, absence of any commercial or financial relationships that could be construed as a 2029–2040. doi:10.1002/bit.26982 potential conflict of interest. van der Heijden, R. T. J. M., Romein, B., Heijnen, J. J., Hellinga, C., and Luyben, K. C. A. M. (1994). Linear Constraint Relations in Biochemical Reaction Systems: Publisher’s Note: All claims expressed in this article are solely those of the authors II. Diagnosis and Estimation of Gross Errors. Biotechnol. Bioeng. 43, 11–20. and do not necessarily represent those of their affiliated organizations, or those of doi:10.1002/bit.260430104 the publisher, the editors, and the reviewers. Any product that may be evaluated in Vanleeuw, E., Winderickx, S., Thevissen, K., Lagrain, B., Dusselier, M., Cammue, B. this article, or claim that may be made by its manufacturer, is not guaranteed or P. A., et al. (2019). Substrate-Specificity of Candida rugosa Lipase and its endorsed by the publisher. Industrial Application. ACS Sust. Chem. Eng. 7, 15828–15844. doi:10.1021/ acssuschemeng.9b03257 Copyright © 2022 Gasset, Garcia-Ortega, Garrigós-Martínez, Valero and Verheijen, P. (2009). “Data Reconciliation and Error Detection,” in The Metabolic Montesinos-Seguí. This is an open-access article distributed under the terms of Pathway Engineering Handbook (Boca Raton, FL: CRC Press), 8. doi:10.1201/ the Creative Commons Attribution License (CC BY). The use, distribution or 9781439802977.ch8 reproduction in other forums is permitted, provided the original author(s) and Wang, N. S., and Stephanopoulos, G. (1983). Application of Macroscopic Balances the copyright owner(s) are credited and that the original publication in this journal is to the Identification of Gross Measurement Errors. Biotechnol. Bioeng. 25, cited, in accordance with accepted academic practice. No use, distribution or 2177–2208. doi:10.1002/bit.260250906 reproduction is permitted which does not comply with these terms. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 12 January 2022 | Volume 10 | Article 818434