Garrigós‑Martínez et al. Microb Cell Fact https://doi.org/10.1186/s12934-021-01564-9 (2021) 20:74 Microbial Cell Factories RESEARCH Open Access Bioprocess performance analysis of novel methanol‑independent promoters for recombinant protein production with Pichia pastoris Javier Garrigós‑Martínez1, Kiira Vuoristo2, Miguel Angel Nieto‑Taype1, Juha Tähtiharju2, Jaana Uusitalo2, Pauliina Tukiainen2, Christian Schmid5,6, Ilya Tolstorukov3, Knut Madden4, Merja Penttilä2, José Luis Montesinos‑Seguí1, Francisco Valero1, Anton Glieder5,6* and Xavier Garcia‑Ortega1 Abstract Background: Pichia pastoris is a powerful and broadly used host for recombinant protein production (RPP), where past bioprocess performance has often been directed with the methanol regulated AOX1 promoter (­ PAOX1), and the constitutive GAP promoter ­(PGAP). Since promoters play a crucial role in an expression system and the bioprocess efficiency, innovative alternatives are constantly developed and implemented. Here, a thorough comparative kinetic characterization of two expression systems based on the commercial PDF and UPP promoters (­ PPDF, ­PUPP) was first conducted in chemostat cultures. Most promising conditions were subsequently tested in fed-batch cultivations. These new alternatives were compared with the classical strong promoter ­PGAP, using the Candida antarctica lipase B (CalB) as model protein for expression system performance. Results: Both the P ­ PDF and ­PUPP-based expression systems outperformed similar ­PGAP-based expression in chemostat cultivations, reaching ninefold higher specific production rates (qp). CALB transcription levels were drastically higher when employing the novel expression systems. This higher expression was also correlated with a marked upregula‑ tion of unfolded protein response (UPR) related genes, likely from an increased protein burden in the endoplasmic reticulum (ER). Based on the chemostat results obtained, best culture strategies for both P ­ PDF and ­PUPP expression sys‑ tems were also successfully implemented in 15 L fed-batch cultivations where qp and product to biomass yield (YP/X*) values were similar than those obtained in chemostat cultivations. Conclusions: As an outcome of the macrokinetic characterization presented, the novel ­PPDF and P ­ UPP were observed to offer much higher efficiency for CalB production than the widely used ­PGAP-based methanol-free alternative. Thus, both systems arise as highly productive alternatives for P. pastoris-based RPP bioprocesses. Furthermore, the different expression regulation patterns observed indicate the level of gene expression can be adjusted, or tuned, which is interesting when using Pichia pastoris as a cell factory for different products of interest. Keywords: Komagataella phaffii (Pichia pastoris), Recombinant protein production, Methanol-free bioprocesses, Expression system characterisation, Promoter regulation, Bioprocess development *Correspondence: a.glieder@tugraz.at 5 Institute of Molecular Biotechnology, Graz University of Technology, Petersgasse 14, 8010 Graz, Austria Full list of author information is available at the end of the article © The Author(s) 2021. 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, visit http://​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. Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 2 of 12 Background 31], and/or alternative systems that allow controlled sub- The non-conventional yeast Komagataella phaffii, widely strate delivery. Using bioreactors, production processes known under the former name Pichia pastoris, is a distin- can be carried out applying optimal ranges of the key bio- guished host for recombinant protein production (RPP) process parameters [32, 33]. Chemostat systems, where [1–7] and metabolite production [8]. Among the many cultures are maintained at non-dynamic, steady-state positive features that make P. pastoris a good choice for conditions, have become a valuable tool for bioprocess RPP, and historically one of the most relevant, is strong characterization and further optimization [34]. In this and tightly regulated expression based on the alcohol oxi- way, a full kinetic characterization of the candidate cell dase 1 promoter ­(PAOX1) [9–12]. When using the ­PAOX1 factories can be performed. Furthermore, interestingly, promoter, induction occurs in the presence of metha- sampling for ‘omics’ analyses can be reliably carried out nol, whereas glycerol or glucose fully repress expression on cells from chemostat, where cultures have constant [13]. De-repression is not sufficient for significant gene key process parameters, and the cell population is highly expression; therefore, a simple recombinant protein pro- homogeneous [34]. duction process is typically divided into two phases. First, Studies including precise strain characterization by a glucose/glycerol-based batch phase, where a relatively chemostat cultivations, have revealed how the specific high amount of biomass is generated without recom- growth rate (µ) significantly affects recombinant protein binant protein production. Subsequently, the metha- production (RPP) rates [6, 15, 35–40]. Importantly, the nol feeding phase triggers strong P ­ AOX1-driven protein relationship between µ and the specific production rate production. However, such tightly controlled induction (qp), also called production kinetics, is dependent on both and strong expression levels by ­PAOX1 causes operational the expression system used and the recombinant protein drawbacks due to the use of methanol, including high expressed. In previous studies, García-Ortega et al. [37], oxygen requirements and heat production, as well as and Nieto-Taype et al. [38] described a linear µ-qp rela- increased costs derived from methanol storage and han- tionship when producing a human Fab, as well as the dling [14, 15]. To address these challenges, mutated pro- Candida rugose lipase 1 (Crl1), respectively, both under moter variants [16] or co-substrate feeding strategies had the control of the constitutive ­PGAP. The same trend was been employed [17]. observed for the production of Lipase B from Candida In order to open new opportunities, innovative alter- antarctica using the constitutive PGK promoter [41]. natives are constantly developed, evaluated and imple- These authors concluded that since the constitutive P ­ GAP mented. In terms of promoter strength, the other has a pivotal role in the growth-associated glycolysis, methanol inducible promoters such as DAS1 and DAS2 therefore, one should expect the RPP to be growth-cou- ­(PDAS1; ­PDAS2), demonstrate similar strength [18]. In pled. On the other hand, curved/non-linear µ-qp trends addition to the numerous attempts that had been made were observed [21, 39, 42], suggesting non-coupled tran- to modify ­PAOX1 regulation by mutagenesis or synthetic scriptional regulation, or bottlenecks in the protein pro- fusions [19], the co-expression of transcription factors cessing pathway. In particular, Garrigós-Martínez et al. was demonstrated as an interesting alternative to induc- [39] remarked that the µ-qp bell-shaped trend observed tion by methanol. In addition, numerous methanol-inde- in the P­ AOX1 regulated production of Crl1 was probably pendent expression systems have been developed and caused by an alternative transcriptional regulation. This tested with promising results such as P ­ GTH1, ­PCAT1, ­PTHI11 conclusion was based on the determination that at differ- ­PHpFMD or ­PTEF, among others [20–25]. ent µ, target protein production profiles and the relative Historically, RPP improvements have been mainly transcripts did not present the usual linearity of growth- obtained through strain and promoter system engineer- coupled expression systems. ing [16, 26–29]. Multiple clones with different expression In this work, the performance characterization of two cassettes or random integration variants with the same novel expression systems for RPP with P. pastoris are expression cassette are tested and compared in parallel in based on: (1) The new PDF promoter (­PPDF, a commer- shake flasks or microtiter plates. This approach is consid- cial variant of the Hansenula polymorpha FMD promoter ered fast and cost-effective; however, most testing plat- [25, 43], which drives strong transcription by simple forms do not allow control of key bioprocess parameters methanol-free de-repression and can be also further such as dissolved oxygen, pH, as well as growth and feed induced with methanol), and (2) UPP promoter ­(PUPP, rates. Since these parameters affect target protein expres- a constitutive commercial variant of a Pichia promoter sion, selection of a “best performing” clone might not called GCW14, [23]). Both promoter systems have been always be optimal. Accordingly, the performance of the thoroughly studied and compared with P ­ GAP, the most production clones candidates should be compared in cul- well characterized constitutive promoter, considered a tivation platforms such as bench-top bioreactors [22, 30, reference standard for methanol-free expression systems. Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 3 of 12 Expression strains for the lipase B from Candida antarc- tica (CalB) were constructed, all with the same parental strain, gene dosage and identical vectors, except for the promoter sequence driving CALB gene expression. To compare these expression systems, a set of chemostat cultivations designed to assess the effect of different µ values on the production kinetics was performed. Fur- thermore, CALB transcript levels were determined and compared to the expression levels for each condition tested in chemostat. Finally, working at the µ ranges that generated the best results in chemostat mode, the same selected clones were cultivated in 15 L fed-batch pro- cesses to evaluate their performance in this operational mode. Results and discussion Strain generation, screening and gene dosage Isogenic clones were generated to compare the per- formance of the promoters ­PUPP, ­PPDF and ­PGAP for the CalB expression as a model recombinant protein. Con- sidering the potential clonal variability often observed in Pichia clone generation methods, care was taken to select a clone for each expression system with a single expression cassette integrated into the genome [29, 44]. Subsequently, around 90 individual transformants were Fig. 1 Physiological state indicators of Pichia pastoris CalB producer analyzed in a high-throughput screen based on deep well clones-GAP-C, PDF-C, UPP-C—in chemostat cultivations. a Specific plate (DWP) system, to develop a “landscape” of expres- glycerol consumption rate (qs), overall glycerol-to-biomass yield (YX/S*). b Specific oxygen uptake rate (qO2), specific carbon dioxide sion data for clone characterization according to Weis production rate (qCO2) and respiratory quotient (RQ). Error bars et al. [45]. Putative single-copy integration transformants represent the standard deviation of two biological replicates for each of the different promoter constructs were picked from the majority of transformants which showed very similar lipase activity in the supernatant after cultivation influence on CalB recombinant expression, and not due and induction in 96-DWP and an initial screen measur- to a different gene dosage. ing CalB activity of secreted reporter enzyme. Among the discarded clones, secreting higher amounts of CalB were suspected to be associated with multicopy or random Physiological state comparison of the P. pastoris clones integration events; while lower activity observed in some harboring different expression system clones might be from with detrimental effects exerted by Chemostat cultivations were performed with one ectopic integration [46]. Subsequently, a second round of selected CalB production clone for each of the three dif- DWP screening among the potential single copy integra- ferent expression systems (GAP-C, PDF-C and UPP-C). tion candidates was performed, which were tested at least This comparison test was performed at three different the biological triplicates (data presented on Additional dilution rates (D): 0.05 ­h−1, 0.10 ­h−1 and 0.15 ­h−1. This file 1: Figure S1A, B, C). Therefore, for the candidate characterization allowed to determine the range of dilu- clones, gene dosage was determined by droplet digital tion rate to significantly improve CalB production in sub- PCR (ddPCR, data presented on Additional file 2: Tables sequent fed-batch (FB) cultivations. S2A, B) to confirm each construct only contained a sin- PUPP and P ­ PDF clones had significantly higher expres- gle-copy of the respective expression vector in the Pichia sion levels than those based on P ­GAP. Furthermore, genome. Confirmed single copy clones for each expres- high levels of recombinant protein expression have sion system presenting an average CalB production were been shown to cause a burden on the protein secre- therefore selected to start the expression systems char- tion machinery likely due to an overload of the process- acterization, thus ensuring that production differences ing capacity [47, 48]. Therefore, an impact of the three are only a result of the effect of each promoter’s specific expression systems on the physiological state was tested Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 4 of 12 in chemostat cultivations by analyzing glycerol and ­O2 consumption rates, and ­CO2 production rates (Fig. 1). In this regard, no significant differences were evident across the D tested. The specific glycerol consumption rate (qs) and overall biomass-to-substrate yield (YX/S*) were rather similar. As expected, qs increases linearly over the D, whereas YX/S* values were constant (only slight differences could be observed at the highest D). All the clones presented similar specific ­CO2 produc- tion rates (qCO2) and specific O ­ 2 consumption rates (qO2) and followed standard linear trends. Consequently, simi- lar respiratory quotient values were exhibited by all the clones studied (RQ, i.e. about 0.62, see Fig. 1b). Based on the analysis at macrokinetic level, it can be stated that the higher CalB production provided by the new generation expression systems based on the promoters ­PPDF or P ­ UPP did not alter any of the studied physiological parameters compared to the GAP-C, which presents lower CALB expression levels. Novel expression systems outperform P ­ GAP‑based CalB production Compared to P ­ GAP, both of the new promoters resulted in notably higher qp values, between 4 and ninefold higher at any D (Table 1). UPP-C also had qp values sig- nificantly higher than PDF-C at the lowest and middle D. At the highest D, UPP-C was similar to PDF-C, with only Fig. 2 Comparison of the overall CalB product-to-biomass yield slightly higher qp. (YP/X*) in chemostat cultivations for the three expression systems tested. Error bars represent the standard deviation of two biological Different production kinetic profiles, qp at different D, replicates were obtained for all three expression systems compared (Fig. 2a). UPP-C presented a bell-shape profile with a maximum at mid D, 0.10 ­h−1. On the other hand, PDF-C generated a clearly saturated profile. And, GAP-C had a Table 1 Comparison of the main production parameters relatively linear pattern, with a slight saturation trend at obtained in chemostat and fed-batch cultures with the producer higher D values. This result differs from other examples clones at different specific growth rates (µ) reported using ­PGAP in which qp clearly increases linearly Clone Operational Nominal µ Experimental qp YP/X* with D [37, 38]. Thus, these results indicate that produc- mode µ tion kinetics, in most cases, are protein dependent. h−1 h−1 AU ­gx−1 ­h−1 AU ­gx−1 Another important key parameter to be considered is the overall product to biomass yield (YP/X*); it deter- GAP-C Chemostat 0.050 0.046 1.16 24.9 mines the overall capacity of cells to produce recombi- 0.100 0.103 2.24 21.9 nant protein under certain conditions. The ­PUPP and ­PPDF 0.150 0.149 2.74 18.4 expression systems are similar to other systems where PDF-C Chemostat 0.050 0.047 7.20 153 increasing D is detrimental for ­YP/X* [38, 39], as shown 0.100 0.100 10.8 108 in Fig. 3. As observed with qp, the biggest difference 0.150 0.155 10.8 70.1 between UPP-C and PDF-C YP/X* values was demon- Fed-batch 0.050 0.042 9.20 219 strated at the lowest D, yet similar at higher D. Impor- 0.100 0.087 13.1 150 tantly, the highest YP/X* values for UPP-C and PDF-C UPP-C Chemostat 0.050 0.052 10.3 197 were notably higher than those obtained with the GAP-C 0.100 0.106 12.9 122 (i.e. up to 8.9-fold higher with UPP-C at 0.05 ­h−1) (Fig. 3). 0.150 0.156 11.4 72.7 Based on CalB production-related parameters qp and Fed-batch 0.050 0.051 11.1 217 YP/X*, both novel expression systems should be consid- 0.100 0.084 9.95 102 ered good candidate promoters to produce recombinant Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 5 of 12 Fig. 4 Effect of dilution rate on the qp ratio and the CALB differential Fig. 3 a CalB production kinetics (qP vs D) and CALB relative RTL calculated as ­log2 fold change relative to GAP-C values. GAP-C transcription levels determined in chemostat cultivations for the values were used as a control for ratio calculations. P-values three expression systems studied. Transcript levels were normalized (t-test) were calculated in order to determine the CALB expression to the levels of the MTH1 transcript, which was used as housekeeping significance between producer clones (* significance level P ≤ 0.05; ** gene for the analysis. Error bars represent the standard deviation of significance level P ≤ 0.01) two biological replicates. b Percentage of CALB relative transcription levels respect to the maximum observed for the corresponding expression system observed among the different D tested. Therefore, the regulation of CALB expression under ­ PUPP control proteins. Different conditions could be optimal depend- should be considered growth independent. Strikingly, ing on the objective: to reach higher product titer, or the CALB mRNA expression patterns do not correlate maximal productivity. If the objective is the highest pro- with the bell-shaped qp profile described in the previ- tein titer, the lowest D should be selected [38, 39, 49], ous section (highest at a 0.10 D, Fig. 2a). For the PDF-C, especially in the case of UPP-C, where YP/X* reductions RTL presents a bell-shape trend, while the qp-D profile with D were more pronounced than with PDF-C. On presents a saturation pattern. Therefore, according to the other hand, in order to maximize qp, the conditions the RTL results (Fig. 2b), the ­PPDF-based expression recommended for high production should be at a D of system exhibits a growth-rate dependent regulation, 0.10 ­h−1, observed most markedly with PDF-C, where D which thus can be considered a system with a promis- has a bigger impact (Fig. 2a). ing tunable expression pattern. As presented in Fig. 4, the comparison of CALB expression regulated by ­PUPP and ­PPDF, relative to the The new promoters enable increased tunability ­PGAP, illustrates an interesting contrasting behavior. The of recombinant protein expression processes in Pichia weaker, growth-coupled, P ­ GAP-based expression system pastoris performs better at higher µ; demonstrated here with mRNA levels are not always directly correlated with the the single-copy expression strain. In this case, the high- level of recombinant protein production obtained [22, est target transcription levels may result in a recombi- 38, 39], especially since high CALB transcript levels nant protein “burden” can still be sorted, or processed, can cause enhanced cellular stress [50]. In the present properly in the ER. On the other hand, with both novel work, variable CALB transcript levels were observed in promoters, CALB transcription levels and specially qp all three expression systems compared: GAP-C, UPP- ratios, decrease over D, thus indicating that high spe- C, and PDF-C. As shown in Fig. 2b, a linear profile of cific growth rates are detrimental for these more pro- CALB relative transcript levels (RTL) was observed ductive systems. Both UPP-C and PDF-C generated across D for GAP-C, confirming the widely reported CALB transcripts, as well as secreted protein at signifi- constitutive and growth-coupled regulation pattern cantly higher levels than GAP-C. The higher transcript with ­PGAP. For this GAP-C, even both RTL-D and qp-D levels may be overwhelming the secretory pathway, present similar profiles, a slight saturation trend of qp triggering the unfolded protein response (UPR). There- can be observed at high D (Fig. 2a), likely because the fore, UPP-C and PDF-C might be better at low and low production rates observed with ­PGAP are not likely medium µ, when most of transcription can generate to overload of the processing and secretory capacity. protein, as is demonstrated by the higher productivity For UPP-C, only slight differences in CALB RTL were Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 6 of 12 rates. Consequently, suitable conditions to balance both growth and protein production are needed to improve production protocols. UPR influence on CalB production In order to assess potential endoplasmic reticulum (ER) stress derived from the excessive heterologous pro- tein production, the expression of key UPR genes were analyzed. The reporters selected were two well-known ER-resident chaperones, KAR2 and ERO1, and a gene product generally considered to be an UPR master regu- lator, HAC1 [2, 51, 52]. The relative transcription levels of these three UPR related genes were measured in UPP-C and PDF-C and compared to the GAP-C. Data is pre- sented as a l­og2 fold change relative to the GAP-C levels (Fig. 5). Overall, the expression of all three of the UPR reporter genes, HAC1, KAR2 and ERO1, had similar expression patterns across all the D rates tested. Their expression was the highest when the clones were cultivated at inter- mediate D of 0.10 ­h−1, whereas only moderate expres- sion increases were found at D 0.15 ­h−1. These results are in-line with other published work reporting that not only recombinant protein production, but also µ, hav- ing a significant impact on UPR induction [36, 53]. The cited studies are also similar to the data presented here, showing unremarkable UPR levels at lower µ, whereas µ increases led to UPR upregulation. Strikingly, the growth-coupled expression of CALB by GAP-C was the only case in which the CALB RTLs pre- Fig. 5 Transcriptional levels of three different UPR-related genes sented a similar pattern to the qp. The UPP-C qp values at the different dilution rates. a PDF-C, and b UPP-C (presented as mimic the UPR sensor gene expression profiles across a comparison of transcript levels with the levels from the control the D tested, regardless of CALB transcription rates, GAP-C, ­log2 scale). P-values (t-test) were calculated for all the genes and conditions in order to determine the gene expression which were rather similar for all the D rates. Therefore, significance between producer clones (* significance level P ≤ 0.05) UPR might have an influence in subsequent steps of CalB processing and secretion. The UPR impact on CALB expression is demonstrated by comparing qp and CALB Together, these analyses indicate that qp for CalB in RTL values at D 0.10 ­h−1. In the Fig. 5, UPR-associated P. pastoris is influenced by several factors: heterolo- gene expression is higher for UPP-C (Fig. 5b) than PDF-C gous gene transcription rates, recombinant protein- (Fig. 5a) and may explain why UPP-C qp is higher than associated UPR, and D-associated UPR. Comparing the PDF-C qp at this D despite UPP-C presenting 84% less different alternatives for methanol-free expression pre- CALB RTL. In this sense, it has been also described that sented in this work, it could be observed that the new the co-expression of protein disulfide isomerase, which generation constructs, based on the PDF and UPP pro- is also upregulated at higher UPR, enhances active lipase moters, allowed to achieve CALB transcription levels of production by P. pastoris [54]. Lastly, the comparison up to eightfold higher than with ­PGAP-regulated expres- between two D conditions for PDF-C in continuous culti- sion, for all the D tested. However, at the high target vations supports this hypothesis. At both D 0.05 ­h−1 and protein expression levels, a direct correlation between D 0.15 ­h−1, the CALB RTL levels are rather similar. How- CALB RTL and qp was not observed. In contrast, higher ever, the UPR-related gene expression is growth coupled, qp values were usually observed at D conditions with enhanced at higher D. The increased UPR could be con- enhanced expression of UPR-related genes, suggesting tributing the 50% higher qp observed at the highest D, a relevant impact of UPR on CalB production. even though the target gene RTL levels are rather similar. Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 7 of 12 Fed‑batch cultures for further scalable bioprocess development Through chemostat cultivations, a systematic characteri- zation of the three expression systems studied at differ- ent dilution rates was carried out, generating information both at macrokinetic, stoichiometric and transcriptional level. From these results, the range of µ that improves CalB production were determined in order to achieve higher product yields and/or productivities. Ideally, the best µ values found in chemostat cultivations should be implemented to fed batch cultures (FB), which is cur- rently the most widely used scalable operational mode for industrial production of recombinant proteins. How- ever, production kinetics may present relevant differ- ences between the different operational modes [38, 39]. Therefore, fed-batch cultivations were also conducted to confirm CalB production kinetics pattern determined in chemostat cultivations. Carbon-limited fed-batch culti- vations, the culture strategy usually considered as most efficient with P. pastoris methanol free processes, were performed with the UPP-C and PDF-C to obtain biomass and CalB production profiles (Fig. 6). Accordingly, based on the results obtained in chemostat cultures in which strong expression systems perform better at mid-low µ, the µ of 0.15 ­h−1 was discarded for further fed-batch implementation. As expected, biomass production of all the cul- tures presented the targeted exponential profile, reaching a maximum between 90 and 100 g ­ L−1 of Fig. 6 Fed-batch culture time profiles of Biomass (a) and CalB dry cell weight (DCW), which is considered a stand- production (b) expressed as cell concentration (g ­L−1) of dry ard endpoint for Pichia high-cell density fed-batch cell weight concentration (DCW) and total activity units (kAU), (Fig. 6a). CalB production, expressed in activity units respectively. Vertical lines separate batch and fed-batch phases (kAU), also increased pseudo-exponentially over time (Fig. 6b). Product titers obtained at the lowest µ tested (~ 0.05 ­h−1) were substantially higher than with the intermediate µ (~ 0.10 ­h−1), being 38% higher with PDF-C and 67% with UPP-C. Therefore, YP/X* val- in the SDS-PAGE presented, no signs of CalB degrada- ues were also markedly higher at the low µ cultures tion is observed, indicating protein quality is similar (Table 1). Comparing product-related parameters between the different expression systems and fed-batch between chemostat and FB cultivations, PDF-C per- cultivations (Additional file 3: Figure S3). formed better in FB mode (24% higher, on average qp, At the end of the batch phase, a twofold higher CalB and 41% on average higher YP/X* at the equivalent µ, titer was obtained with UPP-C compared to PDF-C. Table 1). On the other hand, UPP-C presented smaller This points to a difference in the expression systems differences of qp and YP/X* values at low µ, on average regulation in presence of excess carbon, a situation that below 10%. Strikingly, for the range µ 0.10 ­h−1, UPP-C only occurs during the batch phase. Whereas, ­PUPP pre- presented performance parameters were significantly sents a constitutive and strong expression in excess of worse in FB cultures relative to the chemostat cultures C-source, ­PPDF-regulated expression was repressed and (Table 1). With respect to product quality and purity, it increase under C-limiting conditions, such as at the is worth to mention that the supernatant impurity fin- end of the batch phase, and during the fed-batch phase. gerprints were similar between all four fed-batch culti- Thus, ­PPDF generates an interesting tunable expression vations; all had very low levels of Pichia native proteins system, allowing simple uncoupling of biomass growth secreted to the culture broth. Additionally, as observed and achieving different transcription levels without altering the carbon source. Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 8 of 12 Conclusions to establish further strategies to be finally industrially In this study focused on developing methanol-free implemented at large scale. Furthermore, these produc- alternatives for RPP with P. pastoris, a macrokinetic tion conditions should also be taken into consideration characterization of two promising expression systems for further improvements of the host at a molecular were conducted in chemostat cultures, P ­ UPP and P ­ PDF, level, closing the duty cycle of synthetic biology based benchmarking their performance with ­PGAP. In terms of on iterative Design-Build-Test-Learn steps (DBTL). The substrate and respiration-related parameters, all three successful results obtained in this work, are expected to expression systems behaved similarly, suggesting that the make this approach transferable to other processes based potential differences in CalB production does not signifi- on other expression systems and also different microbial cantly alter the yeast homeostasis in chemostat cultiva- cell factories towards rationally improve the efficiency of tions. Overall, the CalB production kinetics with the two bioprocesses. novel expression systems generated significantly higher levels of recombinant protein than the reference, GAP- Materials and methods C, up to ninefold higher in terms of qp. The differences Clone construction, selection and expression testing in product-related parameters were primarily attrib- The parental strain, P. pastoris BSYBG11(aox1-/MutS), uted to the significantly higher CALB transcription lev- which is a BioGrammatics (Carlsbad, CA) Pichia pasto- els. Interestingly, under carbon-limiting conditions, the ris BG11 strain, deposited at Bisy in Austria, was trans- ­PPDF-based expression system showed a D-dependent formed with each of the expression vectors, only differing tunable expression, while ­PUPP-regulated expression was in the respective promoters. This strain is derived from more constant, independent of the growth rates tested. the BioGrammatics K. phaffii (P. pastoris) strain BG10 Furthermore, an UPR up-regulation was noted most but with a slower methanol metabolization phenotype, markedly with the UPP-C at D = 0.1 ­h−1. At this dilution MutS [55]. The transformation method to express Can- rate, the three UPR reporter genes monitored were at dida antarctica lipase B (CalB) under the regulation of their highest level. Notable is that the highest qp was also ­PGAP, ­PUPP and ­PPDF has been described elsewhere [56]. at this D condition. Recombinant vectors with the selected promoters were The chemostat results were used to design strategies based on pPpT4_Alpha_S vector [57]. To avoid biasing based on the µ ranges that provide the best production of the results by transformant variability, low amounts of results for its further implementation in fed-batch culti- the linearized plasmid DNA (< 1 µg of DNA) were used vations. Thus, both novel expression systems, based on to avoid multi copy expression cassette integration as ­PUPP, ­PPDF, were also tested in this operational mode. The described elsewhere [29, 44]. difference in regulation patterns was reproduced in a fed- Candidate clones expression screening was done at batch mode as UPP-C had around twofold higher CalB microscale cultures in deep well plates (DWP´s), as production than PDF-C at the end of batch phase, illus- described by Krainer et al. [58] with minor modifications. trating a strong constitutive CALB expression under P ­ UPP During the induction phase, methanol from BMM2 and regulation. On the other hand, highest CALB expression BMM10 was replaced with glycerol as the carbon source in the PDF-C was obtained under C-limiting conditions, with 1% and 5% glycerol (w/v), respectively. All media in which the expression is derepressed, thus presenting were buffered at pH 7.0. an interesting tunable expression pattern. Concerning An initial screen in DWP’s allowed to select the pro- the production kinetics, UPP-C showed much better per- ducer candidates from the clonal variability after trans- formance at low µ in fed-batch, as the qp at this µ out- formation. Furthermore, a second screening employing performed the levels obtained at mid µ by 11%. PDF-C DWP’s with seven replicates per clone was used to vali- expression, on the other hand, was enhanced at mid µ, date previous results with biological replicates towards as qp was significantly higher than under low µ condi- the selection of producer clones that integrated only tions. In all cases, no significant difference in the prod- one copy of the desired expression cassettes. These uct quality was observed among the different fed-batch final clones were used in the chemostat and fed-batch cultivations performed, presenting for all cases low lev- cultivations. els of protein impurity fingerprints and no signs of CalB degradation. Gene dosage determination This work, testing alternative promoter designs, vali- The gene dosage/copy number was determined for date the approach that confirms the transferability from each selected producer clone using droplet digital PCR small scale screenings to its characterization in chemo- (ddPCR) as described elsewhere [38, 39]. ddPCR was stats. Most importantly, outcomes obtained during the performed with primers to amplify the CALB gene pre- characterization should be considered as highly relevant sent in the expression cassette, as well as with primers Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 9 of 12 for amplification of the Actin 1 gene (ACT1) as a refer- monitored from the third residence time until the fifth ence. The ACT1 gene had been demonstrated before, and one, where samples were taken. thus to be a reference for single copy gene of the haploid P. pastoris. The list of primers used is presented as Addi- Fed‑batch cultivations tional file 2: Table S2A. Fed-batch cultivations were performed in duplicate, all in New Brunswick BioFlo 510 bioreactor (Eppendorf, Total RNA extraction, cDNA synthesis and transcript level Germany), connected to BioCommand Control soft- determination ware. Batch media composition was the same as that RNA was isolated from 1 mL culture samples taken from used in the chemostat runs, stoichiometrically identical the chemostat culture broth as defined by Landes et al. to that used in prior work [49], except glycerol, instead [22]; RNA was prepared using the SV Total RNA Isola- of glucose, was used in the fed-batch feeding. The cul- tion System (Promega, Madison, Wisconsin, US) follow- tures were grown at 25 °C under overpressure (0.2 bar) ing the manufacturer’s instructions. and had a 7.5 L starting volume, including 1 L seed. The RNA integrity was checked by agarose electrophoresis pH was kept at 5.0 by the automated addition of 12.5% 2000 (Thermo Scientifc™, Waltham, MA, US). and RNA concentration was measured with Nanodrop ­NH4OH. Dissolved oxygen (DO) was maintained above cDNA was synthetized using iScript™ cDNA Synthesis 30% of air saturation with the automatic control of stirrer speed (400–700 rpm), constant airflow 10 L ­min−1, and kit (Bio-Rad, Hercules, CA, USA), following the manu- enriched with ­O2 when needed (0–50% of total inlet flow facturer’s instructions. Primers were designed to analyze rate). the relative transcript levels (RTL) of the following target Glycerol feeding was started upon depletion of batch genes by qPCR: CALB, KAR2 and ERO1 (two ER-resident medium glycerol. An exponential pre-programming chaperones), and HAC (UPR master regulator). MTH1 feeding rate was performed to maintain the specific gene was used as housekeeping gene for the transcrip- growth rate (0.05 ­h−1 or 0.10 ­h−1) constant at the selected tional analysis. The qPCR procedure, including equip- set-point. All cultivations were grown under carbon- ment, qPCR master mix solution and housekeeping gene limiting conditions. The procedure is described in detail is detailed in prior work [39]. The annealing-extension elsewhere [49]. temperature was adjusted to 59 °C. The list of primers used is presented as Additional file 2: Table S2A. Biomass determination Biomass concentrations were measured in triplicate as Chemostat cultivations DCW values, as described elsewhere [59]. The relative Chemostat cultivations were performed in duplicate, all standard deviation (RSD) of the measurements was about in 2 L Biostat B plus Bioreactors (Sartorius Stedim, Goet- 4%. tingen, Germany) according to García Ortega et al. [37]. Batch and chemostat media compositions are stoichio- Quantification of the carbon source and byproducts metrically identical to the detailed in the reference [37]; The concentration of both glycerol and the potential fer- however, the concentrations were reduced by half. mentation byproducts, were measured by HPLC. The Cultivation conditions were monitored and controlled column and the program used are described elsewhere at the following set points: pH, 5.0 with addition of 15% [60]. RSD was invariably less than 1%. (v/v) ammonium hydroxide; temperature, 25 °C; stir- ring rate, 700 rpm; air flow, 0.8 vvm and ­pO2 values Off‑gas analyses were variable depending on the dilution rate. p ­ O2 values A BlueInOne Cell gas analyzer (BlueSens, Herten, Ger- were above 20% in all conditions tested. An exhaust gas many) was used for monitoring the ­CO2 and ­O2 molar condenser with cooling water at 4 °C minimized mass fraction of the chemostat cultivations off-gas. Off-gas loses by water evaporation and other possible volatile pressure and humidity measurements were used to cal- compounds. culate the oxygen uptake rate (OUR), carbon dioxide evo- A broad range of dilution rates were covered for the lution rate (CER), and their corresponding specific rates three expression systems tested. Taking into considera- ­(qO2 and ­qCO2) and respiratory quotient (RQ), as previ- tion that 0.19 ­h−1 was the P. pastoris µmax of GAP-C at ously described. RSD was less than 5% in all cases. 25 °C (data not shown), the following dilution rates were used: 0.05 ­h−1, 0.10 ­h−1 and 0.15 ­h−1 (dilution rates were Enzymatic analyses tested as low, middle and high growth rate conditions, CalB activity was determined by an esterase activity respectively). In order to ensure that the steady state was assay, similar to the assay described by Krainer et al. [61]. reached, the stability of the parameters of interest were Briefly, 100 μL of culture supernatant was mixed with Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 10 of 12 900 μL of fresh assay solution containing 4 mM p-NPB concentration ­(OD600). S1A–PGAP-based clone GAP-C; S1B–PPDF-based in 300 mM Tris–HCl, pH 7.4, 1% acetone. The absorb- clone PDF-C; S1C–PUPP-based clone UPP-C. ance increase at 405 nm was monitored at 30 °C for 2 min Additional file 2: Table S2.A, table listing the primer pairs used for gene (Specord 200 Plus spectrophotometer from Analytic Jena dosage analyses and relative transcription levels (RTL) determination by Germany). One activity unit was defined as the amount means of ddPCR and qPCR, respectively. B, table that presents the gene dosage determination of CalB producer clones by digital droplet PCR of enzyme needed to release 1 μmol of p-nitrophenol per (ddPCR). Analyses were performed by triplicates, using Actin gene (ACT1) minute under assay conditions. RSD was less than 4%. as reference. Two positive controls which contain 3 and 5 copies of the SDS-PAGE analysis were performed with the culture expression cassette for Candida rugosa lipase 1 (CRL1) were also analyzed as controls. supernatants collected during the bioreactor cultiva- Additional file 3: Figure S3. Figure that presents the analysis of the prod‑ tions, which were diluted into water prior to loading uct quality by SDS-PAGE. Samples from fed-batch cultivations—PDF-C into precast 4–20% kD Criterion TGX Stain-Free Gel and UPP-C—run at A: µsp = 0.10 ­h−1 and B: µsp = 0.05 ­h−1 were analyzed. (Biorad, Hercules, CA, USA). Fifteen µL of the diluted Different samples obtained at different feeding time (FT) supernatants were loaded on SDS-PAGE. BSA standards at different concentrations were samples were mixed with 5 µL of 4 × loading buffer (20% also loaded in SDS-PAGE as reference (lanes 1–4). A. µsp = 0.10 ­h−1.PDF 7: glycerol, 4% SDS, 0.3 mM bromophenol blue (Merck), 17h FT; PDF 8: 18h FT; PDF 9: 20h FT; PDF 10: 21h FT; UPP 8: 19h FT; UPP 9: 10% β-mercaptoethanol, 0.1 M Tris, pH 6.8), and incu- 21h FT; UPP 10: 22h FT. B. µsp = 0.05 ­h−1. PDF 1: Batch end, PDF 5: 19h FT; PDF 7: 24h FT; PDF 8: 27h FT; PDF 9: 36h FT; PDF 10: 41h FT; PDF 11: 43h FT; bated at 95 °C for 5 min. Samples were cooled, centri- UPP 1: Batch end, UPP 7: 27h FT; UPP 8: 37h FT; UPP 9: 40h FT; UPP 10: 42h fuged quickly, and 19 µL loaded into 4–20% kD Criterion FT; UPP 11:44h FT. TGX Stain-Free Gel together with Precision Plus Protein molecular weight marker (Biorad, Hercules, CA, USA). Acknowledgements The visualization of the gel was performed using the Gel The Spanish group is member of 2017-SGR-1462 and the Reference Network Doc EZ (Biorad, Hercules, CA, USA). in Biotechnology (XRB, Generalitat de Catalunya). JGM acknowledges a PIF scholarship from the Universitat Autònoma de Barcelona. MAN acknowledges award by the National Council of Science Technology and Technological Innovation (CONCYTEC) through its executing unit and the National Fund for Process parameter determination, consistency test Scientific, Technological and Technological Innovation Development (FON‑ and data reconciliation DECYT). This work was performed within the EU project IBISBA1.0 including Mass balance and stoichiometric equations the two IBISBA partners UAB and VTT in partnership with the bisy GmbH. EU- IBISBA obtained an ESFRI status in 2018, with the aim to accelerate biotechnol‑ All equations used to calculate yields and rates are based ogy and synthetic biology activities in Europe through service infrastructure on mass balances at continuous and fed-batch opera- and know-how. EU-IBISBA is gratefully acknowledged for facilitating part of tion and can be found elsewhere [11, 37]. The P. pasto- the work in this study through its trans-national access that received funding from the European Union’s Horizon 2020 research and innovation program ris elemental composition grown on glycerol as the sole (IBISBA 1.0 project) under grant N° 730,976. C-source was determined as previously reported [62]. Carbon and electron balances were checked and less than Authors’ contributions JG-M: Chemostat cultivations, tech transfer between UAB and VTT, data 5% of deviation was observed prior to reconciliation. treatment for UAB and VTT fermentations, transcriptional analysis, manu‑ script drafting; KV: design and execution of fed-batch fermentations at VTT, enzymatic assays, interpretation of data and contributed to the drafting of Consistency test and data reconciliation manuscript; MAN-T: Strain construction, clone screening and transcriptional Measurement consistency was checked by using the analysis; JT: tech transfer and fed-batch fermentations at VTT, data analysis; JU: participated in the design of experiments and assisted drafting manuscript; PT: standard test with carbon and electron balances as con- Coordination of partners and commented manuscript draft; CS: participated straints. Both online and offline measurements enabled in the strain engineering of CalB producer clones from Bisy; IT: Promoter and the calculation of five key specific rates in the black-box strain design and qPCR results interpretation; KM: qPCR results interpretation and manuscript preparation; MP: participated in the interpretation of results process model: biomass generation (μ), glucose uptake and assisted drafting manuscript; JLM-S: Assistance on data treatment, data rate (qs), product generation rate (qp), oxygen uptake interpretation and manuscript drafting; FV: Data interpretation and assistance rate (qO2) and carbon dioxide production rate (qCO2). The on manuscript drafting; AG: Conceived the study, participated in the interpre‑ tation of results and assisted manuscript drafting; XG-O: Conceived the study, method used for this purpose is described in detail else- participated in the interpretation of results and assisted manuscript drafting. where [11]. All authors read and approved the final manuscript. Funding Supplementary Information This work was funded by the EU-IBISBA through its trans-national access The online version contains supplementary material available at https://​doi.​ funded by the European Union’s Horizon 2020 research and innovation pro‑ org/​10.​1186/​s12934-​021-​01564-9. gram (IBISBA 1.0 project) under grant N° 730,976. Availability of data and materials Additional file 1: Table S1. Figures that present as landscapes the The datasets and materials used and/or analysed during the current study are second round of clone screening results performed for the expression available from the corresponding author on reasonable request. systems studied. Mean values for each and standard deviations are plotted and expressed in terms of activity units (AU) normalized per biomass Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 11 of 12 Declarations 15. Çalık P, Ata Ö, Güneş H, Massahi A, Boy E, Keskin A, Örtürk S, Zerze GH, Özdamar TH. Recombinant protein production in Pichia pastoris under Ethics approval and consent to participate glyceraldehyde-3-phosphate dehydrogenase promoter: from carbon Not applicable. source metabolism to bioreactor operation parameters. Biochem Eng J. 2015;95:20–36. Consent for publication 16. Hartner FS, Ruth C, Langenegger D, Johnson SN, Hyka P, Lin-Cereghino Not applicable. GP, Lin-Cereghino J, Kovar K, Cregg JM, Glieder A. Promoter library designed for fine-tuned gene expression in Pichia pastoris. Nucleic Acids Competing interests Res. 2008;36:e76–e76. The authors of Bisy GmbH and BioGrammatics declare that they have interests 17. Berrios J, Flores MO, Díaz-Barrera A, Altamirano C, Martínez I, Cabrera Z. A to commercialize the promoters used for this recombinant protein production comparative study of glycerol and sorbitol as co-substrates in methanol- study. The authors of UAB and VTT have no any competing interests. induced cultures of Pichia pastoris: temperature effect and scale-up simulation. J Ind Microbiol Biotechnol. 2017;44:407–11. Author details 18. Tschopp JF, Brust PF, Cregg JM, Stillman CA, Gingeras TR. Expression of 1 Department of Chemical, Biological and Environmental Engineering, Univer‑ the lacZ gene from two methanol-regulated promoters in Pichia pastoris. sitat Autònoma de Barcelona, 08193 Bellaterra (Cerdanyola del Vallès), Spain. Nucleic Acids Res. 1987;15:3859–76. 2 Solutions for Natural Resources and Environment, VTT Technical Research 19. Portela RMC, Vogl T, Ebner K, Oliveira R, Glieder A. Pichia pastoris Alcohol Centre of Finland Ltd, Tietotie 2, Espoo, Finland. 3 Keck Graduate Institute, 535 Oxidase 1 (AOX1) core promoter engineering by high resolution system‑ Watson Dr, Claremont, CA, USA. 4 Biogrammatics Inc, 2794 Loker Ave, West, atic mutagenesis. Biotechnol J. 2018;13:e1700340. Suite 103, Carlsbad, CA 92010, USA. 5 Institute of Molecular Biotechnology, 20. Rajamanickam V, Metzger K, Schmid C, Spadiut O. A novel bi-directional Graz University of Technology, Petersgasse 14, 8010 Graz, Austria. 6 Bisy GmbH, promoter system allows tunable recombinant protein production in Wuenschendorf 292, 8200 Hofstaetten/Raab, Austria. Pichia pastoris. Microb Cell Fact. 2017;16:152. 21. Prielhofer R, Reichinger M, Wagner N, Claes K, Kiziak C, Gasser B, Mat‑ Received: 21 October 2020 Accepted: 12 March 2021 tanovich D. Superior protein titers in half the fermentation time: Promoter and process engineering for the glucose-regulated GTH1 promoter of Pichia pastoris. Biotechnol Bioeng. 2018;115:2479–88. 22. Landes N, Gasser B, Vorauer-Uhl K, Lhota G, Mattanovich D, Maurer M. The vitamin-sensitive promoter ­PTHI11 enables pre-defined autonomous References induction of recombinant protein production in Pichia pastoris. Biotech‑ 1. Cereghino GPL, Cereghino JL, Ilgen C, Cregg JM. Production of recombi‑ nol Bioeng. 2016;113:2633–43. nant proteins in fermenter cultures of the yeast Pichia pastoris. Curr Opin 23. Wang J, Zhang T, Li Y, Li L, Wang Y, Yang B, Wang Y. High-level expression Biotech. 2002;13:329–32. of Thermomyces dupontii thermo-alkaline lipase in Pichia pastoris under 2. Ahmad M, Hirz M, Pichler H, Schwab H. Protein expression in Pichia the control of different promoters. Biotech. 2019;9:33. pastoris: recent achievements and perspectives for heterologous protein 24. Shirvani R, Yazdanpanah S, Barshan-tashnizi M, Shahali M. A Novel production. Appl Microbiol Biotechnol. 2014;98:5301–17. Methanol-Free Platform for Extracellular Expression of rhGM-CSF in Pichia 3. Yang Z, Zhang Z. Engineering strategies for enhanced production of pastoris. Mol Biotechnol. 2019;61:521–7. protein and bio-products in Pichia pastoris: a review. Biotechnol Adv. 25. Vogl T, Fischer JE, Hyden P, Wasmayer R, Sturmberger L, Glieder A. Orthol‑ 2018;36:182–95. ogous promoters from related methylotrophic yeasts surpass expression 4. Juturu V, Wu JC. Heterologous protein expression in Pichia pastoris: latest of endogenous promoters of Pichia pastoris. AMB Express. 2020;10:38. research progress and applications. ChemBioChem. 2018;19:7–21. 26. Liu J, Han Q, Cheng Q, Chen Y, Wang R, Li X, Liu Y, Yan D. Efficient expres‑ 5. Vogl T, Hartner FS, Glieder A. New opportunities by synthetic biology for sion of human lysozyme through the increased gene dosage and co- biopharmaceutical production in Pichia pastoris. Curr Opin Biotechnol. expression of transcription factor Hac1p in Pichia pastoris. Curr Microbiol. 2013;24:1094–101. 2020;77:846–54. 6. Looser V, Bruhlmann B, Bumbak F, Stenger C, Costa M, Camattari A, Foti‑ 27. Ben Azoun S, Belhaj AE, Göngrich R, Gasser B, Kallel H. Molecular optimi‑ adis D, Kovar K. Cultivation strategies to enhance productivity of Pichia zation of rabies virus glycoprotein expression in Pichia pastoris. Microb pastoris: a review. Biotechnol Adv. 2015;33:1177–93. Biotechnol. 2016;9:355–68. 7. Cregg JM, Tolstorukov I, Kusari A, Sunga J, Madden K, Chappell T. Expres‑ 28. Yang J, Lu Z, Chen J, Chu P, Cheng Q, Liu J, Ming F, Huang C, Xiao A, Cai H, sion in the yeast Pichia pastoris. Methods Enzymol. 2009;13:169–89. Zhang L. Effect of cooperation of chaperones and gene dosage on the 8. Peña DA, Gasser B, Zanghellini J, Steiger MG, Mattanovich D. Metabolic expression of porcine PGLYRP-1 in Pichia pastoris. Appl Microbiol Biotech‑ engineering of Pichia pastoris. Metab Eng. 2018;50:2–15. nol. 2016;100:5453–65. 9. Barrigón JM, Montesinos JL, Valero F. Searching the best operational 29. Vogl T, Sturmberger L, Kickenweiz T, Wasmayer R, Schmid C, Hatzl AM, strategies for Rhizopus oryzae lipase production in Pichia pastoris Mut+ Gerstmann MA, Pitzer J, Wagner M, Thallinger GG, Geier M, Glieder A. A phenotype: Methanol limited or methanol non-limited fed-batch cul‑ toolbox of diverse promoters related to methanol utilization: functionally tures? Biochem Eng J. 2013;75:47–54. verified parts for heterologous pathway expression in Pichia pastoris. ACS 10. Barrigon JM, Valero F, Montesinos JL. A macrokinetic model-based Synth Biol. 2016;5:172–86. comparative meta-analysis of recombinant protein production by Pichia 30. Cámara E, Landes N, Albiol J, Gasser B, Mattanovich D, Ferrer P. Increased pastoris under AOX1 promoter. Biotechnol Bioeng. 2015;112:1132–45. dosage of AOX1 promoter-regulated expression cassettes leads to tran‑ 11. Ponte X, Montesinos-Seguí JL, Valero F. Bioprocess efficiency in Rhizopus scription attenuation of the methanol metabolism in Pichia pastoris. Sci oryzae lipase production by Pichia pastoris under the control of P ­ AOX1 is Rep. 2017;7:44302. oxygen tension dependent. Process Biochem. 2016;51:1954–63. 31. Cámara E, Monforte S, Albiol J, Ferrer P. Deregulation of methanol 12. Ponte X, Barrigón JM, Maurer M, Mattanovich D, Valero F, Montesinos- metabolism reverts transcriptional limitations of recombinant Pichia pas- Seguí JL. Towards optimal substrate feeding for heterologous protein toris (Komagataella spp) with multiple expression cassettes under control production in Pichia pastoris (Komagataella spp) fed-batch processes of the AOX1 promoter. Biotechnol Bioeng. 2019;116:1710–20. under ­PAOX1 control: a modeling aided approach. J Chem Technol Bio‑ 32. Krause M, Ukkonen K, Haataja T, Ruottinen M, Glumoff T, Neubauer technol. 2018;93:3208–18. A, Neubauer P, Vasala A. A novel fed-batch based cultivation method 13. Vogl T, Glieder A. Regulation of Pichia pastoris promoters and its conse‑ provides high cell-density and improves yield of soluble recombinant quences for protein production. N Biotechnol. 2013;30:385–404. proteins in shaken cultures. Microb Cell Fact. 2010;9:11. 14. Heyland J, Fu J, Blank LM, Schmid A. Quantitative physiology of Pichia 33. Panula-Perälä J, Šiurkus J, Vasala A, Wilmanowski R, Casteleijn MG, Neu‑ pastoris during glucose-limited high-cell density fed-batch cultivation for bauer P. Enzyme controlled glucose auto-delivery for high cell density recombinant protein production. Biotechnol Bioeng. 2010;107:357–68. cultivations in microplates and shake flasks. Microb Cell Fact. 2008;7:31. Garrigós‑Martínez et al. Microb Cell Fact (2021) 20:74 Page 12 of 12 34. Nieto-Taype MA, Garcia-Ortega X, Albiol J, Montesinos-Seguí JL, Valero F. 49. Garcia-Ortega X, Ferrer P, Montesinos JL, Valero F. Fed-batch operational Continuous cultivation as a tool toward the rational bioprocess develop‑ strategies for recombinant Fab production with Pichia pastoris using the ment with Pichia Pastoris cell factory. Front Bioeng Biotechnol. 2020;8:632. constitutive GAP promoter. Biochem Eng J. 2013;79:172–81. 35. Potvin G, Ahmad A, Zhang Z. Bioprocess engineering aspects of heter‑ 50. Theron CW, Vandermies M, Telek S, Steels S, Fickers P. Comprehensive ologous protein production in Pichia pastoris: a review. Biochem Eng J. comparison of Yarrowia lipolytica and Pichia pastoris for production of 2012;64:91–105. Candida antarctica lipase B. Sci Rep. 2020;10:1–9. 36. Rebnegger C, Graf AB, Valli M, Steiger MG, Gasser B, Maurer M, Mat‑ 51. Puxbaum V, Mattanovich D, Gasser B. Quo vadis? The challenges of tanovich D. In Pichia pastoris, growth rate regulates protein synthesis and recombinant protein folding and secretion in Pichia pastoris. Appl Micro‑ secretion, mating and stress response. Biotechnol J. 2014;9:511–25. biol Biotechnol. 2015;99:2925–38. 37. Garcia-Ortega X, Adelantado N, Ferrer P, Montesinos JL, Valero F. A step 52. Zhu T, Guo M, Zhuang Y, Chu J, Zhang S. Understanding the effect of for‑ forward to improve recombinant protein production in Pichia pastoris: eign gene dosage on the physiology of Pichia pastoris by transcriptional from specific growth rate effect on protein secretion to carbon-starving analysis of key genes. Appl Microbiol Biotechnol. 2011;89:1127–35. conditions as advanced strategy. Process Biochem. 2016;51:681–91. 53. Raschmanová H, Zamora I, Borčinová M, Meier P, Weninger A, Mächler D, 38. Nieto-Taype MA, Garrigós-Martínez J, Sánchez-Farrando M, Valero F, Gar‑ Glieder A, Melzoch K, Knejzlík Z, Kovar K. Single-cell approach to monitor cia-Ortega X, Montesinos-Seguí JL. Rationale-based selection of optimal the unfolded protein response during biotechnological processes with operating strategies and gene dosage impact on recombinant protein Pichia pastoris. Front Microbiol. 2019;10:335. production in Komagataella phaffii (Pichia pastoris). Microb Biotechnol. 54. Sha C, Yu XW, Lin NX, Zhang M, Xu Y. Enhancement of lipase r27RCL pro‑ 2020;13:315–27. duction in Pichia pastoris by regulating gene dosage and co-expression 39. Garrigós-Martínez J, Nieto-Taype MA, Gasset-Franch A, Montesinos-Seguí with chaperone protein disulfide isomerase. Enzyme Microb Technol. JL, Garcia-Ortega X, Valero F. Specific growth rate governs AOX1 gene 2013;53:438–43. expression, affecting the production kinetics of Pichia pastoris (Komaga- 55. Vogl T, Gebbie L, Palfreyman RW, Speight R. Effect of plasmid design and taella phaffii) ­PAOX1-driven recombinant producer strains with different type of integration event on recombinant protein expression in Pichia target gene dosage. Microb Cell Fact. 2019;18:187. pastoris. Appl Environ Microbiol. 2018;84:e02712-e2717. 40. García-Ortega X, Cámara E, Ferrer P, Albiol J, Montesinos-Seguí JL, Valero 56. Lin-Cereghino J, Wong WW, Xiong S, Giang W, Luong LT, Vu J, Johnson SD, F. Rational development of bioprocess engineering strategies for recom‑ Lin-Cereghino GP. Condensed protocol for competent cell preparation binant protein production in Pichia pastoris (Komagataella phaffii) using and transformation of the methylotrophic yeast Pichia pastoris. Biotech‑ the methanol-free GAP promoter Where do we stand? New Biotechnol. niques. 2005;38:44. 2019;53:24–34. 57. Näätsaari L, Mistlberger B, Ruth C, Hajek T, Hartner FS, Glieder A. Deletion 41. De Macedo RJ, Garcia-Ortega X, Montesinos-Seguí JL, Guimaraes Freire of the Pichia pastoris ku70 homologue facilitates platform strain genera‑ DM, Valero F. Continuous operation, a realistic alternative to fed-batch tion for gene expression and synthetic biology. PLoS ONE. 2012;7:e39720. fermentation for the production of recombinant lipase B from Candida 58. Krainer FW, Gerstmann MA, Darnhofer B, Birner-Gruenberger R, Glieder A. antarctica under the constitutive promoter PGK in Pichia pastoris. Bio‑ Biotechnological advances towards an enhanced peroxidase production chem Eng J. 2019;147:39–47. in Pichia pastoris. J Biotechnol. 2016;233:181–9. 42. Canales C, Altamirano C, Berrios J. Effect of dilution rate and metha‑ 59. Cos O, Serrano A, Montesinos JL, Ferrer P, Cregg JM, Valero F. Combined nol-glycerol mixed feeding on heterologous Rhizopus oryzae lipase effect of the methanol utilization (Mut) phenotype and gene dosage on production with Pichia pastoris Mut+ phenotype in continuous culture. recombinant protein production in Pichia pastoris fed-batch cultures. J Biotechnol Prog. 2015;31:707–14. Biotechnol. 2005;116:321–35. 43. Fischer JE, Hatzl A-M, Weninger A, Schmid C, Glieder A. Methanol 60. Garcia-Ortega X, Valero F, Montesinos-Seguí JL. Physiological state Independent Expression by Pichia Pastoris Employing De-repression as transferable operating criterion to improve recombinant protein Technologies. J Vis Exp. 2019. https://​doi.​org/​10.​3791/​58589. production in Pichia pastoris through oxygen limitation. J Chem Technol 44. Vogl T, Ruth C, Pitzer J, Kickenweiz T, Glieder A. Synthetic core promoters Biotechnol. 2017;92:2573–82. for Pichia pastoris. ACS Synth Biol. 2014;3:188–91. 61. Krainer FW, Dietzsch C, Hajek T, Herwig C, Spadiut O, Glieder A. Recom‑ 45. Weis R, Luiten R, Skranc W, Schwab H, Wubbolts M, Glieder A. Reliable binant protein expression in Pichia pastoris strains with an engineered high-throughput screening with Pichia pastoris by limiting yeast cell methanol utilization pathway. Microb Cell Fact. 2012;11:22. death phenomena. FEMS Yeast Res. 2004;5:179–89. 62. Tomàs-Gamisans M, Ferrer P, Albiol J. Fine-tuning the P pastoris iMT1026 46. Schwarzhans JP, Wibberg D, Winkler A, Luttermann T, Kalinowski J, Friehs genome-scale metabolic model for improved prediction of growth K. Integration event induced changes in recombinant protein productiv‑ on methanol or glycerol as sole carbon sources. Microb Biotechnol. ity in Pichia pastoris discovered by whole genome sequencing and 2018;11:224–37. derived vector optimization. Microb Cell Fact. 2016;15:84. 47. Delic M, Göngrich R, Mattanovich D, Gasser B. Engineering of pro‑ tein folding and secretion-strategies to overcome bottlenecks for Publisher’s Note efficient production of recombinant proteins. Antioxid Redox Signal. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 2014;21:414–37. lished maps and institutional affiliations. 48. Young CL, Yuraszeck T, Robinson AS. Decreased secretion and unfolded protein response upregulation. Methods Enzymol. 2011;491:235–60. 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