This is the accepted version of the review article: Garcia-Ortega, Xavier; Cámara, Elena; Ferrer, Pau; [et al.]. «Rational develop- ment of bioprocess engineering strategies for recombinant protein production in Pichia pastoris (Komagataella phaffii) using the methanol-free GAP promoter. Where do we stand?». New Biotechnology, Vol. 53 (November 2019), p. 24-34. DOI 10.1016/j.nbt.2019.06.002 This version is available at https://ddd.uab.cat/record/281244 under the terms of the license 1 Rational development of bioprocess engineering strategies for recombinant protein 2 production in Pichia pastoris (Komagataella phaffi) using the methanol free GAP 3 promoter. Where do we stand? 4 5 Xavier García-Ortega1 6 Elena Cámara1‡ 7 Pau Ferrer1 8 Joan Albiol1 9 José Luis Montesinos-Seguí1 10 Francisco Valero1* 11 12 1Department of Chemical, Biological and Environmental Engineering, School of Engineering, 13 Universitat Autònoma de Barcelona, 08193 Bellaterra (Cerdanyola del Vallès), Spain. 14 ‡Present address: Department of Biology and Biological Engineering, Division of Industrial 15 Biotechnology, Chalmers University of Technology, 412 96 Gothenburg, Sweden. 16 17 18 19 *Correspondence: Dr Francisco Valero, Departament d’Enginyeria Química, Biològica i 20 Ambiental, Universitat Autònoma de Barcelona, 08193 Bellaterra (Cerdanyola del Vallès), Spain. 21 E-mail: francisco.valero@uab.cat 22 23 24 Keywords: Bioprocess Engineering, Bioprocess efficiency, Methanol-free alternatives, 25 Industrial recombinant protein production, Komagataella phaffi, Pichia pastoris. 26 1 27 List of abbreviations 28 - Antibody fragment (Fab) 29 - Compound annual growth rate (CAGR) 30 - Constitutive GAP promoter (glyceraldehyde-3-phosphate dehydrogenase, PGAP) 31 - Dissolved oxygen concentration (pO2) 32 - Food and Drug Administration (FDA) 33 - Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) 34 - Heat yield (YQ/X) 35 - Methanol inducible AOX1 promoter (alcohol oxidase 1, PAOX1). 36 - Oxygen to biomass yield (YO2/X) 37 - Post-translational modifications (PTM) 38 - Product to biomass yield (YP/X) 39 - Product to substrate yield (YP/S) 40 - Recombinant protein production (RPP) 41 - Respiratory quotient (RQ) 42 - Single-cell protein (SCP) 43 - Space-time-yield (STY) 44 - Specific ethanol production rate (qethanol) 45 - Specific growth rate (μ) 46 - Specific product formation rate (qP) 47 - Specific substrate uptake rate (qS) 48 - Temperature (T) 49 - Unfolded protein responses (UPR) 50 - Volumetric oxygen mass transfer coefficient (kLa) 51 - Volumetric productivity (QP) 52 2 53 Abstract 54 The increasing demand for recombinant proteins for a wide range of applications, from 55 biopharmaceutical protein complexes to industrial enzymes, is leading to important growth in this 56 market. Among the different efficient host organism alternatives commonly used for protein 57 production, the yeast Pichia pastoris (Komagataella phaffi) is currently considered to be one of 58 the most effective and versatile expression platforms. The promising features of this cell factory 59 are giving rise to interesting studies covering the different aspects that contribute to improving 60 the bioprocess efficiency, from strain engineering to bioprocess engineering. The numerous 61 drawbacks of using methanol in industrial processes are driving interest towards methanol-free 62 alternatives, among which the GAP promoter-based systems stand out. The aim of this work is to 63 present the most promising innovative developments in operational strategies based on rational 64 approaches through bioprocess engineering tools. This rational design should be based on 65 physiological characterization of the producing strains under bioprocess conditions and its 66 interrelation with specific rates. This review focuses on understanding the key factors that can 67 enhance recombinant protein production in Pichia pastoris; they are the basis for a further 68 discussion on future industrial applications with the aim of developing scalable alternative 69 strategies that maximize yields and productivity. 70 71 Introduction 72 73 Currently, proteins are used in a wide range of applications for industries, including 74 pharmaceuticals, diagnostics, food, detergents, biofuels, textiles, polymers, and paper and pulp 75 [1,2]. By using recombinant DNA technology, proteins of interest can be produced in different 76 cell factories obtaining expression levels hundreds of times higher than those produced naturally 77 [3]. The market for recombinant proteins, including from biopharmaceutical protein complexes 78 to industrial enzymes, is growing steadily. The market for biopharmaceuticals is expected to reach 79 $386.7 billion by the end of 2019 at a compound annual growth rate (CAGR) of 10.6% [4,5], 3 80 while for industrial enzymes, it is predicted to reach $6.3 billion in 2021 at a CAGR of 4.7% [6]. 81 High demand and continuous market expansion provide incentives to improve the protein 82 production platforms, both to enable the production of novel proteins and to reduce the 83 manufacturing costs of the existing processes [7]. 84 85 The choice of a suitable cell factory for the production of a recombinant protein is therefore very 86 important and probably one of the most challenging and critical steps that should be addressed at 87 the beginning of the bioprocess design [8,9]. In a perfect scenario, one universal expression 88 system would enable the expression of all possible recombinant genes in a fast, cheap, and 89 accurate manner with respect to yield, folding and biological activity [10]. However, the 90 limitations of the current existing systems require empirical and specific determination of the 91 most suitable cell factory among the potential candidates. The selection of the expression system 92 must be based on production parameters, such as costs, yield, production timescale, scale up 93 capacity and downstream processes, as well as on the properties and use of the product [11]. 94 Different applications have different requirements for both quantity and quality. The 95 characteristics of the target protein must also be taken into account in terms of structure and the 96 requirement to incorporate post-translational modifications (PTM) [10]. Glycosylation represents 97 the most complex and widespread PTM, being associated with 40% of all approved products [12]. 98 Selecting a wrong expression host can result in the protein being misfolded or poorly expressed, 99 lacking the necessary PTM’s or containing unsuitable modifications. 100 101 In general, bacteria provide an excellent expression system for proteins that do not require 102 synthesis in a glycosylated or extensively modified form, allowing fast and inexpensive 103 production processes [8]. Among the different alternatives, Escherichia coli has been considered 104 the workhorse for recombinant protein production (RPP), but Bacillus species, as well as other 105 bacterial organisms, such as Pseudomonas fluorescens, Staphylococcus carnosus and 106 Streptomyces lividans, are also efficiently used as cell factories for RPP [13,14]. Interestingly, 4 107 significant progress has been made over the past recent years towards the development of E. coli 108 strains able to perform eukaryotic-like PTM’s such as glycosylation, as well as engineering 109 efficient protein secretion systems. In contrast to bacteria, mammalian cell lines should be able 110 to overcome most of the limitations of producing recombinant eukaryotic proteins. Consequently, 111 due to the capacity to perform human PTM’s, as well as to correctly fold and assemble human 112 proteins, they have become the dominant recombinant production system for medical applications 113 [16]. They still present important drawbacks, such as low growth rates and product yield, high 114 costs derived from the use of complex media and sera or potential viral contaminations [17]. In 115 addition, in terms of bioprocessing, the operational mode, process control and scaling-up steps 116 are also challenging [18]. Consequently the specific yields obtained with these bioprocesses based 117 on mammalian cell lines are often low [17]. 118 119 In this context, eukaryotic microorganisms such as yeast emerge as an ideal intermediate. They 120 combine the eukaryotic ability for protein processing, such as folding, assembling and introducing 121 PTMs, with important microbial advantages such as the capacity to grow quickly to high cell 122 densities in chemically defined media. In addition, yeasts can secrete recombinant proteins into 123 the extracellular media, greatly facilitating downstream processing [19]. Yeasts were first 124 implemented as a recombinant protein platform in 1981 using Saccharomyces cerevisiae [20]. 125 Since then, other yeasts have been developed into alternative cell factories presenting some clear 126 advantages, with prime examples including Pichia pastoris, recently reclassified as Komagataella 127 phaffi, as well as other yeasts, such as Hansenula polymorpha [21] or Kluyveromyces lactis [22]. 128 P. pastoris is currently considered to be one of the most effective and versatile systems for the 129 production of heterologous proteins [23]. This methylotrophic organism was developed as a host 130 system in 1985 [24]. Its primary advantages are summarized in Table 1. Currently, over 5,000 131 recombinant proteins have been expressed in the P. pastoris system reaching expression levels 132 approximately 80% of the total secreted protein or up to 30% of the total cell protein [23] 133 [www.pichia.com]. Due to its positive features, over 300 industrial processes have been licensed 5 134 and more than 70 commercial products made in P. pastoris are currently on the market. The 135 range of application fields is very wide from industrial enzymes (nitrate reductase), to animal feed 136 additives (phytase) and biopharmaceutical proteins. Interestingly, two biopharmaceuticals have 137 been approved by the US Food and Drug Administration (FDA) and the European Commission 138 (Kalbitor® and Jetrea®) [25] and over 20 therapeutic product candidates are still in the clinical 139 pipeline [15,26]. Due to the high potential of P. pastoris, several reviews have been published 140 recently [27–32]. However, none covers specifically novel aspects focused on bioprocess 141 engineering to improve the current fermentation strategies, especially those focused on avoiding 142 the use of methanol, since this compound involves numerous drawbacks for industrial application. 143 Up-to-date fermentation technology has not reached yet the same maturity as traditional chemical 144 processes, which is essential to increase the product accessibility in low and middle resource 145 countries [33,34]. 146 147 This review aims to update and discuss the most promising innovative developments in methanol- 148 free P. pastoris cultivation strategies, mostly focused on constitutive GAP promoter 149 (glyceraldehyde-3-phosphate dehydrogenase, PGAP), and based on rational approaches through the 150 efficient application of bioprocess engineering tools. Special attention has been given to the 151 potential scalability and industrial implementation feasibility towards improvement of the overall 152 bioprocess efficiency and profit. 153 154 Operational strategies for P. pastoris bioprocesses: from conventional approaches to 155 rational innovative strategies 156 157 Even though the current interest in P. pastoris as a cell factory for different production processes 158 is rapidly increasing [2], the fermentation technology developed for this yeast is still far from 159 achieving the maturity of other common organisms used in the microbial fermentation industry, 6 160 such as E. coli or S. cerevisiae [35]. As with other production processes, a fermentation aims to 161 produce the maximum amount of product in the minimum process time (space-time-yield, STY) 162 and volumetric productivity, (QP [35]). For the design and implementation of innovative 163 bioprocess strategies, it is important to select reference parameters independent of process- 164 specific settings, such as specific growth rate (μ), specific substrate uptake rate (qS), specific 165 production rate (qP) or yields. Instead, other parameters non-comparable between different 166 systems, such as biomass concentration, reactor volume, process time or addition feed rates 167 should be avoided [35]. In general, the efficiency of the fermentation processes is importantly 168 influenced by different operational parameters including temperature (T), pH, osmolality, 169 dissolved oxygen concentration (pO2) and medium composition. However, excluding pO2 [36– 170 38], the selection of the optimal set-point as well as the control strategies for these parameters 171 along cultivation are often considered straightforward for P. pastoris bioprocesses. The best 172 ranges are widely known, extensively reviewed and used by most of the authors [39,40]. 173 174 Promoter selection plays an essential role in the bioprocess development, since it critically 175 determines the fermentation strategy. From the different promoters available for P. pastoris 176 described in the literature [41,42], which are in continuous rapid evolution, the most commonly 177 used in industrial processes are still constitutive PGAP and the methanol inducible AOX1 (alcohol 178 oxidase 1, PAOX1). The advantages of the former are primarily related to the non-utilization of 179 methanol as the carbon source and/or inducer; thus, it can be substituted by alternative standard 180 carbon sources, such as glucose or glycerol. Problems related to the storage, delivery and safety 181 of large quantities of methanol would be avoided [40], as well as operational problems associated 182 with methanol metabolism, which increases the complexity of the culture strategies that allow it 183 to reach optimal production performances [43,44]. The PGAP expression system, therefore, 184 contributes greatly to the development of cost-effective bioprocesses for large-scale RPP [45]. In 185 the following sections, the primary operational strategies related to PGAP are summarized. 186 7 187 Carbon source selection. How do the alternatives affect the process? 188 189 The two most frequently used carbon sources for PGAP-based systems are glucose and glycerol. 190 However, fructose, sorbitol, mannitol, ethanol, methanol and trehalose also allow the expression 191 of PGAP-driven proteins. In general, in the literature similar production rates have been achieved 192 using both substrates or slightly higher using glucose [46]. As an example, when comparing both 193 alternatives systematically to produce a human antibody fragment (Fab) in batch process, similar 194 production was obtained. However, growing in conditions of glucose excess, YX/S was lower and 195 by-products, such as ethanol, arabitol and acetate, were detected in the culture broth [47]. Thus, 196 since glycerol does not generate by-products, it should be recommended for batch cultures. On 197 the other hand, when comparing glucose and glycerol in carbon-limited fed-batch strategy at 198 constant µ, no differences in yields and productivities were observed. However, due to both the 199 lower heat yield (YQ/X) and YO2/X for glucose, this substrate should be recommended for the fed- 200 batch phase [48]. For industrial bioprocesses, economic factors such as substrate costs and the 201 potential low-cost alternatives derived from industry by-products should also be taken into 202 consideration. For yeast cell factories, molasses from the sugar industry, glycerol from the 203 biodiesel industry, corn steep liquor from the starch industry, spent sulphite liquor from the forest 204 products industry and cheese whey from the dairy industry are considered possible alternatives 205 [49]. 206 207 Chemostat and fed-batch as tools for physiological characterization 208 209 Currently, in bioprocess engineering, a physiological characterization of the producing strains 210 that includes production rates and yields is considered essential for a proper bioprocess 211 development [35,50,51]. In this regard, is essential to know how changes in environmental culture 212 factors affect the cell factory, including the rearrangement of central carbon, amino acid 213 metabolism and other basic cell functions. These effects can have a direct impact on cell growth, 214 folding stress and vesicular transport, with possible implications in protein secretion, and other 8 215 aspects affecting the production rates and yields [52–55]. To determine these relationships, fed- 216 batch or continuous cultivations are generally performed. In general, continuous mode is the 217 strategy most frequently used to obtain physiological data, due to the robustness that allows 218 working in the stationary state and thus very stable culture conditions, as well as providing an 219 equivalent physiological state for all the cells in the culture. Due to these features, the chemostat 220 is currently considered an excellent tool for systems biology to provide mechanistic explanations 221 for the effect of different process variables in cell factory performance [56]. Although interesting 222 studies have been recently published, fed-batch cultures are often considered less robust and more 223 laborious and time consuming [57–59]. Interestingly, the use of automated micro-fermentation 224 platforms using either microtitre plates as a bioreactor, or micro-bioreactors on a mL scale, is 225 emerging as an interesting potential alternative to minimize the time consuming steps of 226 bioprocess physiological characterization when comparing different strains or producer clones 227 [60–64]. 228 229 Fed-batch, the most frequently used operating mode for recombinant protein production 230 processes 231 232 To date, the fed-batch mode has been widely used for recombinant production processes, since it 233 allows extension of process time, as well as attaining higher biomass and product titre by gradual 234 feeding of the selected substrate(s) [65]. Currently, simple nutrient feeding strategies for the fed- 235 batch phase using pulse additions, constant rates, ramp and/or step-based profiles, which have 236 been commonly used in the basic protocols, may be considered obsolete. The implementation of 237 these simple feeding strategies during the fed-batch phase is not designed to cover the 238 physiological requirements of the cells and thus gives rise to bioprocesses which are far from their 239 optimal performance. A successful feeding strategy for fed-batch cultivation should aim to allow 240 requirements for cell growth alongside the bioprocess. Therefore, physiological knowledge of the 241 host cell is essential to increase its performance in order to reach optimal production rates and 9 242 yields. For instance, exponential feeding profiles provide a straightforward and effective strategy 243 to achieve pseudo-stationary state conditions. In this case, culture reaches a constant and 244 controlled specific growth rate, as well as fairly constant substrate(s) specific consumption and 245 product production rates, i.e. providing a controlled environment to cover cell growth 246 requirements and RPP [40,48]. 247 248 Effect of specific growth rates on protein production 249 250 “Production kinetics”, which correlate specific product formation (qP) and specific biomass 251 growth rate (µ), is usually considered a crucial factor for bioprocess efficiency, regardless of the 252 selected criterion for bioprocess optimization [66]. Thus, since the involved rates are non- 253 extensive variables, they are comparable at different fermentation systems. The product formation 254 kinetics reflect the equilibrium between the various steps until the product is secreted and are 255 subject to numerous physiological factors [39]. Each cell factory producing a specific 256 recombinant protein in a given fermentation mode has a kinetic profile that should be taken into 257 consideration to reach an optimum bioprocess production [67]. Thus, when developing a 258 production process, the qP vs μ relationship has to be empirically determined. 259 260 Compromise between yield and productivity is essential during bioprocess development in order 261 to reach optimal performance. Different approaches are proposed to address this scenario in order 262 to maximize the goal criteria. Usually, the feeding profile is optimized as a decision variable, 263 while in other applications the goal of optimization is to find the best values for µ or qS in which 264 volumetric productivity or space-time yield (STY) is maximized. To apply more precise and robust 265 control schemes, the use of a closed-loop control and software sensors, as well as predictive and 266 adaptive control, generally requires multiple on-line measurements to follow the optimal time 267 profiles of the specific growth rate, biomass or substrate concentration [68–72]. However, if an 268 advanced control system is implemented in an industrial application, it must increase performance 10 269 while keeping the cost lower than the derived benefit and also do this without significantly 270 increasing process complexity and operational and labour costs [73]. Specific case-studies of 271 monitoring and direct feed-back control in P. pastoris are available in the literature [40,74,75]. 272 As an example, a human Fab production applying the exponential feeding strategy showed that 273 the implementation of a constant high μ resulted into the maximum values for yield and 274 productivity [48]. Alternatively, an optimal decreasing μ trajectory was implemented in order to 275 improve production of the same Fab [76]. When comparing the alternatives, as time proceeds 276 during the cultivation process, optimal μ profiles allow maximizing volumetric productivity, 277 while also providing suitable product yields. However, in terms of specific productivity, a high 278 constant μ achieved two-fold higher rates due to the lower fermentation time. 279 280 An updated revision of the different operational fed-batch strategies under PGAP presented in 281 previous studies [23,40,67] is shown in Table 2. For processes based on the PAOX1, different 282 alternatives to optimize the production of some recombinant proteins can also be found in the 283 literature [77–80]. For this system, the non-monotonic product formation kinetics may be coupled 284 to the methanol metabolic burden or to secretory pathway saturation due to its outstanding 285 expression levels [35,81,82]. Thus, using optimal profiles is advisable when the production 286 kinetics are nonlinear [83]. 287 288 Exploiting host physiology knowledge for rational development of bioprocess 289 engineering strategies 290 291 Current trends in bioprocess engineering are directed towards a conceptual approach that aims to 292 design processes based on the properties of both the physiology of the host cell system and the 293 product, as well as the characteristics of the bioreactor equipment available [58,84]. In practice, 294 improved cultivation strategies are usually rationally designed from the physiological 295 characterization of producer strains [35,85]. As a simple but successful example of bioprocess 11 296 engineering development based on knowledge of host cell physiology is the activation of cell 297 stress responses [11]. The recent development of these strategies derives from several reported 298 cases in which increasing cell stress may couple with overexpression of the protein [26]. 299 Therefore, when identifying a stress condition that enhances protein expression, a systematic 300 study on the cellular machinery of cell factories can lead to determination of optimal conditions 301 that allow protein overexpression. 302 303 Currently, detailed transcriptomic, proteomic and metabolomic studies of cellular reactions to 304 environmental stress factors have been performed and published with different microorganisms, 305 including P. pastoris and S. cerevisiae [53,86]. In P. pastoris cultivations, the effects of different 306 media osmolality, temperature, specific growth rate and oxygen supply were compared at the 307 transcriptome and proteome levels, revealing strong transcriptional and expression regulation of 308 core metabolic genes, and in turn, also of the recombinant protein of interest driven by PGAP 309 [53,86–89]. From those, the effects of biomass growth and oxygen supply have been specifically 310 studied in order to develop innovative cultivation strategies in both chemostat and fed-batch 311 modes that exploit the stress conditions leading to recombinant protein overexpression. 312 313 Carbon-starving conditions to enhance recombinant protein production 314 315 It was first reported in 2007 that a significant increase in PGAP –driven protein expression occurred 316 on short-time glucose depletion periods in shake-flask cultures. This work did neither hypothesize 317 about the causes of overexpression nor aimed to exploit this phenomena in large-scale cultures 318 [90]. In a later study, different time combinations of carbon feeding and starvation periods were 319 compared in order to develop a strategy to increase the production in fed-batch cultures. All the 320 strategies compared achieved important increases in terms of product titre and yields (YP/S and 321 YP/X) without affecting growth parameters. However, due to increased bioprocess times, in terms 322 of productivity values, the results were not always better than the conventional exponential 323 profile. Overall, best results were obtained when combining both the shortest feeding and carbon- 12 324 starving periods. The hypothesis proposed to describe the mechanisms leading to the protein 325 overexpression was based on the described yeast response adaptation to changing nutritional 326 states [91,92]. These rapid and complex mechanisms of transcription and translation regulation 327 promote adaptation to new environments, which may include PGAP-regulated genes [93,94]. A 328 similar strategy was previously proposed for the yeast Hansenula polymorpha, which also 329 enhanced RPP compared to conventional feeding profiles [95]. Based on the favourable results, 330 as well as the simplicity of the new operational strategy, the development of feeding profiles that 331 include carbon-starving periods in order to promote target gene overexpression, should be 332 considered highly promising for industrial yeast-based RPP. 333 334 Implementation of oxygen-limiting conditions to increase RPP 335 336 Oxygen availability has been widely studied in P. pastoris recombinant production processes, 337 because its limited transfer capacity can become a bioprocess-limiting factor. Different studies 338 have assessed the dissolved oxygen effect on protein production observing slight differences 339 [36,96]. However, for PGAP-based bioprocesses, it has been reported that oxygen-limited cultures 340 can lead to up to 3-fold increases in terms of qP, [51,53,87]. Interestingly, oxygen-limiting 341 conditions led to similar improvements in a PAOX1 system and S. cerevisiae [36,97–99]. A primary 342 test assessing oxygen-limiting effect in both chemostat and fed-batch cultivation was carried out 343 for a PGAP-driven Fab production [87]. This work described a decrease in biomass production, the 344 generation of ethanol and a significant increase in both qP and volumetric productivity, QP. This 345 environmental stress led to a shift from a respiratory to respiro-fermentative metabolism resulting 346 in the physiological effect mentioned. The metabolic impact increased progressively as oxygen 347 availability decreased. Inter-disciplinary systems biology studies, including transcriptomic, 348 proteomic and metabolomic analyses, were subsequently carried out to understand the cell 349 responses to hypoxic conditions [53,55,100]. Recently, a publication further discussed how to 13 350 implement practically the desired oxygen-limiting conditions for different bioreactors regardless 351 of oxygen transfer capacity [51]. 352 Maximum levels of protein production were always achieved for the most severely oxygen- 353 limiting conditions prior to respiratory pathway collapse. The target gene overexpression may be 354 due to the overall increased transcriptional levels observed for glycolytic pathway genes. In 355 addition, overexpression may also be coupled to further hypoxic effects such as increased 356 transcription of the unfolded protein responses (UPR) genes and changes in membrane fluidity. 357 Overall, the metabolic shift to respiro-fermentative pathways leads to a decrease in biomass 358 yields, generation of secreted by-products (primarily ethanol), and an increase in qS, qCO2 and 359 respiratory quotient (RQ). Specifically for hypoxic Pichia cultures, the trends of specific ethanol 360 production rate (qethanol) and RQ respect to oxygen availability could lead to feasible indirect 361 reporting parameters of oxygen availability for the cells [51,101]. Importantly, it would allow 362 implementation of oxygen-limiting conditions to different fermenters regardless of their oxygen 363 transfer capacity. At the practical level, since the control of qethanol would require a more complex 364 estimation, a priori the RQ on-line determination should be considered more feasible. 365 366 The effect of lack of essential nutrients on RPP 367 368 Both alternatives previously discussed demonstrate a common factor, namely the induction of an 369 environmental stress to the cells caused by the lack of essential nutrients for growth. PGAP natively 370 regulates glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a key NAD-binding enzyme in 371 the glycolysis and gluconeogenesis pathways, and thus the central carbon metabolism. Adaptation 372 responses to stress are expected to regulate yeast central carbon metabolism, the metabolic fluxes 373 through the glycolysis pathway, and, in turn, transcription of the PGAP-regulated genes. Thus, it is 374 not surprising that environmental stress conditions have a direct impact on the levels of RPP. 375 However, an excessive impact of both environmental stress conditions studied can even lead to a 376 detrimental effect on product formation. Therefore, deep and systematic studies to characterize 14 377 the impact on host cell physiology and its productivity are necessary to exploit overexpression of 378 the desired product (Figure 1). In turn, bioprocess engineering will aim to design, develop and 379 implement new operational cultivation strategies that allow maximizing the recombinant 380 production yields and product. 381 382 Towards industrial implementation of P. pastoris bioprocesses 383 384 Scale up of complex bioprocesses is always challenging, since this step in a bioprocess 385 development is not straightforward. Usually, important variation in bioprocess performance and 386 efficiencies are observed, as described in several studies [33,38,75,102–104]. One of the most 387 important factors that can lead to significant loss of bioprocesses efficiency is that, while lab- 388 scale bioreactors are considered perfectly mixed, lack of homogeneity is an issue in large-scale 389 cultivations. The difference in mixing often leads to significant differences in mass and heat 390 transfer in the processes [105]. As a result, spatial gradients of important parameters, such as 391 dissolved gases, pH, temperature or concentration of substrates are all likely to emerge, leading 392 to potential conditions of oxygen limitation or other essential nutrient starvation [106]. Spatial 393 heterogeneity, in turn, leads to important difficulties in monitoring and controlling large-scale 394 cultivation, because the process sensors can only display an average of the whole system [33]. 395 These environmental conditions may impose stress conditions on the cells, which can often 396 importantly affect biomass growth and product generation of industrial-scale bioreactors and 397 consequently bioprocess efficiency being positive or negative. Specifically hypoxia, transient 398 anoxia and nutrient starvation are of great importance in process optimization [38,50,51]. 399 Therefore, the most commonly selected parameters to keep constant between scales for these 400 highly aerobic and high cell density systems are impeller tip speed, volumetric power input, 401 oxygen transfer rates, volumetric oxygen mass transfer coefficient (kLa) or minimum dissolved 402 oxygen concentration. Although bioprocess engineering development of PGAP systems is still in 15 403 its infancy compared with PAOX1 [40], scale up under PGAP should be easier due to the use of 404 glucose/glycerol as main substrate instead of methanol. 405 406 Why to avoid the use of methanol on the way to large-scale production 407 408 From an industrial point of view, a priori primary advantages of using alternative systems to 409 methanol-inducible bioprocesses can be anticipated. Primarily, problems related to storage, 410 delivery and safety of large quantities of methanol would be avoided [40]. In addition, using 411 methanol as a main substrate involves operational problems at high-cell density cultures derived 412 from the methanol catabolism, in particular, high values for oxygen requirements and heat 413 production [44,48,107]. Avoiding the use of methanol also reduces the accumulation of the by- 414 products of methanol metabolism, as well as cell lysis and subsequent release of native protein 415 and potential proteolysis of secreted products of interest [108,109]. Furthermore, accounting for 416 downstream issues, additional steps in order to achieve the total elimination of methanol in the 417 final product, would be rendered necessary [45]. 418 Thus, alternative methanol-free systems are expected to contribute greatly to the development of 419 cost-effective solutions for industrial RPP [40]. Recent strain engineering approaches that exploit 420 the exceptional capacity features of the PAOX1, but in simpler methanol-free de-repressed 421 bioprocesses, could become promising alternatives. Up-to-date conventional methanol inducible 422 strains have been retrofitted into methanol-free systems regulated by glucose/glycerol. For 423 instance, it has been achieved by the derepressed overexpression of key identified PAOX1-related 424 transcription factors (Mxr1, Mit1 and Prm1), which by themselves are suitable to activate the 425 PAOX1-based RPP machinery strongly [110–112]. However, although some reports are starting to 426 address basic parameters such as substrate comparison and production kinetics determination, the 427 bioprocesses developed for these strains are currently still not widely described in the literature 428 [59,113]. Finally, although methanol involves operational drawbacks from an industrial point of 16 429 view, the productivity of the bioprocess, product quality and its corresponding analysis of 430 economic viability will determine the decision about which the expression system to be used. 431 432 Design criteria in bioprocess optimization 433 434 The current trend in bioprocess optimization is to move away from standard protocols towards 435 concepts that enable an operator to adapt particular recommendations to any specific clones/strain 436 or bioreactor [67]. Strain and process engineering strategies should be combined in order to 437 achieve optimal results, avoiding complex media to reduce costs and facilitate downstream 438 processes [23,114], with the aim of increasing economic viability of the bioprocess. However, 439 bottlenecks are generally product-specific and need to be identified in each particular case taking 440 into account economic considerations [29]. 441 442 When optimizing a bioprocess efficiency, usually YP/S, QP and the final titre are selected as design 443 criteria. Significant reduction in capital and operating costs can be achieved when reaching high 444 values of these indexes. These criteria are also importantly affected by the field of application of 445 the product. Usually, YP/S is selected to obtain low added value products, whereas QP and titre are 446 often prioritized to produce high added value products [36,115]. In fermentation, in order to boost 447 bioprocess efficiency it is essential to produce the maximum product amount within the minimum 448 time (QP or space time yield, STY). This goal will be achievable by reaching the maximum 449 biomass amount jointly with high qP within the shortest process time required to recover a product 450 that meets quality specifications demanded for its application [35]. In addition, for high value- 451 added products, the requirement for a high final titre are usually also crucial, since it allows the 452 reduction of costs for downstream processing. Product quality takes into account target attributes, 453 such as biological activity, half-life, immunogenicity and safety, protein glycosylation, and 454 aggregation [67,116]. These may be compromised by release of intracellular proteases and/or host 455 proteins which occurs when the culture suffers cell lysis or death [39,117]. Correct folding and 17 456 functionality of the product are primarily defined by the cell physiological characteristics but can 457 be modified by cultivation parameters [88,118,119]. 458 459 Since in most industrial fermentation processes for RPP product formation is growth-associated, 460 biomass is often considered an unavoidable but crucial waste by-product. In fact, the optimal qP 461 is usually reached when working at high µ [35]. Furthermore, due to the different biological and 462 different physical restrictions (primarily heat and mass transfer), it must also be taken into 463 consideration that the maximum amount of biomass reachable in the bioprocesses must be limited 464 as much as possible. Consequently, higher product amounts may be generated when performing 465 the bioprocesses at lower growth rates, which therefore lead to longer fermentation times before 466 the process is stopped due to system limitations. In this context, for careful optimization of 467 bioprocess performance indicators such as STY or product yield, the trade-off between qP and µ 468 turns to be crucial [50]. In addition, the use of biomass as single-cell protein (SCP), in order to 469 obtain value-added products for food and agriculture industries, can become a viable and 470 interesting circular-economy practice [120,121] 471 472 The industrial trend: From fed-batch to continuous production processes? 473 474 Currently, most protein production processes using yeast as cell factories are still carried out with 475 fed-batch fermentations, which allows higher biomass and product concentration, yields and 476 productivity by preventing substrate inhibition and catabolite repression [122]. P. pastoris 477 cultivations in industrial processes are often carried out in large stirred tank bioreactors (up to 10 478 m3), achieving high amounts of biomass and product [123]. In general, the primary costs of yeast 479 fermentations are usually related to substrate and energy. When reaching very high cell densities, 480 the aeration requirements are coupled with the need for efficient cooling systems. However, 481 considering the overall bioprocess, purification represents a very high proportion of the total 18 482 bioproduct cost, especially for those with high added-value [123]. Finally, the biomass separation 483 from high cell density cultures is also a challenging task in downstream processing [124]. 484 485 Although until now the protein production processes using yeasts are primarily conducted in batch 486 or fed-batch mode, the current trend, as in other industries, is leading to the transition from batch 487 to continuous manufacturing. This trend aims to exploit the benefits of continuous production 488 such as flexible operation, higher productivity and quality, decreased cost, smaller facilities and 489 the integration and the simplification of the processes. However, the use of this cultivation mode 490 still presents important drawbacks, including losses of productivity due to genetic instability of 491 the cells [125], higher risk of contamination and poor short-term flexibility to handle multiple 492 products due to long run times [126–128]. Continuous cultivations are currently being used for 493 industrial recombinant human insulin production, carried out using S. cerevisiae as a cell factory 494 [129]. For P. pastoris, several recombinant proteins have been successfully produced in 495 continuous cultures at laboratory bench-scale for both constitutive (PGAP) and methanol-inducible 496 (PAOX1) expression [38,88,130–133]. A summary of several continuous processes under PGAP is 497 presented in Table 3 [76,134–136]. The change of operational mode from fed-batch to 498 continuous should be considered an effective strategy for improving the bioprocess efficiency. 499 Indeed, the FDA has encouraged the development of continuous processing for 500 biopharmaceuticals manufacturing [126,137,138]. 501 502 Acknowledgements 503 This work was supported by the project CTQ2016-74959-R (MINECO/FEDER, UE). The group 504 is a member of 2017-SGR-1462 and the Reference Network in Biotechnology (XRB) (Generalitat 505 de Catalunya). 506 507 Conflict of interest 508 The authors declare no financial or commercial conflict of interest. 19 509 510 References 511 512 [1] Demain AL, Vaiashnav P. Production of recombinant proteins by microbes and higher 513 organisms. Biotechnol Adv 2009;27:297–306. doi:10.1016/j.biotechadv.2009.01.008. 514 [2] Bill RM. Playing catch-up with Escherichia coli: using yeast to increase success rates in 515 recombinant protein production experiments. Front Microbiol 2014;5:1–5. 516 doi:10.3389/fmicb.2014.00085. 517 [3] Adrio JL, Demain AL. Recombinant organisms for production of industrial products. 518 Bioeng Bugs 2010;1:116–31. doi:10.4161/bbug.1.2.10484. 519 [4] Highsmith J. Biological Therapeutic Drugs: Technologies and Global Markets. BCC Res 520 2015. 521 [5] Walsh G. Biopharmaceutical benchmarks 2014. Nat Biotechnol 2014;32:992–1000. 522 doi:10.1038/nbt.3040. 523 [6] Dewan SS. Global markets for enzymes in industrial use this report to : BCC Res 524 2017;7215. 525 [7] Rosano GL, Ceccarelli EA. Recombinant protein expression in Escherichia coli: Advances 526 and challenges. Front Microbiol 2014;5:1–17. doi:10.3389/fmicb.2014.00172 527 [8] Overton TW. Recombinant protein production in bacterial hosts. Drug Discov Today 528 2014;19:590–601. doi:10.1016/j.drudis.2013.11.008. 529 [9] Matthews CB, Wright C, Kuo A, Colant N, Westoby M, Love JC. Reexamining 530 opportunities for therapeutic protein production in eukaryotic microorganisms. Biotechnol 531 Bioeng 2017;114:2432–444. doi:10.1002/bit.26378. 532 [10] Sørensen HP. Towards universal systems for recombinant gene expression. Microb Cell 533 Fact 2010;9:27. doi:10.1186/1475-2859-9-27. 534 [11] Huang M, Bao J, Nielsen J. Biopharmaceutical protein production by Saccharomyces 535 cerevisiae : current state and future prospects 2014;2:167–82. doi:10.2217/PBP.14.8. 536 [12] Walsh G. Biopharmaceutical benchmarks 2010. Nat Biotechnol 2010;28:917–924. 537 doi:10.1038/nbt0910-917. 538 [13] Adrio J, Demain A. Microbial Enzymes: Tools for Biotechnological Processes. 539 Biomolecules 2014;4:117–39. doi:10.3390/biom4010117. 540 [14] Westers L, Westers H, Quax WJ. Bacillus subtilis as cell factory for pharmaceutical 541 proteins: a biotechnological approach to optimize the host organism. Biochim Biophys 542 Acta - Mol Cell Res 2004;1694:299–310. doi:10.1016/j.bbamcr.2004.02.011. 543 [15] Corchero JL, Gasser B, Resina D, Smith W, Parrilli E, Vázquez F, et al. Unconventional 544 microbial systems for the cost-efficient production of high-quality protein therapeutics. 545 Biotechnol Adv 2013;31:140–53. doi:10.1016/j.biotechadv.2012.09.001. 546 [16] Bandaranayake AD, Almo SC. Recent advances in mammalian protein production. FEBS 547 Lett 2014;588:253–60. doi:10.1016/j.febslet.2013.11.035. 548 [17] Zhu J. Mammalian cell protein expression for biopharmaceutical production. Biotechnol 549 Adv 2012;30:1158–70. doi:10.1016/j.biotechadv.2011.08.022. 550 [18] Spadiut O, Capone S, Krainer F, Glieder A, Herwig C. Microbials for the production of 551 monoclonal antibodies and antibody fragments. Trends Biotechnol 2014;32:54–60. 552 doi:10.1016/j.tibtech.2013.10.002. 553 [19] Porro D, Sauer M, Branduardi P, Mattanovich D. Recombinant protein production in 554 yeasts. Mol Biotechnol 2005;31:245–59. doi:10.1385/MB:31:3:245. 555 [20] Hitzeman RA, Hagie FE, Levine HL, Goeddel DV, Ammerer G, Hall BD. Expression of a 556 human gene for interferon in yeast. Nature 1981;293:717–22. doi:10.1038/293717a0. 557 [21] Ubiyvovk VM, Ananin VM, Malyshev AY, Kang HA, Sibirny AA. Optimization of 558 glutathione production in batch and fed-batch cultures by the wild-type and recombinant 559 strains of the methylotrophic yeast Hansenula polymorpha DL-1. BMC Biotechnol 560 2011;11:8. doi:10.1186/1472-6750-11-8. 20 561 [22] Çelik E, Çalık P. Production of recombinant proteins by yeast cells. Biotechnol Adv 562 2012;30:1108–18. doi:10.1016/j.biotechadv.2011.09.011. 563 [23] Potvin G, Ahmad A, Zhang Z. Bioprocess engineering aspects of heterologous protein 564 production in Pichia pastoris: A review. Biochem Eng J 2012;64:91–105. 565 doi:10.1016/j.bej.2010.07.017. 566 [24] Cregg JM, Barringer KJ, Hessler AY, Madden KR. Pichia pastoris as a host system for 567 transformations. Mol Cell Biol 1985;5:3376–85. doi:10.1128/MCB.5.12.3376. 568 [25] Ahmad M, Hirz M, Pichler H, Schwab H. Protein expression in Pichia pastoris: recent 569 achievements and perspectives for heterologous protein production. Appl Microbiol 570 Biotechnol 2014;98:5301–17. doi:10.1007/s00253-014-5732-5. 571 [26] Gasser B, Prielhofer R, Marx H, Maurer M, Nocon J, Steiger M, et al. Pichia pastoris: 572 protein production host and model organism for biomedical research. Future Microbiol 573 2013;8:191–208. doi:10.2217/fmb.12.133. 574 [27] Love KR, Dalvie NC, Love JC. The yeast stands alone: the future of protein biologic 575 production. Curr Opin Biotechnol 2018;53:50–8. doi:10.1016/J.COPBIO.2017.12.010. 576 [28] Juturu V, Wu JC. Heterologous Protein Expression in Pichia pastoris : Latest Research 577 Progress and Applications. ChemBioChem 2018;19:7–21. doi:10.1002/cbic.201700460. 578 [29] Yang Z, Zhang Z. Engineering strategies for enhanced production of protein and bio- 579 products in Pichia pastoris: A review. Biotechnol Adv 2018;36:182–95. 580 doi: 10.1016/j.biotechadv.2017.11.002. 581 [30] Theron CW, Berrios J, Delvigne F, Fickers P. Integrating metabolic modeling and 582 population heterogeneity analysis into optimizing recombinant protein production by 583 Komagataella (Pichia) pastoris. Appl Microbiol Biotechnol 2018;102:63–80. 584 doi:10.1007/s00253-017-8612-y. 585 [31] Schwarzhans JP, Luttermann T, Geier M, Kalinowski J, Friehs K. Towards systems 586 metabolic engineering in Pichia pastoris. Biotechnol Adv 2017;35:681–710. 587 doi: 10.1016/j.biotechadv.2017.07.009. 588 [32] Zahrl RJ, Peña DA, Mattanovich D, Gasser B. Systems biotechnology for protein 589 production in Pichia pastoris. FEMS Yeast Res 2017;17. doi:10.1093/femsyr/fox068. 590 [33] Formenti LR, Nørregaard A, Bolic A, Hernandez DQ, Hagemann T, Heins AL, et al. 591 Challenges in industrial fermentation technology research. Biotechnol J 2014;9:727–38. 592 doi:10.1002/biot.201300236. 593 [34] Love JC, Love KR, Barone PW. Enabling global access to high-quality 594 biopharmaceuticals. Curr Opin Chem Eng 2013;2:383–90. 595 doi:10.1016/j.coche.2013.09.002. 596 [35] Looser V, Bruhlmann B, Bumbak F, Stenger C, Costa M, Camattari A, et al. Cultivation 597 strategies to enhance productivity of Pichia pastoris: A review. Biotechnol Adv 598 2015;33:1177–93. doi:10.1016/j.biotechadv.2015.05.008. 599 [36] Ponte X, Montesinos-Seguí JL, Valero F. Bioprocess efficiency in Rhizopus oryzae lipase 600 production by Pichia pastoris under the control of PAOX1 is oxygen tension dependent. 601 Process Biochem 2016;51:1954–63. doi:10.1016/j.procbio.2016.08.030. 602 [37] Güneş H, Çalık P. Oxygen transfer as a tool for fine-tuning recombinant protein production 603 by Pichia pastoris under glyceraldehyde-3-phosphate dehydrogenase promoter. 604 Bioprocess Biosyst Eng 2016;39:1061–72. doi:10.1007/s00449-016-1584-y. 605 [38] Carly F, Niu H, Delvigne F, Fickers P. Influence of methanol/sorbitol co-feeding rate on 606 pAOX1 induction in a Pichia pastoris Mut+ strain in bioreactor with limited oxygen 607 transfer rate. J Ind Microbiol Biotechnol 2016;43:517–23. doi:10.1007/s10295-015-1722- 608 6. 609 [39] Cos O, Ramón R, Montesinos JL, Valero F. Operational strategies, monitoring and control 610 of heterologous protein production in the methylotrophic yeast Pichia pastoris under 611 different promoters: A review. Microb Cell Fact 2006;5:17. doi:10.1186/1475-2859-5-17. 612 [40] Çalık P, Ata Ö, Güneş H, Massahi A, Boy E, Keskin A, et al. Recombinant protein 613 production in Pichia pastoris under glyceraldehyde-3-phosphate dehydrogenase promoter: 21 614 From carbon source metabolism to bioreactor operation parameters. Biochem Eng J 615 2015;95:20–36. doi:10.1016/j.bej.2014.12.003. 616 [41] Prielhofer R, Maurer M, Klein J, Wenger J, Kiziak C, Gasser B, et al. Induction without 617 methanol: novel regulated promoters enable high-level expression in Pichia pastoris. 618 Microb Cell Fact 2013;12:5. doi:10.1186/1475-2859-12-5. 619 [42] Vogl T, Sturmberger L, Kickenweiz T, Wasmayer R, Schmid C, Hatzl AM, et al. A 620 Toolbox of Diverse Promoters Related to Methanol Utilization: Functionally Verified Parts 621 for Heterologous Pathway Expression in Pichia pastoris. ACS Synth Biol 2016;5:172–86. 622 doi:10.1021/acssynbio.5b00199. 623 [43] Jungo C, Marison I, von Stockar U. Mixed feeds of glycerol and methanol can improve the 624 performance of Pichia pastoris cultures: A quantitative study based on concentration 625 gradients in transient continuous cultures. J Biotechnol 2007;128:824–37. 626 doi:10.1016/j.jbiotec.2006.12.024. 627 [44] Heyland J, Fu J, Blank LM, Schmid A. Quantitative physiology of Pichia pastoris during 628 glucose-limited high-cell density fed-batch cultivation for recombinant protein production. 629 Biotechnol Bioeng 2010;107:357–68. doi:10.1002/bit.22836. 630 [45] Zhang AL, Luo JX, Zhang TY, Pan YW, Tan YH, Fu CY, et al. Recent advances on the 631 GAP promoter derived expression system of Pichia pastoris. Mol Biol Rep 2009;36:1611– 632 9. doi:10.1007/s11033-008-9359-4. 633 [46] Waterham HR, Digan ME, Koutz PJ, Lair SV, Cregg JM. Isolation of the Pichia pastoris 634 glyceraldehyde-3-phosphate dehydrogenase gene and regulation and use of its promoter. 635 Gene 1997;186:37–44. doi:10.1016/S0378-1119(96)00675-0. 636 [47] Ata Ö, Rebnegger C, Tatto NE, Valli M, Mairinger T, Hann S, et al. A single Gal4-like 637 transcription factor activates the Crabtree effect in Komagataella phaffii. Nat Commun 638 2018;9:4911. doi:10.1038/s41467-018-07430-4. 639 [48] Garcia-Ortega X, Ferrer P, Montesinos JL, Valero F. Fed-batch operational strategies for 640 recombinant Fab production with Pichia pastoris using the constitutive GAP promoter. 641 Biochem Eng J 2013;79:172–81. doi:10.1016/j.bej.2013.07.013. 642 [49] Hahn-Hägerdal B, Karhumaa K, Larsson CU, Gorwa-Grauslund M, Görgens J, van Zyl 643 WH. Role of cultivation media in the development of yeast strains for large scale industrial 644 use. Microb Cell Fact 2005;4:31. doi:10.1186/1475-2859-4-31. 645 [50] Garcia-Ortega X, Adelantado N, Ferrer P, Montesinos JL, Valero F. A step forward to 646 improve recombinant protein production in Pichia pastoris: From specific growth rate 647 effect on protein secretion to carbon-starving conditions as advanced strategy. Process 648 Biochem 2016;51:681–91. doi:10.1016/j.procbio.2016.02.018. 649 [51] Garcia-Ortega X, Valero F, Montesinos-Seguí JL. Physiological state as transferable 650 operating criterion to improve recombinant protein production in Pichia pastoris through 651 oxygen limitation. J Chem Technol Biotechnol 2017;92:2573–82. doi:10.1002/jctb.5272. 652 [52] Zahrl RJ, Peña DA, Mattanovich D, Gasser B. Systems biotechnology for protein 653 production in Pichia pastoris. FEMS Yeast Res 2017;17:1–15. 654 doi:10.1093/femsyr/fox068. 655 [53] Baumann K, Carnicer M, Dragosits M, Graf AB, Stadlmann J, Jouhten P, et al. A multi- 656 level study of recombinant Pichia pastoris in different oxygen conditions. BMC Syst Biol 657 2010;4:141. doi:10.1186/1752-0509-4-141. 658 [54] Dragosits M, Stadlmann J, Albiol J, Baumann K, Maurer M, Gasser B, et al. The effect of 659 temperature on the proteome of recombinant Pichia pastoris. J Proteome Res 2009;8:1380– 660 92. doi:10.1021/pr8007623. 661 [55] Carnicer M, ten Pierick A, van Dam J, Heijnen JJ, Albiol J, van Gulik W, et al. Quantitative 662 metabolomics analysis of amino acid metabolism in recombinant Pichia pastoris under 663 different oxygen availability conditions. Microb Cell Fact 2012;11:83. doi:10.1186/1475- 664 2859-11-83. 22 665 [56] de Hollander JA. Kinetics of microbial product formation and its consequences for the 666 optimization of fermentation processes. Antonie Van Leeuwenhoek 1993;63:375–81. 667 doi:10.1007/BF00871231. 668 [57] Spadiut O, Rittmann S, Dietzsch C, Herwig C. Dynamic process conditions in bioprocess 669 development. Eng Life Sci 2013;13:88–101. doi:10.1002/elsc.201200026. 670 [58] Spadiut O, Herwig C. Dynamics in bioprocess development for Pichia pastoris. 671 Bioengineered 2014;5:401–4. doi:10.4161/bioe.36152. 672 [59] Looser V, Lüthy D, Straumann M, Hecht K, Melzoch K, Kovar K. Effects of glycerol 673 supply and specific growth rate on methanol-free production of CALB by P. pastoris: 674 functional characterisation of a novel promoter. Appl Microbiol Biotechnol 675 2017;101:3163–76. doi:10.1007/s00253-017-8123-x. 676 [60] Wilming A, Bähr C, Kamerke C, Büchs J. Fed-batch operation in special microtiter plates: 677 A new method for screening under production conditions. J Ind Microbiol Biotechnol 678 2014;41:513–25. doi:10.1007/s10295-013-1396-x. 679 [61] Lattermann C, Büchs J. Microscale and miniscale fermentation and screening. Curr Opin 680 Biotechnol 2015;35:1–6. doi:10.1016/j.copbio.2014.12.005. 681 [62] Perez-Pinera P, Han N, Cleto S, Cao J, Purcell O, Shah KA, et al. Synthetic biology and 682 microbioreactor platforms for programmable production of biologics at the point-of-care. 683 Nat Commun 2016;7:1–10. doi:10.1038/ncomms12211. 684 [63] Hemmerich J, Adelantado N, Barrigón J, Ponte X, Hörmann A, Ferrer P, et al. 685 Comprehensive clone screening and evaluation of fed-batch strategies in a microbioreactor 686 and lab scale stirred tank bioreactor system: application on Pichia pastoris producing 687 Rhizopus oryzae lipase. Microb Cell Fact 2014;13:36. doi:10.1186/1475-2859-13-36. 688 [64] Wewetzer SJ, Kunze M, Ladner T, Luchterhand B, Roth S, Rahmen N, et al. Parallel use 689 of shake flask and microtiter plate online measuring devices (RAMOS and BioLector) 690 reduces the number of experiments in laboratory-scale stirred tank bioreactors. J Biol Eng 691 2015;9:1–18. doi:10.1186/s13036-015-0005-0. 692 [65] Hensing MCM, Rouwenhorst RJ, Heijnen JJ, van Dijken JP, Pronk JT. Physiological and 693 technological aspects of large-scale heterologous-protein production with yeasts. Antonie 694 Van Leeuwenhoek 1995;67:261–79. doi:10.1007/BF00873690. 695 [66] Barrigon JM, Valero F, Montesinos JL. A macrokinetic model-based comparative meta- 696 analysis of recombinant protein production by Pichia pastoris under AOX1 promoter. 697 Biotechnol Bioeng 2015;112:1132–45. doi:10.1002/bit.25518. 698 [67] Looser V, Bruhlmann B, Bumbak F, Stenger C, Costa M, Camattari A, et al. Cultivation 699 strategies to enhance productivity of Pichia pastoris: A review. Biotechnol Adv 700 2015;33:1177–93. doi:10.1016/j.biotechadv.2015.05.008. 701 [68] Soons ZITA, Voogt JA, Van Straten G, Van Boxtel AJB. Constant specific growth rate in 702 fed-batch cultivation of Bordetella pertussis using adaptive control. J Biotechnol 703 2006;125:252–68. doi:10.1016/j.jbiotec.2006.03.005. 704 [69] Veloso ACA, Rocha I, Ferreira EC. Monitoring of fed-batch E. coli fermentations with 705 software sensors. Bioprocess Biosyst Eng 2009;32:381–8. doi:10.1007/s00449-008-0257- 706 x. 707 [70] Dabros M, Schuler MM, Marison IW. Simple control of specific growth rate in 708 biotechnological fed-batch processes based on enhanced online measurements of biomass. 709 Bioprocess Biosyst Eng 2010;33:1109–18. doi:10.1007/s00449-010-0438-2. 710 [71] Kuprijanov A, Schaepe S, Simutis R, Lübbert A. Model predictive control made accessible 711 to professional automation systems in fermentation technology. Biosyst Inf Technol 712 2013;2:26–31. doi:10.11592/bit.131101. 713 [72] Craven S, Whelan J, Glennon B. Glucose concentration control of a fed-batch mammalian 714 cell bioprocess using a nonlinear model predictive controller. J Process Control 715 2014;24:344–57. doi:10.1016/j.jprocont.2014.02.007. 23 716 [73] Mears L, Stocks SM, Sin G, Gernaey KV. A review of control strategies for manipulating 717 the feed rate in fed-batch fermentation processes. J Biotechnol 2017;245:34–46. 718 doi:10.1016/j.jbiotec.2017.01.008. 719 [74] Ferreira AR, Ataíde F, von Stosch M, Dias JML, Clemente JJ, Cunha AE, et al. Application 720 of adaptive DO-stat feeding control to Pichia pastoris X33 cultures expressing a single 721 chain antibody fragment (scFv). Bioprocess Biosyst Eng 2012;35:1603–14. 722 doi:10.1007/s00449-012-0751-z. 723 [75] Zhao W, Wang J, Deng R, Wang X. Scale-up fermentation of recombinant Candida rugosa 724 lipase expressed in Pichia pastoris using the GAP promoter. J Ind Microbiol Biotechnol 725 2008;35:189–95. doi:10.1007/s10295-007-0283-8. 726 [76] Maurer M, Kühleitner M, Gasser B, Mattanovich D. Versatile modeling and optimization 727 of fed batch processes for the production of secreted heterologous proteins with Pichia 728 pastoris. Microb Cell Fact 2006;5:37. doi:10.1186/1475-2859-5-37. 729 [77] Kobayashi K, Kuwae S, Ohya T, Ohda T, Ohyama M, Tomomitsu K. High level secretion 730 of recombinant human serum albumin by fed-batch fermentation of the methylotrophic 731 yeast, Pichia pastoris, based on optimal methanol feeding strategy. J Biosci Bioeng 732 2000;90:280–8. doi:10.1016/S1389-1723(00)80082-1. 733 [78] Ohya T, Ohyama M, Kobayashi K. Optimization of human serum albumin production in 734 methylotrophic yeast Pichia pastoris by repeated fed-batch fermentation. Biotechnol 735 Bioeng 2005;90:876–87. doi:10.1002/bit.20507. 736 [79] Barrigón JM, Montesinos JL, Valero F. Searching the best operational strategies for 737 Rhizopus oryzae lipase production in Pichia pastoris Mut+ phenotype: Methanol limited 738 or methanol non-limited fed-batch cultures? Biochem Eng J 2013;75:47–54. 739 doi:10.1016/j.bej.2013.03.018. 740 [80] Zhang W, Sinha J, Smith LA, Inan M, Meagher MM. Maximization of production of 741 secreted recombinant proteins in Pichia pastoris fed-batch fermentation. Biotechnol Prog 742 2005;21:386–93. doi:10.1021/bp049811n. 743 [81] Cámara E, Albiol J, Ferrer P. Droplet digital PCR-aided screening and characterization of 744 Pichia pastoris multiple gene copy strains. Biotechnol Bioeng 2016;113:1542–51. 745 doi:10.1002/bit.25916. 746 [82] Cámara E, Landes N, Albiol J, Gasser B, Mattanovich D, Ferrer P. Increased dosage of 747 AOX1 promoter-regulated expression cassettes leads to transcription attenuation of the 748 methanol metabolism in Pichia pastoris. Sci Rep 2017;7:44302. doi:10.1038/srep44302. 749 [83] Ponte X, Barrigón JM, Maurer M, Mattanovich D, Valero F, Montesinos-Seguí JL. 750 Towards optimal substrate feeding for heterologous protein production in Pichia pastoris 751 (Komagataella spp) fed-batch processes under PAOX1 control: a modeling aided approach. J 752 Chem Technol Biotechnol 2018. doi:10.1002/jctb.5677. 753 [84] Zhang W, Inan M, Meagher MM. Rational Design and Optimization of Fed-Batch and 754 Continuous Fermentations. Methods Mol. Biol., vol. 389, 2007; 43–63. doi:10.1007/978- 755 1-59745-456-8_4. 756 [85] Nocon J, Steiger MG, Pfeffer M, Sohn SB, Kim TY, Maurer M, et al. Model based 757 engineering of Pichia pastoris central metabolism enhances recombinant protein 758 production. Metab Eng 2014;24:129–38. doi:10.1016/j.ymben.2014.05.011. 759 [86] Dragosits M, Stadlmann J, Graf A, Gasser B, Maurer M, Sauer M, et al. The response to 760 unfolded protein is involved in osmotolerance of Pichia pastoris. BMC Genomics 761 2010;11:207. doi:10.1186/1471-2164-11-207. 762 [87] Baumann K, Maurer M, Dragosits M, Cos O, Ferrer P, Mattanovich D. Hypoxic fed-batch 763 cultivation of Pichia pastoris increases specific and volumetric productivity of 764 recombinant proteins. Biotechnol Bioeng 2008;100:177–83. doi:10.1002/bit.21763. 765 [88] Rebnegger C, Graf AB, Valli M, Steiger MG, Gasser B, Maurer M, et al. In Pichia pastoris, 766 growth rate regulates protein synthesis and secretion, mating and stress response. 767 Biotechnol J 2014;9:511–25. doi:10.1002/biot.201300334. 24 768 [89] Rebnegger C, Vos T, Graf AB, Valli M, Pronk JT, Daran-Lapujade P, et al. Pichia pastoris 769 Exhibits High Viability and a Low Maintenance Energy Requirement at Near-Zero 770 Specific Growth Rates. Appl Environ Microbiol 2016;82:4570–83. 771 doi:10.1128/AEM.00638-16. 772 [90] Kern A, Hartner FS, Freigassner M, Spielhofer J, Rumpf C, Leitner L, et al. Pichia pastoris 773 “just in time” alternative respiration. Microbiology 2007;153:1250–60. 774 doi:10.1099/mic.0.2006/001404-0. 775 [91] Gancedo JM. Yeast carbon catabolite repression. Microbiol Mol Biol Rev 1998;62:334– 776 61. 777 [92] Zaman S, Lippman SI, Zhao X, Broach JR. How Saccharomyces responds to nutrients. 778 Annu Rev Genet 2008;42:27–81. doi:10.1146/annurev.genet.41.110306.130206. 779 [93] Ashe MP, De Long SK, Sachs AB. Glucose Depletion Rapidly Inhibits Translation 780 Initiation in Yeast. Mol Biol Cell 2000;11:833–48. doi:10.1091/mbc.11.3.833. 781 [94] Arribere JA, Doudna JA, Gilbert WV. Reconsidering movement of eukaryotic mRNAs 782 between polysomes and P bodies. Mol Cell 2011;44:745–58. 783 doi:10.1016/j.molcel.2011.09.019. 784 [95] Heo JH, Ananin V, Kang HA, Rhee SK, Kim CH. Feeding strategies for the enhanced 785 production of recombinant human serum albumin in the fed-batch cultivation of Hansenula 786 polymorpha. Process Biochem 2008;43:918–24. doi:10.1016/j.procbio.2008.04.017. 787 [96] Jazini M, Cekici G, Herwig C. Quantifying the Effects of Frequency and Amplitude of 788 Periodic Oxygen-Related Stress on Recombinant Protein Production in Pichia pastoris. 789 Bioengineering 2013;1:47–61. doi:10.3390/bioengineering1010047. 790 [97] Tang H, Bao X, Shen Y, Song M, Wang S, Wang C, et al. Engineering protein folding and 791 translocation improves heterologous protein secretion in Saccharomyces cerevisiae. 792 Biotechnol Bioeng 2015;112:1872–82. doi:10.1002/bit.25596. 793 [98] Liu L, Zhang Y, Liu Z, Petranovic D, Nielsen J. Improving heterologous protein secretion 794 at aerobic conditions by activating hypoxia-induced genes in Saccharomyces cerevisiae. 795 FEMS Yeast Res 2015;15:fov070. doi:10.1093/femsyr/fov070. 796 [99] Berdichevsky M, d’Anjou M, Mallem MR, Shaikh SS, Potgieter TI. Improved production 797 of monoclonal antibodies through oxygen-limited cultivation of glycoengineered yeast. J 798 Biotechnol 2011;155:217–24. doi:10.1016/j.jbiotec.2011.06.021. 799 [100] Carnicer M, Baumann K, Töplitz I, Sánchez-Ferrando F, Mattanovich D, Ferrer P, et al. 800 Macromolecular and elemental composition analysis and extracellular metabolite balances 801 of Pichia pastoris growing at different oxygen levels. Microb Cell Fact 2009;8:65. 802 doi:10.1186/1475-2859-8-65. 803 [101] Alexeeva S, Hellingwerf KJ, Teixeira de Mattos MJ. Quantitative assessment of oxygen 804 availability: perceived aerobiosis and its effect on flux distribution in the respiratory chain 805 of Escherichia coli. J Bacteriol 2002;184:1402–6. doi:10.1128/JB.184.5.1402-1406.2002. 806 [102] Zalai D, Golabgir A, Wechselberger P, Putics A, Herwig C. Advanced Development 807 Strategies for Biopharmaceutical Cell Culture Processes. Curr Pharm Biotechnol 808 2015;16:983–1001. doi:10.2174/1389201016666150724100450. 809 [103] Liu WC, Gong T, Wang QH, Liang X, Chen JJ, Zhu P. Scaling-up Fermentation of Pichia 810 pastoris to demonstration-scale using new methanol-feeding strategy and increased air 811 pressure instead of pure oxygen supplement. Sci Rep 2016;6:1–13. doi:10.1038/srep18439. 812 [104] Sin G, Gernaey KV, Lantz AE. Good modelling practice (GMoP) for PAT applications: 813 Propagation of input uncertainty and sensitivity analysis. Biotechnol Prog 2009;25:1043– 814 53. doi:10.1021/bp.166. 815 [105] Takors R. Scale-up of microbial processes: Impacts, tools and open questions. J Biotechnol 816 2012;160:3–9. doi:10.1016/j.jbiotec.2011.12.010. 817 [106] Garcia-Ochoa F, Gomez E. Bioreactor scale-up and oxygen transfer rate in microbial 818 processes: An overview. Biotechnol Adv 2009;27:153–76. 819 doi:10.1016/j.biotechadv.2008.10.006. 25 820 [107] Jungo C, Urfer J, Zocchi A, Marison I, von Stockar U. Optimisation of culture conditions 821 with respect to biotin requirement for the production of recombinant avidin in Pichia 822 pastoris. J Biotechnol 2007;127:703–15. doi:10.1016/j.jbiotec.2006.08.001. 823 [108] Mattanovich D, Graf A, Stadlmann J, Dragosits M, Redl A, Maurer M, et al. Genome, 824 secretome and glucose transport highlight unique features of the protein production host 825 Pichia pastoris. Microb Cell Fact 2009;8:29. doi:10.1186/1475-2859-8-29. 826 [109] Hartner FS, Glieder A. Regulation of methanol utilisation pathway genes in yeasts. Microb 827 Cell Fact 2006;5:39. doi:10.1186/1475-2859-5-39. 828 [110] Shen W, Xue Y, Liu Y, Kong C, Wang X, Huang M, et al. A novel methanol-free Pichia 829 pastoris system for recombinant protein expression. Microb Cell Fact 2016;15:178. 830 doi:10.1186/s12934-016-0578-4. 831 [111] Wang J, Wang X, Shi L, Qi F, Zhang P, Zhang Y, et al. Methanol-Independent Protein 832 Expression by AOX1 Promoter with trans-Acting Elements Engineering and Glucose- 833 Glycerol-Shift Induction in Pichia pastoris. Sci Rep 2017;7:41850. 834 doi:10.1038/srep41850. 835 [112] Vogl T, Sturmberger L, Fauland PC, Hyden P, Fischer JE, Schmid C, et al. Methanol 836 independent induction in Pichia pastoris by simple derepressed overexpression of single 837 transcription factors. Biotechnol Bioeng 2018;115:1037–50. doi:10.1002/bit.26529. 838 [113] Rajamanickam V, Metzger K, Schmid C, Spadiut O. A novel bi-directional promoter 839 system allows tunable recombinant protein production in Pichia pastoris. Microb Cell Fact 840 2017;16:152. doi:10.1186/s12934-017-0768-8. 841 [114] Sreekrishna K. Pichia, Optimization of Protein Expression. Encycl. Ind. Biotechnol., 842 Hoboken, NJ, USA: John Wiley & Sons, Inc.; 2010. doi:10.1002/9780470054581.eib480. 843 [115] Wechselberger P, Herwig C. Model-based analysis on the relationship of signal quality to 844 real-time extraction of information in bioprocesses. Biotechnol Prog 2012;28:265–75. 845 doi:10.1002/btpr.700. 846 [116] Eon-Duval A, Broly H, Gleixner R. Quality attributes of recombinant therapeutic proteins: 847 An assessment of impact on safety and efficacy as part of a quality by design development 848 approach. Biotechnol Prog 2012;28:608–22. doi:10.1002/btpr.1548. 849 [117] Jahic M, Veide A, Charoenrat T, Teeri T, Enfors SO. Process technology for production 850 and recovery of heterologous proteins with Pichia pastoris. Biotechnol Prog 851 2006;22:1465–73. doi:10.1021/bp060171t. 852 [118] Costa AR, Rodrigues ME, Henriques M, Oliveira R, Azeredo J. Glycosylation: impact, 853 control and improvement during therapeutic protein production. Crit Rev Biotechnol 854 2014;34:281–99. doi:10.3109/07388551.2013.793649. 855 [119] Wu D, Chu J, Hao YY, Wang YH, Zhuang YP, Zhang SL. Incomplete protein disulphide 856 bond conformation and decreased protein expression result from high cell growth during 857 heterologous protein expression in Pichia pastoris. J Biotechnol 2012;157:107–12. 858 doi:10.1016/j.jbiotec.2011.08.032. 859 [120] Gervasi T, Pellizzeri V, Calabrese G, Di Bella G, Cicero N, Dugo G. Production of single 860 cell protein (SCP) from food and agricultural waste by using Saccharomyces cerevisiae. 861 Nat Prod Res 2018;32:648–53. doi:10.1080/14786419.2017.1332617. 862 [121] Yunus F un N, Nadeem M, Rashid F. Single-cell protein production through microbial 863 conversion of lignocellulosic residue (wheat bran) for animal feed. J Inst Brew 864 2015;121:553–7. doi:10.1002/jib.251. 865 [122] Meyer HP, Minas W, Schmidhalter D. Industrial Biotechnology. Weinheim, Germany: 866 Wiley;VCH Verlag GmbH; Co. KGaA; 2017. doi:10.1002/9783527807833. 867 [123] Mattanovich D, Branduardi P, Dato L, Gasser B, Sauer M, Porro D. Recombinant protein 868 production in yeasts. Methods Mol Biol 2012;824:329–58. doi:10.1007/978-1-61779-433- 869 9_17. 870 [124] Balasundaram B, Harrison S, Bracewell DG. Advances in product release strategies and 871 impact on bioprocess design. Trends Biotechnol 2009;27:477–85. 872 doi:10.1016/j.tibtech.2009.04.004. 26 873 [125] Curvers S, Linnemann J, Klauser T, Wandrey C, Takors R. Recombinant Protein 874 Production with Pichia pastoris in Continuous Fermentation – Kinetic Analysis of Growth 875 and Product Formation. Eng Life Sci 2002;2:229. doi:10.1002/1618- 876 2863(20020806)2:8<229::AID-ELSC229>3.0.CO;2-9. 877 [126] Gernaey K V, Cervera-Padrell AE, Woodley JM. Development of continuous 878 pharmaceutical production processes supported by process systems engineering methods 879 and tools. Future Med Chem 2012;4:1371–4. doi:10.4155/fmc.12.77. 880 [127] Croughan MS, Konstantinov KB, Cooney C. The future of industrial bioprocessing: Batch 881 or continuous? Biotechnol Bioeng 2015;112:648–51. doi:10.1002/bit.25529. 882 [128] Walthe J, Godawat R, Hwang C, Abe Y, Sinclair A, Konstantinov K. The business impact 883 of an integrated continuous biomanufacturing platform for recombinant protein 884 production. J Biotechnol 2015:1–10. doi:10.1016/j.jbiotec.2015.05.010. 885 [129] Kazemi Seresht A, Cruz AL, de Hulster E, Hebly M, Palmqvist EA, van Gulik W, et al. 886 Long-term adaptation of Saccharomyces cerevisiae to the burden of recombinant insulin 887 production. Biotechnol Bioeng 2013;110:2749–63. doi:10.1002/bit.24927. 888 [130] Baumann K, Adelantado N, Lang C, Mattanovich D, Ferrer P. Protein trafficking, 889 ergosterol biosynthesis and membrane physics impact recombinant protein secretion in 890 Pichia pastoris. Microb Cell Fact 2011;10:93. doi:10.1186/1475-2859-10-93. 891 [131] Jordà J, Jouhten P, Cámara E, Maaheimo H, Albiol J, Ferrer P. Metabolic flux profiling 892 of recombinant protein secreting Pichia pastoris growing on glucose:methanol mixtures. 893 Microb Cell Fact 2012;11:57. doi:10.1186/1475-2859-11-57. 894 [132] Canales C, Altamirano C, Berrios J. Effect of dilution rate and methanol-glycerol mixed 895 feeding on heterologous Rhizopus oryzae lipase production with Pichia pastoris Mut + 896 phenotype in continuous culture. Biotechnol Prog 2015;31:707–14. 897 doi:10.1002/btpr.2069. 898 [133] Berrios J, Flores MO, Díaz-Barrera A, Altamirano C, Martínez I, Cabrera Z. A 899 comparative study of glycerol and sorbitol as co-substrates in methanol-induced cultures 900 of Pichia pastoris: temperature effect and scale-up simulation. J Ind Microbiol Biotechnol 901 2017;44:407–11. doi:10.1007/s10295-016-1895-7. 902 [134] Tang S, Reiche A, Potvin G, Zhang Z. Modeling of Phytase Production by Cultivation of 903 Pichia pastoris Under the Control of the GAP Promoter. Int J Chem React Eng, 2010; 8.1 904 doi:10.2202/1542-6580.2144[ 905 [135] Liu ZW, Yin HX, Yi XP, Zhang AL, Luo JX, Zhang TY, et al. Constitutive expression of 906 barley α-amylase in Pichia pastoris by high-density cell culture. Mol Biol Rep 907 2012;39:5805–10. doi:10.1007/s11033-011-1390-1. 908 [136] Khasa YP, Khushoo A, Srivastava L, Mukherjee KJ. Kinetic studies of constitutive human 909 granulocyte-macrophage colony stimulating factor (hGM-CSF) expression in continuous 910 culture of Pichia pastoris. Biotechnol Lett 2007;29:1903–8. doi:10.1007/s10529-007- 911 9473-8. 912 [137] Gernaey K V., Woodley JM, Sin G. Introducing mechanistic models in Process Analytical 913 Technology education. Biotechnol J 2009;4:593–9. doi:10.1002/biot.200800323. 914 [138] Nasr MM, Krumme M, Matsuda Y, Trout BL, Badman C, Mascia S, et al. Regulatory 915 Perspectives on Continuous Pharmaceutical Manufacturing: Moving From Theory to 916 Practice. J Pharm Sci 2017;106:3199–206. doi:10.1016/j.xphs.2017.06.015. 917 [139] Várnai A, Tang C, Bengtsson O, Atterton A, Mathiesen G, Eijsink VG. Expression of 918 endoglucanases in Pichia pastoris under control of the GAP promoter. Microb Cell Fact 919 2014;13:57. doi:10.1186/1475-2859-13-57. 920 [140] Zhu T, Sun H, Li P, Xue Y, Li Y, Ma Y. Constitutive expression of alkaline β-mannanase 921 in recombinant Pichia pastoris. Process Biochem 2014;49:2025–9. 922 doi:10.1016/j.procbio.2014.08.014. 923 [141] Qian K, Gong XH, Guan B, Wu SP, Zhang JJ, Qian J, et al. Efficient expression of 924 glucagon-like peptide-1 analogue with human serum albumin fusion protein in Pichia 27 925 pastoris using the glyceraldehyde-3-phosphate dehydrogenase promoter. Biotechnol 926 Bioprocess Eng 2015;20:694–700. doi:10.1007/s12257-014-0818-6. 927 [142] Wang X, Sun Y, Ke F, Zhao H, Liu T, Xu L, et al. Constitutive expression of Yarrowia 928 lipolytica lipase LIP2 in Pichia pastoris using GAP as promoter. Appl Biochem Biotechnol 929 2012;166:1355–67. doi:10.1007/s12010-011-9524-4. 930 [143] Zhang AL, Zhang TY, Luo JX, Chen SC, Guan WJ, Fu CY, et al. Constitutive expression 931 of human angiostatin in Pichia pastoris by high-density cell culture. J Ind Microbiol 932 Biotechnol 2007;34:117–22. doi:10.1007/s10295-006-0175-3. 933 [144] Gasser B, Maurer M, Gach J, Kunert R, Mattanovich D. Engineering of Pichia pastoris for 934 improved production of antibody fragments. Biotechnol Bioeng 2006;94:353–61. 935 doi:10.1002/bit.20851 936 [145] Pal Y, Khushoo A, Mukherjee KJ. Process optimization of constitutive human granulocyte- 937 macrophage colony-stimulating factor (hGM-CSF) expression in Pichia pastoris fed-batch 938 culture. Appl Microbiol Biotechnol 2006;69:650–7. doi:10.1007/s00253-005-0018-6. 939 [146] Stadlmayr G, Mecklenbräuker A, Rothmüller M, Maurer M, Sauer M, Mattanovich D, et 940 al. Identification and characterisation of novel Pichia pastoris promoters for heterologous 941 protein production. J Biotechnol 2010;150:519–29. doi:10.1016/j.jbiotec.2010.09.957. 942 [147] Müller JM, Bruhn S, Flaschel E, Friehs K, Risse JM. GAP promoter-based fed-batch 943 production of highly bioactive core streptavidin by Pichia pastoris. Biotechnol Prog 944 2016;32:855–64. doi:10.1002/btpr.2283. 945 [148] Yu XW, Yang M, Jiang C, Zhang X, Xu Y. N-Glycosylation Engineering to Improve the 946 Constitutive Expression of Rhizopus oryzae Lipase in Komagataella phaffii. J Agric Food 947 Chem 2017;65:6009–15. doi:10.1021/acs.jafc.7b01884. 948 [149] Goodrick JC, Xu M, Finnegan R, Schilling BM, Schiavi S, Hoppe H, et al. High-level 949 expression and stabilization of recombinant human chitinase produced in a continuous 950 constitutive Pichia pastoris expression system. Biotechnol Bioeng 2001;74:492–7. 951 doi:10.1002/bit.1140. 952 [150] Schilling BM, Goodrick JC, Wan NC. Scale-up of a High Cell Density Continuous Culture 953 with Pichia pastoris X-33 for the Constitutive Expression of rh-Chitinase. Biotechnol 954 Prog 2001;17:629–33. doi: 10.1021/bp010041e 955 956 28 957 Figure captions 958 959 Figure 1: Scheme to exploit the adaptive cell responses to environmental stress conditions to 960 boost the recombinant protein production in P. pastoris. 961 29 1 Tables 2 3 Table 1. Summary of the main advantages of P. pastoris as a recombinant protein expression 4 system. (GRAS = Generally Recognised as Safe. DCW = Dry Cell Weight). Genetic Engineering Protein processing Bioprocess engineering Simple and stable genetic Eukaryotic capacity for folding, GRAS organism lacks detectable manipulation assembling and performing PTM’s endotoxins Numerous tools available for genetic Protein processing and secreting manipulation including Non-fermentative growth on glucose pathway similar to high eukaryotes CRISPR/Cas9 system. Ability to secrete efficiently target Different strong and efficient Growth in chemically defined proteins combined with low levels of promoters medium up to 120 g DCW L-1 secretion for native proteins Reduced hyperglycosilation and Well-established large-scale Reported high yield and stable reported human-like glycoengineered production and downstream producing strains strains processing 1 Table 2. Summary of recombinant protein production processes in fed-batch cultures under GAP promoter in P. pastoris. Promoter Operational strategy Volume μ Production Qp qp YX/S YP/X Final DCW Protein expressed PGAP vs PAOX1 Strain Ref. (L) (h-1) (mg L-1) (mg L-1 h-1) (mg gx-1 h-1) (g g-1) (mg g-1) (g/L) comparison Between Fungal Rather Pseudo Constant X33 1.03* [139] endoglucanase similar feeding rate 15 n.d. 86 0.5 37.5 160 5000-7000 n.d. 960 10 β-Mannanase AOX1 > GAP GS115 Constant feeding rate 1 n.d. Equiv. Equiv n.d. n.d. n.d. n.d [140] 2980 UL-1 31 U L-1 h-1 Constant glucose GGH / GLP-1 GAP > AOX1 X33 5 n.d. 3.8 0.1 5.9 41.2 [141] concentration 5 g L-1 246 -- rhIK-2-HSA AOX1 > GAP X33 DO depending 5 n.d. 250 2.5 -- -- -- 75 [142] Angiostatin GAP > AOX GS115 Continuous feeding 20 n.d. 176 4.9 -- -- -- -- [143] Fab fragment (anti-HIV --- X33 Constant feeding rate 1.75 Variable 41.2 -- 0.02 -- -- -- [144] antibody 2F5) Fab fragment Fed-batch optimal (anti-HIV --- X33 2 Variable 45 0.67 0.04 --- 0.47 94 [76] profile antibody 2F5) Fab fragment Exponential feeding Controlled (anti-HIV --- X33 5 0.05-0.15 25 0.73 0.09 0.45 0.25 100 [48] rate antibody 2F5) Glucose DO control 4500 126 --- X33 3 -- -- -- 0.22 44 [37] isomerase 15% U L-1 U g-1 Exponential feeding hGM-CSF --- GS115 2 0.2 250 7.35 0.07 1.05 2.4 98 [145] rate X33 Fed-batch optimal hSA --- 3.5 Variable 44 2.2 0.06-0.02 -- -- 98 [146] SMD1168H profile Streptavidin --- X33 Constant feeding rate 0.5 -- 221 3.06 -- -- -- 30 [147] Rhizopus oryzae 2600 30.95 --- GS115 Constant feeding rate 7 -- -- -- -- 140 [148] lipase U mL-1 U mL-1 h-1 *Some of the presented values were estimated from the published available data. 2 Table 3. Summary of recombinant protein production processes in continuous mode under GAP promoter in P. pastoris. Final Protein Volume μ Production Qp qp YX/S YP/X Strain DCW Ref. expressed (L) (h-1) (mg L-1) (mg L-1 h-1) (mg gx-1 h-1) (g g-1) (mg g-1) (g/L) Serum albumin 30.04 SMD1168H 1 0.015-0.15 --- 0.15 0.52 1.16 --- [88] (human) (μ = 0.15) Fab fragment (anti-HIV X33 2 0.0086-0.2 --- --- 0.046 0.56 --- --- [76] antibody 2F5) Fab fragment (anti-HIV X33 1 0.025-0.15 --- --- 0.057 0.43 --- --- [50] antibody 2F5) hGM-CSF X33 1 0.025-0.2 82 --- 0.5 0.8 2.5 35 [136] (human) X33 rh-chitinase 1.5 0.05 360 18 0.15 0.33 3.6 100 [149] SMD1168 rh-chitinase X33 15 0.042 250 6 0.055 0.37 2.3 110 [150] phytase X33 2 0.05-0.035 --- --- 1,4 kU/g·h -- -- -- [134] α-amylase GS115 20 125 --- --- --- --- --- [135] *Some of the presented values were estimated from the published available data. 3