Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Functional cross-kingdom conservation of mammalian and moss ( Physcomitrella patens ) transcription, translation and secretion machineries

Functional cross-kingdom conservation of mammalian and moss ( Physcomitrella patens )... <h1>Introduction</h1> The moss Physcomitrella patens , unique for its high rate of homologous recombination, generic codon usage, haploidy, simple body plan, physiologic properties and exclusive phylogenetic position ( Quatrano et al ., 2007 ; Rensing et al ., 2008 ), is gathering momentum for the biopharmaceutical manufacture of protein therapeutics because of favourable bioprocess and downstream processing economics ( Decker and Reski, 2007 ). In vitro cultivation of P. patens throughout its complete life cycle ( Frank et al ., 2005 ), transgenic protonema and transient protoplast cultures ( Baur et al ., 2005a ), stirred-tank and tubular photo-bioreactors ( Decker and Reski, 2007 ), the generation of moss mutants devoid of immunogenic product protein glycosylation ( Huether et al ., 2005 ), and the production of human antibodies with improved antibody-dependent cellular cytotoxicity (ADCC) activity ( Nechansky et al ., 2007 ) have been important milestones in establishing the moss as a promising biopharmaceutical manufacturing platform ( Decker and Reski, 2007 ). Transgenic mammalian cell cultures are currently the most successful platform for the production of biopharmaceuticals, as most of the protein therapeutics on the market originate from mammalian cell bioprocesses ( Wurm, 2004 ). However, as the global product pipeline exceeds the worldwide manufacturing capacity, alternative host cell systems for biopharmaceutical manufacturing are on the increase ( Hamilton et al ., 2006 ; Decker and Reski, 2007 ). Since the mid-1980s, the productivity of mammalian cells cultivated in bioreactors has reached the gram per litre range, an over 100-fold yield improvement over titres achieved for the first commercial bioprocesses ( Wurm, 2004 ). Part of this success is based on the development of sophisticated expression technologies and metabolic engineering strategies ( Umana et al ., 1999 ; Hartenbach and Fussenegger, 2005 ). The latest generation of expression vectors harbour: (i) compact strong constitutive promoters for the high-level transcription of product genes ( Hartenbach and Fussenegger, 2006 ); (ii) multicistronic expression units enabling one-vector-based selection and expression of multiprotein complexes ( Fux et al ., 2004 ); and (iii) regulated expression systems for the production of difficult-to-express protein therapeutics ( Weber and Fussenegger, 2007 ). In mammalian cells, the initiation of translation is typically managed by a cap structure which is post-transcriptionally attached to the 5′ end of mRNAs ( Kozak, 1989 ). Alternatively, internal ribosome entry sites (IRESs), which adopt a specific secondary RNA structure triggering ribosome assembly and translational initiation, have evolved to ensure a minimal level of protein synthesis for survival during cap-compromising physiological emergency situations (co-ordination of viral defence; cellular IRES) ( Gan and Rhoads, 1996 ) or to redirect the cellular translation machinery to the production of virus proteins (viral IRES) ( Kaufman et al ., 1991 ; Dirks et al ., 1993 ). The tandem arrangement of different transgenes, each preceded by an IRES element, enables the transcription of a multicistronic mRNA producing stoichiometric levels of various proteins. Recently, a sophisticated vector platform (pTRIDENT) has been designed for the multicistronic expression of up to three different transgenes ( Fux et al ., 2004 ). Heterologous mammalian transcription control modalities have been designed in two different configurations: ON-type systems, which are induced following the addition of a trigger molecule, and OFF-type systems, which are repressed after the administration of a regulating compound ( Weber and Fussenegger, 2007 ). ON-type systems typically consist of a transrepressor (optionally containing a silencing domain), which binds to specific (tandem) operator sequences and blocks transcription from upstream constitutive promoters until the transrepressor is released after interaction with the inducer ( Weber et al ., 2002 , 2005 ). Transactivators, which only bind to their operator modules in the presence of the inducer, are also classified as ON-type systems ( Weber et al ., 2004 ; Hartenbach and Fussenegger, 2005 ). OFF-type systems usually consist of a chimeric transactivator, which binds a specific (tandem) operator sequence and triggers transcription from an adjacent minimal promoter until it is released by interaction with the inducer ( Weber et al ., 2002 , 2003 ). A variety of these transcription control systems have been used for basic and applied research ( Weber and Fussenegger, 2007 ). The protein production machineries of mammalian cells and plants are known to be largely incompatible, which requires mammalian expression technology to be specifically modified for use in plant cells and plants ( Frey et al ., 2001 ; Mayfield et al ., 2003 ). The availability of cross-kingdom-compatible protein expression technology would significantly improve the use of plant cells for biopharmaceutical manufacturing. We provide comprehensive evidence that the transcription, translation and secretion machineries of mammalian cells and the non-seed plant P. patens are compatible, pioneer a novel protoplast-based fermentation technology for the production of human glycoproteins, and thus establish P. patens as a valuable host system for synthetic biology, in particular to functionally understand the most conserved molecular devices controlling biological signalling in the different kingdoms. <h1>Results</h1> <h2>Profiling of mammalian promoter activities in P. patens</h2> The swapping of expression units between mammalian and plant cell platforms for gene function analysis has been hampered by incompatibilities in the transcription/translation/secretion machineries. These systems require exclusive genetic elements (promoters, reporter genes, polyadenylation sites) for the expression of transgenes ( Frey et al ., 2001 ). In order to measure the activity of mammalian promoters in P. patens , isogenic, all-mammalian expression vectors were designed harbouring the human placental secreted alkaline phosphatase (SEAP), an easy-to-assay reporter gene, a polyadenylation site derived from simian virus 40 and various mammalian promoters (P hCMV , P SV40 , P GTX , P hEF1α ), including the smallest synthetic promoter P GTX ( Hartenbach and Fussenegger, 2006 ). The polioviral IRES (IRES PV ), known to be devoid of any promoter activity, was used as a negative control. Of the promoters tested, only P hEF1α was not functional in P. patens . Interestingly, the world's smallest synthetic promoter (182 bp) was fully functional, reaching P hCMV -driven expression levels in the moss. The SEAP expression profiles reached using mammalian expression vectors were compared with those of an isogenic plant expression vector encoding SEAP under the control of the cauliflower mosaic virus 35S promoter (P CaMV35S ) ( Figure 1 ). <h2>The secretory machineries of mammalian cells and P. patens are compatible</h2> Previous studies on mammalian promoter compatibility in the moss revealed that human placental SEAP could be secreted by P. patens protoplasts ( Figure 1 ). In order to assess whether product genes containing mammalian secretion signals are generically secreted by P. patens , several product genes, including SEAP, the Bacillus stearothermophilus -derived secreted α-amylase (SAMY) containing an immunoglobulin G (IgG)-derived secretion signal, human vascular endothelial growth factor 121 (VEGF 121 ) and human erythropoietin (EPO), were cloned into isogenic P GTX -driven mammalian and isogenic P CaMV35S -driven plant expression vectors. The product levels were profiled in the supernatant of transfected moss protoplast cultures, as well as in the cytosol, in order to assess the overall product secretion efficiency ( Table 1 ). All of the mammalian product proteins were efficiently secreted by moss protoplasts, whereas the control protein α-amylase (AMY) lacking the IgG secretion signal sequence could only be detected in the plant cytosol ( Table 1 ). In order to characterize the processing of secreted mammalian proteins in P. patens , we N-terminally sequenced human VEGF 121 purified from moss culture supernatants. The finding that the first 10 amino acids of secreted VEGF 121 were A(OH-Pro)MAEGGGQN suggests that secreted mammalian proteins are identically processed in P. patens . <h2>Mammalian cap-independent translation initiation is functional in P. patens</h2> Internal ribosome entry sites are capable of managing cap-independent translation initiation under physiological conditions which compromise classical cap-mediated translation ( Pestova et al ., 2001 ). Non-limiting examples of IRES-mediated translation include: (i) virus infection, during which the virus interferes with the cellular translation machinery and redirects it to translation of its IRES-tagged transcripts (viral IRES) ( Kaufman et al ., 1991 ; Dirks et al ., 1993 ), and (ii) hijacked cells may co-ordinate a molecular defence by translating a set of IRES-containing transcripts (cellular IRES) ( Gan and Rhoads, 1996 ). With the functionality of mammalian promoters and protein secretion established in P. patens , mammalian cell- and virus-derived IRESs were evaluated in P. patens to determine whether they could trigger translation initiation and enable multicistronic expression. We designed a variety of latest generation pTRIDENT vectors containing: (i) a constitutive P hCMV driving the transcription of multicistronic mRNAs; (ii) an artificial polyadenylation site (apA) signalling the terminus of the multicistronic transcript ( Hartenbach and Fussenegger, 2005 ); (iii) two tandem IRES elements of poliovirus (IRES PV ) ( Dirks et al ., 1993 ) or encephalomyocarditis virus (IRES EMCV ) ( Kaufman et al ., 1991 ) and IRES Rbm3 , derived from the human RNA-binding motif protein 3 (Rbm3) ( Chappell and Mauro, 2003 ); (iv) vast multiple cloning sites (MCSs) flanking each IRES element (many of which are targets for rare-cutting, 8-bp-recognizing restriction endonucleases) for complication-free sequential insertion of (v) product genes, including SAMY, VEGF 121 and SEAP; and (vi) P hCMV , SAMY-IRES-VEGF 121 -IRES-SEAP and apA, flanked by rare-cutting homing endonucleases (I- Ceu I, I- Sce I, I- Pop I, PI- Psp I), which enable, together with MCSs, straightforward exchange/swapping of expression modules and transgenes among different members of the pTRIDENT vector family ( Fux et al ., 2004 ). Following the transfection of pTRIDENT45 (P hCMV -SAMY-IRES PV -VEGF 121 -IRES EMCV -SEAP-apA), pTRIDENT46 (P hCMV -SAMY-IRES PV -VEGF 121 -IRES Rbm3 -SEAP-apA) and pTRIDENT47 (P hCMV -SAMY-IRES PV -VEGF 121 -IRES PV -SEAP-apA) into P. patens protoplasts, significant levels of product protein were produced from all positions within the vectors, indicating that mammalian cell/virus-derived IRES elements are functional and enable multicistronic transgene expression in the moss ( Figure 2 ). <h2>Tunable product gene expression in P. patens using mammalian transgene control technology</h2> Transcription control of specific genes by small trigger molecules is essential for gene function analysis ( Malleret et al ., 2001 ), drug discovery ( Weber et al ., 2008 ), the design of complex artificial gene circuits ( Kramer and Fussenegger, 2005 ), precise and timely molecular interventions in gene therapy ( Gersbach et al ., 2006 ), engineering of preferred cell phenotypes for tissue engineering ( Niwa et al ., 2000 ) and biopharmaceutical manufacturing ( Fussenegger et al ., 1998 ). Although a variety of transgene control systems are available for fine-tuning transgene transcription in mammalian cells ( Weber and Fussenegger, 2007 ), the choice for controlling transgene expression in plant cells, in particular in P. patens , is limited ( Saidi et al ., 2005 ). Recently, mammalian transcription control circuits have been designed which are responsive to the butyrolactone 2-(1′-hydroxy-6-methylheptyl)-3-(hydroxymethyl)-butanolide (SCB1) (QuoRex; Q-ON, Q-OFF) ( Weber et al ., 2003 , 2005 ), the macrolide antibiotic erythromycin (E.REX; E ON , E OFF ) ( Weber et al ., 2002 ) and acetaldehyde or ethanol (AIR) ( Weber et al ., 2004 , 2007 ). AIR-controlled transgenes are induced by acetaldehyde/ethanol, whereas E.REX and QuoRex are available in two different design versions, which can either be induced (E ON , Q-ON) or repressed (E OFF , Q-OFF) by the addition of the regulating molecule ( Weber et al ., 2002 , 2003, 2005 ). All mammalian transgene control systems were optimized for regulated SEAP expression and were transfected into P. patens , which was grown in the presence and absence of different trigger molecules at various concentrations. SCB1 was well tolerated by the moss (toxic only above 20 µg/mL, data not shown) and mediated adjustable, up to 15-fold induction (Q-ON; pWW504, P SV40 - scbR -KRAB-pA; pWW162, P SCA ON8-SEAP-pA) and repression (Q-OFF; pWW122, P SV40 - scbR -VP16-pA; pWW124, P SPA -SEAP-pA) of SEAP expression within a concentration range of 0–15 µg/mL ( Figure 3a,b ). The Q-ON system is so sensitive in P. patens that the moss senses the presence of co-cultivated SCB1-producing Streptomyces coelicolor , and participates in St. coelicolor's quorum-sensing cross-talk by adjusting SEAP production in response to the size of the bacterial population ( Figure 3c ). The E ON (pWW43, P SV40 -E-KRAB-pA; pWW56, P ETR ON8-SEAP-pA) and E OFF (pWW35, P SV40 -E-VP16-pA; pWW37, P ETR2 -SEAP-pA) systems were able to induce or repress SEAP expression up to 13-fold using erythromycin levels not exceeding 20 µg/mL (toxic above 30 µg/mL) ( Figure 3d,e ). The AIR system produced 20-fold SEAP expression (AIR; pWW195, P SV40 - alcR -pA; pWW192, P AIR -SEAP-pA) in the moss when induced by 20 µL/mL ethanol ( Figure 4a ). This compares favourably with the regulated performance of plant-specific, ethanol-mediated transgene regulation in tobacco ( Caddick et al ., 1998 ), Arabidopsis thaliana ( Roslan et al ., 2001 ), potato and oilseed rape ( Sweetman et al ., 2002 ). The AIR-controlled system is incredibly sensitive in P. patens , such that SEAP production can be induced by Saccharomyces cerevisiae populations cultivated at a distance. As part of its metabolism, Sa. cerevisiae converts ethanol into gaseous acetaldehyde, which reaches moss cultures ‘over the air’ and induces SEAP production in a distance-dependent manner ( Figure 4b ). The combination of AIR-based transcription control with multicistronic expression technology (pTRIDENT42; P AIR -SAMY-IRES PV -VEGF 121 -IRES EMCV -SEAP-apA) enabled the co-ordinated induction of three different transgenes after the addition of 10 µL/mL ethanol ( Figure 4c–e ). <h2>Autoregulated transgene expression in P. patens</h2> Classic transgene control systems consist of two expression vectors, one harbouring the transrepressor/transactivator and the other encoding the transgene driven by the trigger-inducible promoter ( Weber and Fussenegger, 2007 ). Such a two-vector design is more complex to engineer compared with the latest generation autoregulated one-vector configurations ( Hartenbach and Fussenegger, 2005 ). Capitalizing on the functionality of IRES elements in P. patens , protoplasts were transfected with the ethanol-controlled autoregulated SEAP expression vector pAutoRex8 (P AIR -SEAP-IRES PV - alcR -apA; ( Hartenbach and Fussenegger, 2005 ). pAutoRex8 contains a P AIR -driven dicistronic expression unit encoding SEAP in the first cistron and the acetaldehyde-dependent transactivator alcR in the second cistron. Leaky P AIR -driven transcripts provide sufficient AlcR to kick start maximum SEAP expression in the presence of inducing ethanol concentrations. In the absence of exogenous ethanol, the autoregulated circuit remains silent. The autoregulated AIR-controlled system reaches SEAP induction factors of up to 36-fold when transfected into P. patens ( Figure 4f ). <h2>VEGF 121 -based biopharmaceutical manufacturing using microencapsulated moss protoplasts</h2> The use of P. patens for the biopharmaceutical manufacture of protein therapeutics has been established, but remains challenging. The moss needs to be constantly blended in order to enable mixing in custom-designed, stirred-tank bioreactors ( Decker and Reski, 2007 ), and the plant cell wall potentially compromises the efficient secretion of larger product proteins. As plant protoplasts lack any cell wall and can be grown in single-cell suspension cultures, they would be the ideal plant cell system for biopharmaceutical manufacturing. However, protoplasts are too fragile and shear force-sensitive for use in state-of-the-art bioprocesses. We have pioneered a process to microencapsulate P. patens protoplasts in coherent alginate beads. Alginate bead polymerization is compatible with W5 culture medium, which was also used for the bioprocess. tWT11.51 VEGF -derived protoplasts (4 × 10 7 ) ( Baur et al ., 2005b ) were microencapsulated in 500-µm capsules (165 protoplasts per capsule) using state-of-the-art encapsulation technology, and cultivated for 9 days in a 2-L Wave Bioreactor operated at a culture volume of 1 L ( Figure 5 ). VEGF 121 production reached 53 mg/L in a 9-day process, which compares well with forefront bioprocesses using moss protonema. A fluorescein/trypan blue-based live/dead staining revealed that microencapsulated protoplasts cultivated for 9 days in a Wave Bioreactor were still 74.8% ± 7.2% viable, which represents only a 5% viability decrease compared with a freshly prepared protonema-derived protoplast population. <h1>Discussion</h1> The complete functionality of the central mammalian expression portfolio, which includes various promoters, mRNA processing signals, transcription factors, translation elements and secretion peptides ( Table 2 ), in the moss P. patens suggests that mammalian expression vectors and product proteins are generically compatible with this evolutionary old and simple plant. Interestingly, not only the functionality, but also the relative performance profiles, of different genetic elements in the moss matched those of mammalian cells. Examples of this include: (i) P hCMV being a stronger promoter than P SV40 , with comparable strength to the smallest synthetic promoter P GTX ( Hartenbach and Fussenegger, 2006 ); (ii) IRES PV and IRES EMCV being equally efficient in the triggering of translation initiation and outperforming IRES Rbm3 ( Fux et al ., 2004 ); (iii) terminal IRES-driven translation units showing lower expression levels from multicistronic mRNAs compared with cap-dependent translation initiation; (iv) the AIR-controlled system responsiveness to gaseous acetaldehyde or ethanol being the most sensitive transgene control modality ( Weber et al ., 2004 ); and (v) a one-vector-based, autoregulated expression configuration providing superior regulation performance. This cross-kingdom conservation of mammalian and moss protein production machineries is phylogenetically profound, and has several implications for basic and applied research. Comparative genomics, as well as functional studies, have recently established major differences in metabolic pathways and gene function between flowering plants and P. patens , and have suggested that a substantial moss gene pool is more closely related to mammals than to flowering plants ( Frank et al ., 2007 ; Rensing et al ., 2008 ). In combination with the functional data presented here, these findings may expand our classical view on the molecular division between plants and animals (e.g. in Yamamoto et al ., 2007 ). This differentially expressed gene pool may reveal unique cross-kingdom functionalities useful for future advances in agriculture and human health. With the discovery that two fundamentally different living systems, such as the moss and mammalian cells, can utilize each other's gene expression and protein production machineries may expand the way in which we perceive ecosystems containing different coexisting species, and could, at least theoretically, lead to the exchange of a compatible gene pool. Synthetic ecosystems have recently established the principle of cross-kingdom communication between Sa. cerevisiae or E. coli and mammalian cells, which replicated coexistence patterns as complex as oscillating predator–prey population dynamics ( Weber et al ., 2007 ), thus expanding our view on quorum sensing between bacteria ( Keller and Surette, 2006 ). We have shown here that P. patens harbouring mammalian gene circuits is responsive to quorum-sensing communication initiated by co-cultivated St. coelicolor , as well as to ‘over-the-air’ signalling triggered by Sa. cerevisiae cultivated adjacently. Rational interventions into the quorum-sensing networks may foster unprecedented advances in agriculture, replicating the progress achieved in attenuating host–pathogen interaction in human therapy ( Benghezal et al ., 2006 ). Moreover, our recent findings have established P. patens as a promising host system for synthetic biology, a novel approach in the life sciences that relies on iterative cycles between analysis and synthesis ( Benner and Sismour, 2005 ), utilizing devices of signalling networks in a cross-kingdom manner (e.g. Khandelwal et al ., 2007 ). Several biopharmaceutical production platforms, including E. coli ( Georgiou and Segatori, 2005 ), (glyco-engineered) Sa. cerevisiae ( Hamilton et al ., 2006 ), mammalian cells ( Wurm, 2004 ) and transgenic animals ( Larrick and Thomas, 2001 ), are competing for the industrial production of protein therapeutics ( Fussenegger and Hauser, 2007 ). Mammalian cells have become the dominant system for the production of recombinant protein pharmaceuticals, partly as a result of the availability of a highly advanced portfolio of expression vectors and engineering strategies ( Umana et al ., 1999 ; Wurm, 2004 ; Hartenbach and Fussenegger, 2005 ). With the global mammalian cell-based biopharmaceutical manufacturing capacity plateauing into a bottleneck, this compromises the availability of drugs to patients. Alternative easy-to-implement bioprocessing concepts are urgently needed ( Fussenegger and Hauser, 2007 ). The moss P. patens has recently come into the limelight as an easy-to-handle/engineer organism which could be cultivated in scale-up-compatible bioreactors, and was able to produce ADCC-optimized therapeutic IgGs in a Good Manufacturing Practice (GMP)-approved bioprocess ( Decker and Reski, 2007 ). Utilizing a compatible mammalian expression and engineering toolbox, the moss, as an emerging biopharmaceutical manufacturing platform, could be propelled to an ex-aequo competitor of mammalian cell-based production systems. Major bioprocess advantages of P. patens include the use of an inexpensive salt solution as production medium, which reduces downstream processing challenges and costs, and the availability of an efficient homologous recombination toolkit that provides stable and predictable production cultures ( Kamisugi et al ., 2006 ). Best-in-class production systems include transient protoplast cultures for the rapid evaluation of bioprocess parameters and a scalable stirred-tank photo-bioreactor that uses stable moss protonema. Moss protonema tissue needs to be constantly blended to avoid complications in bioreactor operation, which may hinder large-scale biopharmaceutical manufacturing, and the established cell wall may compromise the secretion of larger proteins. Protoplasts could be an alternative ( Baur et al ., 2005a ), but they are not sufficiently robust to survive long-term bioreactor operation. The microencapsulation protocol established during this study is compatible with the W5 medium, and enables the cultivation of encapsulated protoplasts in a proliferation-inhibited and cell wall-free state. Being protected by a physiologically inert alginate shell, the protoplasts are able to devote all of their metabolic energy to the production of heterologous protein rather than biomass, and, being devoid of any secretion-limiting cell wall, microencapsulated protoplasts cultivated in a standard Wave Bioreactor, equipped with a photosynthesis kit, were able to produce the human growth factor VEGF 121 at titres comparable with those of the highly optimized best-in-class protonema cultures. The use of Wave Bioreactor systems, which can be easily upscaled to 500-L cultures, has recently gathered momentum for use in the pilot production of proteins for clinical trials ( Haldankar et al ., 2006 ). The combination of a novel protoplast-based bioprocess with powerful mammalian expression technology will further enhance the use of P. patens as a complementary and competitive platform for the biopharmaceutical manufacturing of protein therapeutics, and establishes this evolutionary old and simple plant as a valuable host for synthetic biology. <h1>Experimental procedures</h1> <h2>Expression vector design</h2> Table 3 lists all the plasmids used in this study and provides detailed information about their construction. <h2>Cultivation and transformation of P. patens</h2> Physcomitrella patens (Hedw.) B.S.G. was grown axenically in Erlenmeyer flasks or in modified stirred-tank bioreactors (5 L; Applikon, Schiedam, the Netherlands) using a 10% modified Knop salt solution (100 mg/L Ca(NO 3 ) 2 ·4H 2 O, 25 mg/L KCl, 25 mg/L KH 2 PO 4 , 25 mg/L MgSO 4 ·7H 2 O and 1.25 mg/L FeSO 4 ·7H 2 O; pH 5.8) ( Reski and Abel, 1985 ). Protoplasts of P. patens were generated by incubation for 2 h in 0.5 m mannitol containing 4% Driselase (Sigma, Buchs, Switzerland), followed by two centrifugation steps (10 min, 50 g ) and resuspension of the protoplast-containing pellet at a desired cell density in 3M medium [87.5 g/L mannitol, 3.1 g/L MgCl 2 ·6H 2 O, 1 g/L 2-( N -morpholino)ethanesulphonic acid hydrate (MES); Sigma; pH 5.6 and 580 mOsm]. Protoplasts (300 000) were chemically transfected with 50 µg/mL DNA (80 µg/mL for transgene control systems), as described previously ( Jost et al ., 2005 ), and cultivated in Knop's regeneration medium [1 g/L Ca(NO 3 ) 2 ·4H 2 O, 250 mg/L KCl, 250 mg/L KH 2 PO 4 , 250 mg/L MgSO 4 ·7H 2 O, 12.5 mg/L FeSO 4 ·7H 2 O, 5% glucose, 3% mannitol, pH 5.7, 540 mOsm]. <h2>Protein production</h2> Protein production was measured 5 days after transformation using standardized assays: (i) human placental SEAP: a p -nitrophenylphosphate-based light-absorbance time course ( Berger et al ., 1988 ; Schlatter et al ., 2002 ); (ii) SAMY and AMY: a blue starch Phadebas ® assay (cat. no. 10-5380-32; Pharmacia Upjohn, Peapack, NJ, USA) ( Schlatter et al ., 2002 ); quantification of intracellular reporter proteins required lysis of the plant cells by four freeze–thaw cycles and elimination of cell debris by centrifugation (2 min at 12 000 g ); (iii) human VEGF 121 : using a VEGF 121 -specific enzyme-linked immunosorbent assay (ELISA) (cat. no. 900-K10, lot. no. 1006010; Peprotech, Rocky Hill, NJ, USA); (iv) human EPO: using an EPO-specific ELISA (Quantikine ® IVD ® , cat. no. DEP00, lot. no. 243030; R & D Systems, Minneapolis, MN, USA). <h2>Transgene regulation</h2> All regulating agents were administered at the indicated concentrations immediately after transformation. The butyrolactone SCB1 was synthesized and purified as described previously ( Weber et al ., 2003 ). Erythromycin (cat. no. 45673, lot. no. 1195447; Fluka, Buchs, Switzerland) was prepared as a stock solution of 1 mg/mL in ethanol. The AIR system was induced by the addition of the indicated volumes of 100% ethanol. <h2>Microencapsulation of P. patens protoplasts</h2> Protoplasts (4 × 10 7 ) generated from transgenic VEGF 121 -producing moss tWT11.51 VEGF ( Baur et al ., 2005b ) were resuspended in 8 mL of 3M medium and stirred gently with 40 mL of 1.5% sodium alginate solution (cat. no. IE1010, lot. no. 060125B1; Inotech Biotechnologies Ltd., Basle, Switzerland). Protoplasts were encapsulated in 500-µm alginate capsules (165 protoplasts per capsule) using an Inotech Encapsulator Research IE-50R (Inotech Biotechnologies Ltd., Basle, Switzerland) set at the following parameters: 0.5 mm nozzle; 853 unit flow rate using a 50-mL syringe; 1250 s −1 nozzle vibration frequency; 1.4 kV for bead dispersion. W5 medium (18.4 g/L CaCl 2 , 8 g/L NaCl 2 , 0.99 g/L glucose, 0.75 g/L KCl; pH 5.8, 600 mOsm) was used as a precipitation solution. It contains sufficient CaCl 2 for the precipitation of alginate beads and enables the direct cultivation of microencapsulated protoplast populations without the need for medium exchange. Microencapsulated tWT11.51 VEGF -derived (a high VEGF 121 producer line harbouring the product gene under the control of the moss actin 5 5′ region) protoplasts (240 000 capsules, 1 L W5 medium) were cultivated in a BioWave 20SPS-F bioreactor (Wave Biotech, Tagelswangen, Switzerland), equipped with 2-L Wave Bags and set at the following parameters: aeration, 100 mL/min sterilized air; rocking rate, 19 min −1 ; rocking angle, 10°. The Wave Bioreactor was placed in an ISF-1-W incubator equipped with a photosynthesis kit set to 25 °C and a day/night cycle of 16 h/8 h (Kuehner, Birsfelden, Switzerland). <h2>Edman sequencing</h2> VEGF 121 was precipitated from tWT11.51 VEGF protoplast culture supernatants for 10 min at 4 °C with 100% w/v trichloroacetic acid (TCA, Sigma) (supernatant : TCA, 9 : 1). The samples were centrifuged for 5 min at 12 000 g , and the VEGF 121 -containing pellet was washed twice in ice-cold acetone, dried and resuspended in 2 × sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) reducing sample buffer [50% glycerol, 250 m m tris(hydroxymethyl)aminomethane (Tris), 10% SDS, 500 m m dithiothreitol, 0.01% bromophenol blue, pH 6.8]. The samples were then denatured for 5 min at 50 °C, and the proteins were size-fractionated on a 12% SDS-polyacrylamide gel and blotted on to a polyvinylidene fluoride membrane (cat. no. IPVH20200; Millipore Corporation, Bedford, MA, USA). VEGF 121 (35 kDa) was N-terminally sequenced on an Applied Biosystems (Foster City, CA, USA) model 492cLC Procise protein/peptide sequencer with an on-line Perkin-Elmer (Waltham, MA, USA) Applied Biosystems Model 140C PTH Amino Acid (phenylthiohydantonin amino acid) Analyser. The PTH amino acids were automatically transferred to a reverse-phase C-18 column (0.8 mm inside diameter) for detection at 269 nm, and identified by comparison with individual runs with a standard mixture of PTH amino acids. <h2>Cultivation of Sa. cerevisiae and St. coelicolor</h2> Saccharomyces cerevisiae [wild-type strain W303, BMA 64, European Sa. cerevisiae archive for functional analysis (EUROSCARF), Frankfurt, Germany] was cultivated on yeast–peptone–dextrose agar (YPD; 1% yeast extract, 2% peptone, 2% dextrose, 1% agar), and St. coelicolor MT1110 (kindly provided by Marc Folcher) was cultivated on mannitol–soy agar (2% soy flour, 2% mannitol, 1.5% agar). For the co-cultivation of St. coelicolor and P. patens , St. coelicolor were pre-cultured in Luria–Bertani (LB) medium to an optical density at 600 nm (OD 600 ) of 340, and the indicated volumes were then transferred to P. patens maintained in regeneration medium. <h2>Viability profiling of microencapsulated P. patens protoplasts</h2> The viability of microencapsulated moss protoplasts was determined by scoring live protoplasts, stained with fluorescein diacetate (FDA), and dead protoplasts, stained with trypan blue, using (fluorescence) microscopy (Leica DM-RB fluorescence microscope; Leica, Heerbrugg, Switzerland). Twenty protoplast-containing alginate beads were incubated for 10 min in a staining solution including 200 µL W5 medium, 20 µL phosphate-buffered saline (PBS; Dulbecco's phosphate-buffered saline, cat. no. 21600-0069; Invitrogen, Basle, Switzerland) containing 0.01% FDA (Sigma) and 40 µL of a 0.4% trypan blue stock solution (lot. no. 1230532; Fluka). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Plant Biotechnology Journal Wiley

Functional cross-kingdom conservation of mammalian and moss ( Physcomitrella patens ) transcription, translation and secretion machineries

Loading next page...
 
/lp/wiley/functional-cross-kingdom-conservation-of-mammalian-and-moss-ldMEVmpqSM

References (58)

Publisher
Wiley
Copyright
Journal compilation © 2009 Blackwell Publishing Ltd
ISSN
1467-7644
eISSN
1467-7652
DOI
10.1111/j.1467-7652.2008.00376.x
pmid
19021876
Publisher site
See Article on Publisher Site

Abstract

<h1>Introduction</h1> The moss Physcomitrella patens , unique for its high rate of homologous recombination, generic codon usage, haploidy, simple body plan, physiologic properties and exclusive phylogenetic position ( Quatrano et al ., 2007 ; Rensing et al ., 2008 ), is gathering momentum for the biopharmaceutical manufacture of protein therapeutics because of favourable bioprocess and downstream processing economics ( Decker and Reski, 2007 ). In vitro cultivation of P. patens throughout its complete life cycle ( Frank et al ., 2005 ), transgenic protonema and transient protoplast cultures ( Baur et al ., 2005a ), stirred-tank and tubular photo-bioreactors ( Decker and Reski, 2007 ), the generation of moss mutants devoid of immunogenic product protein glycosylation ( Huether et al ., 2005 ), and the production of human antibodies with improved antibody-dependent cellular cytotoxicity (ADCC) activity ( Nechansky et al ., 2007 ) have been important milestones in establishing the moss as a promising biopharmaceutical manufacturing platform ( Decker and Reski, 2007 ). Transgenic mammalian cell cultures are currently the most successful platform for the production of biopharmaceuticals, as most of the protein therapeutics on the market originate from mammalian cell bioprocesses ( Wurm, 2004 ). However, as the global product pipeline exceeds the worldwide manufacturing capacity, alternative host cell systems for biopharmaceutical manufacturing are on the increase ( Hamilton et al ., 2006 ; Decker and Reski, 2007 ). Since the mid-1980s, the productivity of mammalian cells cultivated in bioreactors has reached the gram per litre range, an over 100-fold yield improvement over titres achieved for the first commercial bioprocesses ( Wurm, 2004 ). Part of this success is based on the development of sophisticated expression technologies and metabolic engineering strategies ( Umana et al ., 1999 ; Hartenbach and Fussenegger, 2005 ). The latest generation of expression vectors harbour: (i) compact strong constitutive promoters for the high-level transcription of product genes ( Hartenbach and Fussenegger, 2006 ); (ii) multicistronic expression units enabling one-vector-based selection and expression of multiprotein complexes ( Fux et al ., 2004 ); and (iii) regulated expression systems for the production of difficult-to-express protein therapeutics ( Weber and Fussenegger, 2007 ). In mammalian cells, the initiation of translation is typically managed by a cap structure which is post-transcriptionally attached to the 5′ end of mRNAs ( Kozak, 1989 ). Alternatively, internal ribosome entry sites (IRESs), which adopt a specific secondary RNA structure triggering ribosome assembly and translational initiation, have evolved to ensure a minimal level of protein synthesis for survival during cap-compromising physiological emergency situations (co-ordination of viral defence; cellular IRES) ( Gan and Rhoads, 1996 ) or to redirect the cellular translation machinery to the production of virus proteins (viral IRES) ( Kaufman et al ., 1991 ; Dirks et al ., 1993 ). The tandem arrangement of different transgenes, each preceded by an IRES element, enables the transcription of a multicistronic mRNA producing stoichiometric levels of various proteins. Recently, a sophisticated vector platform (pTRIDENT) has been designed for the multicistronic expression of up to three different transgenes ( Fux et al ., 2004 ). Heterologous mammalian transcription control modalities have been designed in two different configurations: ON-type systems, which are induced following the addition of a trigger molecule, and OFF-type systems, which are repressed after the administration of a regulating compound ( Weber and Fussenegger, 2007 ). ON-type systems typically consist of a transrepressor (optionally containing a silencing domain), which binds to specific (tandem) operator sequences and blocks transcription from upstream constitutive promoters until the transrepressor is released after interaction with the inducer ( Weber et al ., 2002 , 2005 ). Transactivators, which only bind to their operator modules in the presence of the inducer, are also classified as ON-type systems ( Weber et al ., 2004 ; Hartenbach and Fussenegger, 2005 ). OFF-type systems usually consist of a chimeric transactivator, which binds a specific (tandem) operator sequence and triggers transcription from an adjacent minimal promoter until it is released by interaction with the inducer ( Weber et al ., 2002 , 2003 ). A variety of these transcription control systems have been used for basic and applied research ( Weber and Fussenegger, 2007 ). The protein production machineries of mammalian cells and plants are known to be largely incompatible, which requires mammalian expression technology to be specifically modified for use in plant cells and plants ( Frey et al ., 2001 ; Mayfield et al ., 2003 ). The availability of cross-kingdom-compatible protein expression technology would significantly improve the use of plant cells for biopharmaceutical manufacturing. We provide comprehensive evidence that the transcription, translation and secretion machineries of mammalian cells and the non-seed plant P. patens are compatible, pioneer a novel protoplast-based fermentation technology for the production of human glycoproteins, and thus establish P. patens as a valuable host system for synthetic biology, in particular to functionally understand the most conserved molecular devices controlling biological signalling in the different kingdoms. <h1>Results</h1> <h2>Profiling of mammalian promoter activities in P. patens</h2> The swapping of expression units between mammalian and plant cell platforms for gene function analysis has been hampered by incompatibilities in the transcription/translation/secretion machineries. These systems require exclusive genetic elements (promoters, reporter genes, polyadenylation sites) for the expression of transgenes ( Frey et al ., 2001 ). In order to measure the activity of mammalian promoters in P. patens , isogenic, all-mammalian expression vectors were designed harbouring the human placental secreted alkaline phosphatase (SEAP), an easy-to-assay reporter gene, a polyadenylation site derived from simian virus 40 and various mammalian promoters (P hCMV , P SV40 , P GTX , P hEF1α ), including the smallest synthetic promoter P GTX ( Hartenbach and Fussenegger, 2006 ). The polioviral IRES (IRES PV ), known to be devoid of any promoter activity, was used as a negative control. Of the promoters tested, only P hEF1α was not functional in P. patens . Interestingly, the world's smallest synthetic promoter (182 bp) was fully functional, reaching P hCMV -driven expression levels in the moss. The SEAP expression profiles reached using mammalian expression vectors were compared with those of an isogenic plant expression vector encoding SEAP under the control of the cauliflower mosaic virus 35S promoter (P CaMV35S ) ( Figure 1 ). <h2>The secretory machineries of mammalian cells and P. patens are compatible</h2> Previous studies on mammalian promoter compatibility in the moss revealed that human placental SEAP could be secreted by P. patens protoplasts ( Figure 1 ). In order to assess whether product genes containing mammalian secretion signals are generically secreted by P. patens , several product genes, including SEAP, the Bacillus stearothermophilus -derived secreted α-amylase (SAMY) containing an immunoglobulin G (IgG)-derived secretion signal, human vascular endothelial growth factor 121 (VEGF 121 ) and human erythropoietin (EPO), were cloned into isogenic P GTX -driven mammalian and isogenic P CaMV35S -driven plant expression vectors. The product levels were profiled in the supernatant of transfected moss protoplast cultures, as well as in the cytosol, in order to assess the overall product secretion efficiency ( Table 1 ). All of the mammalian product proteins were efficiently secreted by moss protoplasts, whereas the control protein α-amylase (AMY) lacking the IgG secretion signal sequence could only be detected in the plant cytosol ( Table 1 ). In order to characterize the processing of secreted mammalian proteins in P. patens , we N-terminally sequenced human VEGF 121 purified from moss culture supernatants. The finding that the first 10 amino acids of secreted VEGF 121 were A(OH-Pro)MAEGGGQN suggests that secreted mammalian proteins are identically processed in P. patens . <h2>Mammalian cap-independent translation initiation is functional in P. patens</h2> Internal ribosome entry sites are capable of managing cap-independent translation initiation under physiological conditions which compromise classical cap-mediated translation ( Pestova et al ., 2001 ). Non-limiting examples of IRES-mediated translation include: (i) virus infection, during which the virus interferes with the cellular translation machinery and redirects it to translation of its IRES-tagged transcripts (viral IRES) ( Kaufman et al ., 1991 ; Dirks et al ., 1993 ), and (ii) hijacked cells may co-ordinate a molecular defence by translating a set of IRES-containing transcripts (cellular IRES) ( Gan and Rhoads, 1996 ). With the functionality of mammalian promoters and protein secretion established in P. patens , mammalian cell- and virus-derived IRESs were evaluated in P. patens to determine whether they could trigger translation initiation and enable multicistronic expression. We designed a variety of latest generation pTRIDENT vectors containing: (i) a constitutive P hCMV driving the transcription of multicistronic mRNAs; (ii) an artificial polyadenylation site (apA) signalling the terminus of the multicistronic transcript ( Hartenbach and Fussenegger, 2005 ); (iii) two tandem IRES elements of poliovirus (IRES PV ) ( Dirks et al ., 1993 ) or encephalomyocarditis virus (IRES EMCV ) ( Kaufman et al ., 1991 ) and IRES Rbm3 , derived from the human RNA-binding motif protein 3 (Rbm3) ( Chappell and Mauro, 2003 ); (iv) vast multiple cloning sites (MCSs) flanking each IRES element (many of which are targets for rare-cutting, 8-bp-recognizing restriction endonucleases) for complication-free sequential insertion of (v) product genes, including SAMY, VEGF 121 and SEAP; and (vi) P hCMV , SAMY-IRES-VEGF 121 -IRES-SEAP and apA, flanked by rare-cutting homing endonucleases (I- Ceu I, I- Sce I, I- Pop I, PI- Psp I), which enable, together with MCSs, straightforward exchange/swapping of expression modules and transgenes among different members of the pTRIDENT vector family ( Fux et al ., 2004 ). Following the transfection of pTRIDENT45 (P hCMV -SAMY-IRES PV -VEGF 121 -IRES EMCV -SEAP-apA), pTRIDENT46 (P hCMV -SAMY-IRES PV -VEGF 121 -IRES Rbm3 -SEAP-apA) and pTRIDENT47 (P hCMV -SAMY-IRES PV -VEGF 121 -IRES PV -SEAP-apA) into P. patens protoplasts, significant levels of product protein were produced from all positions within the vectors, indicating that mammalian cell/virus-derived IRES elements are functional and enable multicistronic transgene expression in the moss ( Figure 2 ). <h2>Tunable product gene expression in P. patens using mammalian transgene control technology</h2> Transcription control of specific genes by small trigger molecules is essential for gene function analysis ( Malleret et al ., 2001 ), drug discovery ( Weber et al ., 2008 ), the design of complex artificial gene circuits ( Kramer and Fussenegger, 2005 ), precise and timely molecular interventions in gene therapy ( Gersbach et al ., 2006 ), engineering of preferred cell phenotypes for tissue engineering ( Niwa et al ., 2000 ) and biopharmaceutical manufacturing ( Fussenegger et al ., 1998 ). Although a variety of transgene control systems are available for fine-tuning transgene transcription in mammalian cells ( Weber and Fussenegger, 2007 ), the choice for controlling transgene expression in plant cells, in particular in P. patens , is limited ( Saidi et al ., 2005 ). Recently, mammalian transcription control circuits have been designed which are responsive to the butyrolactone 2-(1′-hydroxy-6-methylheptyl)-3-(hydroxymethyl)-butanolide (SCB1) (QuoRex; Q-ON, Q-OFF) ( Weber et al ., 2003 , 2005 ), the macrolide antibiotic erythromycin (E.REX; E ON , E OFF ) ( Weber et al ., 2002 ) and acetaldehyde or ethanol (AIR) ( Weber et al ., 2004 , 2007 ). AIR-controlled transgenes are induced by acetaldehyde/ethanol, whereas E.REX and QuoRex are available in two different design versions, which can either be induced (E ON , Q-ON) or repressed (E OFF , Q-OFF) by the addition of the regulating molecule ( Weber et al ., 2002 , 2003, 2005 ). All mammalian transgene control systems were optimized for regulated SEAP expression and were transfected into P. patens , which was grown in the presence and absence of different trigger molecules at various concentrations. SCB1 was well tolerated by the moss (toxic only above 20 µg/mL, data not shown) and mediated adjustable, up to 15-fold induction (Q-ON; pWW504, P SV40 - scbR -KRAB-pA; pWW162, P SCA ON8-SEAP-pA) and repression (Q-OFF; pWW122, P SV40 - scbR -VP16-pA; pWW124, P SPA -SEAP-pA) of SEAP expression within a concentration range of 0–15 µg/mL ( Figure 3a,b ). The Q-ON system is so sensitive in P. patens that the moss senses the presence of co-cultivated SCB1-producing Streptomyces coelicolor , and participates in St. coelicolor's quorum-sensing cross-talk by adjusting SEAP production in response to the size of the bacterial population ( Figure 3c ). The E ON (pWW43, P SV40 -E-KRAB-pA; pWW56, P ETR ON8-SEAP-pA) and E OFF (pWW35, P SV40 -E-VP16-pA; pWW37, P ETR2 -SEAP-pA) systems were able to induce or repress SEAP expression up to 13-fold using erythromycin levels not exceeding 20 µg/mL (toxic above 30 µg/mL) ( Figure 3d,e ). The AIR system produced 20-fold SEAP expression (AIR; pWW195, P SV40 - alcR -pA; pWW192, P AIR -SEAP-pA) in the moss when induced by 20 µL/mL ethanol ( Figure 4a ). This compares favourably with the regulated performance of plant-specific, ethanol-mediated transgene regulation in tobacco ( Caddick et al ., 1998 ), Arabidopsis thaliana ( Roslan et al ., 2001 ), potato and oilseed rape ( Sweetman et al ., 2002 ). The AIR-controlled system is incredibly sensitive in P. patens , such that SEAP production can be induced by Saccharomyces cerevisiae populations cultivated at a distance. As part of its metabolism, Sa. cerevisiae converts ethanol into gaseous acetaldehyde, which reaches moss cultures ‘over the air’ and induces SEAP production in a distance-dependent manner ( Figure 4b ). The combination of AIR-based transcription control with multicistronic expression technology (pTRIDENT42; P AIR -SAMY-IRES PV -VEGF 121 -IRES EMCV -SEAP-apA) enabled the co-ordinated induction of three different transgenes after the addition of 10 µL/mL ethanol ( Figure 4c–e ). <h2>Autoregulated transgene expression in P. patens</h2> Classic transgene control systems consist of two expression vectors, one harbouring the transrepressor/transactivator and the other encoding the transgene driven by the trigger-inducible promoter ( Weber and Fussenegger, 2007 ). Such a two-vector design is more complex to engineer compared with the latest generation autoregulated one-vector configurations ( Hartenbach and Fussenegger, 2005 ). Capitalizing on the functionality of IRES elements in P. patens , protoplasts were transfected with the ethanol-controlled autoregulated SEAP expression vector pAutoRex8 (P AIR -SEAP-IRES PV - alcR -apA; ( Hartenbach and Fussenegger, 2005 ). pAutoRex8 contains a P AIR -driven dicistronic expression unit encoding SEAP in the first cistron and the acetaldehyde-dependent transactivator alcR in the second cistron. Leaky P AIR -driven transcripts provide sufficient AlcR to kick start maximum SEAP expression in the presence of inducing ethanol concentrations. In the absence of exogenous ethanol, the autoregulated circuit remains silent. The autoregulated AIR-controlled system reaches SEAP induction factors of up to 36-fold when transfected into P. patens ( Figure 4f ). <h2>VEGF 121 -based biopharmaceutical manufacturing using microencapsulated moss protoplasts</h2> The use of P. patens for the biopharmaceutical manufacture of protein therapeutics has been established, but remains challenging. The moss needs to be constantly blended in order to enable mixing in custom-designed, stirred-tank bioreactors ( Decker and Reski, 2007 ), and the plant cell wall potentially compromises the efficient secretion of larger product proteins. As plant protoplasts lack any cell wall and can be grown in single-cell suspension cultures, they would be the ideal plant cell system for biopharmaceutical manufacturing. However, protoplasts are too fragile and shear force-sensitive for use in state-of-the-art bioprocesses. We have pioneered a process to microencapsulate P. patens protoplasts in coherent alginate beads. Alginate bead polymerization is compatible with W5 culture medium, which was also used for the bioprocess. tWT11.51 VEGF -derived protoplasts (4 × 10 7 ) ( Baur et al ., 2005b ) were microencapsulated in 500-µm capsules (165 protoplasts per capsule) using state-of-the-art encapsulation technology, and cultivated for 9 days in a 2-L Wave Bioreactor operated at a culture volume of 1 L ( Figure 5 ). VEGF 121 production reached 53 mg/L in a 9-day process, which compares well with forefront bioprocesses using moss protonema. A fluorescein/trypan blue-based live/dead staining revealed that microencapsulated protoplasts cultivated for 9 days in a Wave Bioreactor were still 74.8% ± 7.2% viable, which represents only a 5% viability decrease compared with a freshly prepared protonema-derived protoplast population. <h1>Discussion</h1> The complete functionality of the central mammalian expression portfolio, which includes various promoters, mRNA processing signals, transcription factors, translation elements and secretion peptides ( Table 2 ), in the moss P. patens suggests that mammalian expression vectors and product proteins are generically compatible with this evolutionary old and simple plant. Interestingly, not only the functionality, but also the relative performance profiles, of different genetic elements in the moss matched those of mammalian cells. Examples of this include: (i) P hCMV being a stronger promoter than P SV40 , with comparable strength to the smallest synthetic promoter P GTX ( Hartenbach and Fussenegger, 2006 ); (ii) IRES PV and IRES EMCV being equally efficient in the triggering of translation initiation and outperforming IRES Rbm3 ( Fux et al ., 2004 ); (iii) terminal IRES-driven translation units showing lower expression levels from multicistronic mRNAs compared with cap-dependent translation initiation; (iv) the AIR-controlled system responsiveness to gaseous acetaldehyde or ethanol being the most sensitive transgene control modality ( Weber et al ., 2004 ); and (v) a one-vector-based, autoregulated expression configuration providing superior regulation performance. This cross-kingdom conservation of mammalian and moss protein production machineries is phylogenetically profound, and has several implications for basic and applied research. Comparative genomics, as well as functional studies, have recently established major differences in metabolic pathways and gene function between flowering plants and P. patens , and have suggested that a substantial moss gene pool is more closely related to mammals than to flowering plants ( Frank et al ., 2007 ; Rensing et al ., 2008 ). In combination with the functional data presented here, these findings may expand our classical view on the molecular division between plants and animals (e.g. in Yamamoto et al ., 2007 ). This differentially expressed gene pool may reveal unique cross-kingdom functionalities useful for future advances in agriculture and human health. With the discovery that two fundamentally different living systems, such as the moss and mammalian cells, can utilize each other's gene expression and protein production machineries may expand the way in which we perceive ecosystems containing different coexisting species, and could, at least theoretically, lead to the exchange of a compatible gene pool. Synthetic ecosystems have recently established the principle of cross-kingdom communication between Sa. cerevisiae or E. coli and mammalian cells, which replicated coexistence patterns as complex as oscillating predator–prey population dynamics ( Weber et al ., 2007 ), thus expanding our view on quorum sensing between bacteria ( Keller and Surette, 2006 ). We have shown here that P. patens harbouring mammalian gene circuits is responsive to quorum-sensing communication initiated by co-cultivated St. coelicolor , as well as to ‘over-the-air’ signalling triggered by Sa. cerevisiae cultivated adjacently. Rational interventions into the quorum-sensing networks may foster unprecedented advances in agriculture, replicating the progress achieved in attenuating host–pathogen interaction in human therapy ( Benghezal et al ., 2006 ). Moreover, our recent findings have established P. patens as a promising host system for synthetic biology, a novel approach in the life sciences that relies on iterative cycles between analysis and synthesis ( Benner and Sismour, 2005 ), utilizing devices of signalling networks in a cross-kingdom manner (e.g. Khandelwal et al ., 2007 ). Several biopharmaceutical production platforms, including E. coli ( Georgiou and Segatori, 2005 ), (glyco-engineered) Sa. cerevisiae ( Hamilton et al ., 2006 ), mammalian cells ( Wurm, 2004 ) and transgenic animals ( Larrick and Thomas, 2001 ), are competing for the industrial production of protein therapeutics ( Fussenegger and Hauser, 2007 ). Mammalian cells have become the dominant system for the production of recombinant protein pharmaceuticals, partly as a result of the availability of a highly advanced portfolio of expression vectors and engineering strategies ( Umana et al ., 1999 ; Wurm, 2004 ; Hartenbach and Fussenegger, 2005 ). With the global mammalian cell-based biopharmaceutical manufacturing capacity plateauing into a bottleneck, this compromises the availability of drugs to patients. Alternative easy-to-implement bioprocessing concepts are urgently needed ( Fussenegger and Hauser, 2007 ). The moss P. patens has recently come into the limelight as an easy-to-handle/engineer organism which could be cultivated in scale-up-compatible bioreactors, and was able to produce ADCC-optimized therapeutic IgGs in a Good Manufacturing Practice (GMP)-approved bioprocess ( Decker and Reski, 2007 ). Utilizing a compatible mammalian expression and engineering toolbox, the moss, as an emerging biopharmaceutical manufacturing platform, could be propelled to an ex-aequo competitor of mammalian cell-based production systems. Major bioprocess advantages of P. patens include the use of an inexpensive salt solution as production medium, which reduces downstream processing challenges and costs, and the availability of an efficient homologous recombination toolkit that provides stable and predictable production cultures ( Kamisugi et al ., 2006 ). Best-in-class production systems include transient protoplast cultures for the rapid evaluation of bioprocess parameters and a scalable stirred-tank photo-bioreactor that uses stable moss protonema. Moss protonema tissue needs to be constantly blended to avoid complications in bioreactor operation, which may hinder large-scale biopharmaceutical manufacturing, and the established cell wall may compromise the secretion of larger proteins. Protoplasts could be an alternative ( Baur et al ., 2005a ), but they are not sufficiently robust to survive long-term bioreactor operation. The microencapsulation protocol established during this study is compatible with the W5 medium, and enables the cultivation of encapsulated protoplasts in a proliferation-inhibited and cell wall-free state. Being protected by a physiologically inert alginate shell, the protoplasts are able to devote all of their metabolic energy to the production of heterologous protein rather than biomass, and, being devoid of any secretion-limiting cell wall, microencapsulated protoplasts cultivated in a standard Wave Bioreactor, equipped with a photosynthesis kit, were able to produce the human growth factor VEGF 121 at titres comparable with those of the highly optimized best-in-class protonema cultures. The use of Wave Bioreactor systems, which can be easily upscaled to 500-L cultures, has recently gathered momentum for use in the pilot production of proteins for clinical trials ( Haldankar et al ., 2006 ). The combination of a novel protoplast-based bioprocess with powerful mammalian expression technology will further enhance the use of P. patens as a complementary and competitive platform for the biopharmaceutical manufacturing of protein therapeutics, and establishes this evolutionary old and simple plant as a valuable host for synthetic biology. <h1>Experimental procedures</h1> <h2>Expression vector design</h2> Table 3 lists all the plasmids used in this study and provides detailed information about their construction. <h2>Cultivation and transformation of P. patens</h2> Physcomitrella patens (Hedw.) B.S.G. was grown axenically in Erlenmeyer flasks or in modified stirred-tank bioreactors (5 L; Applikon, Schiedam, the Netherlands) using a 10% modified Knop salt solution (100 mg/L Ca(NO 3 ) 2 ·4H 2 O, 25 mg/L KCl, 25 mg/L KH 2 PO 4 , 25 mg/L MgSO 4 ·7H 2 O and 1.25 mg/L FeSO 4 ·7H 2 O; pH 5.8) ( Reski and Abel, 1985 ). Protoplasts of P. patens were generated by incubation for 2 h in 0.5 m mannitol containing 4% Driselase (Sigma, Buchs, Switzerland), followed by two centrifugation steps (10 min, 50 g ) and resuspension of the protoplast-containing pellet at a desired cell density in 3M medium [87.5 g/L mannitol, 3.1 g/L MgCl 2 ·6H 2 O, 1 g/L 2-( N -morpholino)ethanesulphonic acid hydrate (MES); Sigma; pH 5.6 and 580 mOsm]. Protoplasts (300 000) were chemically transfected with 50 µg/mL DNA (80 µg/mL for transgene control systems), as described previously ( Jost et al ., 2005 ), and cultivated in Knop's regeneration medium [1 g/L Ca(NO 3 ) 2 ·4H 2 O, 250 mg/L KCl, 250 mg/L KH 2 PO 4 , 250 mg/L MgSO 4 ·7H 2 O, 12.5 mg/L FeSO 4 ·7H 2 O, 5% glucose, 3% mannitol, pH 5.7, 540 mOsm]. <h2>Protein production</h2> Protein production was measured 5 days after transformation using standardized assays: (i) human placental SEAP: a p -nitrophenylphosphate-based light-absorbance time course ( Berger et al ., 1988 ; Schlatter et al ., 2002 ); (ii) SAMY and AMY: a blue starch Phadebas ® assay (cat. no. 10-5380-32; Pharmacia Upjohn, Peapack, NJ, USA) ( Schlatter et al ., 2002 ); quantification of intracellular reporter proteins required lysis of the plant cells by four freeze–thaw cycles and elimination of cell debris by centrifugation (2 min at 12 000 g ); (iii) human VEGF 121 : using a VEGF 121 -specific enzyme-linked immunosorbent assay (ELISA) (cat. no. 900-K10, lot. no. 1006010; Peprotech, Rocky Hill, NJ, USA); (iv) human EPO: using an EPO-specific ELISA (Quantikine ® IVD ® , cat. no. DEP00, lot. no. 243030; R & D Systems, Minneapolis, MN, USA). <h2>Transgene regulation</h2> All regulating agents were administered at the indicated concentrations immediately after transformation. The butyrolactone SCB1 was synthesized and purified as described previously ( Weber et al ., 2003 ). Erythromycin (cat. no. 45673, lot. no. 1195447; Fluka, Buchs, Switzerland) was prepared as a stock solution of 1 mg/mL in ethanol. The AIR system was induced by the addition of the indicated volumes of 100% ethanol. <h2>Microencapsulation of P. patens protoplasts</h2> Protoplasts (4 × 10 7 ) generated from transgenic VEGF 121 -producing moss tWT11.51 VEGF ( Baur et al ., 2005b ) were resuspended in 8 mL of 3M medium and stirred gently with 40 mL of 1.5% sodium alginate solution (cat. no. IE1010, lot. no. 060125B1; Inotech Biotechnologies Ltd., Basle, Switzerland). Protoplasts were encapsulated in 500-µm alginate capsules (165 protoplasts per capsule) using an Inotech Encapsulator Research IE-50R (Inotech Biotechnologies Ltd., Basle, Switzerland) set at the following parameters: 0.5 mm nozzle; 853 unit flow rate using a 50-mL syringe; 1250 s −1 nozzle vibration frequency; 1.4 kV for bead dispersion. W5 medium (18.4 g/L CaCl 2 , 8 g/L NaCl 2 , 0.99 g/L glucose, 0.75 g/L KCl; pH 5.8, 600 mOsm) was used as a precipitation solution. It contains sufficient CaCl 2 for the precipitation of alginate beads and enables the direct cultivation of microencapsulated protoplast populations without the need for medium exchange. Microencapsulated tWT11.51 VEGF -derived (a high VEGF 121 producer line harbouring the product gene under the control of the moss actin 5 5′ region) protoplasts (240 000 capsules, 1 L W5 medium) were cultivated in a BioWave 20SPS-F bioreactor (Wave Biotech, Tagelswangen, Switzerland), equipped with 2-L Wave Bags and set at the following parameters: aeration, 100 mL/min sterilized air; rocking rate, 19 min −1 ; rocking angle, 10°. The Wave Bioreactor was placed in an ISF-1-W incubator equipped with a photosynthesis kit set to 25 °C and a day/night cycle of 16 h/8 h (Kuehner, Birsfelden, Switzerland). <h2>Edman sequencing</h2> VEGF 121 was precipitated from tWT11.51 VEGF protoplast culture supernatants for 10 min at 4 °C with 100% w/v trichloroacetic acid (TCA, Sigma) (supernatant : TCA, 9 : 1). The samples were centrifuged for 5 min at 12 000 g , and the VEGF 121 -containing pellet was washed twice in ice-cold acetone, dried and resuspended in 2 × sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) reducing sample buffer [50% glycerol, 250 m m tris(hydroxymethyl)aminomethane (Tris), 10% SDS, 500 m m dithiothreitol, 0.01% bromophenol blue, pH 6.8]. The samples were then denatured for 5 min at 50 °C, and the proteins were size-fractionated on a 12% SDS-polyacrylamide gel and blotted on to a polyvinylidene fluoride membrane (cat. no. IPVH20200; Millipore Corporation, Bedford, MA, USA). VEGF 121 (35 kDa) was N-terminally sequenced on an Applied Biosystems (Foster City, CA, USA) model 492cLC Procise protein/peptide sequencer with an on-line Perkin-Elmer (Waltham, MA, USA) Applied Biosystems Model 140C PTH Amino Acid (phenylthiohydantonin amino acid) Analyser. The PTH amino acids were automatically transferred to a reverse-phase C-18 column (0.8 mm inside diameter) for detection at 269 nm, and identified by comparison with individual runs with a standard mixture of PTH amino acids. <h2>Cultivation of Sa. cerevisiae and St. coelicolor</h2> Saccharomyces cerevisiae [wild-type strain W303, BMA 64, European Sa. cerevisiae archive for functional analysis (EUROSCARF), Frankfurt, Germany] was cultivated on yeast–peptone–dextrose agar (YPD; 1% yeast extract, 2% peptone, 2% dextrose, 1% agar), and St. coelicolor MT1110 (kindly provided by Marc Folcher) was cultivated on mannitol–soy agar (2% soy flour, 2% mannitol, 1.5% agar). For the co-cultivation of St. coelicolor and P. patens , St. coelicolor were pre-cultured in Luria–Bertani (LB) medium to an optical density at 600 nm (OD 600 ) of 340, and the indicated volumes were then transferred to P. patens maintained in regeneration medium. <h2>Viability profiling of microencapsulated P. patens protoplasts</h2> The viability of microencapsulated moss protoplasts was determined by scoring live protoplasts, stained with fluorescein diacetate (FDA), and dead protoplasts, stained with trypan blue, using (fluorescence) microscopy (Leica DM-RB fluorescence microscope; Leica, Heerbrugg, Switzerland). Twenty protoplast-containing alginate beads were incubated for 10 min in a staining solution including 200 µL W5 medium, 20 µL phosphate-buffered saline (PBS; Dulbecco's phosphate-buffered saline, cat. no. 21600-0069; Invitrogen, Basle, Switzerland) containing 0.01% FDA (Sigma) and 40 µL of a 0.4% trypan blue stock solution (lot. no. 1230532; Fluka).

Journal

Plant Biotechnology JournalWiley

Published: Jan 1, 2009

Keywords: biopharmaceutical manufacturing; bioreactor; gene regulation; microencapsulation; multicistronic expression; quorum sensing; synthetic biology

There are no references for this article.