A cost-effective ELP-intein coupling system for recombinant protein purification from plant production platform

Background

Plant bioreactor offers an efficient and economical system for large-scale production of recombinant proteins. However, high cost and difficulty in scaling-up of downstream purification of the target protein, particularly the common involvement of affinity chromatography and protease in the purification process, has hampered its industrial scale application, therefore a cost-effective and easily scale-up purification method is highly desirable for further development of plant bioreactor.

Methodology/Principal Findings

To tackle this problem, we investigated the ELP-intein coupling system for purification of recombinant proteins expressed in transgenic plants using a plant lectin (PAL) with anti-tumor bioactivity as example target protein and rice seeds as production platform. Results showed that ELP-intein-PAL (EiP) fusion protein formed novel irregular ER-derived protein bodies in endosperm cells by retention of endogenous prolamins. The fusion protein was partially self-cleaved in vivo, but only self-cleaved PAL protein was detected in total seed protein sample and deposited in protein storage vacuoles (PSV). The in vivo uncleaved EiP protein was accumulated up to 2–4.2% of the total seed protein. The target PAL protein could be purified by the ELP-intein system efficiently without using complicated instruments and expensive chemicals, and the yield of pure PAL protein by the current method was up to 1.1 mg/g total seed protein.

Conclusion/Significance

This study successfully demonstrated the purification of an example recombinant protein from rice seeds by the ELP-intein system. The whole purification procedure can be easily scaled up for industrial production, providing the first evidence on applying the ELP-intein coupling system to achieve cost-effective purification of recombinant proteins expressed in plant bioreactors and its possible application in industry.     

Protein production from recombinant protein bodies

In this study, we describe a process for protein expression and purification from plants and insect cells based on the accumulation of recombinant proteins in protein bodies. This technology is using Zera^®, which sequence has the capacity to trigger in vivo the formation of dense, non-secretory storage protein body-like organelles derived from the endoplasmic reticulum (ER). With this method, recombinant human growth hormone (hGH) was expressed and purified from protein bodies accumulated in plants (Nicotiana benthamiana) and in insect cells (Spodoptera frugiperda). We found that recombinant Zera-hGH are stored in large quantity inside those proteins bodies and can be easily recovered during a one-step process from plant and insect cell biomass. After solubilization of recombinant protein bodies and cleavage of Zera tag from the fusion protein, active hGH was finally purified by a single chromatography step. These results indicate that recombinant proteins derived from Zera-fusion could provide both an efficient protein production system and eased purification downstream process.     

Analysis of ER stress in developing rice endosperm accumulating β‐amyloid peptide

The common neurodegenerative disorder known as Alzheimer’s disease is characterized by cerebral neuritic plaques of amyloid β (Aβ) peptide. Plaque formation is related to the highly aggregative property of this peptide, because it polymerizes to form insoluble plaques or fibrils causing neurotoxicity. Here, we expressed Aβ peptide as a new causing agent to endoplasmic reticulum (ER) stress to study ER stress occurred in plant. When the dimer of Aβ_1–42 peptide was expressed in maturing seed under the control of the 2.3‐kb glutelin GluB‐1 promoter containing its signal peptide, a maximum of about 8 μg peptide per grain accumulated and was deposited at the periphery of distorted ER‐derived PB‐I protein bodies. Synthesis of Aβ peptide in the ER lumen severely inhibited the synthesis and deposition of seed storage proteins, resulting in the generation of many small and abnormally appearing PB bodies. This ultrastructural change was accounted for by ER stress leading to the accumulation of aggregated Aβ peptide in the ER lumen and a coordinated increase in ER‐resident molecular chaperones such as BiPs and PDIs in Aβ‐expressing plants. Microarray analysis also confirmed that expression of several BiPs, PDIs and OsbZIP60 containing putative transmembrane domains was affected by the ER stress response. Aβ‐expressing transgenic rice kernels exhibited an opaque and shrunken phenotype. When grain phenotype and expression levels were compared among transgenic rice grains expressing several different recombinant peptides, such detrimental effects on grain phenotype were correlated with the expressed peptide causing ER stress rather than expression levels.     

Protein accumulation and rumen stability of wheat γ‐gliadin fusion proteins in tobacco and alfalfa

The nutritional value of various crops can be improved by engineering plants to produce high levels of proteins. For example, because methionine deficiency limits the protein quality of Medicago Sativa (alfalfa) forage, producing alfalfa plants that accumulate high levels of a methionine‐rich protein could increase the nutritional value of that crop. We used three strategies in designing methionine‐rich recombinant proteins that could accumulate to high levels in plants and thereby serve as candidates for improving the protein quality of alfalfa forage. In tobacco, two fusion proteins, γ‐gliadin‐δ‐zein and γ‐δ‐zein, as well as δ‐zein co‐expressed with β‐zein, all formed protein bodies. However, the γ‐gliadin‐δ‐zein fusion protein accumulated to the highest level, representing up to 1.5% of total soluble protein (TSP) in one transformant. In alfalfa, γ‐gliadin‐δ‐zein accumulated to 0.2% of TSP, and in an in vitro rumen digestion assay, γ‐gliadin‐δ‐zein was more resistant to microbial degradation than Rubisco. Additionally, although it did not form protein bodies, a γ‐gliadin‐GFP fusion protein accumulated to much higher levels, 7% of TSP, than a recombinant protein comprised of an ER localization signal fused to GFP in tobacco. Based on our results, we conclude that γ‐gliadin‐δ‐zein is a potential candidate protein to use for enhancing methionine levels in plants and for improving rumen stability of forage protein. γ‐gliadin fusion proteins may provide a general platform for increasing the accumulation of recombinant proteins in transgenic plants.     

Secretory type of recombinant thioredoxin h induces ER stress in endosperm cells of transgenic rice

Thioredoxin h (TRX h) functions as a reducing protein and is present in all organisms. As a new approach for inducing the endoplasmic reticulum (ER) stress, TRX h (OsTRX23) was expressed as a secretory protein using the endosperm-specific glutelin GluB-1 promoter and a signal peptide. In transgenic rice seeds, the majority of the recombinant TRX h accumulated in the ER but some was also localized to the protein body IIs (PB-IIs). The rice grain quality was dependent on the TRX h accumulation level. Increased TRX h expression resulted in aberrant phenotypes, such as chalky and shriveled features, lower seed weight and lower seed protein content. Furthermore, the accumulation of some seed storage proteins (SSPs) was significantly suppressed and the morphology of the protein bodies (PB-Is and PB-IIs) changed according to the level of TRX h. SSPs, such as 13 kDa prolamin and GluA, were specifically modified via the reducing action of TRX h. These changes led to the activation of the ER stress response, which was accompanied by the expression of several chaperone proteins. Specifically, the ER stress markers BiP4 and BiP5 were significantly up-regulated by an increase in the level of TRX h. These results suggest that changes in the conformation of certain SSPs via the action of recombinant TRX h lead to an induced ER stress response in transgenic rice seeds.     

Overproduction of fungal ribotoxin α-sarcin in Escherichia coli: generation of an active immunotoxin

α-Sarcin is a ribonucleolytic protein secreted by the mold Aspergillus giganteus. DNA encoding α-sarcin was isolated from the host and cloned into T7 promoter based E. coli expression vectors. Using bacterial outer membrane protein A (OmpA) signal sequence, properly processed recombinant (re-) protein was secreted into the culture medium while in the absence of a signal sequence protein remained insoluble in the bacterial inclusion bodies. The re-α-sarcin was purified to homogeneity by simple chromatographic techniques both from the insoluble and soluble sources with respective yields of 40–50 μg/ml and 2–3 μg/ml. The re-ribotoxin was functionally as active as the native toxin and preserved its specificity. The re-α-sarcin was used in the construction of an active immunotoxin targeted at the human cancer cells overexpressing transferrin receptor (TFR).     

Generation of a transgenic rice seed-based edible vaccine against house dust mite allergy

As an alternative approach to conventional allergen-specific immunotherapy, transgenic rice seed expressing a major house dust mite (HDM) allergen, Der p 1, was developed as an edible vaccine. The C-terminal KDEL-tagged Der p 1 allergen specifically accumulated in seed endosperm tissue under the control of the endosperm-specific GluB1 promoter. Der p 1 reached a maximum concentration of 58 μg/grain and was deposited in the endoplasmic reticulum (ER)-derived protein body I (PB-I). Plant-derived Der p 1 was posttranslationally modified with high-mannose-type glycan structures. Glycosylated Der p 1 displayed reduced IgE binding capacity in comparison with its unglycosylated counterpart in vitro. Our results indicate that transgenic Der p 1 rice seeds are a safe, potential oral delivery vaccine for the treatment of HDM allergy.     

Recombinant protein yield in rice seed is enhanced by specific suppression of endogenous seed proteins at the same deposit site

Human IL‐10 (hIL‐10) is a therapeutic treatment candidate for inflammatory allergy and autoimmune diseases. Rice seed‐produced IL‐10 can be effectively delivered directly to gut‐associated lymphoreticular tissue (GALT) via bio‐encapsulation. Previously, the codon‐optimized hIL‐10 gene was expressed in transgenic rice with the signal peptide and endoplasmic reticulum (ER) retention signal (KDEL) at its 5′ and 3′ ends, respectively, under the control of the endosperm‐specific glutelin GluB‐1 promoter. The resulting purified hIL‐10 was biologically active. In this study, the yield of hIL‐10 in transgenic rice seed was improved. This protein accumulated at the intended deposition sites, which had been made vacant through the selective reduction, via RNA interference, of the endogenous seed storage proteins prolamins or glutelins. Upon suppression of prolamins that were sequestered into ER‐derived protein bodies (PB‐I), hIL‐10 accumulation increased approximately 3‐fold as compared to rice seed with no such suppression and reached 219 μg/grain. In contrast, reducing the majority of the glutelins stored in protein‐storage vacuoles (PB‐II) did not significantly affect the accumulation of hIL‐10. Considering that hIL‐10 is synthesized in the ER lumen and subsequently buds off in ER‐derived granules called IL‐10 granules in a manner similar to PB‐Is, these results indicate that increases in the available deposition space for the desired recombinant proteins may be crucial for improvements in yield. Furthermore, efficient dimeric intermolecular formation of hIL‐10 by inhibiting interaction with Cys‐rich prolamins also contributed to the enhanced formation of IL‐10 bodies. Higher yield of hIL‐10 produced in rice seeds is expected to have broad application in the future.     

Expression and regulation of vitellogenin messenger RNA in the mosquito,Aedes aegypti

Vitellogenin (yolk protein) gene expression in the mosquito was investigated at the level of mRNA using a subcloned fragment (403-1c) of the vitellogenin DNA derived from an Aedes aegypti genomic library. Message appeared 1–3 hr after a blood meal, peaked at 36 hr and was rapidly degraded thereafter. Fluctuations in levels of 20-hydroxyecdysone after a blood meal coincided with accumulation of vitellogenin message. Blood-fed, decapitated females injected with 5 μg of 20-hydroxyecdysone accumulated up to 75% of the message found in blood-fed controls. Fat bodies from non-blood-fed females incubated with physiological levels of 20-hydroxyecdysone and the juvenile hormone analog methoprene contained twice as much vitellogenin message as those incubated with 20-hydroxyecdysone alone. Methoprene alone had no effect.     

Production of recombinant Chikungunya virus envelope 2 protein in Escherichia coli

Chikungunya, a mosquito-borne viral disease caused by Chikungunya virus (CHIKV), has drawn substantial attention after its reemergence causing massive outbreaks in tropical regions of Asia and Africa. The recombinant envelope 2 (rE2) protein of CHIKV is a potential diagnostic as well as vaccine candidate. Development of cost-effective cultivation media and appropriate culture conditions are generally favorable for large-scale production of recombinant proteins in Escherichia coli. The effects of medium composition and cultivation conditions on the production of recombinant Chikungunya virus E2 (rCHIKV E2) protein were investigated in shake flask culture as well as batch cultivation of Escherichia coli. Further, the fed-batch process was also carried out for high cell density cultivation of E. coli expressing rE2 protein. Expression of rCHIKV E2 protein in E. coli was induced with 1 mM isopropyl-beta-thiogalactoside (IPTG) at ~23 g dry cell weight (DCW) per liter of culture and yielded an insoluble protein aggregating to form inclusion bodies. The final DCW after fed-batch cultivation was ~35 g/l. The inclusion bodies were isolated, solubilized in 8 M urea and purified through affinity chromatography to give a final product yield of ~190 mg/l. The reactivity of purified E2 protein was confirmed by Western blotting and enzyme-linked immunosorbent assay. These results show that rE2 protein of CHIKV may be used as a diagnostic reagent or for further prophylactic studies. This approach of producing rE2 protein in E. coli with high yield may also offer a promising method for production of other viral recombinant proteins.     

Composition and Structure of Protein Bodies and Spherosomes Isolated from Ungerminated Seeds of Sorghum bicolor (Linn.) Moench

Protein bodies and spherosomes have been isolated from mature seeds of Sorghum bicolor (Linn.) Moench by a procedure which successfully disrupts the protein starch complex in the grain. Protein bodies from whole grain are 68% protein and have a distinct border and a monolithic appearance. Those from embryo are 95% protein and have diffuse borders, vacuoles, and appear very granular. Aleurone tissue protein bodies are 46% protein with a structure similar to those from embryo, but possibly are composed of a protein carbohydrate mixture. Spherosomes from all sources are quite similar in composition and structure. They have an average composition of 27% protein, 12% phosphorus, and 8.6% metals. Microscopically, they appear as small vesicles bounded by a wall which is probably composed of protein and the potassiummagnesium salt of phytic acid.     

N(pro) fusion technology to produce proteins with authentic N termini in E. coli

We describe a prokaryotic expression system using the autoproteolytic function of N(pro) from classical swine fever virus. Proteins or peptides expressed as N(pro) fusions are deposited as inclusion bodies. On in vitro refolding by switching from chaotropic to kosmotropic conditions, the fusion partner is released from the C-terminal end of the autoprotease by self-cleavage, leaving the target protein with an authentic N terminus. A tailor-made N(pro) mutant called EDDIE, with increased in vitro and decreased in vivo cleavage rates, has enabled us to express proinsulin, domain-D of staphylococcal protein A, hepcidin, interferon-alpha1, keratin-associated protein 10-4, green fluorescent protein, inhibitorial peptide of senescence-evasion-factor, monocyte chemoattractant protein-1 and toxic gyrase inhibitor, among others. This N(pro) expression system can be used as a generic tool for the high-level production of recombinant toxic peptides and proteins (up to 12 g/l) in Escherichia coli without the need for chemical or enzymatic removal of the fusion tag.     

Unexpected deposition patterns of recombinant proteins in post-endoplasmic reticulum compartments of wheat endosperm

Protein transport within cereal endosperm cells is complicated by the abundance of endoplasmic reticulum (ER)-derived and vacuolar protein bodies. For wheat storage proteins, two major transport routes run from the ER to the vacuole, one bypassing and one passing through the Golgi. Proteins traveling along each route converge at the vacuole and form aggregates. To determine the impact of this trafficking system on the fate of recombinant proteins expressed in wheat endosperm, we used confocal and electron microscopy to investigate the fate of three recombinant proteins containing different targeting information. KDEL-tagged recombinant human serum albumin, which is retrieved to the ER lumen in leaf cells, was deposited in prolamin aggregates within the vacuole of endosperm cells, most likely following the bulk of endogenous glutenins. Recombinant fungal phytase, a glycoprotein designed for secretion, was delivered to the same compartment, with no trace of the molecule in the apoplast. Glycan analysis revealed that this protein had passed through the Golgi. The localization of human serum albumin and phytase was compared to that of recombinant legumin, which contains structural targeting information directing it to the vacuole. Uniquely, legumin accumulated in the globulin inclusion bodies at the periphery of the prolamin bodies, suggesting a different mode of transport and/or aggregation. Our results demonstrate that recombinant proteins are deposited in an unexpected pattern within wheat endosperm cells, probably because of the unique storage properties of this tissue. Our data also confirm that recombinant proteins are invaluable tools for the analysis of protein trafficking in cereals.     

Deposition of a recombinant peptide in ER-derived protein bodies by retention with cysteine-rich prolamins in transgenic rice seed

A 7Crp peptide composed of seven major human T cell epitopes derived from the Japanese cedar pollen allergens Cry j 1 and Cry j 2 is an ideal tolerogen for peptide immunotherapy against Japanese cedar pollinosis. To maximize the accumulation level of the 7Crp peptide in transgenic rice seed, we tested endosperm specific promoters and intracellular localizations suitable for stable accumulation. A 7Crp peptide carrying the KDEL ER retention signal directed by the 2.3-kb promoter of the glutelin GluB-1, which contains a signal peptide, accumulated at the highest level of about 60 μg/grain. Notably, the 7Crp peptide predominantly accumulated in ER-derived protein bodies irrespective of the presence of various sorting signals or expression as a fusion protein with glutelin. We attribute this abnormal pattern of accumulation to the formation of disulfide bonds between the 7Crp peptide and cysteine-rich (Cys-rich) prolamin storage proteins. Furthermore, the formation of these aggregates induced the chaperone proteins BiP and PDI as an ER stress response.     

Origin and development of protein bodies in cotyledons of Vicia faba

Storage proteins of the field bean (Vicia faba L., var. minor, cv. “Fribo”) are synthesized and accumulated in the cotyledons during stage 2 of seed development. Deposition of protein reserves takes place in the protein bodies. The generation of protein bodies was investigated electronmicroscopically using ultra-thin sections as well as the freeze-fracturing technique. During the initial period of storage protein formation, globulins are deposited in large vacuoles which later are transformed to give increasing numbers of small vacuoles with decreasing size. The vacuoles disappear early during the stage of storage protein formation and generate the first protein bodies. During the subsequent period of maximum storage protein formation, which takes place at the rough endoplasmic reticulum (rER), swollen ER strands appear which seem to be entirely filled with protein, and these generate ER-produced protein vacuoles (ERPVAC). The vesicles are transformed in a manner comparable to the vacuoles in the initial period of developmental stage 2 and thus generate the major quantity of protein bodies. Both processes seem to represent only two variants of an uniform mechanism of protein body generation.     

Rapid transient production in plants by replicating and non-replicating vectors yields high quality functional anti-HIV antibody

Background

The capacity of plants and plant cells to produce large amounts of recombinant protein has been well established. Due to advantages in terms of speed and yield, attention has recently turned towards the use of transient expression systems, including viral vectors, to produce proteins of pharmaceutical interest in plants. However, the effects of such high level expression from viral vectors and concomitant effects on host cells may affect the quality of the recombinant product.

Methodology/Principal Findings

To assess the quality of antibodies transiently expressed to high levels in plants, we have expressed and characterised the human anti-HIV monoclonal antibody, 2G12, using both replicating and non-replicating systems based on deleted versions of Cowpea mosaic virus (CPMV) RNA-2. The highest yield (approximately 100 mg/kg wet weight leaf tissue) of affinity purified 2G12 was obtained when the non-replicating CPMV-HT system was used and the antibody was retained in the endoplasmic reticulum (ER). Glycan analysis by mass-spectrometry showed that the glycosylation pattern was determined exclusively by whether the antibody was retained in the ER and did not depend on whether a replicating or non-replicating system was used. Characterisation of the binding and neutralisation properties of all the purified 2G12 variants from plants showed that these were generally similar to those of the Chinese hamster ovary (CHO) cell-produced 2G12.

Conclusions

Overall, the results demonstrate that replicating and non-replicating CPMV-based vectors are able to direct the production of a recombinant IgG similar in activity to the CHO-produced control. Thus, a complex recombinant protein was produced with no apparent effect on its biochemical properties using either high-level expression or viral replication. The speed with which a recombinant pharmaceutical with excellent biochemical characteristics can be produced transiently in plants makes CPMV-based expression vectors an attractive option for biopharmaceutical development and production.     

Characterization of Afp1, an antifreeze protein from the psychrophilic yeast Glaciozyma antarctica PI12

The psychrophilic yeast Glaciozyma antarctica demonstrated high antifreeze activity in its culture filtrate. The culture filtrate exhibited both thermal hysteresis (TH) and ice recrystallization inhibition (RI) properties. The TH of 0.1 °C was comparable to that previously reported for bacteria and fungi. A genome sequence survey of the G. antarctica genome identified a novel antifreeze protein gene. The cDNA encoded a 177 amino acid protein with 30 % similarity to a fungal antifreeze protein from Typhula ishikariensis. The expression levels of AFP1 were quantified via real time-quantitative polymerase chain reaction (RT-qPCR), and the highest expression levels were detected within 6 h of growth at ?12 °C. The cDNA of the antifreeze protein was cloned into an Escherichia coli expression system. Expression of recombinant Afp1 in E. coli resulted in the formation of inclusion bodies that were subsequently denatured by treatment with urea and allowed to refold in vitro. Activity assays of the recombinant Afp1 confirmed the antifreeze protein properties with a high TH value of 0.08 °C.     

Accumulation of Japanese cedar pollen allergen,Cry j 1, in the protein body I of transgenic rice seeds using the promoter and signal sequence of glutelin GluB-1 gene

Recombinant Cry j 1, a Japanese cedar pollen allergen, was produced in rice seeds for potential use for oral immunotherapy. Cry j 1 cDNA was divided into two parts, an N-terminal half and a C-terminal half, and each was fused downstream to glutelin GluB-1 gene containing sequences of the promoter, 5 untranslated region and signal peptide. A gene for green fluorescent protein was also fused to the 3 end of the Cry j 1 fragment. Recombinant Cry j 1 of up to 16.6 g per mg total protein of the seeds was expressed in transgenic rice seeds. Although the recombinant Cry j 1 was expected to be accumulated in protein body II because of the employment of glutelin signal peptide, it was demonstrated to be accumulated exclusively in protein body I. The recombinant Cry j 1 was not shown to react with IgE of allergic patients, indicating the reduction of the risk of anaphylactic reaction. These results demonstrate that the transgenic rice seeds with the recombinant Cry j 1 would be useful for the study of oral immunotherapy.     

Use of a stress-minimisation paradigm in high cell density fed-batch Escherichia coli fermentations to optimise recombinant protein production

Production of recombinant proteins is an industrially important technique in the biopharmaceutical sector. Many recombinant proteins are problematic to generate in a soluble form in bacteria as they readily form insoluble inclusion bodies. Recombinant protein solubility can be enhanced by minimising stress imposed on bacteria through decreasing growth temperature and the rate of recombinant protein production. In this study, we determined whether these stress-minimisation techniques can be successfully applied to industrially relevant high cell density Escherichia coli fermentations generating a recombinant protein prone to forming inclusion bodies, CheY–GFP. Flow cytometry was used as a routine technique to rapidly determine bacterial productivity and physiology at the single cell level, enabling determination of culture heterogeneity. We show that stress minimisation can be applied to high cell density fermentations (up to a dry cell weight of >70 g L^?1) using semi-defined media and glucose or glycerol as carbon sources, and using early or late induction of recombinant protein production, to produce high yields (up to 6 g L^?1) of aggregation-prone recombinant protein in a soluble form. These results clearly demonstrate that stress minimisation is a viable option for the optimisation of high cell density industrial fermentations for the production of high yields of difficult-to-produce recombinant proteins, and present a workflow for the application of stress-minimisation techniques in a variety of fermentation protocols.