Escherichia coli Cytoplasmic Expression of Disulfide-Bonded Proteins: Side-by-Side Comparison Between Two Competing Strategies.

The production of disulfide bond-containing recombinant proteins in Escherichia coli has traditionally been done by either refolding from inclusion bodies or by targeting the protein to the periplasm. However, both approaches have limitations and were developed to allow the production of proteins with disulfide bonds in the cytoplasm of E. coli: i) engineered strains with deletions in the disulfide reduction pathways, e.g. SHuffle, and ii) the co-expression of oxidative folding catalysts, e.g. CyDisCo. However, to our knowledge, the effectiveness of these strategies has not been properly evaluated. Here, we systematically compare the purified yields of 14 different proteins of interest (POI) that contain disulfide bonds in their native state when expressed in both systems. We also compared the effects of different background strains, commonly used promoters, and two media types: defined and rich autoinduction. In rich autoinduction media, POI which can be produced in a soluble (non-native) state without a system for disulfide bond formation were produced in higher purified yields from SHuffle, whereas all other proteins were produced in higher purified yields using CyDisCo. In chemically defined media, purified yields were at least 10x higher in all cases using CyDisCo. In addition, the quality of the three POI tested was superior when produced using CyDisCo.

Komagataella phaffii Erp41 is a protein disulfide isomerase with unprecedented disulfide bond catalyzing activity when coupled to glutathione.

In the methylotrophic yeast Komagataella phaffii, we identified an endoplasmic reticulum-resident protein disulfide isomerase (PDI) family member, Erp41, with a peculiar combination of active site motifs. Like fungal ERp38, it has two thioredoxin-like domains which contain active site motifs (a and a'), followed by an alpha-helical ERp29c C-terminal domain (c domain). However, while the a domain has a typical PDI-like active site motif (CGHC), the a' domain instead has CGYC, a glutaredoxin-like motif which confers to the protein an exceptional affinity for GSH/GSSG. This combination of active site motifs has so far been unreported in PDI-family members. Homology searches revealed ERp41 is present in the genome of some plants, fungal parasites, and a few nonconventional yeasts, among which are Komagataella spp. and Yarrowia lipolytica. These yeasts are both used for the production of secreted recombinant proteins. Here, we analyzed the activity of K. phaffii Erp41. We report that it is nonessential in K. phaffii, and that it can catalyze disulfide bond formation in partnership with the sulfhydryl oxidase Ero1 in vitro with higher turnover rates than the canonical PDI from K. phaffii, Pdi1, but slower activation times. We show how Erp41 has unusually fast glutathione-coupled oxidation activity and relate it to its unusual combination of active sites in its thioredoxin-like domains. We further describe how this determines its unusually efficient catalysis of dithiol oxidation in peptide and protein substrates.

Cytoplasmic production of Fabs in chemically defined media in fed-batch fermentation.

Fragment of antigen-binding region (Fab) of antibodies are important biomolecules, with a broad spectrum of functionality in the biomedical field. While full length antibodies are usually produced in mammalian cells, the smaller size, lack of N-glycosylation and less complex structure of Fabs make production in microbial cell factories feasible. Since Fabs contain disulfide bonds, such production is often done in the periplasm, but there the formation of the inter-molecular disulfide bond between light and heavy chains can be problematic. Here we studied the use of the CyDisCo system (cytoplasmic disulfide bond formation in E. coli) to express two Fabs (Herceptin and Maa48) in the cytoplasm of E. coli in fed-batch fermentation using a generic chemically defined media. We were able to solubly express both Fabs with purified yields of 565 mg/L (Maa48) and 660 mg/L (Herceptin) from low density fermentation. Both proteins exhibited CD spectra consistent with natively folded protein and both were biologically active. To our knowledge this is the first demonstration of high-level production of biological active Fabs in the cytoplasm of E. coli in industrially relevant fermentation conditions.

Biochemical analysis of Komagataella phaffii oxidative folding proposes novel regulatory mechanisms of disulfide bond formation in yeast.

Oxidative protein folding in the endoplasmic reticulum (ER) is driven mainly by protein disulfide isomerase PDI and oxidoreductin Ero1. Their activity is tightly regulated and interconnected with the unfolded protein response (UPR). The mechanisms of disulfide bond formation have mainly been studied in human or in the yeast Saccharomyces cerevisiae. Here we analyze the kinetics of disulfide bond formation in the non-conventional yeast Komagataella phaffii, a common host for the production of recombinant secretory proteins. Surprisingly, we found significant differences with both the human and S. cerevisiae systems. Specifically, we report an inactive disulfide linked complex formed by K. phaffii Ero1 and Pdi1, similarly to the human orthologs, but not described in yeast before. Furthermore, we show how the interaction between K. phaffii Pdi1 and Ero1 is unaffected by the introduction of unfolded substrate into the system. This is drastically opposed to the previously observed behavior of the human pathway, suggesting a different regulation of the UPR and/or possibly different interaction mechanics between K. phaffii Pdi1 and Ero1.

Highly efficient export of a disulfide-bonded protein to the periplasm and medium by the Tat pathway using CyDisCo in Escherichia coli.

High-value heterologous proteins produced in Escherichia coli that contain disulfide bonds are almost invariably targeted to the periplasm via the Sec pathway as it, among other advantages, enables disulfide bond formation and simplifies downstream processing. However, the Sec system cannot transport complex or rapidly folding proteins, as it only transports proteins in an unfolded state. The Tat system also transports proteins to the periplasm, and it has significant potential as an alternative means of recombinant protein production because it transports fully folded proteins. Most of the studies related to Tat secretion have used the well-studied TorA signal peptide that is Tat-specific, but this signal peptide also tends to induce degradation of the protein of interest, resulting in lower yields. This makes it difficult to use Tat in the industry. In this study, we show that a model disulfide bond-containing protein, YebF, can be exported to the periplasm and media at a very high level by the Tat pathway in a manner almost completely dependent on cytoplasmic disulfide formation, by other two putative Tat SPs: those of MdoD and AmiC. In contrast, the TorA SP exports YebF at a low level.

Introduction of a More Glutaredoxin-like Active Site to PDI Results in Competition between Protein Substrate and Glutathione Binding.

Proteins in the thioredoxin superfamily share a similar fold, contain a -CXXC- active site, and catalyze oxidoreductase reactions by dithiol-disulfide exchange mechanisms. Protein disulfide isomerase (PDI) has two -CGHC- active sites. For in vitro studies, oxidation/reduction of PDI during the catalytic cycle is accomplished with glutathione. Glutathione may act as electron donor/acceptor for PDI also in vivo, but at least for oxidation reactions, GSSG probably is not the major electron acceptor and PDI may not have evolved to react with glutathione with high affinity, but merely having adequate affinity for both glutathione and folding proteins/peptides. Glutaredoxins, on the other hand, have a high affinity for glutathione. They commonly have -CXFC- or -CXYC- active site, where the tyrosine residue forms part of the GSH binding groove. Mutating the active site of PDI to a more glutaredoxin-like motif increased its reactivity with glutathione. All such variants showed an increased rate in GSH-dependent reduction or GSSG-dependent oxidation of the active site, as well as a decreased rate of the native disulfide bond formation, with the magnitude of the effect increasing with glutathione concentration. This suggests that these variants lead to competition in binding between glutathione and folding protein substrates.

Mutation in protein disulfide isomerase A3 causes neurodevelopmental defects by disturbing endoplasmic reticulum proteostasis.

Recessive gene mutations underlie many developmental disorders and often lead to disabling neurological problems. Here, we report identification of a homozygous c.170G>A (p.Cys57Tyr or C57Y) mutation in the gene coding for protein disulfide isomerase A3 (PDIA3, also known as ERp57), an enzyme that catalyzes formation of disulfide bonds in the endoplasmic reticulum, to be associated with syndromic intellectual disability. Experiments in zebrafish embryos show that PDIA3C57Y expression is pathogenic and causes developmental defects such as axonal disorganization as well as skeletal abnormalities. Expression of PDIA3C57Y in the mouse hippocampus results in impaired synaptic plasticity and memory consolidation. Proteomic and functional analyses reveal that PDIA3C57Y expression leads to dysregulation of cell adhesion and actin cytoskeleton dynamics, associated with altered integrin biogenesis and reduced neuritogenesis. Biochemical studies show that PDIA3C57Y has decreased catalytic activity and forms disulfide-crosslinked aggregates that abnormally interact with chaperones in the endoplasmic reticulum. Thus, rare disease gene variant can provide insight into how perturbations of neuronal proteostasis can affect the function of the nervous system.

High-resolution Crystal Structure of Human pERp1, A Saposin-like Protein Involved in IgA, IgM and Integrin Maturation in the Endoplasmic Reticulum.

The folding of disulfide bond containing proteins in the endoplasmic reticulum (ER) is a complex process that requires protein folding factors, some of which are protein-specific. The ER resident saposin-like protein pERp1 (MZB1, CNPY5) is crucial for the correct folding of IgA, IgM and integrins. pERp1 also plays a role in ER calcium homeostasis and plasma cell mobility. As an important factor for proper IgM maturation and hence immune function, pERp1 is upregulated in many auto-immune diseases. This makes it a potential therapeutic target. pERp1 belongs to the CNPY family of ER resident saposin-like proteins. To date, five of these proteins have been identified. All are implicated in protein folding and all contain a saposin-like domain. All previously structurally characterized saposins are involved in lipid binding. However, there are no reports of CNPY family members interacting with lipids, suggesting a novel function for the saposin fold. However, the molecular mechanisms of their function remain elusive. To date, no structure of any CNPY protein has been reported. Here, we present the high-resolution (1.4 Å) crystal structure of human pERp1 and confirm that it has a saposin-fold with unique structural elements not present in other saposin-fold structures. The implications for the role of CNPY proteins in protein folding in the ER are discussed.

Production of Extracellular Matrix Proteins in the Cytoplasm of E. coli: Making Giants in Tiny Factories.

Escherichia coli is the most widely used protein production host in academia and a major host for industrial protein production. However, recombinant production of eukaryotic proteins in prokaryotes has challenges. One of these is post-translational modifications, including native disulfide bond formation. Proteins containing disulfide bonds have traditionally been made by targeting to the periplasm or by in vitro refolding of proteins made as inclusion bodies. More recently, systems for the production of disulfide-containing proteins in the cytoplasm have been introduced. However, it is unclear if these systems have the capacity for the production of disulfide-rich eukaryotic proteins. To address this question, we tested the capacity of one such system to produce domain constructs, containing up to 44 disulfide bonds, of the mammalian extracellular matrix proteins mucin 2, alpha tectorin, and perlecan. All were successfully produced with purified yields up to 6.5 mg/L. The proteins were further analyzed using a variety of biophysical techniques including circular dichroism spectrometry, thermal stability assay, and mass spectrometry. These analyses indicated that the purified proteins are most likely correctly folded to their native state. This greatly extends the use of E. coli for the production of eukaryotic proteins for structural and functional studies.

Applications of catalyzed cytoplasmic disulfide bond formation.

Disulfide bond formation is an essential post-translational modification required for many proteins to attain their native, functional structure. The formation of disulfide bonds, otherwise known as oxidative protein folding, occurs in the endoplasmic reticulum and mitochondrial inter-membrane space in eukaryotes and the periplasm of prokaryotes. While there are differences in the molecular mechanisms of oxidative folding in different compartments, it can essentially be broken down into two steps, disulfide formation and disulfide isomerization. For both steps, catalysts exist in all compartments where native disulfide bond formation occurs. Due to the importance of disulfide bonds for a plethora of proteins, considerable effort has been made to generate cell factories which can make them more efficiently and cheaper. Recently synthetic biology has been used to transfer catalysts of native disulfide bond formation into the cytoplasm of prokaryotes such as Escherichia coli. While these engineered systems cannot yet rival natural systems in the range and complexity of disulfide-bonded proteins that can be made, a growing range of proteins have been made successfully and yields of homogenously folded eukaryotic proteins exceeding g/l yields have been obtained. This review will briefly give an overview of such systems, the uses reported to date and areas of future potential development, including combining with engineered systems for cytoplasmic glycosylation.

Molecular analysis of human Ero1 reveals novel regulatory mechanisms for oxidative protein folding.

Oxidative protein folding in the ER is driven mainly by oxidases of the endoplasmic reticulum oxidoreductin 1 (Ero1) family. Their action is regulated to avoid cell stress, including hyperoxidation. Previously published regulatory mechanisms are based on the rearrangement of active site and regulatory disulfides. In this study, we identify two novel regulatory mechanisms. First, both human Ero1 isoforms exist in a dynamic mixed disulfide complex with protein disulfide isomerase, which involves cysteines (Cys166 in Ero1α and Cys165 in Ero1ß) that have previously been regarded as being nonfunctional. Second, our kinetic studies reveal that Ero1 not only has a high affinity for molecular oxygen as the terminal acceptor of electrons but also that there is a high cooperativity of binding (Hill coefficient >3). This allows Ero1 to maintain high activity under hypoxic conditions, without compromising cellular viability under hyper-hypoxic conditions. These data, together with novel mechanistic details of differences in activation between the two human Ero1 isoforms, provide important new insights into the catalytic cycle of human Ero1 and how they have been fine-tuned to operate at low oxygen concentrations.

Structures of Angptl3 and Angptl4, modulators of triglyceride levels and coronary artery disease.

Coronary artery disease is the most common cause of death globally and is linked to a number of risk factors including serum low density lipoprotein, high density lipoprotein, triglycerides and lipoprotein(a). Recently two proteins, angiopoietin-like protein 3 and 4, have emerged from genetic studies as being factors that significantly modulate plasma triglyceride levels and coronary artery disease. The exact function and mechanism of action of both proteins remains to be elucidated, however, mutations in these proteins results in up to 34% reduction in coronary artery disease and inhibition of function results in reduced plasma triglyceride levels. Here we report the crystal structures of the fibrinogen-like domains of both proteins. These structures offer new insights into the reported loss of function mutations, the mechanisms of action of the proteins and open up the possibility for the rational design of low molecular weight inhibitors for intervention in coronary artery disease.

Efficient soluble expression of disulfide bonded proteins in the cytoplasm of Escherichia coli in fed-batch fermentations on chemically defined minimal media.

BACKGROUND: The production of recombinant proteins containing disulfide bonds in Escherichia coli is challenging. In most cases the protein of interest needs to be either targeted to the oxidizing periplasm or expressed in the cytoplasm in the form of inclusion bodies, then solubilized and re-folded in vitro. Both of these approaches have limitations. Previously we showed that soluble expression of disulfide bonded proteins in the cytoplasm of E. coli is possible at shake flask scale with a system, known as CyDisCo, which is based on co-expression of a protein of interest along with a sulfhydryl oxidase and a disulfide bond isomerase. With CyDisCo it is possible to produce disulfide bonded proteins in the presence of intact reducing pathways in the cytoplasm. RESULTS: Here we scaled up production of four disulfide bonded proteins to stirred tank bioreactors and achieved high cell densities and protein yields in glucose fed-batch fermentations, using an E. coli strain (BW25113) with the cytoplasmic reducing pathways intact. Even without process optimization production of purified human single chain IgA1 antibody fragment reached 139 mg/L and hen avidin 71 mg/L, while purified yields of human growth hormone 1 and interleukin 6 were around 1 g/L. Preliminary results show that human growth hormone 1 was also efficiently produced in fermentations of W3110 strain and when glucose was replaced with glycerol as the carbon source. CONCLUSIONS: Our results show for the first time that efficient production of high yields of soluble disulfide bonded proteins in the cytoplasm of E. coli with the reducing pathways intact is feasible to scale-up to bioreactor cultivations on chemically defined minimal media.

ALS-linked protein disulfide isomerase variants cause motor dysfunction.

Disturbance of endoplasmic reticulum (ER) proteostasis is a common feature of amyotrophic lateral sclerosis (ALS). Protein disulfide isomerases (PDIs) areERfoldases identified as possibleALSbiomarkers, as well as neuroprotective factors. However, no functional studies have addressed their impact on the disease process. Here, we functionally characterized fourALS-linked mutations recently identified in two majorPDIgenes,PDIA1 andPDIA3/ERp57. Phenotypic screening in zebrafish revealed that the expression of thesePDIvariants induce motor defects associated with a disruption of motoneuron connectivity. Similarly, the expression of mutantPDIs impaired dendritic outgrowth in motoneuron cell culture models. Cellular and biochemical studies identified distinct molecular defects underlying the pathogenicity of thesePDImutants. Finally, targetingERp57 in the nervous system led to severe motor dysfunction in mice associated with a loss of neuromuscular synapses. This study identifiesERproteostasis imbalance as a risk factor forALS, driving initial stages of the disease.

Systematic screening of soluble expression of antibody fragments in the cytoplasm of E. coli.

BACKGROUND: Disulfide bonds are the most common structural, post-translational modification found in proteins. Antibodies contain up to 25 disulfide bonds depending on type, with scFv fragments containing two disulfides and Fab fragments containing five or six disulfide bonds. The production of antibody fragments that contain native disulfide bonds can be challenging, especially on a large scale. The protein needs to be targeted to prokaryotic periplasm or the eukaryotic endoplasmic reticulum. These compartments are specialised for disulfide bond formation, but both compartments have limitations. RESULTS: Here we show that the introduction into the cytoplasm of a catalyst of disulfide bond formation and a catalyst of disulfide bond isomerization allows the efficient formation of natively folded scFv and Fab antibody fragments in the cytoplasm of Escherichia coli with intact reducing pathways. Eleven scFv and eleven Fab fragments were screened and ten of each were obtained in yields of >5 mg/L from deep-well plates. Production of eight of the scFv and all ten of the Fab showed a strong dependence on the addition of the folding factors. Yields of purified scFv of up to 240 mg/L and yields of purified Fab fragments of up to 42 mg/L were obtained. Purified fragments showed circular dichroism spectra consistent with being natively folded and were biologically active. CONCLUSIONS: Our results show that the efficient production of soluble, biologically active scFv and Fab antibody fragments in the cytoplasm of E. coli is not only possible, but facile. The required components can be easily transferred between different E. coli strains.

Disulfide bond formation in the cytoplasm.

SIGNIFICANCE: Disulfide bond formation is critical for biogenesis of many proteins. While most studies in this field are aimed at elucidating the mechanisms in the endoplasmic reticulum, intermembrane space of mitochondria, and prokaryotic periplasm, structural disulfide bond formation also occurs in other compartments including the cytoplasm. Such disulfide bond formation is essential for biogenesis of some viruses, correct epidermis biosynthesis, thermal adaptation of some extremophiles, and efficient recombinant protein production. RECENT ADVANCES: The majority of work in this new field has been reported in the past decade. Within the past few years very significant new data have emerged on the catalytic and noncatalytic mechanisms for disulfide bond formation in the cytoplasm. This includes the crystal structure of a key component of viral oxidative protein folding, identification of a missing component in cytoplasmic disulfide bond formation in hyperthermophiles, and introduction of de novo dithiol oxidants in engineered oxidative folding pathways. CRITICAL ISSUES AND FUTURE DIRECTIONS: While a broad picture of cytoplasmic disulfide bond formation has emerged many critical questions remain unanswered. The individual components in the natural systems are largely known, but the molecular mechanisms by which these processes occur are largely deduced from the mechanisms of analogous components in other compartments. This prevents full understanding and manipulation of these systems, including the potential for novel anti-viral drugs based on the unique features of their sulfhydryl oxidases and the generation of more efficient cell factories for the large-scale production of therapeutic and industrial proteins.

Two endoplasmic reticulum PDI peroxidases increase the efficiency of the use of peroxide during disulfide bond formation.

Disulfide bond formation in the endoplasmic reticulum by the sulfhydryl oxidase Ero1 family is thought to be accompanied by the concomitant formation of hydrogen peroxide. Since secretory cells can make substantial amounts of proteins that contain disulfide bonds, the production of this reactive oxygen species could have potentially lethal consequences. Here, we show that two human proteins, GPx7 and GPx8, labeled as secreted glutathione peroxidases, are actually endoplasmic reticulum-resident protein disulfide isomerase peroxidases. In vitro, the addition of GPx7 or GPx8 to a folding protein along with protein disulfide isomerase and peroxide enables the efficient oxidative refolding of a reduced denatured protein. Furthermore, both GPx7 and GPx8 interact with Ero1α in vivo, and GPx7 significantly increases oxygen consumption by Ero1α in vitro. Hence, GPx7 and GPx8 may represent a novel route for the productive use of peroxide produced by Ero1α during disulfide bond formation.

The role of dehydroascorbate in disulfide bond formation.

Dehydroascorbate (DHA) is a higher oxidation state of ascorbate formed through its action as an intracellular antioxidant. The recycling of DHA back to ascorbate is thought to be catalyzed by a variety of enzymes, including protein disulfide isomerase (PDI), linking ascorbate metabolism to oxidative protein folding in the endoplasmic reticulum (ER). Here we examine the possible role of PDI as a dehydroascorbate reductase. We find the reaction too slow to be the major route for reduction of DHA in the ER, with a second-order rate constant for the reaction of reduced PDI with DHA of only 12.5 M(-1)s(-1). Rates of a similar order of magnitude were obtained for other thioredoxin-superfamily members. However, glutaredoxin was able to catalyze DHA reduction more rapidly through a monothiol mechanism. In addition, DHA can rapidly react with many other dithiol systems, including dithiols in unfolded or partially folded proteins in a PDI-independent manner, with second-order rate constants of up to 186 M(-1)s(-1). Furthermore, we identify borate as a potent inhibitor of catalyzed and noncatalyzed DHA reduction in vitro. This study both provides insights into the link between ascorbate metabolism and oxidative protein folding and suggests a novel link between ascorbate metabolism and borate toxicity.

The C-terminal active site cysteine of Escherichia coli glutaredoxin 1 determines the glutathione specificity of the second step of peptide deglutathionylation.

Glutaredoxins are oxidoreductases specialized in reducing glutathione-protein mixed disulfides. In the first step of deglutathionylation, glutaredoxins form a mixed disulfide with glutathione, releasing reduced peptide. The specificity of this reaction is based on the unusual amide linkage formed between the gamma-carboxylate of the N-terminal glutamic acid and the alpha-amino group of the cysteine present in glutathione. In the second step of deglutathionylation, glutathione reduces the glutaredoxin-glutathione mixed disulfide. Here we show that the specificity of this second reaction for Escherichia coli Grx1, but not for human or yeast Grx1, also is based on the unusual gamma-linkage present in glutathione. Mutating Tyr13, Thr58, and/or Asp74 to alanine in E. coli Grx1 results in the glutaredoxin-peptide mixed disulfide being thermodynamically favored over the glutaredoxin-glutathione mixed disulfide in the first step of the reaction. An increased propensity to form glutaredoxin-protein mixed disulfides was observed in vivo for these same mutants. Furthermore, we demonstrate that all mutations studied in Cys14, the C-terminal active site cysteine, abolish the specificity of E. coli Grx1 for glutathione over the corresponding tripeptide Glu-Cys-Gly, which has a normal peptide bond linking Glu-Cys instead of the gamma-linkage present in glutathione, in the second step of deglutathionylation.

Efficient peroxide-mediated oxidative refolding of a protein at physiological pH and implications for oxidative folding in the endoplasmic reticulum.

The majority of secreted and outer membrane eukaryotic proteins contain disulfide bonds, formed by complex interdependent pathways in the endoplasmic reticulum. The current model for the major route of disulfide formation is the regulated flow of oxidizing equivalents from molecular oxygen to the membrane-associated enzyme Ero1 to protein disulfide isomerase, and hence to substrate proteins. One molecule of hydrogen peroxide is produced by Ero1 per disulfide bond made. This peroxide is usually considered to be a dangerous by-product. Here we show that peroxide, added to a refolding buffer or generated enzymatically in situ, results in the efficient refolding of a model protein to the native state. At pH 7.0, the kinetics of obtaining the native folded state are more efficient using peroxide than by the use of a glutathione redox buffer. Disulfide bond formation by peroxide is kinetically favored over oxidation of cysteine to cysteine sulfinic acid and over the oxidation of other amino acids in the proteins such as methionine. Hence, unless peroxides are added in excess, oxidative damage to the folding protein is minimal. Our results offer insights into potential mechanisms for disulfide bond formation in vivo.