PEDS Advance Access originally published online on May 23, 2006
Protein Engineering Design and Selection 2006 19(8):385-390; doi:10.1093/protein/gzl018
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Short Communication |
A system for concomitant overexpression of four periplasmic folding catalysts to improve secretory protein production in Escherichia coli
Lehrstuhl für Biologische Chemie, Technische Universität München 85350 Freising-Weihenstephan, Germany
1To whom Correspondence should be addressed. E-mail: skerra{at}wzw.tum.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
Although Escherichia coli is in wide use for preparative protein expression, problems with the folding of the recombinant gene product and protein aggregation are frequently encountered. Apart from cytoplasmic expression, this is also true for secretion into the bacterial periplasm, the method of choice for the production of proteins that carry structural disulfide bonds. Here we report the construction of the helper plasmid pTUM4, which effects overexpression of four established periplasmic chaperones and folding catalysts: the thiol-disulfide oxidoreductases DsbA and DsbC that catalyze the formation and isomerization of disulfide bridges and the peptidyl-prolyl cis/trans-isomerases with chaperone activity, FkpA and SurA. pTUM4 carries a p15a origin of replication and a chloramphenicol resistance gene and, thus, it is compatible with many conventional expression vectors that use the ColEI origin and an ampicillin resistance. Its positive effects on the yield of soluble recombinant protein and the homogeneity of disulfide pattern are illustrated here using the human plasma retinol-binding protein as well as the extracellular carbohydrate recognition domain of the dendritic cell membrane receptor DC-SIGN. Hence, pTUM4 represents a novel helper vector which complements existing cytosolic chaperone coexpression plasmids and should be useful for the functional secretion of various recombinant proteins with hampered folding efficiency.
Keywords: chaperone/disulfide isomerase/DsbA/DsbC/FkpA/folding catalyst/peptidyl-prolyl cis/trans-isomerase/SurA
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
The periplasm of the Gram-negative bacterium Escherichia coli is of particular interest for the heterologous expression of eukaryotic secretory proteins as its oxidizing environment favors the formation of structural disulfide bonds (Georgiou and Segatori, 2005
). Thus, periplasmic secretion has been used for the functional production of a variety of recombinant proteins (Mergulhao et al., 2005), including antibody fragments as probably the most widely investigated class of eukaryotic proteins (Skerra, 1993).
However, the folding efficiency of such recombinant gene products in the bacterial periplasm varies and depends significantly on their individual properties. While some proteins can be almost quantitatively recovered in an acitve state others are to a large fraction deposited as incorrectly folded and aggregated material, which limits the yield of easily purifiable product (Skerra and Plückthun, 1991; Mergulhao et al., 2005). Even though there seems to be no strict correlation (Knappik et al., 1993), inefficient protein folding in the periplasm is often accompanied by a toxic effect on the bacterial host cell, which leads to reduced growth after induction of foreign gene expression and an increased tendency of lysis, thus hampering selective cell fractionation. Studies of protein folding in the periplasm of E. coli received early attention (Wülfing and Plückthun, 1994), but only more recently the presence of molecular chaperones and of several folding catalysts that support disulfide bond formation and peptidyl-prolyl cis/trans isomerization has been recognized (Baneyx and Mujacic, 2004; Choi and Lee, 2004).
Folding of proteins that carry more than one disulfide bridge involves two types of reactions, which both may be rate-limiting for the folding process: (i) introduction of disulfide bonds into the nascent polypeptide chain via pairwise oxidation of Cys residues and (ii) isomerization of intermediary non-physiological disulfide bonds to attain the native connectivity. These two processes are catalyzed by the thiol-disulfide oxidoreductases DsbA (Bardwell et al., 1991) and DsbC (Missiakas et al., 1994; Shevchik et al., 1994), respectively. Both enzymes are members of the thioredoxin superfamily and exhibit a catalytic disulfide bond as part of a Cys-Xaa-Xaa-Cys consensus sequence in their active sites (Collet and Bardwell, 2002).
DsbA, a monomeric protein of 21.1 kDa, constitutes the generic dithiol oxidase in the periplasm of E. coli. After accepting two electrons from a pair of Cys residues in the substrate polypeptide it is recycled as an oxidant by the inner membrane protein DsbB (Bardwell et al., 1993). DsbB subsequently becomes reoxidized by molecular oxygen via the electron transport chain in the bacterial plasma membrane (Bader et al., 2000). DsbA transfers a disulfide bond from its oxidized active site to reduced, unfolded polypeptide chains in an extremely rapid reaction while merely showing weak disulfide isomerase activity.
The latter task is fulfilled by DsbC, a homodimer of 23.4 kDa subunits that comprises two thioredoxin domains (Zapun et al., 1995). DsbC is kept in its catalytically active, reduced state by the inner membrane protein DsbD, which receives electrons from the cytoplasmic NADPH pool via thioredoxin reductase and thioredoxin (Rietsch et al., 1997). Thus, DsbA and DsbC provide a specific thiol oxidant and reductant, respectively, in a mechanistically separated manner (Collet and Bardwell, 2002). Previous studies have shown that overproduction of one or both of the enzymes can greatly improve proper disulfide bond formation in cysteine-rich recombinant proteins, for example in human tissue plasminogen activator (Qiu et al., 1998) or in T-cell receptor fragments (Wülfing and Plückthun, 1994).
Another well-known step along the folding pathway of proteins that often requires catalysis in vivo is the cis/trans isomerization of prolyl-iminopeptide bonds (Fischer et al., 1998). While the trans configuration of Xaa-Pro bonds is favored in newly synthesized polypeptides, 
5% of all prolyl-peptide bonds occur in a cis configuration in the natively folded protein. The trans to cis isomerization of Xaa-Pro bonds is catalyzed by peptidyl-prolyl cis/trans-isomerases (PPIases, sometimes also called rotamases) (Fanghanel and Fischer, 2004). So far, four PPIases have been described for the periplasm of E. coli, including the parvulins SurA (Rouviere and Gross, 1996) and PpiD (Dartigalongue and Raina, 1998), the cyclophilin PpiA (also known as CypA or RotA) (Liu and Walsh, 1990) and the FKBP (FK506-binding protein) FkpA (Horne and Young, 1995).
Among these, the 26.2 kDa homodimeric protein FkpA exhibits generic folding enhancer activity with a broad substrate range in addition to its PPIase function (Baneyx and Mujacic, 2004). For example, FkpA was described to suppress the formation of insoluble aggregates in case of a folding-defective mutant of the maltose-binding protein, MalE31 (Saul et al., 2003), and it was shown to improve the yield of single-chain antibody fragments secreted into the periplasm of E. coli (Bothmann and Plückthun, 2000; Zhang et al., 2003).
SurA (for ‘survival’) is known to be involved in the folding and assembly of endogenous outer membrane porins (Lazar and Kolter, 1996; Hennecke et al., 2005). Furthermore, SurA can promote the folding of unstable proteins (Missiakas and Raina, 1997), e.g. a ß-lactamase fusion with protein A, and aggregation-prone proteins, e.g. MalE31 or a truncated periplasmic AnsB. Although having PPIase activity, it was recently shown that the major function of SurA is that of a chaperone, an activity which is apparently linked to a distinct structural domain (Bitto and McKay, 2002).
Contrasting with these two proteins, no effect on the kinetics of periplasmic protein folding and outer membrane protein assembly was found when investigating an E. coli deletion mutant of PpiA, indicating that this PPIase does not play a crucial role (Kleerebezem et al., 1995). Similarly, the PPIase PpiD seems to be redundant, especially in the context of SurA (Dartigalongue and Raina, 1998). Hence, SurA and FkpA appear to be most promising with respect to supporting the efficient periplasmic folding of recombinant proteins.
We describe in this article a plasmid, pTUM4, that is largely compatible with existing expression vectors and combines the oxidoreductase, PPIase and general chaperone activities of DsbA, DsbC, FkpA and SurA to improve the folding of recombinant proteins in the periplasm of E. coli. The applicability of pTUM4 is exemplified with two recombinant proteins that carry several disulfide bonds and exhibit hampered folding efficiency, a human lipocalin and a soluble extracellular domain of an immunological membrane receptor.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
Construction of pTUM3 and pTUM4
pTUM3 was constructed starting from pASK61, a previously developed coexpression vector for DsbA and PpiA/RotA (Müller and Skerra, 1993). The dsbA gene, including its constitutive promoter, was first amplified from pASK61 using the phosphorothioate primers (Skerra, 1992) dsbA-1 (5'-TAATGATCTAGAAGCTTATGAAGAATTTAGCG-3'; underlined: XbaI recognition site; italics: HindIII recognition site) and dsbA-2 (5'-ATTATTTTTTCTCGGACAGATATTTCACTG-3'). In this way a ‘T’ nucleotide was added at the 3' end of the dsbA gene, directly after the TAA stop codon, thus giving rise to a second stop codon (TAG) upon insertion into the polylinker region upstream of the dsbC structural gene on pASK75-DsbC (Schmidt et al., 1998). To this end, pASK75-DsbC was first digested with HindIII and the sticky ends were filled in by T4 DNA polymerase. After a second cut with XbaI the vector fragment was ligated with the XbaI-digested PCR product, yielding pASK75-DsbA/C. From this plasmid the resulting dicistronic operon, with the intergenic sequence and the lipoprotein transcription terminator downstream of the dsbC gene (Schmidt et al., 1998), was amplified as a whole using primers dsbA-1 and LPTT-1 (5'-TCCGTGATGCATTAGCGGTAAACGG-3'; underlined: NsiI recognition site). The PCR product was cut with HindIII and NsiI and ligated with the likewise cut plasmid pASK61, thereby substituting the dsbA gene and yielding pTUM3a.
Unfortunately, double-stranded DNA sequencing of the cloned dicistronic operon on pTUM3a revealed that the TAGC sequence at the 5' end of the intergenic region (5'-TAATAGCTTCGGGAAGATTTATG-3'; italics: second stop codon of dsbA; underlined: start codon of dsbC) was missing, probably resulting from an incomplete fill-in reaction during construction of pASK75-DsbA/C. Notably, tests with pTUM3a showed that while DsbA was overexpressed DsbC was not readily detectable, most likely due to inefficient translational initiation for the dsbC cistron. Consequently, the intergenic sequence on pASK75-DsbA/C was replaced with the translational initiation site that had been developed for the generic expression vector pASK75 (Skerra, 1994), 5'-TAACGAGGGCAAAAAATG-3' (italics: stop codon of dsbA; underlined: start codon of dsbC), using site-directed mutagenesis (Geisselsoder et al., 1987) with the oligodeoxynucleotide 5'-CAACATAAAACCTTTTTTCATTTTTTGCCCTCGTTATTTTTTCTCGGACAG-3'. The optimized dicistronic operon was then amplified and cloned on pASK61 as before, finally yielding pTUM3.
In the next step, the ppiA/rotA gene on pTUM3 was replaced by the structural genes for FkpA and SurA, which were arranged in a dicistronic operon under common transcriptional control of the constitutive fkpA promoter (Dartigalongue et al., 2001). To this end, the fkpA gene including its promoter region was PCR-amplified from genomic DNA of E. coli K-12 W3110 (Bachmann, 1972) using phosphorothioate primers FkpA-1 (5'-GCATGAGTGCCCTCTTTTGTCGAATGGTCG-3'; underlined: Bme1580I recognition site) and FkpA-2 (5'-TTATTTTTTAGCAGAATCTGCGG-3'; 5'-phosphorylated). The surA gene, including the 16 bp upstream region encompassing its ribosomal binding site, was likewise amplified using primers SurA-1 (5'-TTGAAATGGAAAAAGTATGAAGAACTGG-3'; 5'-phosphorylated) and SurA-2 (5'-GTTTTAAAGCTTAGTTGCTCAGGATTTTAACGTAG-3'; underlined: HindIII recognition site). After cutting the purified PCR fragments with Bme1580I and HindIII, respectively, they were combined in a three-fragment ligation (involving blunt end ligation between the two PCR fragments) with the large fragment of pTUM3 that was obtained by digestion with the same restriction enzymes (cutting upstream and downstream of ppiA, respectively), thus yielding pTUM4. All DNA sequences were confirmed by restriction analysis and double-stranded DNA sequencing (ABI-PrismTM310 Genetic analyzer, Perkin-Elmer Applied Biosystems, Weiterstadt, Germany) using the BigDyeTM terminator kit.
Recombinant protein production and purification
Expression analysis of folding catalysts encoded on pTUM3 and pTUM4 was performed at 30°C in 50 ml LB medium (Sambrook et al., 1989) supplemented with 30 µg/ml chloramphenicol. After cell harvest at OD550 = 1.0 by centrifugation, periplasmic protein extraction was performed by resuspending the cell pellet with ice-cold 500 µl 500 mM sucrose, 1 mM EDTA, 100 mM Tris/HCl pH 8.0 containing 100 µg/ml lysozyme and incubating for 30 min on ice. The spheroplasts were removed by repeated centrifugation (Skerra and Schmidt, 2000) and the supernatant was recovered as periplasmic cell fraction.
Human RBP (22230 Da; 193 amino acids, including Strep-tag II) as well as the extracellular carbohydrate recognition domain (CRD) of DC-SIGN (15896 Da; 136 amino acids, including Strep-tag II; comprising residues 254–381 of the mature membrane protein with a H254Q substitution) were produced at 22°C in E. coli JM83 (Yanisch-Perron et al., 1985) harboring the pASK75-based (Skerra, 1994) expression plasmids phRBP-A (D.Breustedt, N.Bischoff and A.Skerra, unpublished data) and pDC1 (A.Holla and A.Skerra, unpublished data) using 2 l LB cultures supplemented with 100 µg/ml ampicillin and, when co-transformed with pTUM4, 30 µg/ml chloramphenicol. Induction of foreign gene expression with anhydrotetracycline at OD550 = 0.5 for 3 h (typically resulting in OD550 
1.0 at harvest) and periplasmic extraction in the presence of 500 mM sucrose, 1 mM EDTA and 100 mM Tris/HCl, pH 8.0, were performed as described (Skerra and Schmidt, 2000). The recombinant proteins were purified via the Strep-tag II by means of streptavidin affinity chromatography (Skerra and Schmidt, 2000) and analyzed by SDS–PAGE (15% w/v), followed by staining with Coomassie Brilliant Blue R-250.
Protein concentrations were determined according to the absorption at 280 nm using calculated extinction coefficients (Gill and von Hippel, 1989) of 38 870 M–1cm–1 for the recombinant human RBP and of 59 640 M–1cm–1 for the extracellular domain of DC-SIGN, respectively. Yields of purified proteins [mg/l/OD] were normalized to 1 l of bacterial culture and an optical density at harvest of 1.0.
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
Design of the helper plasmids pTUM3 and pTUM4
Two different low-copy number plasmids, pTUM3 and pTUM4, were constructed on the basis of the generic vector pACYC184 (Chang and Cohen, 1978) to achieve the constitutive expression of several protein folding catalysts. pTUM3 encodes DsbA, DsbC and PpiA/RotA whereas pTUM4 encodes DsbA, DsbC, FkpA and SurA. Both plasmids are compatible with commonly used bacterial secretion vectors for recombinant proteins (Skerra et al., 1991; Skerra, 1994; Sorensen and Mortensen, 2005). An important aspect for the concomitant overexpression of the entire set of encoded protein folding catalysts was that on the bacterial chromosome the structural genes for DsbA, FkpA as well as PpiA/RotA are equipped with their own constitutive promoters whereas the dsbC and surA genes are part of larger operons and lie further downstream of their corresponding promoter regions.
Hence, starting from the simpler helper plasmid pASK61 (Müller and Skerra, 1993), which encodes DsbA and PpiA\ RotA, we first constructed an artificial dicistronic operon where the dsbC gene was inserted directly downstream of the cloned dsbA gene, thus forming a transcriptional fusion under common control of the constitutive dsbA promoter. The resulting plasmid, pTUM3, carries the p15a origin of replication (Selzer et al., 1983), the chloramphenicol resistance gene (Alton and Vapnek, 1979), the dicistronic operon encoding DsbA and DsbC, followed by the strong lpp transcription terminator, as well as the gene for PpiA/RotA.
As previous studies indicated that PpiA/RotA does not play a major role in assisting endogenous protein folding (Kleerebezem et al., 1995), its expression cassette on pTUM3 was subsequently replaced by a second artificial dicistronic operon coding for FkpA and SurA. To this end, the fkpA gene, together with its constitutive promoter, and the surA structural gene were amplified in parallel from bacterial genomic DNA of E. coli K-12 W3110 and inserted instead of the ppiA/rotA gene on pTUM3, thus resulting in the plasmid pTUM4 (Figure 1).
|
To analyze overexpression of the plasmid-encoded protein folding catalysts, E. coli K-12 strain JM83 harboring pTUM4, pTUM3 or pTUM2 (a truncated derivative of pACYC184 without the Tcr gene, serving as negative control) was cultivated in the presence of chloramphenicol at 30°C to an OD550 = 1. Cells harboring pTUM3 and pTUM4 showed slightly slower growth kinetics compared with pTUM2, which was obviously due to the strong overexpression of the encoded folding catalysts. Nevertheless, overnight cultivation resulted in very similar stationary OD550 values of 3.4 and 3.6 for the cultures harboring either pTUM4 or pTUM2 as well as pTUM3, respectively.
Especially in the case of pTUM4 all four cloned folding catalysts, DsbA, DsbC, FkpA and SurA, gave rise to prominent bands in a Coomassie-stained SDS polyacrylamide gel upon analysis of the periplasmic cell fraction after cell harvest at OD550 = 1 (Figure 2). Interestingly, DsbC, SurA and FkpA were quantitatively released from the periplasm only when the cell fractionation was assisted by digesting the peptidoglycan with lysozyme (Schmidt et al., 1998) (see below). While these three proteins showed electrophoretic mobilities as expected from their calculated molecular weights (21.1, 26.2 and 45.0 kDa, respectively), DsbC with its molecular weight of 23.4 kDa showed slightly retarded mobility, thus giving rise to an apparent size >25 kDa, which had also been noted before (Missiakas et al., 1994).
|
Similarly, pTUM3 led to detectable bands—under identical experimental conditions—for the three cloned folding catalysts DsbA, DsbC, and RotA. However, their relative amounts, though mutually identical for DsbA and DsbC, were significantly lower when compared with pTUM4 (estimated as
25%). This concerted drop in expression efficiency might be due to upstream transcriptional effects, topological differences between the plasmid DNA supercoil forms or to changes in the copy number between pTUM3 and pTUM4. Nevertheless, it appears that the genes of the artifical dicistronic operons are overexpressed on each of the two plasmids, thus yielding elevated concentrations of the soluble periplasmic folding catalysts. Effect of coexpression of folding helper proteins on human RBP
To investigate the effect of the protein folding catalysts on the soluble periplasmic production of a recombinant protein we first chose the human plasma retinol-binding protein (RBP), a small monomeric protein of 183 amino acids that carries three disulfide bonds (Cowan et al., 1990). RBP belongs to the lipocalin family of proteins and exhibits a central eight-stranded anti-parallel ß-barrel, which carries the binding pocket for vitamin A at its open end. RBP is known to form non-native disulfide isomers upon secretion in E. coli and to cause toxic effects on the bacterial host cell (Müller and Skerra, 1993; Schmidt et al., 1998).
The recombinant human RBP was secreted into the periplasm of E. coli by means of the OmpA signal peptide according to a previously published strategy for RBP from pig (Müller and Skerra, 1993). To this end the expression plasmid phRBP-A was employed, based on the generic cloning and expression vector pASK75, where its gene expression is tightly controlled by the chemically inducible tetA promoter/operator (Skerra, 1994). The Strep-tag II fused to its C-terminus permitted one step purification via streptavidin affinity chromatography (Skerra and Schmidt, 2000).
When E. coli JM83 harboring phRBP-A was grown at 22°C to a mid-log phase and heterologous gene expression was induced by the addition of anhydrotetracycline, the optical density of the culture started to decrease 2.5 h after induction, thus indicating cell lysis caused by the recombinant gene product. Although RBP could be purified from the periplasmic cell fraction at a yield of 150 µg/l/OD, its analysis by SDS–PAGE under non-reducing conditions revealed the presence of a second prominent band, which should correspond either to a fraction of reduced protein or to a non-native disulfide bond isomer (Figure 3). Notably, both the toxicity effect and the inhomogeneity of the disulfide-related protein pattern were abolished when expression of RBP was performed—under otherwise identical conditions—in the presence of either pTUM3 or pTUM4.
|
The quantitative formation of disulfide bonds in the recombinant human RBP is obviously owing to overexpression of the thiol-disulfide oxidoreductases DsbA and/or DsbC, which are encoded both on pTUM3 and pTUM4. It is known that the ability of the bacterial oxidative folding machinery to correctly fold both endogenous and heterologous proteins that carry disulfide bonds is limited. This is probably the reason for the repeated observation of at least one non-native disulfide bond isomer of serum RBP observed in the absence of catalytic support (Müller and Skerra, 1993
). In this earlier study the formation of disulfide isomers was prevented by combined overexpression of DsbA and addition of the thiol reductant N-acetyl cysteine to the culture medium. Meanwhile, it has been shown that the formation and isomerization of disulfide bonds in recombinant proteins can be boosted by overexpression of DsbA together with DsbC (Qiu et al., 1998; Schmidt et al., 1998; Maskos et al., 2003).
The concomitant reduction in the periplasmic amount of misfolded proteins may also have a positive influence on cell viability, even though the growth rates of the bacterial cells haboring pTUM3 or pTUM4 were slightly reduced as a consequence of the simultaneous overexpression of the encoded folding catalysts. Nevertheless, the yield of purifyable RBP from the culture also harboring pTUM3 or pTUM4 was 0.4 mg/l/OD or 0.6 mg/l/OD, respectively. Compared with the culture harboring phRBP-A alone, coexpression of pTUM3 or pTUM4 raised the yield of correctly folded RBP by a factor of 3 and 4, respectively. Hence, the additional overexpression of the two PPIases SurA and FkpA with their chaperone functions (Arie et al., 2001; Behrens et al., 2001) seems to promote recombinant protein folding in the periplasm of E. coli and to lead to higher yields of soluble RBP.
Effect of PTUM4 on the secretion of the CRD of DC-SIGN
The effect of folding helper coexpression was further investigated in the production of an extracellular fragment of the human dendritic cell membrane receptor DC-SIGN. DC-SIGN is a cell surface lectin which can bind gp120 on HIV and therefore constitutes an interesting protein in biomedical research (Su et al., 2003). The cDNA sequence encoding the extracellular CRD of DC-SIGN (Curtis et al., 1992), which carries three disulfide bonds, was expressed from pDC1. This plasmid, again a derivative of pASK75, codes for a fusion protein of the CRD with an N-terminal OmpA signal peptide and the Strep-tag II at its C-terminus.
Production of the DC-SIGN CRD in E. coli JM83 transformed with pDC1, under similar conditions as described for RBP above, caused dramatic cell lysis starting 2 h after induction of gene expression. In contrast, after co-transformation with pTUM4 no indication of lysis was observed, even if production of the CRD was extended overnight (data not shown). Analysis of the periplasmic extracts, which were prepared via mild osmotic shock and EDTA treatment from pelleted cells 3 h after induction, by SDS–PAGE (Figure 4) revealed distinct protein patterns, indicating more selective cell fractionation in the presence of pTUM4. In contrast, in its absence the periplasmic cell extract was highly contaminated with cytoplasmic proteins, which apparently resulted from cell lysis during the fractionation step.
|
Following streptavidin affinity chromatography under gentle conditions, the recombinant CRD was nearly 100% pure when produced in the presence of pTUM4, whereas a series of contaminating host cell proteins was still visible in the preparation obtained without the helper plasmid (Figure 4). Apart from the positive influence on cell viability, co-transformation with pTUM4 increased the yield of soluble protein almost 10-fold, resulting in 0.4 mg/l/OD of the purified CRD versus 45 µg/l/OD in the situation with pDC1 alone.
|
Conclusion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
The construction of pTUM4 as a generally applicable helper plasmid for the production of recombinant proteins in the periplasm of E. coli enabled the overexpression of DsbA, DsbC, FkpA and SurA as soluble proteins in high amounts. Our data show that this system can (i) prevent inefficient disulfide bond formation, (ii) raise the yield of correctly folded, soluble recombinant protein and (iii) abolish toxic effects of heterologous protein secretion on the bacterial host cell.
The effects of FkpA and SurA coexpression on the production of human RBP or the extracellular fragment of human DC-SIGN are mechanistically less obvious than for DsbA and DsbC. However, comparison between pTUM4 and pTUM3, which lacks these two PPIase genes, indicates that overexpression of FkpA and SurA not only increases the yield of recombinant protein but also significantly improves bacterial cell viability. Both effects may be due to their cis/trans isomerization activity on Xaa-Pro peptide bonds and/or to their general chaperone activities with respect to preventing periplasmic protein aggregation.
Although more detailed mechanistic analyses would be needed to pinpoint the individual influences of the encoded folding catalysts and chaperones on the folding pathway of each recombinant protein, the examples presented in the present study, together with several other successful applications of this system in our laboratory (Breustedt et al., 2006; Nasreen et al., 2006), indicate that pTUM4 provides an appropriate combination of these helper proteins. Hence, this plasmid constitutes a generally useful tool for the periplasmic secretion of functional recombinant proteins in E. coli.
|
Footnotes |
|---|
Edited by Thomas Kiefhaber
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
The authors wish to thank Klaus Wachinger for the technical assistance and Daniel Breustedt, Nicole Bischoff, Andrea Holla and Jan-Peter Mayer for kindly providing the plasmids phRBP-A, pDC1 and pTUM2.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionConclusionAcknowledgementsReferences |
|---|
Alton N.K. and Vapnek D. (1979) Nature 282:864–869.[CrossRef][Medline]
Arie J.P., Sassoon N., Betton J.M. (2001) Mol. Microbiol. 39:199–210.[CrossRef][ISI][Medline]
Bachmann B.J. (1972) Bacteriol. Rev. 36:525–557.
Bader M.W., Xie T., Yu C.A., Bardwell J.C. (2000) J. Biol. Chem. 275:26082–26088.
Baneyx F. and Mujacic M. (2004) Nat. Biotechnol. 22:1399–1408.[CrossRef][ISI][Medline]
Bardwell J.C., McGovern K., Beckwith J. (1991) Cell 67:581–589.[CrossRef][ISI][Medline]
Bardwell J.C., Lee J.O., Jander G., Martin N., Belin D., Beckwith J. (1993) Proc. Natl Acad. Sci. USA 90:1038–1042.
Behrens S., Maier R., de Cock H., Schmid F.X., Gross C.A. (2001) EMBO J. 20:285–294.[CrossRef][Medline]
Bitto E. and McKay D.B. (2002) Structure 10:1489–1498.[Medline]
Bothmann H. and Plückthun A. (2000) J. Biol. Chem. 275:17100–17105.
Breustedt D.A., Schönfeld D.L., Skerra A. (2006) Biochim. Biophys. Acta. 1764:161–173.[Medline]
Chang A.C. and Cohen S.N. (1978) J. Bacteriol. 134:1141–1156.
Choi J.H. and Lee S.Y. (2004) Appl. Microbiol. Biotechnol. 64:625–635.[CrossRef][Medline]
Collet J.F. and Bardwell J.C. (2002) Mol. Microbiol. 44:1–8.[CrossRef][ISI][Medline]
Cowan S.W., Newcomer M.E., Jones T.A. (1990) Proteins 8:44–61.[CrossRef][ISI][Medline]
Curtis B.M., Scharnowske S., Watson A.J. (1992) Proc. Natl Acad. Sci. USA 89:8356–8360.
Dartigalongue C. and Raina S. (1998) EMBO J. 17:3968–3980.[CrossRef][ISI][Medline]
Dartigalongue C., Missiakas D., Raina S. (2001) J. Biol. Chem. 276:20866–20875.
Fanghanel J. and Fischer G. (2004) Front. Biosci. 9:3453–3478.[Medline]
Fischer G., Tradler T., Zarnt T. (1998) FEBS Lett. 426:17–20.[CrossRef][Medline]
Geisselsoder J., Whitney F., Yuckenberg P. (1987) Biotechniques 5:786–790.
Georgiou G. and Segatori L. (2005) Curr. Opin. Biotechnol. 16:538–545.[Medline]
Gill S.C. and von Hippel P.H. (1989) Anal. Biochem. 182:319–326.[CrossRef][ISI][Medline]
Hennecke G., Nolte J., Volkmer-Engert R., Schneider-Mergener J., Behrens S. (2005) J. Biol. Chem. 280:23540–23548.
Horne S.M. and Young K.D. (1995) Arch. Microbiol. 163:357–365.[Medline]
Kleerebezem M., Heutink M., Tommassen J. (1995) Mol. Microbiol. 18:313–320.[CrossRef][Medline]
Knappik A., Krebber C., Plückthun A. (1993) Biotechnology 11:77–83.[Medline]
Lazar S.W. and Kolter R. (1996) J. Bacteriol. 178:1770–1773.
Liu J. and Walsh C.T. (1990) Proc. Natl Acad. Sci. USA 87:4028–4032.
Maskos K., Huber-Wunderlich M., Glockshuber R. (2003) J. Mol. Biol. 325:495–513.[CrossRef][ISI][Medline]
Mergulhao F.J., Summers D.K., Monteiro G.A. (2005) Biotechnol. Adv. 23:177–202.[CrossRef][Medline]
Missiakas D. and Raina S. (1997) J. Bacteriol. 179:2465–2471.
Missiakas D., Georgopoulos C., Raina S. (1994) EMBO J. 13:2013–2020.[ISI][Medline]
Müller H.N. and Skerra A. (1993) J. Mol. Biol. 230:725–732.[CrossRef][Medline]
Nasreen A., Vogt M., Kim H.J., Eichinger A., Skerra A. (2006) Protein Sci. 15:190–199.
Qiu J., Swartz J.R., Georgiou G. (1998) Appl. Environ. Microbiol. 64:4891–4896.
Rietsch A., Bessette P., Georgiou G., Beckwith J. (1997) J. Bacteriol. 179:6602–6608.
Rouviere P.E. and Gross C.A. (1996) Genes Dev. 10:3170–3182.
Sambrook J., Fritsch E.F, Maniatis T. (1989) Molecular Cloning: A Laboratory Manual 2nd edition (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY).
Saul F.A., Mourez M., Vulliez-Le Normand B., Sassoon N., Bentley G.A., Betton J.M. (2003) Protein Sci. 12:577–585.
Schmidt A.M., Bloß I., Skerra A. (1998) Protein Eng. 11:601–607.
Selzer G., Som T., Itoh T., Tomizawa J. (1983) Cell 32:119–129.[CrossRef][ISI][Medline]
Shevchik V.E., Condemine G., Robert-Baudouy J. (1994) EMBO J. 13:2007–2012.[Medline]
Skerra A. (1992) Nucleic Acids Res. 20:3551–3554.
Skerra A. (1993) Curr. Opin. Immunol. 5:256–262.[CrossRef][ISI][Medline]
Skerra A. (1994) Gene 151:131–135.[CrossRef][ISI][Medline]
Skerra A. and Plückthun A. (1991) Protein Eng. 4:971–979.
Skerra A. and Schmidt T.G. (2000) Methods Enzymol. 326:271–304.[ISI][Medline]
Skerra A., Pfitzinger I., Plückthun A. (1991) Biotechnology 9:273–278.[Medline]
Sorensen H.P. and Mortensen K.K. (2005) J. Biotechnol. 115:113–128.[CrossRef][ISI][Medline]
Su S.V., Gurney K.B., Lee B. (2003) Curr. HIV Res. 1:87–99.[CrossRef][Medline]
Wülfing C. and Plückthun A. (1994) Mol. Microbiol. 12:685–692.[ISI][Medline]
Yanisch-Perron C., Vieira J., Messing J. (1985) Gene 33:103–119.[CrossRef][ISI][Medline]
Zapun A., Missiakas D., Raina S., Creighton T.E. (1995) Biochemistry 34:5075–5089.[CrossRef][Medline]
Zhang Z., et al. (2003) Biotechniques 35:1032–1038.[Medline]
Received March 28, 2006; accepted April 14, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N. S. Outchkourov, W. Roeffen, A. Kaan, J. Jansen, A. Luty, D. Schuiffel, G. J. van Gemert, M. van de Vegte-Bolmer, R. W. Sauerwein, and H. G. Stunnenberg Correctly folded Pfs48/45 protein of Plasmodium falciparum elicits malaria transmission-blocking immunity in mice PNAS, March 18, 2008; 105(11): 4301 - 4305. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||