The Periplasmic Escherichia coli Peptidylprolyl cis,trans-Isomerase FkpA
2000; Elsevier BV; Volume: 275; Issue: 22 Linguagem: Inglês
10.1074/jbc.m910234199
ISSN1083-351X
AutoresKathrin Ramm, Andreas Plückthun,
Tópico(s)Signaling Pathways in Disease
ResumoWe recently identified FkpA by selecting for the increased yield of antibody single-chain Fv (scFv) fragments in phage display, even of those not containing cis-prolines. We have now investigated the properties of FkpA in vitro. The peptidylprolyl cis-trans-isomerase activity of FkpA was found to be among the highest of any such enzyme with a protein substrate, yet FkpA is not able to enhance the proline-limited refolding rate of the disulfide-free hu4D5-8 scFv fragment, probably due to inaccessibility of Pro-L95. Nevertheless, the yield of the soluble and functional scFv fragment was dramatically increasedin vitro in the presence of FkpA. Similar effects were observed for an scFv fragment devoid of cis-prolines. We are thus forced to conclude that the observed folding-assisting function is independent of the isomerase activity of the protein. The beneficial effect of FkpA was found to be due to two components. First, FkpA interacts with early folding intermediates, thus preventing their aggregation. Additionally, it has the ability to reactivate inactive protein, possibly also by binding to a partially unfolded species that may exist in equilibrium with the aggregated form, which may thus be released on a productive pathway. These in vitromeasurements therefore fully reflect the in vivo results from periplasmic overexpression of FkpA. We recently identified FkpA by selecting for the increased yield of antibody single-chain Fv (scFv) fragments in phage display, even of those not containing cis-prolines. We have now investigated the properties of FkpA in vitro. The peptidylprolyl cis-trans-isomerase activity of FkpA was found to be among the highest of any such enzyme with a protein substrate, yet FkpA is not able to enhance the proline-limited refolding rate of the disulfide-free hu4D5-8 scFv fragment, probably due to inaccessibility of Pro-L95. Nevertheless, the yield of the soluble and functional scFv fragment was dramatically increasedin vitro in the presence of FkpA. Similar effects were observed for an scFv fragment devoid of cis-prolines. We are thus forced to conclude that the observed folding-assisting function is independent of the isomerase activity of the protein. The beneficial effect of FkpA was found to be due to two components. First, FkpA interacts with early folding intermediates, thus preventing their aggregation. Additionally, it has the ability to reactivate inactive protein, possibly also by binding to a partially unfolded species that may exist in equilibrium with the aggregated form, which may thus be released on a productive pathway. These in vitromeasurements therefore fully reflect the in vivo results from periplasmic overexpression of FkpA. peptidylprolyl cis,trans-isomerase single-chain Fv enzyme-linked immunosorbent assay guanidinium chloride bovine serum albumin FK506-binding protein FkpA is one of three soluble peptidylprolylcis,trans-isomerases (PPIase)1 in the periplasm ofEscherichia coli, in addition to PpiA (RotA) and SurA (1.Missiakas D. Betton J.-M. Raina S. Mol. Microbiol. 1996; 21: 871-884Crossref PubMed Scopus (295) Google Scholar). It was first described as being similar to the macrophage infectivity potentiator of Legionella species (2.Horne S.M. Young K.D. Arch. Microbiol. 1995; 163: 357-365Crossref PubMed Scopus (85) Google Scholar). It is homologous to the FK506-binding proteins and was subsequently shown to belong to the ςE regulon and thus to be inducible under stress conditions (Refs. 3.Rouvière P.E. De Las Peñas A. Mecsas J. Lu C.Z. Rudd K.E. Gross C.A. EMBO J. 1995; 14: 1032-1042Crossref PubMed Scopus (264) Google Scholar, 4.Raina S. Missiakas D. Georgopoulos C. EMBO J. 1995; 14: 1043-1055Crossref PubMed Scopus (244) Google Scholar, 5.Danese P.N. Silhavy T.J. Genes Dev. 1997; 11: 1183-1193Crossref PubMed Scopus (217) Google Scholar; reviewed in Ref. 6.Danese P.N. Silhavy T.J Annu. Rev. Genet. 1998; 32: 59-94Crossref PubMed Scopus (190) Google Scholar). Although thefkpA deletion mutant has been shown to be viable (1.Missiakas D. Betton J.-M. Raina S. Mol. Microbiol. 1996; 21: 871-884Crossref PubMed Scopus (295) Google Scholar, 2.Horne S.M. Young K.D. Arch. Microbiol. 1995; 163: 357-365Crossref PubMed Scopus (85) Google Scholar), FkpA has been proposed to have a general folding-assisting function in the periplasm (Refs. 1.Missiakas D. Betton J.-M. Raina S. Mol. Microbiol. 1996; 21: 871-884Crossref PubMed Scopus (295) Google Scholar and 5.Danese P.N. Silhavy T.J. Genes Dev. 1997; 11: 1183-1193Crossref PubMed Scopus (217) Google Scholar; reviewed in Ref. 6.Danese P.N. Silhavy T.J Annu. Rev. Genet. 1998; 32: 59-94Crossref PubMed Scopus (190) Google Scholar). This idea was based on an fkpA deletion causing up-regulation of ςE and thus stimulation of degP transcription and, in addition, FkpA's own induction by extracytoplasmic stress (5.Danese P.N. Silhavy T.J. Genes Dev. 1997; 11: 1183-1193Crossref PubMed Scopus (217) Google Scholar) and its ability to restore near-normal ςE-dependent response when overexpressed (1.Missiakas D. Betton J.-M. Raina S. Mol. Microbiol. 1996; 21: 871-884Crossref PubMed Scopus (295) Google Scholar). In the accompanying paper (7.Bothmann H. Plückthun A. J. Biol. Chem. 2000; 275: 17100-17105Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar), FkpA was identified in a selection system as a factor improving the functional expression of antibody scFv fragments in the bacterial periplasm. This was achieved by overexpressing a library of E. coli proteins in a phagemid, on which a poorly expressing antibody was encoded for phage display, and panning for functional antibody. In vivo overexpression experiments provided further evidence that the yield of functional antibody scFv fragments produced in the periplasm is increased upon coexpression of FkpA. Interestingly, the effect was also observed with antibody fragments devoid of cis-prolines (7.Bothmann H. Plückthun A. J. Biol. Chem. 2000; 275: 17100-17105Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar). In antibody scFv fragments, the variable domain of the heavy chain (VH) is connected via a linker to the variable domain of the light chain (VL). Antibody VH domains have never been found to carry a cis-proline in their framework, 2A. Honegger, personal communication. whereas antibody VL domains come in two types, κ and λ. κ domains have two cis-prolines, at positions L8 and L95 (consensus numbering of Kabat et al. (8.Kabat E.A. Wu T.T. Perry H.M. Gottesman K.S. Foeler C. NIH Publication No. 91-3242. National Institutes of Health, Bethesda, MD1991Google Scholar)). In contrast, λ domains do not carry any cis-prolines in the framework in any known structure in the Protein Data Bank.2 The process of structure formation during in vitro refolding of an scFv fragment has been described in some detail for the scFv fragments McPC603 (9.Freund C. Honegger A. Hunziker P. Holak T.A. Plückthun A. Biochemistry. 1996; 35: 8457-8464Crossref PubMed Scopus (35) Google Scholar, 10.Freund C. Gehrig P. Baici A. Holak T.A. Plückthun A. Fold. Des. 1997; 3: 39-49Abstract Full Text Full Text PDF Scopus (15) Google Scholar, 11.Freund C. Gehrig P. Holak T.A. Plückthun A. FEBS Lett. 1997; 407: 42-46Crossref PubMed Scopus (19) Google Scholar, 12.Jäger M. Plückthun A. FEBS Lett. 1997; 418: 106-110Crossref PubMed Scopus (39) Google Scholar, 13.Jäger M. Plückthun A. J. Mol. Biol. 1999; 285: 2005-2019Crossref PubMed Scopus (43) Google Scholar) and hu4D5-8 (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar). 3M. Jäger, P. Gehrig, and A. Plückthun, submitted for publication. It was found that a state with considerable structure within the domains is formed on the millisecond time scale, regardless of the peptide conformation at the proline residues. However, a complication is introduced by the slow isomerization of the peptide bond preceding Pro-L95, which must becis for the native VH/VL interface to form. As a consequence, the final correct docking of VHand VL (and thus, the subsequent stabilization of the structure) is limited in rate by this prolinecis,trans-isomerization (10.Freund C. Gehrig P. Baici A. Holak T.A. Plückthun A. Fold. Des. 1997; 3: 39-49Abstract Full Text Full Text PDF Scopus (15) Google Scholar, 12.Jäger M. Plückthun A. FEBS Lett. 1997; 418: 106-110Crossref PubMed Scopus (39) Google Scholar, 13.Jäger M. Plückthun A. J. Mol. Biol. 1999; 285: 2005-2019Crossref PubMed Scopus (43) Google Scholar). In the scFv format, presumably through the increased effective domain concentration mediated by the interdomain peptide linker, the domains have been shown to associate prematurely instead of folding rapidly and independently into native-like structures,3 thus forming off-pathway intermediates that are potentially prone to aggregation. We have chosen two different model systems to investigate the effect of FkpA on scFv fragment refolding in vitro. The disulfide-free variant (15.Wörn A. Plückthun A. FEBS Lett. 1998; 427: 357-361Crossref PubMed Scopus (112) Google Scholar) of the hu4D5-8 scFv fragment (termed 4D5− −) to indicate the absence of both disulfide bonds (16.Carter P. Kelley R.F. Rodrigues M.L. Snedecor B. Covarrubias M. Velligan M.D. Wong W.L. Rowland A.M. Kotts C.E. Carver M.E. Bio/Technology. 1992; 10: 163-167Crossref PubMed Scopus (320) Google Scholar), an antibody scFv fragment binding the extracellular domain of p185HER2, was chosen for its low stability and reduced reversibility of refolding. With refolding yields of 75% at 10 °C under optimal conditions, this scFv fragment leaves room for a possible improvement by FkpA (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar). The anti-GCN4 scFv fragment, which binds the transcriptional activator Gcn4p (17.Hanes J. Jermutus L. Weber-Bornhauser S. Bosshard H.R. Plückthun A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 14130-14135Crossref PubMed Scopus (277) Google Scholar), was chosen because it has a VL domain of the λ type and thus nocis-prolines in its native structure.2 Since, however, the wild-type anti-GCN4 fragment displays excellent folding properties, we used a destabilized variant for this study containing the H-R66K mutation, where H refers to the heavy chain (18.Wörn A. Auf der Maur A. Escher D. Honegger A. Barberis A. Plückthun A. J. Biol. Chem. 2000; 275: 2795-2803Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar). Arg at the heavy chain position H66 was shown previously to result in higher protein stability than Lys (19.Proba K. Wörn A. Honegger A. Plückthun A. J. Mol. Biol. 1998; 275: 245-253Crossref PubMed Scopus (222) Google Scholar, 20.Wörn A. Plückthun A. Biochemistry. 1999; 38: 8739-8750Crossref PubMed Scopus (130) Google Scholar). The obtained mutant is indeed considerably less stable (18.Wörn A. Auf der Maur A. Escher D. Honegger A. Barberis A. Plückthun A. J. Biol. Chem. 2000; 275: 2795-2803Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar) and shows reduced folding reversibility due to aggregation at elevated temperatures. The experiments with the anti-GCN4 (H-R66K) scFv fragment could therefore be carried out at physiologically relevant temperatures (37 °C) under conditions allowing for the observation of an effect of FkpA. In this study, we have attempted a thorough analysis of the biochemical properties of FkpA in vitro and its effect on scFv fragments. We investigated whether FkpA can accelerate the rate of refolding or increase the refolding yield of antibody fragmentsin vitro and whether these two effects are connected. Particularly, we also determined the effect of FkpA on an scFv fragment with a λ domain, devoid of cis-prolines. We measured the rate of acceleration in folding and the amount of soluble monomeric as well as functional scFv protein and also tested the dependence on the time of PPIase addition. We came to the conclusion that the effect on increasing the yield is independent of the PPIase activity of the enzyme. Protein expression and purification of the disulfide-free hu4D5-8 scFv fragment (abbreviated 4D5− −) and the destabilized version of the anti-GCN4 scFv fragment carrying the heavy chain mutation H-R66K were carried out essentially as described (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar, 15.Wörn A. Plückthun A. FEBS Lett. 1998; 427: 357-361Crossref PubMed Scopus (112) Google Scholar, 18.Wörn A. Auf der Maur A. Escher D. Honegger A. Barberis A. Plückthun A. J. Biol. Chem. 2000; 275: 2795-2803Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar, 17.Hanes J. Jermutus L. Weber-Bornhauser S. Bosshard H.R. Plückthun A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 14130-14135Crossref PubMed Scopus (277) Google Scholar). FkpA was expressed at 37 °C in E. coli JM83, transformed with the phagemid pHB602 or pHB602-His, containing FkpA (without and with the histidine tag) under control of its own promoter. Both phagemids also carried the scFv fragment 4-4-20 under control of the lac promoter. TheE. coli cells were harvested after 5 h of induction of the scFv fragment with isopropyl-β-d-thiogalactopyranoside. For purification of the His-tagged version, the soluble cell extract was purified by immobilized metal ion affinity chromatography at pH 7, followed by S/H cation-exchange chromatography at pH 6. FkpA without a His tag was purified by Q-Sepharose anion exchange chromatography at pH 8, followed by S/H cation-exchange chromatography at pH 6 and gel filtration at pH 7. All ELISA experiments were performed exclusively with FkpA without a His tag, to avoid its interference with the detection system, while the activity was determined for both variants, as were the 4D5− − fluorescence kinetics of folding, whereas the gel filtration experiments were mostly carried out with the His-tagged version. Fluorescence measurements were performed with a PTI Alpha Scan spectrofluorometer (Photon Technologies, Inc.) at 20 °C using excitation wavelengths of 295 and 280 nm as indicated and an emission wavelength of 350 nm. The buffer was 50 mm Tris, pH 7.0, and 50 mm NaCl (filtered and degassed). The denaturation curve was prepared by incubating the native protein at the respective GdmCl concentrations at 20 °C overnight. Intensity values have been corrected for the fluorescence of the buffer, and GdmCl concentrations were determined by the refractive index. Wild-type RNase T1 (Sigma) was unfolded by incubating the protein in 5.6 m GdmCl, pH 7, at 10 °C overnight. Refolding was initiated by 80-fold dilution to a final concentration of 0.2 μm in 50 mm Tris, pH 7, 50 mm NaCl, and the indicated concentrations of FkpA at 10 °C. RCM-RNase T1 (S54G/P55N) is a variant in which both disulfide bonds have been reduced and S-carboxymethylated. Its refolding has been shown to be a monoexponential process, but the protein requires 2 m NaCl for stability (21.Dolinski K. Scholz C. Muir R.S. Rospert S. Schmid F.X. Cardenas M.E. Heitman J. Mol. Biol. Cell. 1997; 8: 2267-2280Crossref PubMed Scopus (30) Google Scholar, 22.Scholz C. Schindler T. Dolinski K. Heitman J. Schmid F.X. FEBS Lett. 1997; 414: 69-73Crossref PubMed Scopus (29) Google Scholar). The wild-type protein, which has two slow isomerization-limited refolding phases, does not require 2 m NaCl for stability (23.Mücke M. Schmid F.X. J. Mol. Biol. 1994; 239: 713-725Crossref PubMed Scopus (53) Google Scholar), and the measurement could therefore be carried out in the standard buffer. The folding reaction was followed with a PTI Alpha Scan spectrofluorometer at 323 nm after excitation at 295 nm. Kinetic traces were evaluated with double exponential functions using Kaleidagraph software (Synergy Software, Reading, United Kingdom), and the rates of the faster phase were plotted to determinek cat/K m. Fluorescence measurements were performed with a PTI Alpha Scan spectrofluorometer at 10 °C using excitation and emission wavelengths of 295 and 326 nm, respectively. The buffer was 50 mm Tris, pH 7.0, and 50 mm NaCl (filtered and degassed). Final protein concentrations were 0.4 μm for both proteins, calculated for a dimer of FkpA. Refolding was initiated by rapid 20-fold dilution of the unfolded 4D5− − scFv fragment in buffer alone or buffer containing FkpA. Kinetic traces were corrected for the fluorescence of FkpA and evaluated using Kaleidagraph software with single exponential functions. Analytical gel filtrations were carried out on a SMART system with a Superose 12 column in 50 mm Tris, pH 7.0, and 150 mm NaCl with 0.005% Tween 20. Proteins were injected in a volume of 50 μl at the concentrations indicated in the figure legends. The concentrations given are theoretical concentrations with 100% refolding yield. The samples were passed through a filter with 0.22-μm pore size prior to injection to remove large aggregates and to prevent clogging of the column. The column was calibrated with β-amylase (200 kDa), aldolase (158 kDa), bovine serum albumin (66 kDa), carbonic anhydrase (29 kDa), and cytochrome c (12.4 kDa) as molecular mass standards. Refolding experiments prior to gel filtration were initiated by rapid 20-fold dilution of the unfolded scFv fragments at 10 °C (4D5− − scFv ) or 37 °C (anti-GCN4 scFv) in buffer alone (50 mm Tris, pH 7.0, and 50 mm NaCl) or buffer containing FkpA (or BSA and PpiA as controls) in equimolar amounts, calculated for a dimer for FkpA. The samples were injected after incubation overnight in the case of the 4D5− −scFv fragment and after 1–2 h in the case of the anti-GCN4 scFv fragment. ELISA plates (Nunc) were coated at 4 °C overnight with 100 μl of Tris-buffered saline containing either 1.5 μg/ml HER2 extracellular domain (for the 4D5− − scFv fragment ) or BSA-GCN4 conjugate (for the anti-GCN4 scFv fragment) at a dilution of 1:5000 (17.Hanes J. Jermutus L. Weber-Bornhauser S. Bosshard H.R. Plückthun A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 14130-14135Crossref PubMed Scopus (277) Google Scholar). Blocking was in Tris-buffered saline containing 0.005% Tween 20 with 5% milk at room temperature for 1 h. Binding took place in the presence of 3% milk at room temperature for 45 min. Detection was carried out by incubation for 45 min at room temperature with either mouse anti-His tag IgG (anti-GCN4 scFv) or mouse anti-Myc tag IgG (4D5− − scFv) and subsequent incubation with a goat anti-mouse IgG-peroxidase conjugate for 45 min. After addition of the substrate (BM Blue POD substrate, soluble; Roche Molecular Biochemicals), the reaction was stopped with 0.1 m HCl, and the plates were measured at 405 nm. The results are averages of at least duplicate measurements per plate that have been reproduced at least twice in independent experiments. In all ELISA experiments, wild-type FkpA without any tags was used. All native protein was centrifuged at 16,000 × g for 20 min at 4 °C before incubation or denaturation. Proteins were denatured by incubation in 3 m GdmCl (4D5− −scFv) or 4 m GdmCl (anti-GCN4 scFv), pH 7, at 4 °C overnight. Refolding reactions for ELISA measurement were initiated by 40–60-fold dilution of the unfolded protein to a final concentration of 0.2 μm in 50 mm Tris, pH 7, and 50 mm NaCl at the indicated temperatures and incubated for 5 h or overnight (4D5− − scFv) and for 1 h (anti-GCN4 scFv). Before addition to the ELISA wells, samples were centrifuged at 16,000 × g for 20 min at 4 °C to remove aggregates and diluted to the concentrations indicated at 4 °C. Probes with native protein were treated identically. For the controls, FkpA was added after the final centrifugation, before dilution and addition to the ELISA wells. Inhibition was achieved with a 10-fold molar excess of the antigen over the scFv fragment present in the respective dilution. FkpA could be obtained in a soluble and functional form in high amounts after expression under the control of its own promoter, as described under "Materials and Methods." A molecular mass of 26,225.4 Da (theoretical mass of 26,223.6 Da) was confirmed by mass spectrometry. In analytical gel filtration, however, FkpA eluted slightly after BSA (66 kDa), more consistent with being a dimer, but not rigorously excluding a trimer (data not shown). A monomer peak was never observed under any condition tested (1–10 μm), nor was the peak shifted to higher values. Native FkpA therefore seems to form dimers, and the monomeric form does not appear to be populated under the conditions used in gel filtration experiments. FkpA contains only one tryptophan residue in its sequence, which is highly quenched in the native state and becomes solvent-exposed upon unfolding of the protein. Upon excitation at 295 nm, the native protein therefore exhibited very low fluorescence, which increased dramatically upon denaturation, providing a convenient means to follow unfolding transitions (Fig. 1 A). The unfolding transition (Fig. 1 B) revealed a rather low thermodynamic stability of FkpA with a midpoint of the equilibrium transition below 1 m GdmCl at 20 °C. No stabilizing effect of increasing protein concentrations could be observed in a 5-fold concentration range (0.4–2 μm), suggesting that the limiting factor is the stability of the monomer, whereas any potential extrinsic contribution of a dimer interface to stability must be small. As it is currently unclear whether the transition is two-state or which molecular changes occur in the sloped pre-transition base line, we do not report ΔG values. FkpA has long been suggested to possess PPIase activity (2.Horne S.M. Young K.D. Arch. Microbiol. 1995; 163: 357-365Crossref PubMed Scopus (85) Google Scholar) based on its high similarity to the eukaryotic FK506-binding proteins. Itsk cat/K m was estimated by a protease-coupled assay using chymotrypsin and a peptide substrate to be 9 × 104m−1 s−1(1.Missiakas D. Betton J.-M. Raina S. Mol. Microbiol. 1996; 21: 871-884Crossref PubMed Scopus (295) Google Scholar). Since we noted, however, that FkpA is highly susceptible to very rapid digestion by the protease at the concentrations used in this assay (data not shown), FkpA activity was estimated directly from a protein folding assay. In this assay (24.Stoller G. Rücknagel K.P. Nierhaus K.H. Schmid F.X. Fischer G. Rahfeld J.-U. EMBO J. 1995; 14: 4939-4948Crossref PubMed Scopus (229) Google Scholar), the enhancement of the proline-limited refolding rate of RNase T1 is observed as a function of enzyme concentration. This assay has been shown in several instances to give a more reliable estimate of the enzymatic activity for those PPIases that are prone to rapid digestion in the standard assay (24.Stoller G. Rücknagel K.P. Nierhaus K.H. Schmid F.X. Fischer G. Rahfeld J.-U. EMBO J. 1995; 14: 4939-4948Crossref PubMed Scopus (229) Google Scholar,21.Dolinski K. Scholz C. Muir R.S. Rospert S. Schmid F.X. Cardenas M.E. Heitman J. Mol. Biol. Cell. 1997; 8: 2267-2280Crossref PubMed Scopus (30) Google Scholar). In the case of FkpA, the faster of the two slow RNase T1 refolding phases was 13-fold accelerated (Fig. 2) when 10 nm FkpA was present, comparable to the 22-fold acceleration observed in the presence of 50 nm Cpr3 (22.Scholz C. Schindler T. Dolinski K. Heitman J. Schmid F.X. FEBS Lett. 1997; 414: 69-73Crossref PubMed Scopus (29) Google Scholar) in a comparable assay involving RCM-RNase T1 (see "Materials and Methods"). The k cat/K m was estimated from the slope of the plot in the inset of Fig. 2to be ∼4 × 106m−1s−1 and thus ∼2 orders of magnitude higher than the value estimated from the protease-coupled colorimetric assay. FkpA is thus a very efficient isomerase, comparable to other FKBPs (FKBP12k cat/K m = 8 × 105m−1 s−1 (21.Dolinski K. Scholz C. Muir R.S. Rospert S. Schmid F.X. Cardenas M.E. Heitman J. Mol. Biol. Cell. 1997; 8: 2267-2280Crossref PubMed Scopus (30) Google Scholar) and trigger factor k cat/K m = 7.4 × 105m−1s−1 (24.Stoller G. Rücknagel K.P. Nierhaus K.H. Schmid F.X. Fischer G. Rahfeld J.-U. EMBO J. 1995; 14: 4939-4948Crossref PubMed Scopus (229) Google Scholar)) or the periplasmic E. colicyclophilin PpiA (k cat/K m = 6 × 106m−1 s−1(25.Compton L.A. Davis J.M. Macdonald J.R. Bachinger H.P. Eur. J. Biochem. 1992; 206: 927-934Crossref PubMed Scopus (85) Google Scholar)). Due to the very dramatic effect of coexpression of FkpA on the yield of soluble and active scFv fragments in vivo (7.Bothmann H. Plückthun A. J. Biol. Chem. 2000; 275: 17100-17105Abstract Full Text Full Text PDF PubMed Scopus (149) Google Scholar), we decided to carry out in vitro experiments to further characterize the mode of action of FkpA, focusing on the relationship between its positive effect on in vivo folding on the one hand and its peptidylprolyl cis,trans-isomerase activity on the other. Two model systems were used for these investigations: first, the disulfide-free variant of the scFv fragment hu4D5-8 (abbreviated 4D5− − to indicate the missing disulfides), and second, a destabilized variant of the anti-GCN4 scFv fragment. The 4D5− − scFv fragment was chosen for the reasons that the rate-limiting steps on its folding pathway are well characterized and known to be proline-limited (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar)3 and that it shows a relatively high tendency to aggregate upon refolding, with yields of 75% native protein at the most. Due to the low intrinsic fluorescence of FkpA in its native state compared with the scFv protein (Fig.3 A), FkpA does not disturb the fluorescence measurements, even if present in stoichiometric amounts. As has been described before (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar), refolding of the 4D5− − scFv fragment is characterized by an initial burst phase. This burst phase is apparent in Fig. 3 B from the amplitude reached already in the dead time of manual mixing, followed by a much slower conversion to the native state.3The initial phase cannot be resolved by stopped-flow mixing techniques and has been shown to involve the formation of an early structured intermediate, indicated by the fast protection of exchangeable protons. As has been shown by short-term unfolding experiments (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar),3 the rate-limiting phase observed by fluorescence spectroscopy (Fig. 3 B) is a slow proline isomerization-limited folding reaction (k = 0.001 s−1). Even in the presence of equimolar amounts of FkpA (0.4 μm, calculated for a dimer) in the refolding buffer, this rate-limiting folding step was not accelerated (k = 0.001 s−1). FkpA, although an efficient isomerase, is thus not able to efficiently catalyze the limiting cis,trans-isomerization in the refolding of this scFv fragment. This probably reflects the lack of accessibility of the relevant proline in the scFv format, as short-term denaturation experiments clearly do lead to a rate acceleration (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar). For comparison, at 10 nm FkpA, the refolding of RNase T1 was already accelerated 13-fold (Fig. 2 A). An acceleration of the isomerization of Pro-L8, which is not overall limiting in refolding, can, however, not be excluded. Despite this apparent lack of a catalytic effect on isomerization, aggregation was reduced, and the yield of native protein increased substantially in the presence of FkpA as judged by fluorescence intensity. Some light scattering may contribute to the total intensity of fluorescence measurements in the absence of FkpA, as inferred from the decline of the signal at long times and the small lag phase. The rate should therefore be treated only as an estimate, even though it is very similar to the one obtained in numerous other measurements (14.Ramm K. Gehrig P. Plückthun A. J. Mol. Biol. 1999; 290: 535-546Crossref PubMed Scopus (50) Google Scholar) under slightly different conditions, where aggregation is less severe (data not shown). To further investigate the positive effect of FkpA on the refolding yield of the 4D5− − scFv fragment, we carried out gel filtration experiments to determine the amount of soluble monomeric scFv fragment obtained after refolding in the absence and presence of FkpA. As mentioned before, FkpA eluted from the column at an elution volume corresponding roughly to that of BSA under all conditions tested (Fig.4 A), whereas the native 4D5− − scFv fragment eluted at the volume expected for a monomer, thus allowing clear separation and identification of the two peaks. No formation of stable complexes was observed, neither upon co-incubation nor after refolding of the scFv fragment in the presence of FkpA. This indicates that any interaction must be low affinity and possibly transient in nature. If refolding of the 4D5− − scFv fragment was carried out in the absence of FkpA at the concentrations necessary for gel filtration experiments (2–10 μm), hardly any monomer could be detected; and instead, a broad peak eluted at the exclusion volume of the column, most likely consisting of higher molecular mass aggregates of the scFv protein, which were not removed by the filtration step prior to gel chromatography. In contrast, if FkpA was present in the refolding buffer in stoichiometric amounts (2 μm dimer) from the onset of the reaction, a significant amount of soluble monomeric protein was obtained, whereas the formation of high molecular mass aggregates was not observed (Fig. 4 A,trace 4). FkpA thus has a highly beneficial effect on the yield of soluble monomeric protein. Controls with BSA and PpiA at the same concentrations had no effect on the refolding yield and thus demonstrate that the observed increase in the presenc
Referência(s)