Exogenous stresses including heat shock and GdnHCl 131415 often results in the misfolding and aggregation of proteins within the cell. Therefore, it would seem reasonable to presume that intracellular proteins which exist in their native soluble forms even under stress conditions have an intrinsic ability to remain efficiently folded compared to proteins that unfold or aggregate under the same stresses. GdnHCl stress facilitates the release of periplasmic proteins through increased membrane permeability 161718 and hence could be regarded as an osmolyte. We investigated changes of the proteome profile of E. coli when the stress reagent, guanidine hydrochloride, was added to growing bacterial cultures, by applying 2-dimensional gel electrophoresis (2-DE). Compared to non-stress condition, the total number of intracellular soluble proteins was reduced to 731 from 788, which indicates that many proteins aggregated under the stress condition induced by GdnHCl. What was surprising is that the RpoS expression level (i.e. protein spot intensity estimated through analysis of 2-DE gel images, Fig. 1. and Table 1) increased 6-fold compared to the non-stress situation. In addition to RpoS, we found several other proteins (Table 1), the expression level of which more than 1.9-fold increased in response to the stressor GdnHCl, including HSP60 chaperonin (57 kD), chaperone HtpG (71 kD), transcription elongation protein NusA (55 kD), DNA gyrase subunit A (97 kD), formate acetyltransferase 1 (85 kD), succinate dehydrogenase flavoprotein subunit (64 kD), etc. Since the synthesis yield of target protein can be significantly reduced, the molecular mass of proteins to be used as fusion expression partner should not be too high. That is, although the large amount of fusion protein is synthesized, the actual amount of the target fusion-free protein could be very low if the fusion partner is too big. RpoS (38 kD) is a relatively small protein among the stress-responsive proteins we found and was highly effective in enhancing the solubility of target proteins.

RpoS, stress responsive RNA polymerase sigma factor is a well-known universal stress regulator controlling many proteins under various stress conditions, e.g. the onset of the stationary phase 19, and carbon starvation 20. RpoS was previously reported to be induced by osmotic stress 16 and heat shock 21. We also observed that the RpoS expression level increased 3- to 5-fold in response to heat shock (data not shown). Muffler et al. 21 reported that the duration of stability of RpoS in response to heat shock stress is maintained by the direct- or indirect-binding of DnaK. Bound DnaK appears to assist the effective folding of RpoS and also protect RpoS against the action of ClpP, a protease that degrades RpoS 21. From this point of view, RpoS may serve as a solubility enhancer in E. coli cytoplasm, when used as a fusion expression partner upon the expression of aggregation-prone heterologous proteins.

We used RpoS as an N-terminus fusion expression partner for the synthesis of numerous heterologous proteins [mp-INS, EGF, ppGRN, hIL-2, AID, GAD_448–585, CUT, hFTN-L, G-CSF, and NACHT (Fig. 2 and Table 2 for expression system construction)]. Stenström et al. 22 reported that the codon following an AUG start triplet (+2 codon) significantly affects gene expression in E. coli, and the second codon starting with A is most advantageous to achieve an enhanced expression level. The second codon of RpoS is AGT (encoding Ser) that seems to be favorable second codon based on the report of Stenström et al. 22. As shown in Figure 3B and Table 3, all heterologous proteins expressed directly without RpoS fusion formed insoluble inclusion bodies resulting in nearly negligible solubility. Compared to these results, it seems surprising that when the same heterologous proteins were expressed with the fusion of RpoS, the solubility of these foreign proteins dramatically increased (Fig. 3A and Table 3), thereby indicating that E. coli RpoS is a highly effective solubility enhancer for aggregation-prone heterologous proteins. Table 3 compares the effect of RpoS- and GST fusion on the solubility enhancement for heterologous proteins. In the fusion expression of CUT, GAD_448–585, and mpINS, the effect of RpoS fusion was much higher, whereas GST fusion was significantly more effective in the expression of AID, NACHT, and hFTN-L. This result indicates that there are no universal helpers to solve all protein solubility issues. Table 3 also shows that the results of direct- and fusion expression of heterologous proteins are highly reproducible.

Moreover, the plasmid vector for the expression of polyhistidine-tagged fusion mutant of G-CSF [(His)₆::RpoS::(D₄K)::G-CSF] was constructed (Fig. 2) to purify fusion-free G-CSF. After (His)₆::RpoS::(D₄K)::G-CSF was bound onto the ProBond resin (Ni⁺²) column, the enterokinase proteolysis was carried out in a batch mode, and subsequently the digested product was collected and centrifuged. SDS-PAGE and Western blot analyses of the supernatant show that the recombinant G-CSF was easily released from the Rpos-fusion protein and was present in the form of soluble protein in the supernatant (Fig. 4).

Cutinase has been used as a lipolytic enzyme in the composition of laundry and dishwashing detergents to more efficiently remove immobilized fats 2324. In addition, the oleochemistry industries 25, and pollutant degradation 2627 represent other potential uses of cutinase. In recent years, the esterification and transesterification properties of cutinase have been intensively exploited and could be applied usefully in other chemical synthesis processes 28. Because of these extensive potential applications, we were particularly interested in the production of a bioactive recombinant cutinase. Thus, we cloned the cutinase gene from the genome of Pseudomonas putida and expressed it in E. coli using the fusion of RpoS. Cutinase is known for its hydrolytic activity for a variety of esters ranging from soluble p-nitrophenyl esters to insoluble long-chain triglycerides. The hydrolytic activity of cutinase, especially on p-nitrophenyl esters of fatty acids, is extremely sensitive to fatty acid chain length. Previously, Lin et al. 29 reported that microbial cutinase lacked a large hydrophobic surface around the active site, in contrast to other lipases and esterases. This structural characteristic of cutinase may be strongly related to high substrate-specificity, i.e. the extremely low activity on p-nitrophenyl ester of a long chain fatty acid such as p-nitrophenyl palmitate (PNP).

We assayed the enzymatic activity of our fusion mutant of cutinase (RpoS::CUT) and demonstrated the same selective bioactivity as native cutinase to degrade p-nitrophenyl butyrate (PNB) but not to degrade PNP (Fig. 5A and 5B). We also purified RpoS::CUT-His₆ (Fig. 6A) through Ni²⁺ affinity chromatography and analyzed the purified RpoS::CUT-His₆ using reversed phase HPLC (Fig. 6B). As shown in Figure 6B, RpoS::CUT-His₆ was analyzed as a single peak, which seems to indicate that the crafted mutant molecules of cutinase have uniform and correctly folded conformation due probably to the help of fusion partner, E. coli RpoS.

E. coli strain BL21(DE3) (F^-ompT hsdS_B(rB^- mB^-)) was selected under both non-stress and GdnHCl-stress conditions for 2-dimensional gel electrophoresis analysis. Through PCR amplification using appropriate primers (Table 2), the genes encoding mp-INS, EGF, ppGRN, hIL-2 30, AID, GAD_448–585 31, CUT, hFTN-L 32, G-CSF, and NACHT were cloned using previously cloned heterologous genes, except for CUT gene that was cloned from chromosomal DNA of Pseudomonas putida (ATCC 53552). The gene clones of mp-INS, ppGRN, AID, G-CSF, and NACHT were kindly donated by other researchers, as acknowledged (see Acknowledgements). The EGF gene was cloned using pCMV6-XL vector (OriGene Technologies, USA). Each of the recombinant genes above and various fusion/hybrid genes [rpoS(or GST gene)::(each heterologous gene)] were inserted into the NdeI-HindIII site of the same plasmid pT7-7 to construct the expression vectors. All the heterologous genes above were fused directly to the RpoS gene (cloned from the chromosomal DNA of E. coli BL21(DE3)) or GST gene [cloned from the GST-fusion vector pET42a(+) (Novagen, USA)] without any linker sequence. Therefore, each expression vector has no enzymatic cleavage site between rpoS (or GST gene) and heterologous gene, except for the case of G-CSF. For the purification of fusion-free recombinant G-CSF, the D₄K sequence for enterokinase digestion was inserted between RpoS and G-CSF genes to synthesize the polyhistidine-tagged fusion mutant of G-CSF, i.e. (His)₆::RpoS::(D₄K)::G-CSF. For the purification of RpoS::CUT, hexahistidine (His₆) was added to the C-terminus of CUT by PCR (Fig. 2). For this, the sequence coding for N-RpoS::CUT-His₆-C was inserted into the NdeI-HindIII site of plasmid pT7-7.

After complete DNA sequencing of all gel-purified hybrid plasmids, the E. coli strain BL21(DE3) (F^- ompT hsdS_B(rB^- mB^-)) was transformed with the hybrid plasmids, and ampicillin-resistant transformants were subsequently selected using LB-agar plates supplemented with ampicillin (100 mg/l).

For shake flask experiments, 250 ml Erlenmeyer flasks containing 50 ml LB media and ampicillin at 100 mg/l of culture (37°C and 150 rpm) was used. When the culture turbidity (OD₆₀₀) reached 0.5, gene expression was induced with the addition of IPTG (1 mM), and after a further 4 h of cultivation the all recombinant cells (80 mg wet cell mass) were harvested by centrifugation (13,000 rpm (MICRO17TR, Hanil Science Industrial, Korea) × 5 min) and cell pellets resuspended in 5 ml lysis buffer (10 mM Tris-HCl, pH 7.5, 10 mM EDTA). Cell disruption was achieved using a Branson Sonifier (Branson Ultrasonics Corp., Danbury, CT). The cell-free supernatant and insoluble protein aggregates were separated at 13,000 rpm (MICRO17TR, Hanil Science Industrial, Korea) for 10 min. The isolated inclusion bodies, if any, were washed twice with 1 % Triton X-100. Cell-free supernatants and the washed inclusion bodies were subjected to polyacrylamide (14%) gel electrophoresis (PAGE) analysis. Coomassie-stained protein bands were ultimately scanned, and the intensity of each recombinant protein band was estimated using densitometry (Duoscan T1200, Bio-Rad, Hercules, CA). The solubility of recombinant proteins was determined by analyzing the fraction of the soluble recombinant protein compared to the synthesized total (soluble + insoluble) recombinant protein. Average and standard deviation values of the solubility were calculated based on the results of repeated triplicate experiments.

The purification of recombinant G-SCF was accomplished using metal affinity chromatography. That is, polyhistidine-tagged fusion mutant of G-CSF [(His)₆::RpoS::(D₄K)::G-CSF] (Fig. 4) were loaded onto ProBond resin (Ni⁺²) column. Prior to sample loading, the resin was washed twice with 10 column volumes of binding buffer (50 mM potassium phosphate, 300 mM KCl, 20 mM imidazole, pH 7.0). Binding buffer contains 20 mM imidazole to minimize non-specific binding of untagged protein contaminants, and binding was carried out in a batch mode at 4°C. Afterwards the resin was washed twice with 5–8 ml Tris-HCl (10 mM Tris, pH 8.0) prior to enterokinase digestion step. The enterokinase digestion was carried out in a batch mode at 4°C for 10 h using 5-unit enterokinase (Invitrogen, CA, USA). Then, the proteolytic product was collected and centrifuged [3,000 rpm (MICRO17TR, Hanil Science Industrial, Korea) for 10 min], and the supernatant fraction was subjected to being analyzed by SDS-PAGE and western blotting.

For the purification of RpoS::CUT-His₆, the soluble fraction was separated after cell disruption by centrifugation (13,000 rpm (MICRO17TR, Hanil Science Industrial, Korea)) for 30 min. The cell-free supernatant containing RpoS::CUT-His₆ was loaded onto the ProBond resin (Ni²⁺) column (Invitrogen) for affinity purification. Before the sample loading, the resin was washed twice with ten column volumes of binding buffer (pH 8.0, 50 mM sodium phosphate, 300 mM NaCl, 10 mM imidazole). Binding was carried out in a batch mode at 4°C. Afterwards, the resin was washed twice with 8 ml washing buffer (pH 8.0, 50 mM sodium phosphate, 300 mM NaCl, 50 mM imidazole) and eluted with elution buffer (pH 8.0, 50 mM sodium phosphate, 300 mM NaCl, 250 mM imidazole). The elution buffer in the eluted solution containing the purified RpoS::CUT-His₆ was changed with PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4) using Amicon Ultra-4 centrifugal filter (Millipore, Ireland). Analysis of the purified RpoS::CUT-His₆ was performed with high-performance liquid chromatography (HPLC) (LC-20A Prominence, Shimadzu Co. Ltd., Japan). A Shim-pack CLC-NH2 column (60 × 150 mm, Shimadzu Co. Ltd., Japan) was equilibrated with 10 % acetonitrile containing 0.1 % trifluoroacetic acid at a flow rate of 1 ml/min. The sample was loaded onto the column and eluted with 75 % acetonitrile. The elution profile was monitored at 215 nm.

Flask culture conditions were identical to those for recombinant gene expressions. Cells were grown at 37°C and then 100 mM GdnHCl was added for the GdnHCl-induced proteome response when culture turbidity (OD₆₀₀) reached 0.5 in the LB media. After further 3 h cultivation, the cells were harvested by centrifugation at 6,000 rpm (MEGA17R, Hanil Science Industrial, Korea)for 15 min (4°C) and then washed twice with 40 mM Tris buffer (pH 8.0). Cell pellets were resuspended in 500 μl of lysis buffer (8 M urea, 4% (w/v) CHAPS, 40 mM Tris, protease inhibitor cocktail; Roche Diagnostics GmbH, Mannheim, Germany) and disrupted by sonication. After sonication (Branson Sonifier 450, USA), the cell debris and the aggregated proteins were removed by centrifugation at 12,000 rpm (MICRO17TR, Hanil Science Industrial, Korea) for 60 min (4°C), and only soluble proteins were obtained. 2-dimensional gel electrophoresis was performed as described previously Kim et al. 33. Stained gels were scanned using a UMAX powerlook 1100 scanner and Image Master software v 4.01 (Amersham Biosciences, Uppsala, Sweden) was used for gel image analysis, including quantification of spot intensities performed on volume bases (i.e. values calculated from the integration of spot optical intensity over spot area).

Samples for MALDI-TOF MS analysis were prepared through the extraction from silver-stained protein spots according to the previous protocol 33. Enzymatic digestion was performed with 10 mg/ml of sequencing grade modified trypsin (Promega, WI, USA) in 25 nM ammonium bicarbonate (pH 8.0) for overnight at 37°C in a stationary incubator. The trypsin-digested protein spots was analyzed using a MALDI-TOF-MS system (Voyager DE-STR, PE Biosystem, Framingham, MA, U.S.A.) by the Korea Basic Science Institute (Seoul, Republic of Korea), and peptide mass fingerprinting for protein identification was performed using MS-Fit 34. Spectra were calibrated using a matrix and tryptic autodigestion ion peaks as internal standards. Peptide mass fingerprints were analyzed using the MS-Fit 34. The identification of a protein with respect to theoretical parameters (pI, molecular mass, etc.) was accepted if the peptide mass matched within a mass tolerance of 10 ppm.

A hydrolytic enzyme activity of the recombinant cutinase fusion mutant was assessed as described below. The hydrolysis reactions occurred in 96-well microplates at 37°C for 15 min where each well contained 200 μl enzyme/substrate solution comprised of 106.7 μl phosphate buffer (0.1 M, pH 8.0), 13.3 μl Triton X-100 solution (4 g/l), cutinase solution. The reaction was initiated by adding 66.7 μl of substrate reagent solution to each well in the 96-well microplate. Absorbance changes (ΔA_415nm per min) were measured using a Bio-Rad microplate reader (Tecan, Austria), and solution A with no enzyme solution was used as blank. From the absorbance changes measured at each column, an average absorbance for a specific reaction condition could then be calculated.