Expression and Characterization of a Single-Chain Variable Fragment

J. Microbiol. Biotechnol. (2017), 27(5), 965–974 https://doi.org/10.4014/jmb.1702.02007 Research Article Review jmb

Expression and Characterization of a Single-Chain Variable Fragment against Human LOX-1 in Escherichia coli and Brevibacillus choshinensis Wei Hu1†, Jun-Yan Xiang1†, Ping Kong1, Ling Liu1, Qiuhong Xie1,2,3*, and Hongyu Xiang1,2,3*

1School of Life Science, 2Key Laboratory for Molecular Enzymology and Engineering, Ministry of Education, 3National Engineering Laboratory for AIDS Vaccine, School of Life Science, Jilin University, Changchun, Jilin 130012, P.R. China

Received: February 6, 2017 Revised: February 28, 2017 The single-chain variable fragment (scFv) against lectin-like oxidized low-density lipoprotein Accepted: March 9, 2017 receptor-1 (LOX-1) is a promising molecule for its potential use in the diagnosis and immunotherapy of atherosclerosis. Producing this scFv in several milligram amounts could be First published online the starting point for further engineering and application of the scFv. In this study, the March 9, 2017 abundant expression of the anti-LOX-1 scFv was attempted using Escherichia coli (E. coli) and *Corresponding authors Brevibacillus choshinensis (B. choshinensis). The scFv had limited soluble yield in E. coli, but it Q.X. Phone: +86-431-85153832; was efficiently secreted by B. choshinensis. The optimized fermentation was determined using Fax: +86-431-85153832; the Plackett-Burman screening design and response surface methodology, under which the E-mail: [email protected] yield reached up to 1.5 g/l in a 5-L fermentor. Moreover, the properties of the scFvs obtained H.X. Phone: +86-431-85153832; from the two expression systems were different. The antigen affinity, transition temperature, Fax: +86-431-85153832; and particle diameter size were 1.01E-07 M, 55.2 ± 0.3ºC, and 9.388 nm for the scFv expressed E-mail: [email protected] by B. choshinensis, and 4.53E-07 M, 52.5 ± 0.3ºC, and 13.54 nm for the scFv expressed by E. coli. †These authors contributed This study established an efficient scale-up production methodology for the anti-LOX-1 scFv, equally to this work. which will boost its use in LOX-1-based therapy. pISSN 1017-7825, eISSN 1738-8872

Copyright© 2017 by Keywords: Lectin-like oxidized low-density lipoprotein receptor-1, single-chain variable The Korean Society for Microbiology fragment antibody, Brevibacillus choshinensis, Escherichia coli, production methodology and Biotechnology

Introduction (MAbs) [5]. Moreover, LOX-1 can be used for plaque imaging using MAbs and selective delivery of anti-atherosclerotic Lectin-like oxidized low-density lipoprotein receptor-1 agents [6]. (LOX-1), a multi-ligand scavenger receptor that was originally MAbs occupy a large portion of biopharmaceutical identified as the primary receptor for oxLDL, mediates a proteins [7]. There are mainly two disadvantages of using series of proatherogenic cellular responses implicated in the MAb molecules; one is the large molecular size, which the pathogenesis of atherosclerosis-related diseases [1, 2]. limits their penetration into diseased areas, and the other is These diseases continue to be a major cause of morbidity the binding of the crystallizable fragment (Fc) domain of and mortality in the world. LOX-1 basal expression is the MAbs to the cell surface Fc receptor, which limits their relatively low, and it is substantially up-regulated in circulation and mobility [8]. As a substitute for large intact multiple disease states [3]. LOX-1 expressed on the cell MAbs, scFvs have tremendous potential for use in surface can be proteolytically cleaved and released into the diagnosis and targeted therapy because of their specific circulation in a soluble form (sLOX-1). sLOX-1 has proved binding affinity to the antigen, their small size, superior to be a sensitive diagnostic biomarker of vascular pathology biodistribution, and blood clearance; additionally, they can [4]. In addition, recent evidence indicates that targeting be easily engineered or modified [9, 10]. Recently, an anti- LOX-1 is a reliable strategy for the therapy of vascular LOX-1 scFv has been developed for targeting LOX-1 [11], disease and that LOX-1-mediated proatherogenic effects which we believe has rich potential as a LOX-1 targeting can be inhibited by anti-LOX-1 monoclonal antibodies vector for immunodiagnostics, drug delivery, and potential

May 2017 ⎪ Vol. 27⎪ No. 5 966 Hu et al.

immunotherapy, that requires further research. However, extract, 0.5% beef extract, 0.001% FeSO4·7H2O, 0.001% MnSO4·4H2O, the minor yield of this scFv produced using the Free Style and 0.0001% ZnSO4·7H2O), and LB. 293 expression system limited its further engineering and application. Therefore, it is urgent to develop efficient Plasmid Construction methods for producing this scFv. The anti-LOX-1 scFv was designed on the basis of sequences of the variable heavy (V ) and light (V ) chains of the anti-human ScFv can be produced using a wide range of platforms, H L LOX-1 antibody [11], in the form of V −(Gly Ser) −V with a C- including production in the periplasm of prokaryotes [12]; H 4 3 L terminal peptide encoding the myc (myc) epitope and a in the endoplasmic reticulum of eukaryotes [13]; and in polyhistidine (6×His) metal-binding tag to facilitate the purification cell-free expression systems [14]. Of these expression and immunodetection of the gene product. The gene fragment systems, Escherichia coli is a widely used host that is was chemically synthesized and cloned into a pUC19 vector by suitable for expressing small non-glycosylated recombinant Beijing Genomic Institute (BGI, China), and the plasmid was antibody fragments, including scFv [15, 16]. In recent years, named pUC19-scFv. a promising host, Brevibacillus choshinensis, was developed Plasmid pUC19-scFv was used as the template for the PCR. To for antibody fragment production. Using B. choshinensis, we append the EcoRI and SalI restriction sites for cloning into the have successfully obtained secreted matrix metalloproteinase vectors pET-21a(+) and pET-23a(+), the primer sets 5’-CGG 26 [17], which expressed as inclusion bodies in E. coli hosts. AATTCGAAGTCAAACTGCTGGAATCTGG-3’ (forward) and 5’- In this study, the abundant expression of the anti-LOX-1 GCGTCGACTCATTAGTGGTGGTGATGATGGTGAGC-3’ (reverse) were used. For cloning into the vectors pET-20b(+) and pET- scFv was attempted using E. coli and B. choshinensis. The 27b(+), primer sets 5’-CGGAATTCCGAAGTCAAACTGCTGGAA optimization study was performed using Plackett-Burman TCTGG-3’ (forward) and the same reverse primer as mentioned screening design and the response surface methodology above were used. For cloning into the vector pNCMO2, primer (RSM). After that, the scFvs obtained from the two sets 5’-GCGTCGACGAAGTCAAACTGCTGGAATCTGG-3’ (forward) expression systems were purified using a two-step column and 5’-CGGAATTCTCATTAGTGGTGGTGATGATGGTGAGC-3’ chromatography purification progress. Finally, properties (reverse) were used. The resulting PCR products were digested of the purified anti-LOX-1 scFv proteins were studied. As with EcoRI and SalI, and then they were cloned into the vectors an ideal antibody for use in therapeutic purposes, it must digested using the corresponding restriction enzymes to form the have good pharmaceutical properties, such as antigen affinity expression plasmids. These expression plasmids were correspondingly and thermal stability. Results showed that B. choshinensis named pET21-scFv, pET23-scFv, pET20-scFv, pET27-scFv, and can act as an efficient host for the secreted expression of the pNC-scFv. anti-LOX-1 scFv in abundance. Expression of the Anti-LOX-1 scFv in E. coli Hosts E. coli strain BL21(DE3), Origami, Rosetta-gami 2(DE3), and Materials and Methods Rosetta-gami 2(DE3)pLysS were respectively transformed with the expression plasmids pET21-scFv, pET23-scFv, pET20-scFv, Bacterial Strains, Vectors, and Media and pET27-scFv, and then the transformants were cultured in LB The E. coli strains (Novagen, USA) BL21(DE3), Origami, Rosetta- medium containing the required antibiotic(s). The culture was gami 2(DE3), and Rosetta-gami 2(DE3)pLysS, and B. choshinensis shaken at 37ºC until the cell density (OD600) reached 0.4-0.6. Then, SP3 (Takara, Japan) were used for scFv expression. The vector the induction was conducted by adding 0.5 mM isopropyl-1-thio- pNCMO2, a B. choshinensis-E. coli shuttle vector obtained from β-D-galactopyranoside (IPTG). For optimization, cultivations were Takara, was used for the expression in B. choshinensis. The vectors performed using different IPTG concentrations (0.25, 0.5, 1, or (Novagen) pET-20b(+), pET-21a(+), pET-23a(+), and pET-27b(+) 2 mM) and temperatures (18ºC, 25ºC, 30ºC, or 37ºC). The cells were were used for expressing the scFv in E. coli. harvested by centrifugation, and the cell pellet was resuspended LB medium was used to culture E. coli. B. choshinensis strains in 50 mM sodium phosphate buffer (PBS, pH 7.4) containing 0.2 M used for the expression and secretion of anti-LOX-1 scFv was NaCl. For cell lysis, the cells were homogenized and disrupted by respectively grown in medium 5YC (3% glucose, 3% polypeptone, sonication. The lysates were centrifuged at 27,000 ×g for 30 min at 0.2% yeast extract, 0.01% MgCl ·6H O, 0.01% CaCl ·2H O, 0.001% 2 2 2 2 4ºC. FeSO4·7H2O, 0.001% MnSO4·4H2O, and 0.0001% ZnSO4·7H2O), B2 (0.5% glucose, 1% soytone, 2.5% yeast extract, 2.5% NaCl, and Secreted Expression of Anti-LOX-1 scFv in the B. choshinensis Host 0.1% K HPO ), 2SY (2% glucose, 4% soytone, 0.5% yeast extract, 2 4 B. choshinensis SP3 cells harboring the plasmid pNC-scFv were and 0.015% CaCl ·2H O), BTY (1% glucose, 1% polypeptone, 0.5% 2 2 cultured overnight at 37ºC in 2SY, BTY, LB, 5YC, B2, and/or TM yeast extract, 0.001% FeSO ·7H O, 0.001% MnSO ·4H O, and 0. 4 2 4 2 medium plates containing 10 mg/l neomycin (Sigma-Aldrich, 001% ZnSO ·7H O), TM (1% glucose, 1% polypeptone, 0.2% yeast 4 2 USA). Shaker-flask studies were carried out in 250 ml shaker

J. Microbiol. Biotechnol. Expression and Characterization of Anti-LOX-1 scFv 967

flasks containing 100 mL of corresponding medium at 30ºC or performed at a scan rate of 10 nm/min and at a response time of 4 37ºC. The secreted proteins from each culture supernatant were sec using the path-length of the cell (0.2 mm). The changes in the collected by centrifugation at 5,400 ×g for 20 min at 4ºC and then mean residue ellipticities at 218 nm upon increased temperatures analyzed by reducing SDS-PAGE. The gel was scanned, and the were monitored, and the transition point was estimated. target protein expression level was analyzed densitometrically using Quantity One software (Bio-Rad Laboratories, USA). Intrinsic Fluorescence Spectroscopy Analysis The intrinsic fluorescence of scFv proteins was measured using Optimization of Culture Medium a RF-5301 fluorescence spectrophotometer (Shimadzu, Japan) A Plackett-Burman screening project was developed to screen with a 1.0 cm quartz cuvette. The fluorescence emission spectrum the formulation variables that have an important effect on the of each antibody sample at the concentration of 0.1 g/l was anti-LOX-1 scFv yield. RSM was applied to optimize the key monitored. The emission spectra, excited at the wavelength of factors involved in anti-LOX-1 scFv production. Design-Expert 295 nm, in the range of 300-400 nm, were recorded. Scans were software (ver. 8.0.6; Stat-Ease, Inc., USA) was used for the performed in triplicates, and the mean values are presented. The experimental design and for making RSM plots. Finally, bioreactor parameter mean I320/365 on the temperature of each scFv protein cultivation was performed in a 5-L fermentor containing 2 L of sample was determined as described previously [20]. optimal medium. The fermentation temperature was maintained at 30ºC, and the medium pH was maintained at 7.2 by adding 2 M Dynamic Light Scattering (DLS)

H2SO4 and 5 M NaOH when required. The particle size of each purified scFv at 0.5 g/l in PBS containing 0.2 M NaCl (pH 7.4) was analyzed using DLS with a Purification of the Anti-LOX-1 scFv Proteins NANO ZS90 system (Melvin, UK). The experiments were performed The target protein, which was secreted in the medium or at room temperature. present in the supernatant of the centrifuged lysate, was ultra- filtered, concentrated, and loaded onto a 5 ml Ni-NTA affinity Results and Discussion column (GE Healthcare, USA), and the column was washed with washing buffer (PBS containing 0.2 M NaCl and 25 mM imidazole, Expression and Purification of the Anti-LOX-1 scFv in pH 7.4). The bound proteins were eluted by gradually increasing E. coli Strains the imidazole concentration to 0.5 M. The peak fractions were Plasmids pET21-scFv and pET23-scFv expressed an pooled and immediately applied to a Hiload 26/60 Superdex abundant amount of insoluble anti-LOX-1 scFv protein in 200 pg (GE Healthcare, USA). The purified scFv from E. coli was the four E. coli strains. Plasmids pET20-scFv and pET27- called scFv-EC, and scFv-BC was obtained from the B. choshinensis host. The protein samples of each purification step were analyzed scFv with a pelB signal sequence fused in front of the anti- by reducing SDS-PAGE, and the concentration was measured LOX-1 scFv expressed a small fraction of soluble anti-LOX- using the BCA Protein Assay Kit (Pierce, USA). 1 scFv in both the Rosetta-gami 2(DE3) and Rosetta-gami 2(DE3)pLysS strains (data not shown). Rosetta-gami Afﬁnity Analysis by Biolayer Interferometry (BLI) 2(DE3) containing pET20-scFv reached the highest soluble Experiments were performed using the purified scFv proteins yield (2.3 mg/l) under the optimized induction conditions on an Octet RED96 system (ForteBio, USA) at 30ºC. For the Octet with 0.1 mM IPTG at 18ºC for 16 h. After induction, cell analysis, streptavidin (SA) biosensors and a kinetic buffer (50 mM pellets from 4 L cultures were collected and underwent PBS, containing 0.2 M NaCl, pH 7.4) were used. One milligram of ultrasonic disruption. The supernatants were concentrated human LOX-1 protein (Sino Biological Inc., China) was biotinylated and subjected to Ni-NTA affinity chromatography and at a 1:1 ratio with EZLink Sulfo-NHS-Biotin (Pierce), so that it subsequent gel filtration chromatography for purification. would bind to the SA biosensor. The assays were measured in a At each step, the samples were analyzed by SDS-PAGE 96-well format according to a method from a previous study [18]. Kinetic responses were fitted to a 1-site binding model to obtain (Fig. 1A). Although these four E. coli hosts tested here had been reported to be efficient for heterologous protein the values for the kinetic association (kon) and dissociation (koff) expression [21], they do not seem to work well for rates and the equilibrium dissociation constant (KD). expressing soluble anti-LOX-1 scFv protein. Perhaps more Far-UV CD Spectral Analysis efficient expression of scFv proteins can be achieved using Far-UV CD spectral measurements were performed as described other E. coli host strains with the corresponding expression previously [19]. The CD spectra in the range of 200-240 nm were vectors; however, this requires laborious attempts. Moreover, determined using a J-715 spectropolarimeter (Jasco, Japan) at the refolding of insoluble protein from E. coli requires room temperature. The purified scFv proteins at 0.5 g/l in PBS extensive trial-and-error attempts [22]. (pH 7.4) were used for CD spectral analysis. Four scans were

May 2017 ⎪ Vol. 27⎪ No. 5 968 Hu et al.

Expression and Purification of the Anti-LOX-1 scFv in B. choshinensis B. choshinensis SP3 cells were transformed with pNC-scFv or the empty vector pNCMO2 as the negative control, and cultured at 30ºC for 48 h in TM medium. The cultures were then centrifuged to separate the supernatants and cells. The samples from the total expression, supernatants, and precipitations were analyzed by SDS-PAGE (Fig. 1B). The results showed that the anti-LOX-1 scFv proteins were expressed and existed in the culture supernatants. The expressed scFv showed an approximately 29 kDa protein band as seen using SDS-PAGE, and this is consistent with its theoretical molecular weight. It reached a yield of 125 mg/l. The supernatants from 200 ml cultures were subjected to purification, and the samples of each purification step were analyzed by SDS–PAGE (Fig. 1C). After purification, the scFv with high electrophoretic purity was obtained. The efficient secretory expression of the scFv protein indicated that the B. choshinensis expression system could, with optimization, become a superior system to E. coli in terms of soluble production of the anti-LOX-1 scFv. B. choshinensis SP3 is a non-sporulating bacterium, lacking extracellular proteases, facilitating disulfide bond formation, and characterized by efficient secretory production, and has been successfully used for the abundant production of heterogenous proteins including antibody fragments [23]. On the other hand, the high-level expression of heterogenous proteins in E. coli hosts often leads to the formation of inclusion bodies, and the secretion efficiency is typically low if expressed using secretion signals [24]. Moreover, the gram-negative E. coli produces toxic endotoxins, whereas B. choshinensis is a gram-positive nonpathogenic bacillus, which is suitable for microbial fermentation in the pharmaceutical industry [25]. Fig. 1. Expression and purification of the anti-LOX-1 scFv in E. coli and B. choshinensis. Expression of the Anti-LOX-1 scFv in B. choshinensis (A) Purification of the scFv from E. coli. Each purification step Cultured in Different Media samples of the suspensions (Sup), flow through (FT), wash with PBS Secreted anti-LOX-1 scFv was observed initially at 24 h in containing 25mM imidazole (W1) and 100 mM imidazole (W2), the culture media TM, BTY, 2SY, and 5YC at 30ºC (Fig. 2A) elution (E) from the Ni-NTA column, and peak samples (pg200) from filtration chromatography were analyzed by SDS-PAGE. (B) and in all of the culture media at 37ºC (Fig. 2B). The highest Expression of the scFv in B. choshinensis. The samples of the total yield was reached at 60 h under 37ºC and 72 h under 30ºC expression of pNCMO2 (Ct) and pNC-scFv (Et), the supernatants of for all of the culture media. A maximum yield of 200 mg/l pNCMO2 (Cs) and pNC-scFv (Es), and the precipitations of pNCMO2 anti-LOX-1 scFv was obtained using the condition of 30ºC, (Cp) and pNC-scFv (Ep) were analyzed by SDS-PAGE. (C) Purification 72 h, and TM medium. The anti-LOX-1 scFv showed a of the scFv from B. choshinensis. The samples of supernatants (Sup), higher expression level in the TM, 5YC, BTY, and 2SY flow through (FT), and wash with PBS containing 25 mM imidazole media, which contained the same carbon source, glucose, (W1) and elution (E) from the Ni-NTA column as well as the peak but different nitrogen sources and metal ions. These chemicals samples (pg200) from filtration chromatography were analyzed by SDS-PAGE. Lane M: protein marker. were subjected to a further systematic optimization.

J. Microbiol. Biotechnol. Expression and Characterization of Anti-LOX-1 scFv 969

acted as nitrogen sources, can provide growth factors, amino acids, and short peptides for the growth of the host strains. During fermentation, the excess nitrogen sources were gradually decomposed and released ammonia into the medium, which can lead to the increase of the medium pH [26]. When the concentration of the beef extract and soytone increased up to 8.2 and 22.4 g/l, the nitrogen source was consumed for the growth of the host strains, which would not lead to intolerable pH changes. However, when excess nitrogen source was supplied, the medium pH increased gradually, which might seriously affect scFv production. As shown in Figs. 3C and 3D, the scFv yield increased as the glucose concentration increased in the range of 5.0 to 31.6 g/l, and nearly remained unchanged in the range from 31.6 to 40 g/l. During fermentation, glucose, acting as the carbon source, can provide energy for the growth of the host strains and the scFv expression, whereas insufficient glucose may lead to lower scFv yield. When the glucose concentration was 31.6 g/l, the carbon source was sufficient, and any more glucose did not make more improvements. The response surface results indicated that the maximum yield of the scFv would be 355 mg/l when the concentrations of glucose, beef extract, and soytone were 31.6, 8.2, and 22.4 g/l, respectively. To verify this prediction, anti-LOX-1 o scFv production was developed under the optimal conditions, Fig. 2. Expression of anti-LOX-1 in various cultures at 30 C (A) and 37ºC (B). and a high yield of 350 mg/l was obtained. Through the TM (solid circle), BTY (hollow circle), LB (solid upper triangle), 2SY systematic optimization, the scFv yield was increased 2.8- (hollow upper triangle), 5YC (solid square), or B2 (hollow square) fold compared with that obtained with the initial culture media were used (see Materials and Methods for compositions). conditions. Based on the optimization results, a batch fermentation strategy was applied in a 5-L fermentor. As Plackett-Burman Screening Design and Optimization of shown in Fig. 4, the growth curve and anti-LOX-1 scFv Key Factors Using RSM expression curve were estimated, and the average maximum A Plackett-Burman screening design was applied to yield (1.5 g/l) can be reached under the optimized screen the 10 independent variables that have an important conditions, which is support for the industrial scale-up effect on the expression yield of anti-LOX-1 scFv. As shown production of this anti-LOX-1 scFv. in Fig. 3A, the Pareto chart revealed that glucose, beef extract, and soytone (p < 0.05) had a significant effect on the LOX-1-Binding Affinity of the Purified scFv Proteins expression yield of anti-LOX-1 scFv. RSM was employed to Assessed by BLI optimize the concentrations of the three key factors. As The association and dissociation reactions of the LOX-1 shown in the results of the 2D contour plots (Fig. 3), the antigen immobilized on a sensor chip were examined against scFv yield increased with increase of the beef extract each scFv sample as a function of the concentrations of 10, concentration from 1.0 to 8.2 g/l and decreased in the 5, 2.5, 1.25, 0.625, and 0.3125 μM. The assays were analyzed range beef extract concentration from 8.2 to 20.0 g/l using Data Analysis 7.0, and the KD value was found to be (Figs. 3B and 3D). Additionally, the scFv yield increased as 1.01E-07 M for the scFv produced from the B. choshinensis the concentration of soytone increased from 5.0 to 22.4 g/l host (scFv-BC) and 4.53E-07 M for the scFv produced from and decreased as the concentration increased from 22.4 to the E. coli host (scFv-EC) (Fig. 5). The anti-LOX-1 scFv 40.0 g/l (Figs. 3B and 3C). Beef extract and soytone, which expressed by B. choshinensis showed a similar antigen

May 2017 ⎪ Vol. 27⎪ No. 5 970 Hu et al.

Fig. 3. Optimization of the culture medium. (A) The Pareto chart for the standardized effect on anti-LOX-1 scFv yield. Two-dimensional contour plots showing the interactions between beef extract and soytone at a glucose concentration of 22.5 g/l (B), glucose and soytone at a beef extract concentration of 10.5 g/l (C), and glucose and beef extrac at a soytone concentration of 22.5 g/l (D).

than that of scFv-EC. A similar result was found in a previous study, where Gaciarz et al. [27] reported that the soluble scFv of Herceptin sourced from E. coli displayed incomplete folding and showed worse activity than that of B. choshinensis.

Structural Characterization of the Purified scFv Proteins Studied by Spectral Analysis The tertiary structural properties of the scFv proteins were probed by fluorescence spectroscopy using an excitation wavelength of 295 nm. Both scFv proteins displayed a peak between 320 and 340 nm (Fig. 6A), indicating a high order Fig. 4. Expression of the anti-LOX-1 scFv in a 5-L fermentor. structure. Compared with scFv-EC, scFv-BC showed a Solid lower triangles and hollow lower triangles indicate the 4 nm blue shift, indicating a more compact folding. The expression level of anti-LOX-1 scFv and the absorbance of the secondary structure properties of scFv-BC and scFv-EC fermentation liquid at 660 nm, respectively. were examined using far-UV CD spectra. The spectra (Fig. 6B) of both scFv proteins showed a negative peak at affinity to that of the Free Style 293 expression system [11] 218 nm, consistent with the β-sheet found in immunoglobulin and displayed a 3.5-fold higher LOX-1-binding affinity protein folding, whereas the whole secondary structure of

J. Microbiol. Biotechnol. Expression and Characterization of Anti-LOX-1 scFv 971

Fig. 5. Antigen affinity of the anti-LOX-1 scFv analyzed by biolayer interferometry (BLI). BLI curves of the scFv produced from the B. choshinensis host (scFv-BC) (A) and of the scFv produced from the E. coli host (scFv-EC) (B) as a function of concentrations of 10 (black), 5 (red), 2.5 (green), 1.25 (blue), 0.625 (purple), and 0.3125 µM (dark yellow) binding with biotinylated LOX-

1 protein. (C) Kinetic parameters of scFvs derived from the BLI experiments. Kon, kinetic association rate; koff, kinetic dissociation rate. KD, equilibrium dissociation constant (koff /kon).

Fig. 6. Structure analysis of the anti-LOX-1 scFv proteins. (A) Intrinsic fluorescence spectra in the wavelength range of 300-400 nm. (B) Far-UV CD spectra over the wavelength range of 200-240 nm. (C) Size distributions in solution as measured by dynamic light scattering. Red and black curves correspond to scFv-EC and scFv-BC, respectively.

scFv-EC was not fairly evident. This may indicate defects VL domains [28, 29], and is an indication of complete in folding. Furthermore, different from scFv-EC, the scFv- folding. The size and homogeneity of the scFv proteins BC spectrum displayed an obvious ellipticity minimum of were analyzed using DLS, which has been widely used to the negative peak at 230 nm, which was described for some characterize micron- and submicron-sized particles in

May 2017 ⎪ Vol. 27⎪ No. 5 972 Hu et al.

solution [30]. As shown in Fig. 6C, the DLS data of scFv-BC showed a monodisperse structure and one peak at 9.388 nm, suggesting a uniform distribution of monomer scFv. For scFv-EC, peaks were detected at 13.54 and >500 nm, indicating an aggregation tendency. The structural analysis proved that compared with scFv-EC, scFv-BC had a more compact folding. The secretory expression style facilitates the proper folding of proteins, especially proteins requiring disulfide bridge formation, as they must pass through a more favorable redox potential in the periplasmic space [31], which may be conducive to the correct folding of anti- LOX-1 scFv.

Thermal Stability Heat denaturation allows for the determination of the thermodynamic parameters of the unfolding process, using intrinsic fluorescence spectra excited at 295 nm. Changes in the value I320/365 [32] of scFv-BC and scFv-EC at temperatures from 25ºC to 65ºC were determined. The I320/365 values of scFv-BC and scFv-EC both decreased slightly as the temperature increased from 25ºC to 50ºC, suggesting that the fluorophores of the two scFv proteins had become more water-accessible during heating. The main transition occurred in the temperature range of 52-56ºC with a midpoint temperature of transition (Tm) value of 55.9 ±

0.2ºC for scFv-BC, and 48-58ºC with a Tm value of 53.2 ± 0.3ºC for scFv-EC (Fig. 7A). As the two anti-LOX-1 scFv samples are both mainly β-sheets with a typical 218 nm minimum ellipticity, the temperature-induced changes in Thermal stability of the anti-LOX-1 scFv proteins. ellipticity at 218 nm allowed us to evaluate the structural Fig. 7. (A) The dependence of I on temperature. (B) Changes in the changes of the scFv samples under increased temperatures 320/365 ellipticity at 218 nm with increase in temperature. Red and black from 40ºC to 65ºC. The Tm value was 54.5 ± 0.4ºC for scFv- curves correspond to scFv-EC and scFv-BC, respectively.

BC and 51.7 ± 0.2ºC for scFv-EC (Fig. 7B). The average Tm value measured by the two methods was 52.5 ± 0.3ºC for scFv-EC and 55.2 ± 0.3ºC for scFv-BC. These results by E. coli host. B. choshinensis proved to be an ideal host for revealed that scFv-BC displayed better thermal stability, efficient expression of the anti-LOX-1 scFv. This study lays which will support its further medical application. a foundation for the scale-up production of anti-LOX-1 In conclusion, we reported that the anti-LOX-1 scFv had scFv, which will boost the utility of the scFv in further limited soluble expression in E. coli but can be efficiently engineering and applications in the diagnosis and therapy secreted by B. choshinensis. After optimization using of atherosclerotic-related diseases. Plackett-Burman screening design and RSM, the secreted scFv yield can reach 350 mg/l in a shaker flask and 1.5 g/l Acknowledgments in a 5-L fermentor. Moreover, the properties of the scFv proteins obtained from the two expression systems were This work was supported by the National Natural different. The antigen affinity, transition temperature, and Science Foundation of China (No. 81072564), the Science particle diameter size were 1.01E-07 M, 55.2 ± 0.3ºC, and and Technology Development Planning of Jilin (No. 9.388 nm for the scFv expressed by B. choshinensis host, and 20140203001YY), and the Jilin Province Development and 4.53E-07 M, 52.5 ± 0.3ºC, and 13.54 nm for the scFv expressed Reform Commission (No. 2014Y080).

J. Microbiol. Biotechnol. Expression and Characterization of Anti-LOX-1 scFv 973

References et al. 2014. Engineering toward a bacterial “endoplasmic reticulum” for the rapid expression of immunoglobulin 1. Sawamura T, Kume N, Aoyama T, Moriwaki H, Hoshikawa proteins. MAbs 6: 671-678. H, Aiba Y, et al. 1997. An endothelial receptor for oxidized 15. Liu A, Ye Y, Chen W, Wang X, Chen F. 2015. Expression of low-density lipoprotein. Nature 386: 73-77. V(H)-linker-V(L) orientation-dependent single-chain Fv 2. Ding Z, Liu S, Wang X, Dai Y, Khaidakov M, Romeo F, antibody fragment derived from hybridoma 2E6 against Mehta JL. 2014. LOX-1, oxidant stress, mtDNA damage, aflatoxin B1 in Escherichia coli. J. Ind. Microbiol. Biotechnol. autophagy, and immune response in atherosclerosis. Can. J. 42: 255-262. Physiol. Pharmacol. 92: 524-530. 16. Chen C, Snedecor B, Nishihara JC, Joly JC, McFarland N, 3. Xu S, Ogura S, Chen J, Little PJ, Moss J, Liu P. 2013. LOX-1 Andersen DC, et al. 2004. High-level accumulation of a in atherosclerosis: biological functions and pharmacological recombinant antibody fragment in the periplasm of modifiers. Cell. Mol. Life Sci. 70: 2859-2872. Escherichia coli requires a triple-mutant (degP prc spr) host 4. Murase T, Kume N, Kataoka H, Minami M, Sawamura T, strain. Biotechnol. Bioeng. 85: 463-474. Masaki T, et al. 2000. Identification of soluble forms of 17. Mu T, Liang W, Ju Y, Wang Z, Wang Z, Roycik MD, et al. lectin-like oxidized LDL receptor-1. Arterioscler. Thromb. 2013. Efficient soluble expression of secreted matrix Vasc. Biol. 20: 715-720. metalloproteinase 26 in Brevibacillus choshinensis. Protein 5. Shaw DJ, Seese R, Ponnambalam S, Ajjan R. 2014. The role Expr. Purif. 91: 125-133. of lectin-like oxidised low-density lipoprotein receptor-1 in 18. Choy CJ, Berkman CE. 2016. A method to determine the vascular pathology. Diab. Vasc. Dis. Res. 11: 410-418. mode of binding for GCPII inhibitors using bio-layer 6. Li D, Patel AR, Klibanov AL, Kramer CM, Ruiz M, Kang interferometry. J. Enzyme Inhib. Med. Chem. 31: 1690-1693. BY, et al. 2010. Molecular imaging of atherosclerotic plaques 19. Liu Y, Meng Z, Shi R, Zhan L, Hu W, Xiang H, et al. 2015. targeted to oxidized LDL receptor LOX-1 by SPECT/CT and Effects of temperature and additives on the thermal stability magnetic resonance. Circ. Cardiovasc. Imaging 3: 464-472. of glucoamylase from Aspergillus niger. J. Microbiol. Biotechnol. 7. Yamada T. 2011. Therapeutic monoclonal antibodies. Keio J. 25: 33-43. Med. 60: 37-46. 20. Gao YS, Su JT, Yan YB. 2010. Sequential events in the 8. Andersen DC, Reilly DE. 2004. Production technologies for irreversible thermal denaturation of human brain-type creatine monoclonal antibodies and their fragments. Curr. Opin. kinase by spectroscopic methods. Int. J. Mol. Sci. 11: 2584- Biotechnol. 15: 456-462. 2596. 9. Tokunaga M, Mizukami M, Yamasaki K, Tokunaga H, 21. Cheng YM, Lu MT, Yeh CM. 2015. Functional expression of Onishi H, Hanagata H, et al. 2013. Secretory production of recombinant human trefoil factor 1 by Escherichia coli and single-chain antibody (scFv) in Brevibacillus choshinensis Brevibacillus choshinensis. BMC Biotechnol. 15: 32. using novel fusion partner. Appl. Microbiol. Biotechnol. 97: 22. Onishi H, Mizukami M, Hanagata H, Tokunaga M, Arakawa T, 8569-8580. Miyauchi A. 2013. Efficient production of anti-fluorescein 10. Ahmad ZA, Yeap SK, Ali AM, Ho WY, Alitheen NB, Hamid and anti-lysozyme as single-chain anti-body fragments M. 2012. scFv antibody: principles and clinical application. (scFv) by Brevibacillus expression system. Protein Expr. Purif. Clin. Dev. Immunol. 2012: 980250. 91: 184-191. 11. Iwamoto S, Fujita Y, Kakino A, Yanagida K, Matsuda H, 23. Hanagata H, Mizukami M, Miyauchi A. 2014. Efficient Yoshimoto R, et al. 2011. An alternative protein standard to expression of antibody fragments with the Brevibacillus measure activity of LOX-1 ligand containing apoB (LAB) - expression system. Antibodies 3: 242-252. utilization of anti-LOX-1 single-chain antibody fused to apoB 24. D’Urzo N, Martinelli M, Nenci C, Brettoni C, Telford JL, fragment. J. Atheroscler. Thromb. 18: 818-828. Maione D. 2013. High-level intracellular expression of 12. Babaei A, Zarkesh-Esfahani SH, Gharagozloo M. 2011. heterologous proteins in Brevibacillus choshinensis SP3 under Production of a recombinant anti-human CD4 single-chain the control of a xylose inducible promoter. Microb. Cell Fact. variable-fragment antibody using phage display technology 12: 12. and its expression in Escherichia coli. J. Microbiol. Biotechnol. 25. Maehashi K, Matano M, Saito M, Udaka S. 2010. Extracellular 21: 529-535. production of riboflavin-binding protein, a potential bitter 13. Shahryari F, Safarnejad MR, Shams-Bakhsh M, Schillberg S, inhibitor, by Brevibacillus choshinensis. Protein Expr. Purif. Nölke G. 2013. Generation and expression in plants of a 71: 85-90. single-chain variable fragment antibody against the 26. Zou C, Duan X, Wu J. 2016. Efficient extracellular immunodominant membrane protein of Candidatus Phytoplasma expression of Bacillus deramificans pullulanase in Brevibacillus aurantifolia. J. Microbiol. Biotechnol. 23: 1047-1054. choshinensis. J. Ind. Microbiol. Biotechnol. 43: 495-504. 14. Groff D, Armstrong S, Rivers PJ, Zhang J, Yang J, Green E, 27. Gaciarz A, Veijola J, Uchida Y, Saaranen MJ, Wang C,

May 2017 ⎪ Vol. 27⎪ No. 5 974 Hu et al.

Hörkkö S, Ruddock LW. 2016. Systematic screening of 283: 15853-15860. soluble expression of antibody fragments in the cytoplasm 30. Panchal J, Kotarek J, Marszal E, Topp EM. 2014. Analyzing of E. coli. Microb. Cell Fact. 15: 22. subvisible particles in protein drug products: a comparison 28. Tsybovsky Y, Shubenok DV, Kravchuk ZI, Martsev SP. 2007. of dynamic light scattering (DLS) and resonant mass Folding of an antibody variable domain in two functional measurement (RMM). AAPS J. 16: 440-451. conformations in vitro: calorimetric and spectroscopic study 31. Jonasson P, Liljeqvist S, Nygren PA, Ståhl S. 2002. Genetic of the anti-ferritin antibody VL domain. Protein Eng. Des. design for facilitated production and recovery of recombinant Sel. 20: 481-490. proteins in Escherichia coli. Biotechnol. Appl. Biochem. 35: 91-105. 29. Baden EM, Owen BA, Peterson FC, Volkman BF, Ramirez- 32. Turoverov KK, Haitlina SY, Pinaev GP. 1976. Ultra-violet Alvarado M, Thompson JR. 2008. Altered dimer interface fluorescence of actin. Determination of native actin content decreases stability in an amyloidogenic protein. J. Biol. Chem. in actin preparations. FEBS Lett. 62: 4-6.

J. Microbiol. Biotechnol.