Introduction

Malaria is a most prevalent vector borne disease worldwide with over 50% of world’s population currently been exposed to malaria [1]. This disease is responsible for nearly 1 million deaths and 250 million clinical cases every year [1], the deaths being primarily due to the infection by Plasmodium falciparum. Development of vaccine against P. falciparum malaria is likely to reduce the severity of infection and death rate significantly [2]. Several parasite antigens that are expressed at liver and blood stages in human have been identified as potential targets for vaccine development and were used for human clinical trials [2, 3]. For vaccine against blood stage of human malaria parasite, P. falciparum Merozoite surface protein-1 (PfMSP-1) is one of the leading candidates [4].

The PfMSP-1 is synthesized as a large (195 kDa) precursor and is present on the surface of the released merozoite in the form of a multi-component protein complex derived via proteolytic processing [5, 6]. The membrane bound C-terminal 42-kDa component (PfMSP-142) of this surface complex is further cleaved into two fragments, PfMSP-133 and PfMSP-119, prior to erythrocyte invasion. Majority of the complex sheds from the parasite surface at the time of invasion, leaving behind only PfMSP-119 attached to merozoite through glycosyl-phosphotidyl-inositol (GPI) anchor [7, 8]. The structural analysis of PfMSP-119 indicates presence of highly conserved, two epidermal growth factor (EGF)-like motifs with 12 cysteines [9, 10].

There have been extensive in vitro and in vivo experimental evidences that support the use of PfMSP-119 as a vaccine candidate against blood stage of malaria parasite [7, 11–15]. For example, monoclonal antibodies directed against PfMSP-119 inhibits erythrocyte invasion by the parasite in vitro [7]. Mice immunized with homologous MSP-119 from rodent malaria parasite P. yoelii expressed in Escherichia coli are protected against an otherwise lethal challenge with this rodent parasite [12]. Vaccination of Aotus monkey with recombinant PfMSP-119 confers protection against malaria [11, 13]. Analysis of sera from malaria-immune individuals using transgenic parasite line suggests that the invasion inhibitory antibodies present in these sera are mainly directed towards the PfMSP-119 [14] and play an important protective role in immunity to blood-stage P. falciparum infection [15]. There are, however, few reports indicating no correlation between anti-PfMSP-119 antibodies and reduced malaria incidence [16, 17].

Recent outcome of clinical trials suggests that a successful malaria vaccine might be a multi-subunit vaccine, as single subunit vaccine could not demonstrate desirable results [18]. We propose PfMSP-119 as one of the component of multi-subunit vaccine. Various expression systems, including E. coli [19], Pichia pastoris [20], baculovirus [21], Saccharomyces cerevisiae [22], had been used in the past to produce PfMSP-119 for use in pre-clinical and clinical studies. Production of PfMSP-119 as a single and correctly folded conformer, an essential feature to obtain invasion inhibitory antibodies [23], was found to be difficult due to presence of two cysteine-rich EGF like structures [24]. The EGF like structure in proteins is characteristic of higher eukaryotes and the machinery to fold this structure efficiently is absent in lower eukaryotes and prokaryotes.

We report here a systematic study to produce recombinant PfMSP-119 with reasonable yield and purity. We used E. coli system for expressing PfMSP-119 because it is the most cost effective and efficient system for production of recombinant proteins. Production of cysteine-rich EGF like structures under correctly folded soluble form in E. coli, though, needed extensive optimization of growth and expression condition. We describe here identification of cultivation conditions for high-level expression of PfMSP-119 in soluble form and development of purification strategy to obtain PfMSP-119 in highly pure form. Characterization of PfMSP-119 indicated that it is highly pure, homogenous and recognized by conformation sensitive monoclonal antibodies.

Materials and methods

Construction of plasmids for expression of recombinant PfMSP-119

A synthetic gene was designed for high-level expression of PfMSP-119 in E. coli. The amino acid sequence corresponding to residues 1526–1619 of Merozoite Surface Protein-1 of P. falciparum Welcome strain (GenBank Accession No. P04933) was back-translated to nucleotide sequence based on E. coli codon frequency table as described before [25]. The codons for 6-histidine tag were introduced at the 5′-region of the synthetic gene to facilitate affinity purification of PfMSP-119. The designed gene was synthesized commercially using overlapping oligomers at Midland Certified Reagent Company, USA. The synthetic gene was cloned at NcoI and BamHI restriction sites of pET28a(+) (Novagen), a kanamycin based vector, to obtain pET28a(+) synPfMSP-119 and transformed in E. coli BLR(DE3) for expressing PfMSP-119.

Native gene encoding the same region of PfMSP-119 as described above was amplified from genomic DNA of Welcome strain of P. falciparum, cloned into BamHI and HindIII site of pQE30 vector and transformed into E. coli M15(pREP4). The native construct was used as control for analyzing expression of PfMSP-119.

Expression of PfMSP-119 at the shake flask level

E. coli strain containing either the native or synthetic gene construct was grown in shake flask culture in LB medium for comparison of PfMSP-119 expression. Overnight grown culture (1%) was inoculated in flask containing 50 ml LB and was grown at 37 °C at 200 rpm. The culture flask was induced at optical density at 600 nm (OD600) of 0.6–0.8 with 1 mM IPTG and grown further for 4 h. Level of expression of PfMSP-119 post-induction was analyzed on 15% SDS-PAGE gel and detected on Western blot using monoclonal antibody specific to histidine tag (Qiagen).

High cell density cultivation to produce PfMSP-119

Batch and fed-batch cultivations were performed to optimize the expression of PfMSP-119. Composition of media used for cultivations are described in Table 1. Optimization of cultivation parameters were performed in a bioreactor with 2–5 l working capacity. The optimized condition was further scaled-up to 10 l cultivation scale to produce biomass for PfMSP-119 purification. The cultivations were carried out in a bioreactor (Applikon AG, The Netherlands) equipped with extensive analytical devices for measurement of pH, temperature, dissolved oxygen, and data sampling was done using Bioexpert software. The batch cultivations were performed with following set-up: air flow rate, 0.5 vvm; stirrer speed, 500 rpm; pH, 6.8; temperature, 37 °C; initial dissolved oxygen level, 100%. Dissolved oxygen (DO2) level was maintained at 40% saturation during the cultivation by controlling agitation rate and air flow rate through proportional–integral–derivative (PID) controller. If required, the inlet air was enriched with pure oxygen. pH was controlled at 6.8 by the addition of 2 N NaOH or 12.5% w/w ammonium hydroxide for cultivation in semi-defined or defined medium, respectively. Sampling was performed at regular interval to measure cell density by taking absorbance at 600 nm and to monitor PfMSP-119 expression by SDS-PAGE. For the cultivations that were carried out only in batch mode, cells were induced at OD600 of ~8.0 with 1 mM IPTG and grown further at 25 °C for 4 h. For the cultivations carried out in fed-batch mode, culture was grown in batch mode until carbon source was exhausted from medium, as indicated by the rise in DO2 value. Glucose feed was then added to achieve either constant DO (DO-stat) or constant specific growth rate as described earlier [26]. The induction was performed with 1 mM IPTG at OD600 of 40 and cells were grown further for 4–5 h at 25 °C. The actual specific growth rate and biomass yield with respect to glucose during cultivation were determined by calculating amount of biomass produced and amount of glucose consumed in a given interval of time.

Table 1 Composition of defined and semi-defined media used for cultivation of E. coli BLR(DE3) pET28a(+) synPfMSP-119
Full size table

Offline measurement

Optical density of the culture was measured at 600 nm (OD600) with Perkin Elmer UV–visible spectrophotometer. Samples with higher values of OD600 were diluted suitably to have OD600 in the range of 0.2–0.8. Dry cell weight (DCW) was determined by weighing 1 ml washed culture pellet dried overnight at 105 °C. OD600 of 1.0 corresponded to approximately 0.35 g l−1 of DCW. Amount of glucose in culture media during fermentation was determined by the DNSA method as described by Miller [27].

Analysis of PfMSP-119 expression level during cultivation

Yield of PfMSP-119 produced during cultivation was determined by quantifying PfMSP-119 following immobilized metal affinity chromatography (IMAC) purification. Wet cell pellet (30 g) was lysed by sonication and supernatant was collected after centrifugation at 10,000 rpm for 45 min at 4 °C. PfMSP-119 present in the supernatant was purified by IMAC using 10 ml of streamline chelating matrix charged with Ni2+. Matrix was washed with 25 mM imidazole and bound proteins were eluted with 250 mM imidazole. Elutes were separated on SDS-PAGE gel to analyze expression of recombinant PfMSP-119. Yield of PfMSP-119 was determined by measuring absorbance at 280 nm with theoretical extinction coefficient of 0.267 mg−1 cm2, calculated based on amino acid sequence.

Purification of PfMSP-119 produced at 10 l cultivation scale

Cells obtained from 10 l cultivation were harvested by centrifugation and were washed with PBS. Washed cell pellet was resuspended in lysis buffer (20 mM Tris/HCl pH 8.0, 500 mM NaCl, 10 mM imidazole, 1% Triton X-100, 25 mg l−1 lysozyme, 5 mM benzamidine HCl) and lysed by passing through Dynomill (KDL, Germany) at the flow rate of 100 ml min−1. Percentage cell lysis was monitored via microscopy. More than 95% cells were lysed after three passages through Dynomill. Lysed culture was clarified through online centrifuge (Carr Pilot centrifuge) at 20,000g and PfMSP-119 present in the supernatant was purified by immobilized metal affinity chromatography (IMAC). The clarified supernatant containing PfMSP-119 was loaded onto Streamline 100 column (GE Healthcare) containing Streamline chelating resin (GE Healthcare) charged with Ni2+ in expanded mode. The resin was washed with wash buffer 1 (20 mM Tris/HCl pH 8.0, 500 mM NaCl, 10 mM imidazole), wash buffer 2 (20 mM Tris/HCl pH 5.5, 500 mM NaCl, 10 mM imidazole) and wash buffer 3 (20 mM Tris/HCl pH 8.0, 10 mM NaCl, 40 mM imidazole) in expanded mode. Elution of protein was performed in packed mode with elution buffer (20 mM Tris/HCl pH 8.0, 10 mM NaCl, 250 mM imidazole). Fractions containing PfMSP-119 with significant purity as observed on 15% SDS-PAGE gel were pooled and loaded on column containing pre-equilibrated Q-Sepharose Fast Flow (GE Healthcare) for ion exchange chromatography using salt gradient. Column was washed with wash buffer (20 mM Tris/HCl pH 7.2, 10 mM NaCl) till base line for absorbance at 280 nm stabilized. Bound proteins were eluted by step gradient of buffers containing 25, 35, 50, 100 and 250 mM of NaCl. Samples obtained from ion exchange chromatography using salt gradient were separated on 15% SDS-PAGE gel and protein fractions of 50 mM NaCl that contained mostly monomeric PfMSP-119 were pooled, buffer exchanged with 20 mM Tris/acetate pH 8.5, 10 mM NaCl and processed further for ion exchange chromatography using pH gradient. An XK-26 column (GE Healthcare) packed with Q-Sepharose FF matrix was used for the chromatography. Protein sample was loaded onto the column matrix pre-equilibrated with 20 mM Tris/acetate pH 8.5 and 10 mM NaCl and PfMSP-119 was eluted from the column using linear gradient of pH from 8.5 to 5.0. The fractions were analyzed on SDS-PAGE gel and relevant fractions containing PfMSP-119 were pooled, filter-sterilized, and stored at −80 °C as 2 ml aliquots until further use. Yield of PfMSP-119 at each step of purification was estimated by measuring absorbance at 280 nm with theoretical extinction coefficient of 0.267 mg−1 cm2.

Characterization of purified PfMSP-119 produced at 10 l scale

PfMSP-119 produced at 10 l cultivation scale and purified by metal affinity and ion exchange chromatographies was characterized for its purity and homogeneity by following methods. PfMSP-119 was evaluated for its purity on 15% SDS-PAGE gel with 5 and 10 μg protein loaded per well. PfMSP-119 was separated under reducing and non-reducing condition on SDS-PAGE gel and was stained with Coomassie to observe shift in mobility due to reduction of cystines involved in disulfide bond formation. Western blot was performed with polyclonal antibodies against purified PfMSP-119 raised in rabbit to confirm identity of full length PfMSP-119. PfMSP-119 was tested for homogeneity by reverse phase chromatography using C-8 column (Supelco). The gradient used for elution was developed using buffer A (0.1% triflouroacetic acid in water) and buffer B (0.05% triflouroacetic acid in methanol). The column was initially equilibrated in 10% buffer B and gradient reached 95% buffer B in 40 min. Endotoxin content was estimated using limulus amebocyte lysate (LAL) gel clot assay (Charles River Endosafe, USA). The assay was performed as per manufacturer instructions. Free thiol groups were estimated by use of Ellman’s reagent (5,5′-dithio-bis-3 nitro benzoic acid). Different concentrations of l-cysteine base were used to plot the standard curve for estimation of free thiol groups in the protein. Protein sample was analyzed for its mass by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS) and N-terminal sequence by Edman degradation followed by HPLC at Michigan State University (USA). CD measurement of PfMSP-119 was performed at a concentration of 8 × 10−6 M in 10 mM phosphate at pH 7.0. The experiment was carried out on a Jasco spectropolarimeter model J-810 in the far-UV range from 190 to 250 nm using a cuvette of path length 0.2 cm and sample volume 1 ml. Spectra were recorded with scanning speed of 100 nm min−1 and band width of 1 nm and averaging six scan. Deconvolution of CD-spectra was performed using Spectra-manager software from Jasco.

Enzyme linked immuno-sorbent assay (ELISA) and indirect fluorescence assay (IFA)

Conformation of PfMSP-119 was analyzed by ELISA using three conformation sensitive monoclonal antibodies, mAb 5.2, 1H4 and 2E10. mAb 5.2 was obtained from MR4 and 1H4 and 2E10 were made commercially (A&G Pharmaceuticals, Columbia, USA) against bacterially expressed highly pure PfMSP-119. ELISA was performed by coating PfMSP-119 on ELISA plate in coating buffer (50 mM carbonate buffer pH 9.6) at a concentration of 0.1 μg per well under folded or denatured form. PfMSP-119 was denatured by boiling 20 μg sample at 1 mg/ml with 20% (v/v) β-mercapto-ethanol for 5 min followed by incubation with 50 μl of 20% (w/v) iodoacetoamide for 30 min at 37 °C. The PfMSP-119 coated on ELISA plate was probed with monoclonal antibodies 5.2, 1H4 and 2E10 at a dilution of 1:2000 in triplicate wells, followed by incubation with HRP conjugated goat anti-mouse immunoglobulin at a dilution of 1:1000. Monoclonal antibody against histidine-tag (Qiagen) at 1:6000 was used as control primary antibody to monitor equal coating of folded and denatured protein. The ELISA was developed by the addition of 100 μl/well of o-phenylene diamine dihydrochloride (1 mg/ml, Sigma) and hydrogen peroxide (Merk) as substrate. The reaction was stopped with 2 N sulfuric acid and the absorbance was measured at 490 nm using an ELISA reader (Molecular Devices). Mean absorbance values and standard deviations for the triplicate wells of each antibody group were calculated and used for plotting graph.

Antibodies were raised against purified PfMSP-119 in rabbits after formulation with Freund’s adjuvant and were used for indirect fluorescence assay (IFA) to detect the native antigen on parasite surface. Enriched late stage schizonts of Plasmodium falciparum were spread on the slide as a thin smear, fixed with methanol and blocked with 1% skimmed milk in PBS. The slides were incubated with polyclonal sera raised against PfMSP-119 at 1:100 dilution at 37 °C for 1 h in humidified chamber. The pre-immune sera at 1:100 dilution was used as control for IFA. Slides were washed with PBS and incubated with FITC labeled goat anti-rabbit immunoglobulin (Sigma) at 1:1000 dilution for 1 h at room temperature. Slides were washed with PBS and incubated with Dapi (Sigma) to stain the parasite DNA for three minutes. After addition of anti-fade flouroguard reagent (Bio-Rad), the slides were sealed with coverslip and observed under fluorescent microscope.

Results

Codon optimization enhances expression of PfMSP-119 in E. coli

A synthetic gene was designed based on E. coli codon usage table for optimal expression of PfMSP-119 in E. coli (Fig. 1a). After modification, the AT content of synthetic gene sequence was 51% as against 64% for the native gene. The designed synthetic gene was cloned in pET28a(+) vector for expression of PfMSP-119 in E. coli and sequenced from both ends to ascertain accuracy of the synthetic gene sequence. The expression level of PfMSP-119 was compared for both synthetic and native gene constructs by growing the recombinant E. coli in shake flasks. The expression of PfMSP-119 from synthetic gene construct was fourfold higher than the native gene construct as determined by densitometry (Fig. 1b). Western blot analysis of crude lysate confirmed the identity of expressed PfMSP-119 and also indicated higher expression level of PfMSP-119 with the synthetic gene construct (Fig. 1c).

Fig. 1
figure 1

Design of synthetic gene coding for PfMSP-119. a Sequence of synthetic gene coding for PfMSP-119 optimized for high-level expression in E. coli. Nucleotides that differ from native sequence are highlighted in black box. b SDS-PAGE gel of total cellular protein showing expression of recombinant PfMSP-119 from synthetic and native gene constructs in E. coli. Protein samples were separated on SDS-PAGE gel (15%) and detected by coomassie staining. Lane 1 molecular weight marker; Lane 2 recombinant PfMSP-119; Lane 3 uninduced cells with synthetic gene construct; Lanes 4, 5, and 6 cells with synthetic gene construct induced for 2, 3 and 4 h, respectively; Lane 7 uninduced cells with native gene construct; Lanes 8, 9, and 10 cells with native gene construct induced for 2, 3 and 4 h, respectively. A distinct band corresponding to 19 kDa get induced in the cells with synthetic gene construct, but poorly visible in the cells with native gene construct. c Protein samples were separated on SDS-PAGE gel and detected by Western blotting using anti-his tag antibody. Lane 1 uninduced cell with native gene construct; Lane 2 induced cells with native gene construct; Lane 3 uninduced cells with synthetic gene construct; Lane 4, induced cells with synthetic gene construct; Lane M, molecular mass marker

Full size image

Effect of media composition on expression of PfMSP-119 in bioreactor

Expression of PfMSP-119 was assessed in the defined and semi-defined media as mentioned in “Materials and methods”. Yeast extract was selected as complex nitrogenous supplement for the semi-defined media as it is free from any animal derived product. Cells grown in semi-defined media displayed faster growth rate and the culture reached inducible cell density in 3.7 h as against 7 h for defined media. This resulted in shorter cultivation duration for the cells grown in semi-defined media (Fig. 2a, b). Expression level of PfMSP-119 was analyzed after immobilized metal affinity chromatography (IMAC) as mentioned in “Materials and methods”. Cultivation in semi-defined media resulted in expression of 110 mg l−1 of PfMSP-119, a sixfold higher value compared to cultivation in defined media (Table 2). Further optimization to improve expression of recombinant PfMSP-119 was performed in semi-defined media using fed-batch approach.

Fig. 2
figure 2

Time profiles of batch cultivation of E. coli BLR(DE3) synPfMSP-119 for the expression of PfMSP-119. Culture was grown in batch mode in (a) defined and (b) semi-defined media and induced with 1 mM IPTG at OD600 of 8.0 (DCW = 2.8 g l−1) for expression of PfMSP-119

Full size image
Table 2 Description of various cultivation strategies performed for optimal expression of PfMSP-119
Full size table

Effect of various feeding strategies on expression of PfMSP-119

Fed-batch cultivation was performed to obtain high biomass and therefore high product concentration. Various feeding strategies employed for fed-batch cultivation were based on either achieving a pre-defined value of specific growth rate of the cells or maintaining a fixed dissolved oxygen concentration (DO-stat) during cell growth. The value for pre-defined specific growth rate was chosen to achieve optimal channelization of glucose through the glycolytic intermediates as discussed in earlier report [28].

When glucose feed was added to achieve a specific growth rate (μ) of 0.2 h−1 during fed-batch cultivation (Online resource-supplementary Fig. 1), the actual specific growth rate was significantly lower than the set value during the cultivation (Table 2). Amount of residual glucose measured at the end of cultivation showed a value of 4.5 g l−1, suggesting that large amount of glucose remained un-utilized due to lower actual specific growth rate. For the efficient utilization of glucose and to avoid acetate accumulation, a two-step feeding strategy was adopted by adding glucose to achieve specific growth rate of 0.12 h−1 before induction and 0.10 h−1 post-induction (Online resource-supplementary Fig. 2) [26, 29]. Here, the actual specific growth rate (0.1 h−1) during growth phase was slightly lower than the set value of 0.12 h−1, but declined to 0.04 h−1 during the induction phase (Table 2). Amount of residual glucose in the media was 2.9 g l−1 suggesting better utilization of glucose compared to the previous feeding strategy. Yield of PfMSP-119 obtained using the two-step feeding strategy was ~1.5 times higher than the yield of PfMSP-119 obtained using the single step feeding strategy (Table 2).

We further examined the effect of DO-stat based feeding strategy on the cell growth and expression of PfMSP-119. Here, glucose feed was added in order to maintain DO2 concentration in the range of 20–60% (Fig. 3). The concentration of residual glucose at the end of cultivation was 1.5 g l−1, suggesting efficient utilization of glucose. The final product concentration of PfMSP-119 post-cultivation was 522 mg l−1, which was nearly 4-fold and 2.5-fold higher than the final product concentration obtained using feeding strategies with specific growth rate of 0.2 h−1 and 0.12 h−1/0.1 h−1, respectively (Table 2).

Fig. 3
figure 3

Time profile of fed-batch cultivation of E. coli BLR(DE3) synPfMSP-119 using DO-stat feeding strategy. Culture was grown in batch mode in semi-defined medium until carbon source was completely exhausted. Glucose feed was then added to keep dissolved oxygen concentration in the range of 20–60%. Figure shows time profile for stirrer speed (rpm), dissolved oxygen concentration (%), air flow rate (l min−1) and dry cell mass (g l−1)

Full size image

Purification of PfMSP-119 produced at 10 l cultivation scale

We selected two best conditions among the batch and fed-batch cultivations described in the previous sections and used them for expression and purification of PfMSP-119 at 10 l cultivation scale. These conditions included batch cultivation in semi-defined media and fed-batch cultivation with DO-stat based addition of feed. Wet cell biomasses obtained from batch and fed-batch cultivations were 240 and 950 g, respectively. Cells were lysed using Dynomill and PfMSP-119 was purified from soluble fraction using IMAC. Post-IMAC yield of PfMSP-119 obtained from fed-batch cultivation was 5630 mg, nearly 4.5-fold higher than the yield obtained from batch fermentation (Fig. 4a) (Table 3). The PfMSP-119 obtained after IMAC was purified further by ion exchange chromatography using salt gradient (Fig. 4b). The PfMSP-119 was eluted in the fractions where NaCl concentration was 50 mM. The yields of ion exchange chromatography for PfMSP-119 obtained from batch and fed-batch cultivations were 16 and 2%, respectively (Table 3). These low yields of ion exchange chromatography signify presence of multiple PfMSP-119 conformers expressed in E. coli during cultivation. IMAC captured all PfMSP-119 conformers through their histidine-tag and when attempt was made to purify PfMSP-119 as homogeneous population by ion exchange chromatography, yield dropped drastically. More profoundly, the fed-batch cultivation, performed to obtain higher biomass and therefore higher final product concentration, actually resulted in lower PfMSP-119 yield following ion exchange chromatography. The PfMSP-119 was purified further by ion exchange chromatography using pH gradient to obtain highly pure protein (Fig. 4c). The final volumetric yield of purified PfMSP-119 after ion exchange chromatography using pH gradient was 12–15 and 7–9 mg l−1 of cultivation volume for batch and fed-batch cultivations, respectively.

Fig. 4
figure 4

Purification profiles of PfMSP-119. a The PfMSP-119 was purified from crude cell lysate obtained from batch and fed-batch fermentation using immobilized metal affinity chromatography (IMAC) and analyzed on 15% SDS-PAGE gel. Lane 1 IMAC purified PfMSP-119 following batch cultivation; Lane 2 IMAC purified PfMSP-119 following fed-batch cultivation; Lane M molecular mass marker. Ion exchange chromatography profiles for purification of PfMSP-119 based on salt gradient (b) and pH gradient (c). The insets in the figures represent SDS-PAGE gel profile of the fractions containing monomeric PfMSP-119

Full size image
Table 3 Purification of PfMSP-119 from batch and fed-batch fermentation at 10 l scale
Full size table

Characterization of PfMSP-119 produced at 10 l scale

Purified PfMSP-119 produced via batch cultivation at 10 l scale was characterized for its purity and homogeneity as follows. SDS-PAGE analysis of PfMSP-119 showed single band corresponding to 19 kDa (Fig. 5a). The difference in mobility on SDS-PAGE gel before and after reduction with β-mercaptoethanol indicated presence of disulphide bonds in PfMSP-119 (Fig. 5a). The purified PfMSP-119 was recognized by anti-PfMSP-119 antibodies on Western blot showing its identity as full length protein (Fig. 5b). The molecular mass of purified PfMSP-119 by MALDI-TOF corresponded to 11.4 kDa, which is similar to the calculated molecular mass of PfMSP-119 without the terminal methionine. This observation suggested that the terminal methionine was processed during PfMSP-119 maturation in E. coli. The N-terminal sequencing data of first thirteen amino acids of PfMSP-119 matched with the predicted sequence as G1 H2 H3 H4 H5 H6 H7 N8 I9 S10 Q11 H12 Q13 and further confirmed the efficient processing of terminal methionine. The RP-HPLC profile of PfMSP-119 showed a single uniform peak, indicating that the final product is homogeneous in nature (Fig. 5c). Analysis of PfMSP-119 on analytical gel permeation chromatography showed a single peak corresponding to 19 kDa, suggesting the presence of only monomeric form of PfMSP-119 (Fig. 5d). The endotoxin content in purified PfMSP-119 was found to be less than 50 EU per mg of protein. Considering a possible human dose of PfMSP-119 in the vaccine as 50 μg, the endotoxin content was 2.5 EU per human dose. Ellman’s test was performed to estimate the presence of any free thiol group in the purified PfMSP-119. This test clearly detected 30 μmol of free cysteine, and no free cysteine was detected when test was performed with 115 μmol of PfMSP-119, indicating that greater than 98% of cysteines in the purified PfMSP-119 are disulfide linked.

Fig. 5
figure 5

Characterization of purified PfMSP-119 produced at 10 l cultivation scale. a Purity of PfMSP-119 was assessed on SDS-PAGE gel after Coomassie staining. Shift in mobility of purified PfMSP-119 after reduction with β-mercaptoethanol (BME) was also analyzed to verify the formation of disulphide bonds. Lane 1 molecular mass marker; lane 2 5 μg of reduced PfMSP-119; lane 3 10 μg of reduced PfMSP-119; lane 4 5 μg of non-reduced PfMSP-119; lane 5 10 μg of non-reduced PfMSP-119. Arrows indicate shift in mobility. b Western blot analysis of PfMSP-119 with polyclonal sera raised against PfMSP-119 in rabbit. c Reverse phase chromatography profile of purified recombinant PfMSP-119. Purified PfMSP-119 eluted as single symmetrical peak indicating that it consists of homogeneous population. d Analytical gel permeation chromatography profile of PfMSP-119 on Superdex S-75 column. PfMSP-119 is eluted as single symmetrical peak at its monomeric position of 19 kDa. Standard protein markers are shown in the background as gray line (Bovine serum albumin, 66 kDa; Carbonic anhydrase, 29 kDa; Cytochrome C, 12.4 kDa)

Full size image

Achieving native conformation of PfMSP-119 has been shown to be critical in eliciting functional immune response [23]. We examined the conformation of PfMSP-119 using conformation specific monoclonal antibodies mAb 5.2, 1H4 and 2E10 by ELISA. All three monoclonal antibodies preferentially recognized folded form of PfMSP-119 (Fig. 6a), indicating that the conformation of purified PfMSP-119 is intact. To ascertain equal coating of folded and denatured PfMSP-119, proteins were also incubated with monoclonal antibody against histidine tag. Histidine tag antibody recognized PfMSP-119 both under folded and denatured condition, with slight preference to denatured form. This preference could be due to better exposure of histidine tag of PfMSP-119 towards the antibody after denaturation. Secondary structure of PfMSP-119 was analyzed using CD-spectroscopy as mentioned in “Materials and methods” (Fig. 6b). Deconvolution of secondary structure by Spectra-manager software using Yang method indicated the presence of 56.5% β-sheets and 43.5% of random coils in recombinant PfMSP-119. The presence of mainly β-sheets and random coils in PfMSP-119 was in agreement with the structure of PfMSP-119 determined earlier [30].

Fig. 6
figure 6

Assessment of conformation and structure of PfMSP-119. a Reactivity of folded and denatured form of PfMSP-119 with conformation specific monoclonal antibodies mAb 5.2, 1H4 and 2E10 by ELISA. The correctly folded PfMSP-119 is recognized preferentially by the conformation sensitive monoclonal antibodies compared to the denatured form of PfMSP-119. The antibody against histidine-tag was used as control to probe the equal coating of native and denatured PfMSP-119 on ELISA plate. b CD spectrum of PfMSP-119 to determine the secondary structure. Spectrum was recorded at a protein concentration of 8 × 10−6 M. Molar Ellipticity is plotted as deg cm2 dmol−1. The spectrum indicates that PfMSP-119 is primarily composed of β-sheets and random coils

Full size image

The ability of recombinant PfMSP-119 to induce antibody response that can recognize native protein on parasite surface was assessed by indirect immunofluorescence assay (IFA). We used high titer antibodies against recombinant PfMSP-119 raised in rabbits to detect native PfMSP-1 on the surface of merozoites of the malaria parasite. The antibodies against recombinant PfMSP-119 recognized the native PfMSP-119 present on the surface of merozoite at late schizont stage with high intensity (Fig. 7).

Fig. 7
figure 7

Indirect immuno-fluorescence assay to detect native protein on parasite surface. Figure shows a late stage segmented P. falciparum schizont in the bright field (a), after staining nucleus with Dapi (b), after surface labeling with anti-PfMSP-119 antibody (c) and merge of Dapi and anti-PfMSP-119 antibody labeling (d). Antibody to recombinant PfMSP-119 recognizes the native PfMSP-1 present on the surface of merozoites

Full size image

Discussion

Production of therapeutically important foreign proteins in E. coli expression system has been highly successful in recent past as this system offers several advantages over the other expression systems, including its fast growth rate, utilization of cheap growth media, high level of expression of foreign proteins, etc. [31]. However, expression of cysteine-rich proteins in E. coli often leads to formation of inclusion bodies due to its reducing cytosolic environment. The C-terminal, 19-kDa domain of P. falciparum MSP-1 (PfMSP-119) has a complex structure with 12 cysteine residues linked through 6 disulfide bridges and two EGF-like domains [10]. In an earlier study, PfMSP-119 was found to accumulate as inclusion bodies in E. coli unless it was tagged with the fusion partner MBP [32]. However, vaccine candidate fused with the bulky tag may not be acceptable for clinical development. Nevertheless, it is important to produce correctly folded PfMSP-119 with appropriate disulfide linkages for inducing protective immune response [23]. Therefore, many researchers opted for other expression systems like yeast [20] or Salmonella typhi [33] for expressing soluble form of PfMSP-119. However, protein heterogeneity in the purified PfMSP-119 from yeast and production of misfolded PfMSP-119 in Salmonella had been some of the issues with these expression systems.

To circumvent the above-mentioned problems, we performed a systemic study to express PfMSP-119 in soluble form using E. coli expression system and purify it to homogeneity. As a first step to achieve our goal, we performed optimization of codons for the gene encoding PfMSP-119 because codon usage of P. falciparum and E. coli is considerably different. Moreover, P. falciparum genome is found to be extremely rich in AT bases [34] and heterologous expression of its genes often leads to poor yields. Codon optimization of DNA sequence encoding PfMSP-119 led to a fourfold improvement in expression of PfMSP-119 in E. coli (Fig. 1).

The optimization of expression condition for PfMSP-119 was performed in the bioreactor under controlled cultivation parameters. We found that the presence of yeast extract in the cultivation media enhanced the level of soluble PfMSP-119 expression (Table 2). This observation was similar to the earlier findings where yeast extract was found to improve the specific production level of recombinant protein by inhibiting proteolysis [35]. When cells were grown to achieve a pre-determined growth rate of 0.2 h−1 during fed-batch cultivation, the actual specific growth rate was much lower than the set value (Table 2). The growth rate further declined at post-induction period. Based on this observation, we devised a two-step feeding strategy where cells were grown to achieve 0.12 h−1 before induction and 0.1 h−1 post-induction. The two-step feeding strategy resulted in twofold higher level of expression of PfMSP-119 compared to single step feeding strategy. We further evaluated feeding based on DO-stat method for expression of PfMSP-119. Expression of PfMSP-119 obtained by the DO-stat cultivation method was 2.5-fold higher than the expression obtained with two-step exponential feeding strategy (Table 2).

We developed purification process for PfMSP-119 to obtain a high quality protein suitable for human clinical use. Purification of PfMSP-119 was performed from the cell biomass obtained from cultivation conditions that yielded highest final product concentration under batch and fed-batch cultivation mode. The immobilized metal affinity chromatography using expanded bed adsorption technique and ion exchange chromatography based on salt gradient followed by pH gradient was devised to obtain highly pure PfMSP-119 from crude cell lysate. The final product concentration of PfMSP-119 following IMAC was significantly higher for fed-batch cultivation as compared to batch cultivation. Improvement in cell density (3-fold higher) and specific product yield (1.6-fold higher) together could be attributed to the 4.5-fold higher final product concentration for fed-batch cultivation as compared to batch cultivation (Table 2). However, there was a substantial drop in the yield of PfMSP-119 following ion exchange chromatography using salt based step gradient. This drop was more significant in case of purification performed with the biomass obtained from fed-batch cultivation. We predict that this low yield of PfMSP-119 following ion exchange chromatography was due to formation of multiple conformers of PfMSP-119 in E. coli having different surface charge properties, which may not be distinguishable on SDS-PAGE gel. Further attempt to purify single conformer of PfMSP-119 based on their surface charge property by ion exchange chromatography resulted in poor yield. This observation also indicates that the higher yield of PfMSP-119 at IMAC stage may not necessarily lead to higher yield of purified PfMSP-119 following subsequent purification step as IMAC will purify all conformers of PfMSP-119 produced in E. coli through histidine tag. Processing of biomass from 10 l batch cultivation yielded ~120 mg of highly pure, correctly folded PfMSP-119. Though this yield is reasonable to proceed for production of cGMP grade PfMSP-119 for use in human clinical trials, efforts are being made to further improve the final yield.

The subsequent characterization of recombinant PfMSP-119 by SDS-PAGE, Western blot, reverse phase chromatography, and analytical gel permeation chromatography demonstrated that PfMSP-119 is highly pure, homogeneous, and monomeric protein. Importantly, the conformation of purified PfMSP-119 was intact as it was recognized by several conformation sensitive antibodies. Upon immunization, PfMSP-119 elicited antibodies that recognized native protein on the surface of P. falciparum.

In conclusion, we have described here a robust and scalable cultivation method to produce recombinant PfMSP-119 as soluble form at 10 l cultivation scale. A simple, three-step purification strategy involving metal affinity, salt gradient ion exchange and pH gradient ion exchange chromatography, was devised to produce PfMSP-119 that was highly pure, homogeneous and recognized by conformation sensitive monoclonal antibodies. The method described here to produce PfMSP-119 may also be useful in producing other complex cysteine-rich proteins in E. coli as soluble form for production at large scale.