Construction of expression vector pVMExp14367

Vector pVMExp14367 is a medium copy number T7 promoter-lac operator-based expression vector containing sequence encoding N-terminal deca-histidine tag (H10), Tobacco Etch Virus (TEV) protease cleavage site, 2.0 Kb SacR-SacB gene cassette encoding levansucrase protein of Bacillus subtilis flanked by two appropriately oriented BsaI sites, and 15 amino acid C-terminal Biotin Acceptor Peptide tag (BAP) (Genbank accession number MG599491). In between the functional tags/protease site, there are glycine-serine rich spacer sequences of appropriate lengths. The vector backbone comprises of the ColE1 origin of replication (ori) with deletion of rop gene, filamentous phage origin of replication (f ori), beta-lactamase gene as a selection marker and lac repressor (lacI). This vector design is compatible with highly efficient restriction enzyme-free cloning of the genes as described before [27]. The vector pVMExp14367 was constructed by assembly of several components encoding T7 promoter-lac operator, H10 tag, TEV protease cleavage site, ampicillin resistance marker; beta-lactamase gene, and ColE1 ori from a set of intermediate vectors, and synthetic DNA sequences available in the laboratory. SacR-SacB gene cassette was obtained as a synthetic gene from Geneart (Thermo Fisher Scientific, Waltham, US), and BAP tag was assembled using duplex of oligonucleotides encoding 15 amino acid sequence. The final recombinant was sequenced using ABI 3730 XL DNA sequencing platform (Applied Biosystems, Thermo Fisher Scientific, Waltham, USA).

Cloning of five M. tuberculosis genes in the expression vector pVMExp14367

DNA encoding five M. tuberculosis H37Rv proteins, namely, MTC28 (Rv0040c), MPT63 (Rv1926c), MPT64 (Rv1980c), Ag85A (Rv3804c), and Ag85B (Rv1886c) was amplified from respective templates (S1A Fig) available in the laboratory using 5’ and 3’ gene-specific primers carrying 7 base extensions to append sequences required for restriction enzyme-free cloning [27]. For cloning, the preparation of inserts was performed using “column method” as described before [27]. The BsaI-digested linearized vector was also prepared as described before [27]. The ligation reaction was set up in 10 μl reaction volume containing 1.0 μl of the T4 DNA polymerase-treated insert (~ 50 ng), 50 ng BsaI-digested linearized and purified pVMExp14367 vector, and 200 U of T4 DNA Ligase (400 U/μl) in 1 x ligation buffer. The ligation reaction was carried out at 16°C for 1 hr and 37°C for 1 hr followed by inactivation of T4 DNA ligase at 65°C for 10 min. The ligation reaction (1 μl) was diluted to 10 μl in distilled water, and 500 pg equivalent ligated vector was electroporated in 25 μl electrocompetent E. coli BL21 (DE3) RIL cells (electroporation efficiency ~ 5 x 10⁸/μg pGEM DNA), and plated on MDAG plates containing 100 μg/ml ampicillin and 30 μg/ml chloramphenicol (MDAGAmp₁₀₀Cm₃₀). Three transformants of each construct were checked by sequencing on ABI 3730 XL sequencing platform (Applied Biosystems, Thermo Fisher Scientific, Waltham, USA), and clones with correct sequence were chosen for expression studies.

Expression and localization of the proteins

Small-scale expression and localization was carried out to determine the yield of soluble protein. For this, protein expression was performed by auto-induction in 50 ml ZYM5052 media [28]. The clones from MDAGAmp₁₀₀Cm₃₀ plates were inoculated in 3 ml MDAGAmp₁₀₀Cm₃₀ liquid media (primary culture) and were grown at 30°C, 250 rpm for 18 hr. The primary culture was diluted 100 fold in 50 ml ZYM5052 media containing 100 μg/ml ampicillin and 30 μg/ml chloramphenicol (ZYM5052Amp₁₀₀Cm₃₀) and grown with shaking at 250 rpm at 30°C for 2 hr, 24°C for 4 hr, and 18°C for 16 hr. The cells were harvested and re-suspended in 25 ml of 1 x TLB (50 mM Tris-HCl, pH 7.5 buffer containing 500 mM NaCl) supplemented with lysozyme at a final concentration of 200 μg/ml and PMSF at a final concentration of 0.1 mM, and incubated on ice for 30 min. Cells were lysed using sonication (Misonix, Model XL 3000). The lysate was centrifuged at 22,000g for 30 min at 4°C to obtain the supernatant (HSS; High Speed Supernatant), which was further centrifuged at 50,000g for 2 hr at 4°C, and the resulting supernatant containing soluble proteins was named as HHSS (High-High Speed Supernatant). For localization of the recombinant proteins, different sub-cellular fractions were analyzed on 0.1% SDS—8–20% polyacrylamide gradient gel under reducing conditions.

Based on the yield of soluble recombinant proteins in HHSS as determined by Coomassie brilliant blue R-250 dye stained gel, an appropriate volume of preparative-scale culture was set up for all the five proteins under the same expression conditions as described above. After auto-induction, the cells were harvested and resuspended in half the volume of 1 x TLB (without the addition of lysozyme and PMSF), followed by homogenization of cells using PANDA high-pressure homogenizer (GEA Niro Soavi) at 800 bars for 6 cycles as per the manufacturer’s instructions. The lysate was centrifuged, and HHSS was prepared as described for the small-scale expression. The HHSS fraction containing soluble proteins was then subjected to three-step chromatography protocol to obtain purified protein preparations.

Purification of the proteins

The entire purification procedure was performed at 4–8°C using appropriate chromatography columns attached to the AKTA Explorer 100 system (GE Healthcare Life Sciences, Uppsala, Sweden). During chromatography, the eluted protein was monitored using absorbance at 280 nm; fractions (of appropriate volumes depending on the type of chromatography) were collected and analyzed using SDS-PAGE under reducing conditions. Based on the purity and yield of the proteins, fractions were pooled.

For the purification of H10-T-MTC28-BAP protein, 280 ml HHSS containing approximately 330 mg of recombinant protein in 1 x TLB with 20 mM imidazole was filtered through 0.45 μ membrane, and applied on 20 ml Ni Sepharose Fast Flow (NiFF) resin packed in HR16/10 column (pre-equilibrated with 1 x TLB containing 20 mM imidazole) at a flow rate of 3 ml/min. After loading of the sample, the column was washed with 60 ml (3 CV) of 1 x TLB containing 20 mM imidazole, followed by 160 ml (8 CV) of 1 x TLB containing 50 mM imidazole at a flow rate of 3.5 ml/min. Finally, the bound protein was eluted with 1 x TLB containing 300 mM imidazole at 2 ml/min and 2 ml fractions were collected. The fractions were analyzed using 0.1% SDS-12.5% PAGE and those containing desired protein were pooled (NiFF pool).

The NiFF pool was further purified by gel-filtration chromatography on 480 ml Superdex 75 column pre-equilibrated with 20 mM Tris-HCl, pH 8.0 containing 50 mM NaCl. The NiFF pool was loaded at a flow rate of 3 ml/min, the column was subsequently developed at a flow rate of 3 ml/min, and 4 ml fractions were collected. The fractions were analyzed using 0.1% SDS-12.5% PAGE and those containing relatively pure protein were pooled (GFC pool).

Next, to obtain the protein devoid of N-terminal H10 tag (i.e., MTC28-BAP from H10-T-MTC28-BAP), the purified protein in the GFC pool was subjected to treatment with H6-TEV protease. For this, approximately 200 mg of H10-T-MTC28-BAP protein was subjected to digestion with 2 mg H6-TEV protease (~ 1:100 ratio of TEV protease:substrate; w/w) in a reaction volume of 45 ml in 1 x TEV reaction buffer (50 mM Tris-HCl, pH 8.0, 0.5 mM EDTA, 1 mM DTT) for 3 hr at 30°C. After completion of the reaction, sample equivalent to ~ 8 μg protein was analyzed on 0.1% SDS-12.5% PAGE to determine the extent of cleavage. After completion of the digestion process, 5 ml of 10 x loading supplement (2.5 M NaCl, 10 mM MgCl₂, 50 mM Tris-HCl pH 8.0, 200 mM imidazole) was added to the reaction. The resultant digestion mixture containing 20 mM imidazole in 1 x TLB, was applied on a 20 ml NiFF resin packed in HR16/10 column (pre-equilibrated with 1 x TLB containing 20 mM imidazole) at a flow rate of 1 ml/min. Flow-through fractions containing MTC28-BAP protein were pooled as NiFF pool after TEV protease Treatment (NiFF-TT pool).

Finally, the NiFF-TT pool containing MTC28-BAP protein was subjected to anion-exchange chromatography. For this, the NiFF-TT pool was desalted using Sephadex G25 (140 ml resin packed in XK 26/40 column), pre-equilibrated with 20 mM Tris-HCl, pH 8.0 (buffer A). For desalting the entire NiFF-TT pool (~ 60 ml), the same column was used twice (30 ml sample was desalted per run) with column re-equilibration in between the two runs. For anion-exchange chromatography, the desalted NiFF-TT pool (~ 80 ml) was applied on 20 ml Q Sepharose High Performance (QHP) resin (packed in HR16/10 column; pre-equilibrated with buffer A) at a flow rate of 4 ml/min. The QHP column was washed with 60 ml (3 CV) of buffer A, elution was performed with 100 ml (5 CV) linear gradient of 0–0.5 M NaCl in buffer A at 4 ml/min, and 2 ml fractions were collected. Fractions were analyzed using 0.1% SDS-12.5% PAGE and fractions containing pure protein were pooled (named as QHP pool). The protein concentration of QHP pool was estimated by measuring absorbance at 280 nm. The remaining four proteins were also purified employing the same process as described above for MTC28-BAP protein.

In vitro biotinylation of purified proteins using E. coli H10-BirA enzyme

For biotinylation of the proteins, the QHP pool containing purified C-terminal BAP-tagged protein was subjected to in vitro biotinylation using recombinant deca-histidine-tagged BirA enzyme (H10-BirA). For each protein, 10 mg of the QHP purified pool (~ 2 ml) was subjected to buffer-exchange using 13 ml Sephadex G25 (fine) syringe column equilibrated in 50 mM Tris-HCl, pH 8.0 buffer. The protein eluted in approximately 4.5–5 ml volume. The in vitro biotinylation reaction was set up in approximately 6 ml volume containing 10 mg of purified C-terminal BAP-tagged protein (substrate), with 10 mM ATP, 10 mM MgCl₂, 50 μM D-biotin, and 0.2 mg of recombinant H10-BirA enzyme (to achieve enzyme to substrate ratio of 1:50, w/w), and incubated at 30°C for 3 hr. After completion, 660 μl of 10 x loading supplement (2.5 M NaCl, 10 mM MgCl₂, 50 mM Tris-HCl pH 8.0, 200 mM imidazole) was added to the reaction and applied on 1.2 ml NiFF column at a flow rate of 0.2 ml/min and 0.5 ml flow-through fractions were collected. This allowed binding of H10-BirA enzyme to the column, and the biotinylated protein collected in the flow-through fractions. The flow-through fractions were pooled and final volume was reduced to approximately 2 ml using amicon-ultra-4 centrifugal filter unit (10 kDa Cut Off; Merck Millipore) followed by desalting on 13 ml Sephadex G25 (fine) syringe column in 1 x PBS buffer (20 mM phosphate, pH 7.5 containing 150 mM NaCl) to obtain purified biotinylated proteins. The protein concentration was estimated by measuring absorbance at 280 nm. Purified biotinylated proteins were analyzed using 0.1% SDS-12.5% PAGE under reducing conditions.

Determination of the extent of biotinylation in biotin-tagged proteins

To determine the extent of biotinylation, 25 μg of each biotin-tagged protein was adsorbed on 50 μl Streptavidin Sepharose HP beads (taken in 3–5 fold excess based on the binding capacity to ensure complete adsorption) by constant mixing for 1 hr at 25°C in a 2 ml tube (click cap, clear, round bottom microtubes, Treff Lab, Switzerland). The supernatant (labeled as ‘after’ adsorption fraction) was collected after centrifugation of mixture at 12,000g for 5 min at 25°C. Separately, a similarly diluted sample of each biotin-tagged protein was prepared and labeled as ‘before’ adsorption fraction. To determine the amount of protein remaining after adsorption, both ‘before’ and ‘after’ adsorption fractions were tested using indirect ELISA with protein-specific monoclonal antibodies to determine the amount of protein. For this, 384 well Nunc Immobilizer streptavidin-coated plate was washed thrice with PBST (1 x PBS with 0.05% Tween-20), and coated with 25 μl each of 7 point 3-fold dilutions of ‘before’ (range ~ 1:100–1:100K) and ‘after’ (range ~ 1:10–1:10K) adsorption fractions (prepared in 1 x PBS) for 2 hr at 25°C. Wells were washed thrice with PBST and blocked with 2% BSA-PBST for 1 hr at 25°C. After blocking, plate was washed thrice with PBST and proteins were probed with 100 ng/ml of respective protein-specific monoclonal antibodies (MAb MTC28-13 for MTC28, MAb MPT63-03 for MPT63, MAb MPT64-33 for MPT64, and MAb Ag85-12 for Ag85A and Ag85B) diluted in 0.1% BSA-PBST for 1 hr at 25°C. The plate was washed thrice with PBST, and HRP-conjugated Goat anti-Mouse IgG (H+L) antibody diluted 1:5000 times in 0.1% BSA-PBST was added for 1 hr at 25°C. Finally, after three washes each with PBST and 1 x PBS, the reaction was revealed by 25 μl TMB substrate (Seramun Diagnostics, Berlin, Germany). Following, incubation in the dark for 15 min at 25°C, the reaction was terminated by addition of 25 μl 1 N H₂SO₄ and absorbance was measured at 450 nm using ELISA plate reader (SpectraMax M5; Molecular Devices, Sunnyvale, CA, USA). The extent of biotinylation was calculated based on the fold reduction in the reactivity after adsorption on streptavidin beads.

Indirect ELISA with biotin-tagged antigens using mouse monoclonal or rabbit polyclonal antibodies

To compare the coating efficiency of biotinylated proteins individually, different dilutions of biotin-tagged proteins (7 point 3-fold dilutions; range ~ 1 μg/ml to ~ 1.3 ng/ml) were prepared in 1 x PBS, and 25 μl of each was added to either 384 well Nunc Immobilizer streptavidin-coated plate (pre-washed thrice with PBST) for 2 hr at 25°C, or 384 well Nunc Maxisorp polystyrene plate for 2 hr at 37°C. Wells were washed thrice with PBST and blocked with 2% BSA-PBST for 1 hr at 25°C. After blocking, the plates were washed thrice with PBST and the proteins were probed either with 100 ng/ml of protein-specific purified mouse monoclonal antibodies (MAb MTC28-13 for MTC28, MAb MPT63-03 for MPT63, MAb MPT64-33 for MPT64, MAb Ag85-14 for Ag85A, and MAb Ag85-11 for Ag85B), or with 100 ng/ml of protein-specific purified rabbit polyclonal antibodies (both diluted in 0.1% BSA-PBST) for 1 hr at 25°C. Following this, the plates were washed thrice with PBST, and HRP-conjugated Goat anti-Mouse IgG (H+L) (diluted 1:5000 times in 0.1% BSA-PBST), or HRP-conjugated Goat anti-Rabbit IgG (H+L) antibodies (diluted 1:10,000 times in 0.1% BSA-PBST) were added for 1 hr at 25°C to probe the bound mouse monoclonal or rabbit polyclonal antibodies, respectively. Remaining steps were performed as described above.

To compare the coating efficiency of a mixture of biotin-tagged proteins, 10 μg each of 5 proteins (MTC28-Bio, MPT63-Bio, MPT64-Bio, Ag85A-Bio and Ag85B-Bio) was mixed (i.e., equal ratio, w/w), and 7 point 3-fold dilutions of resultant mixture (range ~ 5 μg/ml to ~ 7 ng/ml of the mixture, equivalent to ~ 1 μg/ml to ~ 1.3 ng/ml of individual protein) were captured on 384 well Nunc Immobilizer streptavidin-coated plate (pre-washed thrice with PBST) for 2 hr at 25°C, or 384 well Nunc Maxisorp polystyrene plate for 2 hr at 37°C. Remaining steps were performed as described above.

Solid-phase and in-solution affinity selection

Phage-displayed naïve human antibody library comprising of 10 billion clones was employed for affinity selection. For in-solution affinity selection, 4 x 10¹² phages from each of the eight phage sub-libraries were pooled, and the total of 3.2 x 10¹³ phages (~ 3200 fold excess of the library) in a volume of 1.6 ml was mixed with 1.6 ml 4% BSA-PBS. The phage library (3.2 ml) was divided into three tubes (2 ml, click cap, clear, round bottom microtubes, Treff Lab, Switzerland), and preadsorbed on MyOne streptavidin T1 beads (100 μl per tube, washed twice with 2% BSA-PBS) for 1 hr at 25°C on vertical rotator at 5 rpm for end-to-end mixing. After separation of streptavidin beads using magnetic particle concentrator, the preadsorbed library was collected in fresh tubes, and MTC28-Bio protein was added (at a final concentration of 100 nM per tube), followed by incubation for 2 hr at 25°C with rotation at 5 rpm. To each tube, 100 μl of pre-washed M280 streptavidin beads were added, followed by incubation for 30 min at 25°C with rotation at 5 rpm to capture biotinylated antigens along with bound phages. The beads were washed 10 times with PBST (1 x PBS with 0.05% Tween 20), and 10 times with 1 x PBS. Finally, the beads were incubated with 100 mM triethylamine (500 μl per tube) for 10 min at 25°C with rotation at 5 rpm, and phage eluate from 3 tubes was pooled. The eluate (total 1.5 ml) was neutralized with 750 μl 1 M Tris-HCl, pH 7.5 and an aliquot was used to determine the phage titer by infecting E. coli TOP10F’ cells. The remaining pool was used for infecting 20 ml E. coli TOP10F’ cells at 37°C for 30 min. After infection, the cells were harvested by centrifugation at 5000 rpm for 5 min at 25°C and re-suspended in 4 ml LB media. The cells were plated on 4 x 150 mm LBAmp₁₀₀Glu_1% plates (LB media containing 100 μg/ml ampicillin and 1% glucose) to select for phage transductants. Following day, the cells from 150 mm LBAmp₁₀₀Glu_1% plates were scraped in fresh 2 x YTGlu_1% media (2 x YT media containing 1% glucose), mixed with equal volume of 2 x glycerol storage solution (65% glycerol, 0.1 M Tris-HCl, pH 8.0, 25 mM MgSO₄), and stored at -80°C. An aliquot of cells (~ 5 x 10¹⁰ cells) was used for rescue of phages using AGM13 helper phage in 1200 ml volume. The phages were harvested and purified using double PEG-NaCl precipitation [29], and used for the subsequent round of affinity-selection. In total, three rounds of affinity selection were performed using the same protocol as described for the first round with appropriate changes.

For passive immobilization-based solid-phase affinity selection, 16 immunotubes were coated with 40 μg of H10-MTC28 protein (10 μg/ml in 4 ml 1 x PBS) for 16 hr at 4°C on a vertical rotator at 10 rpm for end-to-end coating, followed by 1 hr incubation at 37°C. One tube was kept as no-coat control and contained only 1 x PBS. The tubes were washed thrice with 1 x PBS (no Tween-20), followed by blocking with 2% SM-PBS (2% Skimmed Milk-PBS) for 2 hr at 25°C with rotation at 10 rpm. Blocked immunotubes were washed thrice with 1 x PBS. Separately, approximately 2.4 x 10¹³ phages were diluted in 2% SM-PBS to 64 ml, and pre-adsorbed on skimmed milk-coated immunotubes tubes for 2 hr at 25°C to eliminate milk and plastic binders. Pre-adsorbed phages were added to H10-MTC28-coated tubes followed by incubation for 2 hr at 25°C with rotation at 10 rpm. The unbound phages were removed by aspiration, followed by 10 washes each with 4 ml of PBST (1 x PBS containing 0.05% Tween-20), and 5 washes with 1 x PBS. The bound phages were eluted using 1 ml of 100 mM triethylamine (prepared fresh in distilled water) per tube by rotation on test tube roller for 10 min at 25°C. The eluate was neutralized with 500 μl of 1 M Tris-HCl, pH 7.5. The titer of eluates from the coat and no-coat tubes and input phages was determined by infecting phages in E. coli TOP10F’ cells at appropriate dilutions. The eluate from the coated tube was processed as described for in-solution selection. In total, three rounds of selection were performed using the same protocol as described for the first round with appropriate changes.

Construction of the vector pVMExp14367

A T7 promoter-lac operator-based IPTG/lactose inducible vector pVMExp14367 was employed for expression of recombinant proteins in E. coli carrying N-terminal deca-histidine tag (H10) followed by TEV protease site and C-terminal Biotin Acceptor Peptide (BAP) tag with appropriate glycine-serine rich spacers (Fig 1). This format allows for affinity-based purification of the expressed recombinant proteins using the H10 tag, and its subsequent specific removal by virtue of the TEV protease site to obtain proteins devoid of the N-terminal tag. The C-terminal BAP tag allows for in vitro site-specific biotinylation of the proteins using recombinant E. coli BirA enzyme. The vector design is compatible with highly efficient and high-throughput method for restriction enzyme-free cloning of the genes [27]. In this strategy, the vector is prepared by digestion with Type IIs restriction enzyme BsaI, whose two recognition sites are present in the vector flanking the stuffer in two orientations (Fig 1B) in a manner that allows generation of 4 base 5’-overhangs. The insert for cloning is prepared by PCR amplification using primers carrying 20–23 base long gene-specific sequence with 7 base long additional sequence (5’-CGGCACC-3’ in forward primer and 5’-CTCCACC-3’ in reverse primer) [27]. The resulting insert is then subjected to treatment with T4 DNA polymerase in the presence of dTTP to generate desired 4 base 5’-overhangs compatible with the vector (shown in bold; Fig 1D and 1E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Diagrammatic representation of T7 promoter-lac operator-based pVMExp14367 expression vector to obtain recombinant proteins with N-terminal H10-TEV and C-terminal BAP tag.

Only relevant genes and restriction sites are shown. The maps are not to scale. T7lac, T7 promoter-lac operator; RBS, ribosome-binding site; H10, deca-histidine tag; TEV, Tobacco Etch Virus protease cleavage site; S, glycine-serine rich spacers; SacR-SacB, 2.0 kbp SacR-SacB gene cassette flanked by BsaI sites; BAP, Biotin Acceptor Peptide; T7Tn, T7 transcription terminator; f ori, origin of replication of filamentous phage; Amp^r, beta-lactamase gene; ColE1 ori, origin of replication. The amino acids encoded are shown in single letter code (bold) above the nucleotide sequence (A-D). (A-C) The sequence of the important components of vector including 2 BsaI cloning sites. (D) Sequence flanking the gene of interest (GOI) after PCR amplification. (E) GOI carrying 5’- 4 base overhangs generated after treatment with T4 DNA polymerase in the presence of dTTP.

https://doi.org/10.1371/journal.pone.0191315.g001

Five M. tuberculosis H37Rv secretory proteins, namely, MTC28 (Rv0040c), MPT63 (Rv1926c), MPT64 (Rv1980c), Ag85A (Rv3804c), and Ag85B (Rv1886c) were chosen as model proteins for this study. The DNA encoding these proteins (without signal sequence) was cloned in the vector pVMExp14367. Transformants in E. coli host BL21 (DE3) RIL were screened by sequencing and 100% cloning efficiency was observed (S1C Fig).

Expression and localization of the proteins

Protein expression was performed using auto-induction process in ZYM5052 media [28] at low-temperature conditions that promote solubility of the proteins. All the five proteins showed good expression in total cell fractions (S2 Fig, Lane A), and constituted ~ 15–30% of the total cellular protein. However, the yield of soluble protein in the HHSS cytosolic fraction varied among different proteins (S2 Fig, Lane D). MTC28, MPT63, and MPT64 proteins were nearly 100% soluble, with almost the entire fraction of expressed protein present in the HHSS fraction, whereas the amount of Ag85A and Ag85B proteins in the HHSS fraction was only ~ 5% of the total expressed protein (S2 Fig, Lane D). Based on the yield of the soluble proteins, appropriate volumes of auto-induced cultures were processed (Table 1, Column 1), and the 2 x HHSS fraction containing soluble protein was subjected to purification.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of purification of five recombinant biotin-tagged proteins.

https://doi.org/10.1371/journal.pone.0191315.t001

Purification of the proteins

The recombinant proteins carried H10 tag followed by TEV protease site at the N-terminus, and BAP tag at the C-terminus (i.e., H10-T-POI-BAP; POI-Protein of interest). To purify proteins, a streamlined workflow was developed (Fig 2), which involved two-step purification including Ni-affinity chromatography and gel-filtration chromatography. The purified monomeric/dimeric fraction of the proteins obtained after gel-filtration chromatography was then subjected to treatment with H6-TEV protease to cleave the N-terminal H10 tag, followed by removal of TEV protease and cleaved tag. The proteins (devoid of H10 tag) were finally purified using anion-exchange chromatography and then subjected to in vitro biotinylation using H10-BirA enzyme and purification of the fully biotinylated protein. The purification of the representative protein MTC28 is described in detail.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Workflow for expression, purification, and in vitro biotinylation of five mycobacterial proteins.

https://doi.org/10.1371/journal.pone.0191315.g002

For the purification of H10-T-MTC28-BAP, HHSS containing approximately 330 mg of recombinant protein was applied on Ni Sepharose Fast Flow (NiFF) column in the presence of 20 mM imidazole to allow specific binding of the H10-tagged recombinant protein. The column was washed with 50 mM imidazole to remove loosely bound non-specific proteins followed by the elution of the desired protein in the presence of 300 mM imidazole (Fig 3A). Based on the SDS-PAGE analysis, the extent of purity was >90% in the NiFF pool, and it contained a small amount of both low and high molecular weight contaminants (Fig 3D, Lane 3). The NiFF pool was further subjected to gel-filtration chromatography on Superdex 75 column. The protein eluted between 0.3–0.5 CV in two peaks, which were not well-resolved (Fig 3B). Based on the elution volume of the peaks and that of gel-filtration calibration standards, it was estimated that peak I comprised of very high molecular weight proteins (possibly containing aggregated H10-T-MTC28-BAP or other high molecular weight contaminants), and peak II comprised of ~ 60–70 kDa molecular weight protein. The fractions comprising both the peaks were analyzed using SDS-PAGE under reducing conditions, and peak I was found to contain high molecular weight contaminants (which were also observed in NiFF pool), whereas later part of Peak II contained relatively pure H10-T-MTC28-BAP free of any visible contaminants and was pooled (GFC pool) for further processing (S3A Fig). Since, the molecular weight of the protein in peak II corresponded to ~ 60–70 kDa, we concluded that the H10-T-MTC28-BAP protein (molecular weight 34.2 kDa) was present as a dimer.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Purification of MTC28-BAP protein.

Chromatogram showing (A) Elution profile of H10-T-MTC28-BAP protein on Ni Sepharose Fast Flow (NiFF) affinity column. Fraction numbers 6–17 were pooled (NiFF pool). (B) Elution profile of H10-T-MTC28-BAP protein on Superdex 75 gel-filtration column. Fraction numbers 16–24 were pooled (GFC pool). The GFC pool was treated with H6-TEV protease to cleave H10 tag from the protein followed by removal of cleaved tag and H6-TEV protease using Ni-affinity chromatography. (C) Elution profile of MTC28-BAP protein on Q Sepharose HP column. Fraction numbers 22–27 were pooled (QHP pool). (D) SDS-PAGE analysis of H10-T-MTC28-BAP protein at different stages during purification. The samples were analyzed by 0.1% SDS-12.5% PAGE under reducing conditions. The protein bands were visualized with Coomassie brilliant blue R-250 staining. Lane M, molecular weight marker, broad range (Bio-Rad, Hercules, CA) (shown in kDa); Lane 1, total cell after homogenization; Lane 2, High-High Speed Supernatant; Lane 3, NiFF pool; Lane 4, GFC pool; Lane 5, NiFF-TT pool (after desalting); Lane 6, QHP pool.

https://doi.org/10.1371/journal.pone.0191315.g003

For the removal of H10 tag from purified H10-T-MTC28-BAP protein, the GFC pool was treated with H6-TEV protease (Table 1, Column 4). Following digestion, the reaction mixture was applied on NiFF column to remove H6-TEV protease, cleaved H10 tag, and uncleaved H10-T-MTC28-BAP protein (if any), and the desired MTC28-BAP protein was obtained in the flow-through (NiFF-TT pool). The MTC28-BAP protein in NiFF-TT pool was finally purified using anion-exchange chromatography on QHP column. The protein eluted as one major peak (Fig 3C) and contained highly pure protein (QHP pool; Fig 3D, Lane 6). Approximately 112 mg highly purified MTC28-BAP protein was obtained from 560 ml culture (Table 1, Column 7).

H10-T-MPT63-BAP, H10-T-MPT64-BAP, H10-T-Ag85A-BAP, and H10-T-Ag85B-BAP proteins were purified following a similar procedure as described for H10-T-MTC28-BAP protein. Based on the yield of proteins in the soluble fraction, different volumes of HHSS were processed for NiFF chromatography (Table 1, Columns 1, 2). For H10-T-MPT63-BAP, and H10-T-MPT64-BAP proteins, SDS-PAGE analysis after NiFF chromatography revealed that the extent of purity was >90%, and NiFF pool contained small amounts of both low and high molecular weight contaminants (S4D and S5D Figs, Lane 3). However, in the case of H10-T-Ag85A-BAP and H10-T-Ag85B-BAP proteins, a distinct band at approximately 60 kDa was observed as a major contaminant after NiFF chromatography (S6D and S7D Figs, Lane 3). This band could be a chaperone like GroEL [30], which is often found to be associated with unfolded/partially proteins. Since these two proteins, namely, Ag85A and Ag85B are sparingly soluble in E. coli cytosol (S2 Fig, Lane D), it is likely that these could be only partially folded and might have a tendency to remain associated with molecular chaperones. Thus, we incorporated a gel-filtration chromatography step in our purification pipeline to obtain a highly pure monomeric form of proteins. The NiFF pool of these proteins was purified using gel-filtration chromatography on Superdex 200 (for H10-T-MPT63-BAP and H10-T-MPT64-BAP proteins) or Superdex 75 (for H10-T-Ag85A-BAP and H10-T-Ag85B-BAP proteins) column. All four proteins eluted as two sharp and well-separated peaks (S4B–S7B Figs). For all the four proteins, SDS-PAGE analysis of peak I and peak II obtained in analytical-scale gel-filtration chromatography of the NiFF pool showed that peak I contained some high molecular weight contaminants, whereas Peak II contained pure form of the desired proteins free of any visible contaminants (S3B–S3F Fig). Based on the elution volume of the peaks and that of gel-filtration calibration standards, it was estimated that for all the four proteins, the peak II comprised of monomeric form of respective proteins and was pooled for further processing (GFC pool; S4D–S7D Figs, Lane 4). The GFC pools were treated with H6-TEV protease followed by its removal (as described before for H10-T-MTC28-BAP protein) to obtain proteins devoid of N-terminal H10 tag (Table 1, Columns 5, 6, and S4D–S7D Figs, Lane 5). These proteins were further purified using anion-exchange chromatography on QHP column. All the four proteins showed elution profile with one major peak (S4C–S7C Figs). SDS-PAGE analysis showed that fraction number 27–32 for MPT63-BAP, 24–29 for MPT64-BAP, 31–37 for Ag85A-BAP, and 31–35 for Ag85B-BAP contained highly pure proteins, and were pooled (QHP pool, S4D–S7D Figs, Lane 6). The final yield of each purified protein is summarized in Table 1, Column 7.

In vitro biotinylation of the purified proteins using H10-BirA enzyme

To perform biotinylation of the purified C-terminal BAP-tagged proteins, the QHP pool was subjected to in vitro biotinylation using H10-BirA enzyme in the presence of biotin and ATP. For this, the proteins were desalted in the appropriate buffer, and treated with H10-BirA. Following biotinylation, the H10-BirA enzyme was removed using NiFF column and the flow-through fraction containing biotinylated protein was desalted. The SDS-PAGE analysis revealed that integrity of proteins was maintained during biotinylation process (S8 Fig).

Determination of the extent of biotinylation in biotin-tagged proteins

As a measure of the efficiency of the in vitro biotinylation reaction, the extent of biotinylation in biotin-tagged proteins was determined. As described in methods, a fixed amount of biotin-tagged proteins was adsorbed on Streptavidin Sepharose HP resin, and total immunoreactive protein was estimated in ‘before’ and ‘after’ adsorption fractions using indirect ELISA on Nunc Immobilizer streptavidin-coated plates by protein-specific monoclonal antibodies. Based on the fold reduction in the reactivity of fractions after adsorption on streptavidin beads (a measure of remaining protein amount), the extent of biotinylation was determined and was found to be greater than 99.5% for all the five biotin-tagged proteins (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Determination of the extent of biotinylation of proteins after in vitro biotinylation using recombinant H10-BirA enzyme.

https://doi.org/10.1371/journal.pone.0191315.t002

Biotin-tagged proteins are superior coating molecules for ELISA

Recombinant immunodominant proteins of infectious organisms are routinely employed in ELISA for the detection of disease-specific antibodies in the body fluids of suspected patients [5, 6, 17, 31–33]. However, immobilization of these proteins on the surface of ELISA plates can be variable and inefficient [11]. We decided to evaluate the performance of biotin-tagged proteins in ELISA using immobilization by passive adsorption on the polystyrene surface, and specific capture on the streptavidin-coated surface. For this, different concentrations of biotin-tagged proteins were coated on the two types of microtiter plate surfaces, and the captured proteins were detected using indirect ELISA with either specific mouse monoclonal antibodies, or rabbit polyclonal antibodies. In this assay, the amount of protein added for coating on the two surfaces to produce equal reactivity was compared.

When detected with monoclonal antibodies, the assay required approximately 5–660 fold less protein to produce comparable signal in ELISA upon specific capture using streptavidin-coated plates, which is due to the efficient capture of biotin-tagged proteins on the streptavidin-coated surface (Fig 4A). This was especially evident in the case of Ag85A-Bio and Ag85B-Bio proteins, where for passive adsorption, much higher protein concentration (greater than 1000 ng/ml) was required to produce a significant signal (A_450nm of ~ 1.0) in comparison to MTC28-Bio, MPT63-Bio, and MPT64-Bio proteins, which were required in much lower concentrations (approximately 100–200 ng/ml) (Fig 4A). This suggested that Ag85A-Bio and Ag85B-Bio proteins tend to coat poorly on the polystyrene surface during passive adsorption. However, upon specific capture on the streptavidin-coated plates, both Ag85A-Bio and Ag85B-Bio proteins produced a significant signal at lower protein concentrations (approximately 1–30 ng/ml), which was comparable to other biotin-tagged proteins (approximately 1–10 ng/ml) (Fig 4A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Comparison of the ELISA reactivity of proteins at different coating concentrations upon passive and specific immobilization.

The five biotin-tagged proteins were immobilized individually or as a mixture (equal, w/w) on Nunc Immobilizer streptavidin-coated (SA) or Nunc Maxisorp (MS) plates at different concentrations. The bound protein was probed with specific mouse monoclonal (MAb) or rabbit polyclonal antibodies (PAb), followed by detection using HRP-conjugated Goat anti-Mouse IgG (H+L) antibody or Goat anti-Rabbit IgG (H+L) antibody. (A) Detection of five biotin-tagged proteins immobilized individually on SA or MS plates with MAb. (B) Detection of five biotin-tagged proteins immobilized individually on SA or MS plates with PAb. (C) Detection of five biotin-tagged proteins immobilized as a mixture on SA or MS plates with MAb. (D) Detection of five biotin-tagged proteins immobilized as a mixture on SA or MS plates with PAb. Identity of the protein and the approximate fold change in the amount of protein required to produce equal reactivity (calculated as the ratio of absorbance readings from the linear region of the graphs of specific versus passive immobilization) is indicated inside the graph panel in red color. The protein concentration values on X-axis for (C) and (D) correspond to the concentration of individual proteins in the mixture. ELISA values shown are mean ± SD of three independent experiments. ND, Not Determined.

https://doi.org/10.1371/journal.pone.0191315.g004

When the detection in ELISA was performed using polyclonal antibodies, efficient protein-coating was observed upon specific capture on the streptavidin-coated plates (Fig 4B). However, the difference between specific and passive immobilization was less pronounced (approximately 3–40 fold) here as compared to the assay performed using monoclonal antibodies (Fig 4A versus 4B). This difference could likely be due to the higher affinity and polyclonal nature of the rabbit antibodies [34–36]. The monoclonal antibodies bind to only one site on the protein (the epitope), and the same may not be accessible on every passively coated molecule, whereas, the polyclonal antibodies bind to multiple epitopes on the same molecule, and thus, even upon passive coating, a larger number of epitopes could be accessible. In contrast, on streptavidin-coated plates, all the epitopes are likely to be exposed, except for those, which are present at the C-terminus of the protein; therefore, even monoclonal antibodies show significant binding.

To mimic the assay conditions, where the coating of proteins as a mixture may be necessary, we also compared the efficiency of specific and passive immobilization of each protein in a mixture. In this case also, the specific immobilization resulted in reduced requirement of proteins to produce comparable reactivity in ELISA upon detection both mouse monoclonal antibodies (~ 6–245 fold), and rabbit polyclonal antibodies (~ 3–26 fold) (Fig 4C and 4D).

This property of efficient capture of biotinylated proteins using streptavidin-coated plates can be used as an effective strategy to improve the ELISA sensitivity, particularly for the proteins like Ag85A and Ag85B that exhibit poor coating characteristics upon passive immobilization. This finding is particularly relevant for developing ELISA-based assays for antibody detection, where it might be necessary to coat multiple proteins with variable surface binding characteristics.

Characterization of biotin-tagged proteins using phage display-based affinity selection

We compared the efficiency for isolating specific antibody binders from a phage-displayed naive human antibody library (10 billion clones in scFv format) against recombinant H10-MTC28 protein by its passive immobilization on Nunc immunotubes or against biotin-tagged MTC28-Bio protein by its specific capture on streptavidin-coated M280 magnetic beads. Three rounds of solid-phase affinity selection were performed on H10-MTC28 protein coated passively on immunotubes at 10 μg/ml, and 48 transductants were screened by phage ELISA. All 48 clones (100%) showed reactivity to MTC28 protein indicating enrichment of specific binders (Table 3). Similarly, three rounds of in-solution affinity selection were performed on biotin-tagged MTC28-Bio at 100 nM concentration (~ 3 μg/ml), and 24 transductants were screened after the third round of selection. All 24 clones (100%) were found to be reactive to the target protein (Table 3). Though the efficiency of the two selection procedures in terms of yielding positive clones after the third round of affinity selection was same, it is worth noting that in-solution selection led to the higher recovery of phages during selection (Table 3, output/input phage ratio). This could have been due to a higher degree of non-specific selection of phages during in-solution selection, but screening of transductants even after a second round of in-solution selection revealed that ~ 75% clones were reactive to the target protein. This confirmed that the higher phage recovery numbers are not due to non-specific enrichment of phages, but should be due to improved availability of the bait. Most importantly, during in-solution affinity selections, the concentration of bait can be carefully controlled and usually reduced in subsequent rounds of selection to obtain binders with higher affinities. In addition, the amount of protein required for specific capture-based in-solution affinity selection was much lower (less than 50 micrograms) as compared to the amount required for passive immobilization-based solid-phase affinity selection (few milligrams).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary of three rounds of solid-phase and in-solution affinity selection of human naïve scFv library on MTC28 antigen.

https://doi.org/10.1371/journal.pone.0191315.t003