For ELISA and immunoblotting tests, detection of recombinant proteins was performed using a specific mouse anti-6× His monoclonal antibody (R&D biosystems). Biotinylated proteins were detected with a polyclonal anti-biotin antibody (Bethyl Laboratories Inc.). Free and fusion sfGFP proteins were detected using anti-GFP polyclonal antibody (Al-Homsi et al., 2012a). Subsequent detection of rabbit or mouse antisera in ELISA or immunoblotting was performed with anti-rabbit or anti-mouse conjugated to horseradish peroxidase (HRP) or to alkaline phosphatase (AP) (Bethyl Laboratories Inc.). Soluble sfGFP was expressed and purified from E. coli BL21-GOLD (DE3) (Stratagene) containing pRSET-sfGFP plasmid according to a previously described method (Al-Homsi et al., 2012a).

Total RNA was extracted from the hepatoma cell line (HepG2) using Illustra RNAspin Mini Kit (GE Life Sciences) following the manufacturer's instructions. Two microgram of RNA were reverse-transcribed to cDNA using Ready-to-Go You-prime first-strand-beads (GE Life Sciences) with oligo-dT¹⁵⁻¹⁸ (Invitrogen). Two microgram of cDNA were used as a template in a PCR with a pair of annexin V specific primers; ANXV-F and ANXV-R (EcoRI/HindIII) (Supporting Material, Table S1). The cDNA was amplified by a high fidelity Taq DNA polymerase (AccuPrime™Kit; Invitrogen) at 55°C annealing temperature resulting in the amplification of 1031 bp DNA fragment. This fragment was cloned in pRSET-S50 (un published data) and pRSET-sfGFP-KMP11 (Meriee et al., 2014) plasmids, using NcoI and EcoRI restriction enzymes in order to construct pRSET-ANXV and pRSET-sfGFP-ANXV plasmid constructs, respectively (Supporting Material, Figure S1). Plasmid constructs were confirmed by sequencing and used to transform a freshly prepared electro-competent E. coli BL21-GOLD (DE3) cells by electroporation, then cells were grown in LB/Amp plate. A single positive colony from the plate was inoculated into 250 mL LB/Amp medium. The culture was grown at 37°C until the OD₆₀₀ reached 0.5–0.7 then IPTG (isopropyl β-D-thiogalactoside, 0.1 mM, Promega) was added to the culture for the induction of protein expression, and the culture was grown overnight at 19°C.

Cells from induced cultures were pelleted by centrifugation and suspended in binding buffer (20 mM Tris-base, 300 mM NaCl, 20 mM imidazole and pH 7.4) then lysed by French–Press (pressure 1.4 bar), centrifuged for 8 min at 10,000 × g to recover the supernatant containing the recombinant proteins (ANXV, sfGFP-ANXV, and sfGFP). Using fast protein liquid chromatography (FPLC) AKTAprime plus system (GE Healthcare), recombinant proteins were purified from the cytoplasmic extracts using a 5 mL nickel-charged column (GE Healthcare). After washing, bound proteins were eluted from the column using elution buffer (20 mM Tris-base, 300 mM NaCl, and 500 mM imidazole). The eluted fractions were concentrated by Vivaspin concentrators with a molecular mass cutoff of 10 kDa (Vivascience). Purified proteins concentrations were determined by Bradford method and adjusted to 1 mg/mL for prolonged storage at −20°C.

Different proteins were labeled with biotin (Biotin N-hydroxysuccinimide ester, BNHS, Sigma) and ANXV was labeled with the fluorescein (6-[Fluorescein-5(6)-carboxamido] hexanoic acid N-hydroxysuccinimide ester, FNHS, λ_ex: 490 nm/λ_em: 514 nm, Sigma), using standard procedure for amine coupling provided by the manufacturer's instructions. The fluorescence, expressed as a relative fluorescent unit (RFU), of the two sfGFP derivatives was monitored by measuring it at different pairs of wavelengths for excitation (λ_ex) and emission (λ_em) using Fluoroskan Ascent® microplate reader (Thermo Labsystems), and the free fluorophore FNHS was used as a control.

SDS-PAGE was performed using Bio-Rad mini-Protean II system following the manufacturer's instructions. Gels were prepared using stacking and running gels 5 and 12% respectively. After electrophoresis, the gel was stained with coomassie blue for 2 h followed by destaining in 5% acetic acid and 10% methanol. For immunoblotting, separation was carried out using 0.25 μg of pure proteins, which then were blotted onto 0.45 μm nitrocellulose membranes (BioRad) using 1× blotting buffer (25 mM Tris-base, 200 mM glycine, 0.1% SDS, and 20% methanol). After blocking with PBS containing 5% skimmed milk, membranes were incubated with the indicated dilutions of the primary antibodies for 1 h at room temperature. After several washes with PBS containing 0.05% Tween-20, separated proteins were detected by an anti-6× His (1:2,000), anti-GFP (1:3,000), and anti-Biotin (1:2,000) antibodies as indicated. Finally, blots were incubated with the secondary antibodies conjugated to AP (1:2,000) for 1 h at room temperature. Band revelation was achieved by adding chromogen substrate (0.05% NBT and 0.025% BCIP; Sigma) in AP buffer (100 mM tris-base, 100 mM NaCl, 5 mM MgCl₂, pH 9.5).

Phospholipids were extracted from 1 g of soya bean (~30 mg PL) or egg yolk (~100 mg PL). Tissues were homogenized in 4 mL methanol before adding 8 mL chloroform then mixing for 30 min on a rotary shaker at 4°C in glass tubes. After centrifugation at 800 × g for 3 min, solvents were filtrated on blotting paper and washed with 2.5 mL NaCl 0.9%. Another centrifugation at 800 × g for 3 min was required to separate the phases before eliminating all of the upper inorganic phase. Finally, PL was dried from chloroform under a stream of N2 for 1 h, then in vacuum rotation overnight. Phospholipids (30 μg/mL) were immobilized onto the wells of polystyrene microtiter plates (TPP) or PVDF membrane (BioRad) in methanol: chloroform mixture (4:1). Wells or membranes were blocked with 5% skimmed milk in HBS (10 mM Hepes, 150 mM NaCl, 2.5 mM CaCl2, pH 7.4), washed with HBS and then allowed to bind purified and biotinylated ANXV, sfGFP-ANXV, and sfGFP (5 μg/mL) in HBS for 1 h at 22°C. After washing with HBS, bound proteins were detected with anti-biotin or anti-GFP antibodies.

Dried PL were suspended in HBS at a final PL concentration of 5 mM (~3.75 mg/mL, MW: ~775). Sonication was performed in ice for 2 min, pulse 15 s/repose 45 s and amplitude 35% using 750 Watt Ultrasonic Processor (Cole Parmer). Then the mixture was centrifuged at 10 k × g for 30 min. Finally, the milky supernatant was filtrated five times through 0.22 μm syringe filter. The size of the prepared liposomes was analyzed using Scanning Electron Microscope (SEM). Liposomes samples were coated with carbon using (K975X turbo evaporator), then the coated samples were observed and examined with the help of SEM and photographs were taken of different liposomes sources.

Annexin V binding to PL was investigated using SR7000DC biosensor (Xantec). Phospholipid liposomes (0.1 mM) in HBS-T buffer (0.005% Tween-20) were passed over flow cells of a research grade biosensor chip coated with phytosphingosine (Xantec) at a flow rate of 50 μl/min forming a lipid bilayer surface according to the manufacturer's instructions. The target amount of immobilization was 1,000 response units (RU). The high binding capacity of liposomes was demonstrated by consecutive injection of five aliquots (100 μl) of liposomes (100 μM) over the left flow cells and followed by a dissociation step of 200 s. All binding experiments were performed at 25°C in HBS-T buffer at a flow rate of 50 μl/min. ANXV, sfGFP-ANXV, and sfGFP, were diluted in HBS-T buffer to obtain concentrations of 100 nM. Then they were injected over immobilized PL and control channels for 120 s (association phase), followed by a dissociation phase of 240 s (buffer alone). The sensorgrams were fitted by subtracting the signal from the reference flow cell and were globally treated using Scrubber 2 software (www.cores.utah.edu). The cells were regenerated with 5 mM EDTA in HBS-T between cycles.

HeLa cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), and 100 U/mL of penicillin and 100 mg/mL of streptomycin (all from Sigma Chemical). Apoptosis was induced in HeLa cells (10⁶ cell/mL) by various stimuli as indicated. Annexin V binding assays were performed using Annexin-V-FITC Apoptosis Kit (CLONTECH Laboratories), following the manufacturer's instructions. Similarly, FNHS-ANXV, sfGFP-ANXV, and sfGFP proteins (2 μg/mL) were used for cell staining by incubation for 15 min in the dark. Apoptotic cells were identified either by direct visualization of green-colored membrane staining under a fluorescent microscope or by flow cytometry. To distinguish cells that had lost membrane integrity, propidium iodide (PI) was added at final concentration of 2 μg/mL before analysis.

Liposomes (0.1 mM) and PI/annexin V stained HeLa cells (10⁶ cell/mL) were diluted with HBS before acquisition on a BD FACS Calibur™ flow cytometer (BD Biosciences) with a 488 nm argon ion laser and a 635 nm red diode laser. The light scatter and fluorescent parameters were set at logarithmic gain, and platelets were identified according to their characteristic forward and side scatter properties, with 10,000 total events per sample acquired. Flow cytometric data acquisition and analysis were conducted by BD Cellquest™ Pro software. UV-treated HeLa cells were analyzed using a fluorescent microscope (AxioImager.Z1 mot, Carl Zeiss Ltd.), and an ISIS imaging system (MetaSystems) was used for images capturing and processing.

Human annexin V is a relatively small protein of 35 kDa composed of 320 amino acids (aa), extending over four functional and well identified domains (Figure 1A). Each of these homolog domains extends over 61 aa forming a conical structure responsible for PS binding in the presence of calcium ions. Annexin V is characteristically devoid of potential glycosylation sites and disulphide bridges, and so is sfGFP, making their expression in a reducing environment (like E. coli cytoplasm) suitable. The human gene for annexin V was cloned in two different plasmids; pRSET-a and pRSET-sfGFP downstream sfGFP gene, resulting in N-terminal 6× His-tagged proteins with theoretical sizes of 40 and 64 kDa, respectively (Figure 1A). The original plasmid pRSET-sfGFP was used for the expression of N-terminal 6× His-tagged sfGFP of 27 kDa in E. coli (Figure 1A; Al-Homsi et al., 2012a; Twair et al., 2014). As inferred from the 3D structure prediction of the fusion protein sfGFP-ANXV, the PS binding side of annexin V appeared to be exposed away from sfGFP and in a favorable direction for binding PS-displaying membranes (Figure 1B, Supporting Material, Data Sheet 1). In addition, the N-terminal 6× His tag seemed to be laterally exposed and suitable for protein purification by nickel affinity chromatography. A similar favorable structure was predicted for the protein ANXV despite the existence of a relatively long sequence (36 aa), initiated with the 6× His tag, at its N-terminal side (Supporting Material, Data Sheet 2).

Expression of sfGFP-ANXV was carried out after transformation of E. coli BL21 (DE3) Gold cells with the confirmed pRSET-ANXV and pRSET-sfGFP-ANXV constructs. A remarkable expression of soluble proteins was observed from both constructs after IPTG induction, and a brief amount of the proteins was found insoluble (data not shown). Furthermore, the 6× His tag of both proteins, ANXV and sfGFP-ANXV, was functional in affinity purification on a nickel-charged column (Figure 2A). Expectedly, the real sizes of both proteins and the control sfGFP were identical to their theoretical ones as shown on SDS-PAGE after blue staining (Figure 2B). Beside their clear greenish color, GFP-containing recombinant proteins (sfGFP and sfGFP-ANXV) were specifically detected by immunoblotting with an anti-GFP antibody (Figure 2B). ANXV was covalently linked to the fluorescein fluorophore (FNHS) through amine conjugation, and the fluorescence characteristics of sfGFP-ANXV and FNHS-ANXV were investigated and compared with sfGFP as well as with the free fluorophore. Apparently, fluorescence spectra of the free and fusion sfGFP seemed to be identical at several pairs of wavelengths for excitation and emission and differ from the spectrum of the fluorescein. On the contrary, chemical conjugation resulted in a remarkable spectrum shifting between the free and bound FNHS, from λ_ex: 485/λ_em: 538 (advised by the supplier λ_ex: 490 nm/λ_em: 514 nm) to λ_ex: 390/λ_em: 485 (Figure 2C).

The capacity of the two purified recombinant forms of annexin V to bind phospholipids from two natural sources; soya bean and egg yolk, was tested by indirect ELISA. To unify the detection method of the different recombinant proteins after binding to the immobilized layer of phospholipids, they were all conjugated with biotin-NHS, and conjugation was confirmed by western blot using anti-biotin-HRP (Figure 3A). Detection of these different proteins in the immunoblotting was also possible using anti-6× His antibody. However, ELISA detection signal using this antibody was extremely low (data not shown). Interestingly, sfGFP-ANXV was able to interact with the immobilized PL from soya bean, but not from egg yolk, and ANXV interacted as well, but to a lesser extent, to soya bean PL (Figure 3B). This interaction was demonstrated and confirmed alternatively, by spotting different amounts of the phospholipids (from soya bean or egg yolk) on PVDF membranes before being detected by biotinylated sfGFP-ANXV or ANXV (Figure 3B, inset).

The difference between sfGFP-ANXV and ANXV in binding to PL from soya bean was demonstrated by SPR after preparing liposomes from these PL. A special sensor chip of phytosphingosine surface immobilized on a 2D carboxymethyldextran hydrogel was used to immobilize the liposomes in a similar way to what have previously been reported (Kim et al., 2015). The binding capacity of the surface was investigated when consecutive pulses of liposomes were injected over the left channel of the sensor chip installed on a dual-channel biosensor. After each injection, a new RFU level was reached and stabilized, then finally the immobilized liposomes were washed away and the initial base line was reconstituted by injecting SDS (Figure 4A). Over the immobilized liposomes, pulses of the three different proteins were injected and the signal was normalized for each protein with the signal from the right channel. Interestingly, only sfGFP-ANXV was able to bind to the immobilized liposomes and not ANXV or sfGFP (Figure 4B).

Prepared liposomes from soya bean and egg yolk PL were seen under the scanning electronic microscope as microspheres of ~400 nm (Figure 5A). The differential capacity of liposomes detection between sfGFP-ANXV and ANXV was tested also by flow cytometry, where liposomes, from soya bean and egg yolk, were stained with each of the three recombinant proteins; sfGFP-ANXV, FNHS-ANXV, and sfGFP (Figure 5B). Interestingly, significant annexin V positive populations were observed in just three combinations; FNHS-ANXV with soya bean liposomes, sfGFP-ANXV with liposomes from soya bean and egg yolk, giving 2.73, 96.1, and 39.9% of total analyzed particles, respectively (Figure 5B).

The behavior of sfGFP-ANXV, FNHS-ANXV toward HeLa cells undergoing apoptosis after different types of stimuli (heat, H₂O₂, and UV) was analyzed by flow cytometry. After staining cells with PI and the fusion protein sfGFP-ANXV, three distinguished populations (living, early, and late apoptotic cells) could be discriminated among the cells before and after treatment (Figure 6A). Living cells (93.3%) in control condition were partially displaced into early-apoptotic cells (37%) in H₂O₂ treated cells. UV radiation resulted in the three distinct populations; living (22.5%), early (27.5%) and late (50.2%) apoptotic cells. In heat-treatment, most of the cells undergo late-apoptosis (75.7%). Identical histograms of fluorescence of UV-treated cells were found after staining with sfGFP-ANXV or FNHS-ANXV but not with sfGFP (Figure 6B). Finally, fluorescent microscopy of UV-treated cells after staining with sfGFP-ANXV and PI has shown three different types of cell; uncolored cells, cells with different degrees of green-stained membrane and a third type of cells with red colored nucleus and green dissected membrane (Figure 6C). This microscopic experiment provided an additional confirmation of the utility of sfGFP-ANXV in marking early and late apoptotic cells, especially when used in combination with PI.

The fusion sfGFP-ANXV was able to bind to the treated (H₂O₂ or UV) HeLa cells in a dose-dependent manner, and living cells decreased as annexin V-positive cells increased gradually (Figure 7A). Using H₂O₂ treatment, early and late apoptotic cells increased in parallel with annexin V-positive cells. Meanwhile, UV treatment with low doses has a similar effect as H₂O₂, but at 2 J, a huge population of early apoptotic cells appeared and directly transformed into late apoptotic cells at 3 J (Figure 7A). Using moderate doses of H₂O₂ (10 mM) and UV (1 J), a comparison between cell labeling with sfGFP, ANXV, sfGFP-ANXV, or FNHS-ANXV was carried on by flow cytometry (Figure 7B). As expected, negative labeling with control proteins; sfGFP and ANXV, did not result in a significant difference in cell percent between control and treated conditions, and most of the cells remained non-fluorescent (>90%). Fluorescent annexins: sfGFP-ANXV or FNHS-ANXV, have marked almost similar percentage of cells in control (7.55 and 11.2%), UV (29.7 and 27.1%), and H₂O₂ (37.8 and 39.2%) treatment conditions. Interestingly, separating of these marked cells between early and late apoptotic portions, because of the additional PI staining, has shown that in all conditions sfGFP-ANXV was accompanied with more early than late apoptotic cells (1.75 ± 0.38 fold), and vice versa in case of FNHS-ANXV (0.54 ± 0.04 fold). Despite this, the same percentage of living cells were left unstained by both fluorescent proteins in all conditions; control (91.2 and 89%), UV (67.7 and 72.9%), and H₂O₂ (59.6 and 60.9%; Figure 7B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.