Ultrasensitive nanostructured immunosensor for stem and carcinoma cell pluripotency gatekeeper protein NANOG

Chemicals & materials

Monoclonal (mouse) primary NANOG antibody (Ab₁ ab76586), biotinylated rabbit polyclonal NANOG antibody (Ab₂ ab84231), NANOG protein (ab50053) and streptavidin protein (HRP ab7403) were obtained from Abcam (Cambridge, UK). Tissue lysate samples were obtained from Protein Biotechnologies (CA, USA). l-gluthathione reduced (99% HAuCl₄ × 3H₂O [99.9%]), sodium borohydride (99%) poly(diallydimethylammonium chloride) (MW 20%) and bovine serum albumin (BSA) were from Sigma-Aldrich (MO, USA). Streptavidin coupled MPs (Dynabeads^® MyOne Streptavidin T1), and biotinylated HRP were from Invitrogen (CA, USA). 1-(3-(dimethylamino) propyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysulfosuccinimide, from Sigma-Aldrich, were dissolved in water immediately before use. Hydrogen peroxide (H₂O₂; 30%) was from Fisher (NJ, USA). Tween-20 was from Sigma-Aldrich. Immunoreagents were dissolved in pH 7.0 phosphate-buffered saline (PBS 0.01 M in phosphate, 0.14 M NaCl, 2.7 mM KCl) unless otherwise noted.

Instrumentation

A CHI 660 electrochemical workstation (CH Instruments, TX, USA) was used for amperometry at ambient temperature (22 ± 2°C). Rotating disk amperometry was done at −0.3 V versus standard calomel reference electrode with electrodes rotating at 3000 rpm.

Preparation of MP-Ab₂-HRP conjugates

Supplier (Invitrogen) protocol was followed to conjugate the biotinylated-Ab₂ and biotinylated-HRP with MPs. A total of 20 µl of MPs (100 mg/ml) were washed twice with 0.1% BSA in PBS buffer (pH 7.4) and resuspended to the same volume. Then 50 µl of biotinylated antibodies (Ab₂ 5 µg/ml) and 30 µl of biotinylated HRP (2.5 mg/ml) were added to the MPs and incubated at room temperature for 30 min using a slow tilt rotator. MP-Ab₂-HRP conjugates were separated magnetically and washed with 0.1% BSA in PBS four-times and later reconstituted in PBS (pH 7.4), and stored at 4°C till further use.

The number of HRP labels per MP was calculated from the measured enzyme activity using 2,2’-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) assay [36] as reactant to be 4 (± 0.9) × 10⁵ per bead, estimated on the total number of magnetic beads in 1 ml, 9 × 10⁹.

Fabrication of the immunosensor

The AuNP sensor platform was fabricated using layer-by-layer alternate electrostatic adsorption [37,38]. Abraded disks (A = 0.14 cm²) of pyrolytic graphite (GE Ceramics, DE, USA) embedded in heat-shrinkable tubing and mounted on steel rods were used as the base electrodes onto which the immunosensors were fabricated. Initially, a 0.5-nm layer of cationic poly(dimethyldiallylammonium) chloride was adsorbed from a 20-µl drop placed onto the (0.14 cm²) pyrolytic graphite disk electrode, followed by washing with water. Next, a negatively charged layer of glutathione-decorated 5 nm AuNPs was deposited from a dispersion onto the poly (diallyldimethylammonium chloride)-coated electrode [27]. The AuNP layer caused a 29% increase in surface area over the underlying pyrolytic graphite surface. Surface area was estimated from cyclic voltammetry using the Randles–Sevcik equation [39] and slope of peak currents of soluble 1 mM ferrocyanide in 0.1 M KCl versus square root of scan rate.

Antibodies for NANOG were anchored to the AuNP-glutathione layer by placing 30 µl of freshly prepared 400 mM ethylcarbodiimide hydrochloride and 100 mM N-hydroxysulfosuccinimide in water onto AuNP electrodes for 10 min, rinsing with water and then incubating the activated electrode surface for 2.5 h with 20 µl of a 3.3 pmol solution (25 µg/ml) of primary NANOG antibody in pH 7.0 PBS. Unbound antibody was removed by rinsing thoroughly with 0.05% Tween-20 in PBS and then PBS for 2 min each. To block the free antibody binding sites on the sensor surface, the immunosensors were incubated with 10 µl of 2% BSA and 0.05% Tween-20 in PBS for 45 min at 37°C, followed by washing with 0.05% Tween-20 in PBS and then PBS for 2 min each.

Sensor operation

For the actual NANOG measurements, the immunosensors were incubated with 10 µl of NANOG standard solutions in pH 7.0 PBS buffer or cell/tissue lysates for 1 h at 37°C. After the washing steps, the sensors were exposed to 10 µl of 1 µg/ml biotinylated-Ab₂ conjugates in buffer containing 0.1% BSA in pH 7.4 PBS buffer for 1 h, followed by washing with 0.05% Tween-20 in PBS and PBS buffer for 2 min each.

Using the moderate sensitivity protocol (Figure 1A), the sensors were next incubated with 10 µl of streptavidin-HRP for 30 min, followed by thorough washing with 0.05% Tween-20 in PBS and then PBS for 2 min each. For the high-sensitivity protocol (Figure 1B), the sensors with captured NANOG from standards or cell/tissue lysates were incubated with multilabel MP conjugates (conjugated with biotinylated-Ab₂ and biotinylated-HRP) for 1 h. Finally, the completed immunosensors were placed in an electrochemical cell containing 1 mM hydroquinone in PBS buffer. Rotating-disk electrode amperometry was carried out by controlling the potential at −0.3 V versus saturated calomel electrode injecting H₂O₂ to obtain 0.1 mM concentration and measuring the amperometric reduction current. When using MP bioconjugates, where the concentration of hydroquinone was 0.25 mM, injected H₂O₂ was adjusted to reach a 0.05 mM concentration.

Sodium dodecyl sulfate gel electrophoresis and immunoblotting

Total cell lysates were obtained from human embryonic stem cells (hESCs) following manufacturer’s instructions (Ambion, TX, USA). In total, 7 µg of proteins or various concentrations of pure NANOG protein were run on 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis gel and transferred onto nitrocellulose membrane (Bio-Rad). Membranes were washed with 1X Tris-buffered saline containing 0.05% Tween-20, blocked in 5% milk in 1X Tris-buffered saline containing 0.05% Tween-20 and incubated with respective primary antibodies at conditions recommended by the manufacturers (Abcam NANOG monoclonal antibody, ab76586, 1:500 and β-actin, Cell Signaling 1:1000). After washing with 1X Tris-buffered saline containing 0.05% Tween-20, membranes were incubated with respective HRP-conjugated Ab₂. Signals were detected with LumiGLO chemiluminescent reagent (Cell Signaling Technology, MA, USA).

Sensor characteristics & detection limits

Nearly complete coverage of AuNPs on the sensor surface and attachment of a high density of capture antibodies have been confirmed previously for this sensor design by atomic force microscopy [27]. Using sensors with antibodies attached, a drop of NANOG standard solution or cell lysate is placed onto the sensor, which is then incubated and washed (see ‘Material & methods’ section). Next, a HRP-labeled Ab₂ or HRP-antibody-MP is used to bind to the captured NANOG to enable the detection step. At this point, the sensor is washed to minimize nonspecific binding and placed into the measuring cell containing buffer, then rotated while a potential of −0.3 V versus saturated calomel electrode is applied. Injection of H₂O₂ into the buffer solution activates the HRP labels to ferryloxy species that then get reduced by a hydroquinone mediator also present in the solution. The role of hydroquinone is to shuttle electrons from the underlying sensor surface to the HRP labels to complete the catalytic cycle and provide an amperometric steady-state current proportional to the amount of NANOG in the sample. Entire assays from placement of the sample on sensor to signal measurement takes approximately 135 min for high sensitivity and 155 min for moderate sensitivity.

Responses for the immunosensors to NANOG protein in the pg/ml range using the moderate sensitivity Ab₂-biotin-strepatavidin-HRP labels (Figure 1A) show a fast rise to the steady-state currents in Figure 2A upon injection of H₂O₂. The control with no NANOG gives a finite steady-state current due to a small amount of residual nonspecific binding of the labeled Ab₂ plus residual direct reduction of the H₂O₂. Steady-state responses increased from 25 to 150 pg/ml NANOG (Figure 2B) with a sensitivity as the slope of the calibration graph of 4.6 nA-ml/cm²/pg NANOG. The DL as the blank response plus three-times the average noise was 25 pg/ml NANOG.

Standardization results for the high sensitivity AuNP immunosensors featuring massively labeled streptavidin-MP-biotin Ab₂-biotin-HRP bioconjugates (Figure 1B) are shown in Figure 3. Good quality, low signal/noise traces with well defined steady-state signals were obtained. Calibration graphs are shown in a linear-log format that covers the range 0.1–20 pg ml⁻¹ as well as a plot of the lowest concentration range. Despite the nonlinearity of these calibrations, sensitivity of the sensor is still applicable for quantitation of NANOG in this range. Using the streptavidin-MP-biotin Ab₂-biotin-HRP bioconjugates, the DL of 100 fg/ml was 250-fold better than that measured with the Ab₂-biotin-streptavidin-HRP system, mainly due to significant signal amplification brought about by the 400,000 HRP labels on each magnetic bead. Sensitivity of the immunosensor in the lower concentration range was 171 nA-ml/cm²/pg NANOG, 175-fold better than the moderate sensitivity approach. Sensitivity was estimated by dividing the slope of the current versus NANOG (from a linear fit in low concentration range [Chikkaveeraiah B, Soldà A, Rusling J, Unpublished Data]) by area of the working electrode (0.14 cm²). Standard deviations show that reproducibility was a little worse than that in the higher concentration range but still acceptable for such low levels. Even a small amount of nonspecific binding of the MP bioconjugates in the amplification strategy causes a significant blank signal as the signal is due to approximately 400,000 HRP labels on the MP, as seen for the zero NANOG blank (Figure 3A).

Quantitative NANOG detection in cell lysates

We evaluated the sensor for determining levels of NANOG protein in cell lysates of testis tissue [17], with liver tissue as a negative control and hESCs as positive controls. We found that NANOG protein expression normalized to total cellular protein was six-times higher in hESCs compared with the testis tissue (Figure 4) and the negative control was very low in NANOG.

Because NANOG expression is known to be downregulated under differentiating conditions [8,10,16], we also examined its expression in differentiating hESCs at different times. The spontaneous method of differentiation was used, which involves culturing the hESCs in the stem cell medium devoid of FGF-2 [40,41]. Our data, as expected, showed that NANOG expression decreased significantly with increasing differentiation time. We confirmed this by sodium dodecyl sulfate-gel electrophoresis and immunoblotting with NANOG-specific antibody that showed a similar trend in decreased intensity with increasing differentiation of the band corresponding to NANOG (Figure 5). However, the differences found by the immunosensor are much more apparent than the western blot analyses, with its low sensitivity being apparent in the relative similarity of the undifferentiated and differentiated samples in Figure 5, compared with more precisely determined differences seen in the immunosensor responses in Figure 4.

The regulation of NANOG expression can be detected at the transcriptional level by quantifying the NANOG transcripts with real-time PCR. Hence, we correlated our NANOG protein expression data with the relative mRNA expression level in the differentiating hESCs. The NANOG transcript expression showed a 21% decrease at D5 compared with control hESCs followed by a 52% decrease at D10. This shows qualitative correlation with immunosensor results (Figure 4) of 40% decrease in actual amounts of NANOG at D5 and 70% decrease in NANOG at D10. Thus, the cellular machinery responsible for NANOG transcripts expression followed a similar trend of decline with the directly observed NANOG levels measured by the immunosensor.