Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jun 1.
Published in final edited form as: Biosens Bioelectron. 2024 Mar 1;253:116185. doi: 10.1016/j.bios.2024.116185

Electrochemical DNA-based sensors for measuring cellgenerated forces

Mahmoud Amouzadeh Tabrizi a,*, Priyanka Bhattacharyya a, Ru Zheng a, Mingxu You a,b,*
PMCID: PMC10947853  NIHMSID: NIHMS1974117  PMID: 38457863

Abstract

Mechanical forces play an important role in cellular communication and signaling. We developed in this study novel electrochemical DNA-based force sensors for measuring cell-generated adhesion forces. Two types of DNA probes, i.e., tension gauge tether and DNA hairpin, were constructed on the surface of a smartphone-based electrochemical device to detect piconewton-scale cellular forces at tunable levels. Upon experiencing cellular tension, the unfolding of DNA probes induces the separation of redox reporters from the surface of the electrode, which results in detectable electrochemical signals. Using integrin-mediated cell adhesion as an example, our results indicated that these electrochemical sensors can be used for highly sensitive, robust, simple, and portable measurements of cell-generated forces.

Keywords: Cellular forces, DNA probe, Electrochemical sensor, Hairpin DNA, Tension gauge tether

1. Introduction

Cell-generated adhesion force is a key component in regulating various cellular processes such as the migration, deformation, invasion, and metastasis of cancer cells (Dalby et al., 2014; Friedl et al., 2011; Janiszewska et al., 2020). During these processes, cell membrane adhesion molecules like integrins and cadherins bind with their target ligands and apply piconewton (pN)-scale forces to mediate cell–extra cellular matrix and cell–cell adhesions (Janiszewska et al., 2020; Keshri et al., 2021). Several techniques have been developed for the detection of these cellular forces including traction force microscopy (Dembo and Wang, 1999), micropillar arrays (Tan et al., 2003), atomic force microscopy (Hinterdorfer et al., 2006), optical/magnetic tweezers (Bustamante et al., 2021), fluorescent tension probes (Blakely et al., 2014; Zhang et al., 2014; Duan et al., 2021; Duan et al., 2023), and photonic cyrstals (Li et al., 2023). Although these powerful approaches have significantly improved our ability to measure cell-generated adhesion forces (Goktas et al., 2017; Schlichthaerle et al., 2021; Ungai-Salánki et al., 2019) the requirement of large/expensive instruments and skilled operators are currently limiting their potential adoption by the broader scientific community. On the other hand, we realized that electrochemical sensors can be more portable, inexpensive, and easier to use, which once incorporated for measuring cellular forces, would potentially start a new chapter in mechanobiology. We aim to test this idea by developing novel electrochemistry-based cellular force sensors in this project.

To engineer electrochemical force sensors, we consider integrating force-sensitive DNA probes as the detection units for cell-generated adhesion forces. DNAs are arguably the most popularly used moieties for designing molecular tension probes, given their modularly tunable force threshold value, high stability, and ease of synthesis and modification (Ma et al., 2019; Tian et al., 2022). DNA tension probes are often designed based on either an irreversible double-stranded mode (e.g., so-called tension gauge tether, TGT, for detecting forces of ~12–56 pN) (Wang et al., 2013) or a reversible hairpin mode (with force thresholdvalue F1/2 in the range of ~2–20 pN) (Brockman et al., 2020; Liu et al., 2016; Rashid et al., 2023). Herein, we want to study the possibility of developing both TGT and hairpin DNA tension probes that can generate electrochemical signals upon experiencing different levels of cellular adhesion forces.

In our design (Fig. 1), the self-assembled DNA tension probes contain (1) a thiol group at one end to tether to the surface of a gold screen-printed electrode (Au-SPE), (2) a cyclic arginine-glycine-aspartic-acid-D-phenylalanine-lysine (cRGDfK or in short RGD) ligand at the other end or in another strand to recognize cell surface integrins, i.e., widely used trans-membrane receptors for cell adhesion, (3) a methylene blue (MB) redox reporter that can undergo reversible electron transfer and release two electrons per redox reaction (Mahlum et al., 2021), and (4) a force-sensitive DNA hairpin (F1/2, 12 pN) or a double-stranded TGT structure with low (~12 pN), medium (~43 pN), or high (~56 pN) threshold forces. In both hairpin and TGT modes, in the absence of cellular forces, the distance of MB from the surface of conductive Au-SPE is minimal and as a result, exhibits a large electron transfer rate and high intensity of electrochemical signals. Once the DNA hairpin or TGT duplex is unfolded due to integrin-mediated cell adhesion, an increased distance between MB and Au-SPE will immediately result in a decreased electron transfer and electrochemical signals. In such a way, specific cell-generated adhesion forces can be converted into detectable electrochemical signals.

Fig. 1.

Fig. 1.

Open in a new tab

Schematic and photo image of the smartphone-based electrochemical DNA-based force sensors. Thiolated tension gauge tether (TGT) or DNA hairpin probes were immobilized on the surface of a gold screen-printed electrode. Cell adhesion force mediated by a specific receptor–ligand pair could separate the double-stranded TGT probes or unfold the DNA hairpin, and as a result, the attached methylene blue reporter exhibited a decreased electrochemical signal.

By attaching the hairpin- or TGT probe-incorporated Au-SPE chip to a smartphone-based electrochemical device, our results indicate that integrin-mediated forces during the adhesion of HeLa cells can indeed be precisely and sensitively detected by these portable electrochemical DNA-based sensors. To the best of our knowledge, this is for the first time that electrochemical sensors can be developed for the detection of molecular forces during cell adhesions. Compared to the other methods used for the detection of cell-generated forces (Goktas et al., 2017; Liu et al., 2016; Wang et al., 2013), the proposed smartphone-based electrochemical sensors are cheaper, smaller, and easier to use, opening a new chapter in the field of electrochemical mechanosensors. We expect these new electrochemical DNA-based force sensors can be powerful complementary tools to existing mechano-biology techniques in improving our understanding of cell-generated forces.

2. Materials and Methods

2.1. Reagents and apparatus

Double deionized water (18.6 MΩ·cm−1) was used throughout the research work. The DNA oligonucleotides were custom synthesized and purified by W. M. Keck Oligonucleotide Synthesis Facility at Yale University School of Medicine. Hanks’ Balanced Salt Solution (HBSS), 4-(2-hydroxyethyl)-1-piperazine ethane sulfonic acid (HEPES), sodium bicarbonate, potassium chloride, sodium chloride, dipotassium phosphate, sulphuric acid, ethylenediaminetetraacetic acid (EDTA), tris(2-carboxyethyl) phosphine (TCEP, Catalog: PI77720), hexaammineruthenium (III) ion (Ru(NH3)63+, Catalog: AC363340050), ferrocene carboxylate (Catalog: 50-533-122), Dulbecco’s Modified Eagle Medium (DMEM), streptavidin (Catalog: 43-430-2), biotinylated cyclic arginine-glycine-aspartic acid (B-cRGDfK) (Catalog: 50-168-6313), 1-hexanethiol (Catalog: AC215271000), and latrunculin B (Catalog: 42-802-01) were purchased from Thermo Fisher Scientific (Agawam, MA) and used without further purification.

Gold screen-printed electrodes (Au-SPE) (Catalog: 250AT, with working electrode made of gold, auxiliary electrode made of platinum, reference electrode and electric contacts made of silver, dimensions: 3.4×1.0×0.05 cm3, length×width×height) and transparent gold electrodes (Catalog: AUTR10, with working electrode made of a thin layer gold of 4 mm diameter on a transparent plastic substrate, auxiliary electrode is made of carbon, reference electrode is made of silver, dimensions: 3.3×1.0×0.0175 cm3, length×width×height) were purchased from Metrohm-DropSens (Llanera, Spain). The cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), and square wave voltammetry (SWV) studies were performed using a Sensit Smart electrochemical device from PalmSens (Houten, Netherlands).

The SWV measurements were performed in a solution containing (v/v) 50% DMEM and 50% phosphate buffer (0.2 M, pH 7.4). The SWV signals were recorded at a step potential of 20 mV, a pulse amplitude of 50 mV, and a frequency of 20 Hz. The EIS measurements were performed in a solution containing (v/v) 50% DMEM, 50% phosphate buffer (0.2 M, pH 7.4), and 5 mM ferrocene carboxylate. The EIS signals were recorded at an AC potential of 5 mV, a DC potential of 0.17 V, and in the frequency range of 100,000–0.1 Hz.

2.2. Preparation of electrochemical TGT sensors

The thiolated anchor strand (0.05 mM) was first reduced using 10 mM tris(2-carboxyethyl)phosphine (TCEP) for 1 h to cleave the disulfide bonds. This solution was then diluted by 0.1 M phosphate buffer (pH 7.4) to obtain 5.0 μM thiolated DNA. Afterward, 100 μL of 5.0 μM thiolated anchor strand was dropped onto the cleaned Au-SPE surface and kept in the refrigerator for 16 h. During this incubation time, the thiolated anchor strand bonded to the Au-SPE surface by self-assembly through the Au–S bonds (Zhang et al., 2011). After rinsing with copious amounts of 0.1 M phosphate buffer (pH 7.4) to wash away nonspecifically adsorbed DNA molecules, 100 μL of 100 μM 1-hexanethiol solution was added on the surface of the electrode and incubated at 37°C for 1 h to passivate the remaining active sites of the electrode surface. Subsequently, 100 μL of 5.0 μM 5’-MB-3’-biotin-modified reporter strand was added to the surface of the electrode and incubated at room temperature for 1 h to generate surface-attached double-stranded DNA. After washing profusely with 0.1 M phosphate buffer (pH 7.4), 100 μL of 5.0 μM streptavidin (in 0.1 M phosphate buffer, pH 7.4) was dropped on the surface of the electrode to interact with the biotin group of DNA duplex at room temperature for 1 h. After again rinsing plentifully with 0.1 M phosphate buffer (pH 7.4), 100 μL of 5.0 μM biotinylated cyclic arginine-glycine-aspartic-acid-D-phenylalanine-lysine (cRGDfK) was dropped on the surface of the electrode to interact with streptavidin at room temperature for 1 h. As a final step, the fabricated TGT sensor was rinsed with 0.1 M phosphate buffer (pH 7.4) and stored at 4°C before usage in the next day.

2.3. Preparation of electrochemical DNA hairpin sensors

100 μL of 50 μM 5’-biotin-3’-thiol-modifiedhairpin anchor strand was first mixed with 100 μL of 5’-methylene blue (MB)-labeled reporter strand for 15 min, and then the mixture was annealed by heating to 75°C for 5 min followed by cooling to 25°C at a rate of −1.3°C/min using a Bio-Rad PCR thermal cycler. During this annealing procedure, two DNA strands hybridized with each other. Then following a similar process as the TGT sensors, the DNA duplex was reduced using 10 mM TCEP for 1 h to cleave the disulfide bonds. This solution was then diluted with 0.1 M phosphate buffer (pH 7.4) to obtain 5.0 μM DNA. Afterward, 100 μL of 5.0 μM thiolated hairpin anchor strand was dropped onto the cleaned Au-SPE surface and kept in the refrigerator for 16 h. During this incubation time, the thiolated anchor strand bonded to the Au-SPE surface by self-assembly through the Au–S bonds. After rinsing with copious amounts of 0.1 M phosphate buffer (pH 7.4) to wash away nonspecifically adsorbed DNA molecules, 100 μL of 100 μM 1-hexanethiol solution was added on the surface of the electrode and incubated at 37°C for 1 h to passivate the remaining active sites of the electrode surface. Subsequently, 100 μL of 5.0 μM streptavidin (in 0.1 M phosphate buffer, pH 7.4) was dropped on the surface of the electrode to interact with the biotin group of DNA probes at room temperature for 1 h. After rinsing plentifully with 0.1 M phosphate buffer (pH 7.4), 100 μL of 5.0 μM biotinylated cyclic arginine-glycine-aspartic-acid-D-phenylalanine-lysine (cRGDfK) was dropped on the surface of the electrode to interact with streptavidin at room temperature for 1 h. As a final step, the fabricated DNA hairpin sensor was rinsed with 0.1 M phosphate buffer (pH 7.4) and stored at 4°C before usage in the next day.

3. Results and Discussion

3.1. Construction and characterization of the electrochemical TGT-based force sensors

We first designed three electrochemical TGT force sensors based on previously characterized DNA duplex structures that can be ruptured by forces at ~12 pN, 43 pN, and 56 pN, respectively (Wang et al., 2013). These TGT probes share the same DNA sequence but have different force application geometries, depending on the relative position of the thiol anchor group on one strand (anchor strand) to that of the RGD ligand on the other (reporter strand). Using the same 5’-MB-3’-RGD-modified reporter strand (Table S1), by inserting the thiol group at the 5’-end (unzipping mode), middle (intermediate mode), or 3’-end (shear mode) of the anchor strand, low-level, medium-level, and high-level forces can be detected, correspondingly. To construct these sensors, the thiolated anchor strand was first added and self-immobilized onto surface-cleaned Au-SPE (Fig. S1a) through the formation of Au–S bonds. After washing away nonspecifically adsorbed anchor DNA and passivating with 1-hexanethiol, 5’-MB-3’-biotin-modified reporter strand was then added and followed by incubation with streptavidin and biotinylatedcyclic arginine-glycine-aspartic-acid-D-phenylalanine-lysine to finally fabricate TGT probes on the electrode surface.

The EIS (Robinson et al., 2023) and the corresponding Nyquist plot were used for the stepwise characterization of this fabrication process. In a Nyquist plot, a semicircular portion at high frequencies and a linear portion at low frequencies is related to the charge-transfer-limited process and diffusion-limited process, respectively. By analyzing the diameter of the semicircle, the charge-transfer resistance (Rct) of the electrode can be obtained (Lazanas et al., 2023). Our EIS measurements were performed by applying 0.17 V as a DC voltage, i.e., a formal potential of ferrocene carboxylate, a redox probe used in this test (Fig. S1b). Using the 12 pN TGT sensor as an example, as can be seen in Fig. 2a, the unmodified Au-SPE exhibited a low Rct of ferrocene carboxylate at ~120 Ω. After the conjugation of the 5’-thiolated anchor strand, the Rct value increased to ~1850 Ω, which is due to the electrostatic repulsion between the negatively charged ferrocene carboxylate and the phosphate groups of DNA oligonucleotides. We further estimated the surface coverage of the DNA strand (θt) to be ~93% by using the equation θt= (1 – Rct/Rct′) ×100%, (Ahmad et al., 2021; Finklea et al., 1993) where Rct and Rct′ is respectively the charge transfer resistance of bare Au-SPE and that after DNA modification. Upon adding 5’-MB-3’-biotin-modified reporter strand, the Rct of the electrode increased to ~2510Ω (Fig. 2a), as the formation of double-stranded TGT caused a further limited mass transfer of ferrocene carboxylate to the electrode surface. These EIS-based stepwise characterizations indicated that the TGT sensors were fabricated successfully.

Fig. 2.

Fig. 2.

Open in a new tab

(a) Nyquist plots of the unmodified Au-SPE (green line), 3’-thiolated anchor strand-modified Au-SPE (red line), 1-hexanethiol-passivated 3’-thiolated anchor strand-modified Au-SPE (blue line), and that after adding 5’-methylene blue-3’-biotin-modified reporter strand (purple line). The electrochemical impedance spectroscopy measurement was performed in a solution containing (v/v) 50% DMEM, 50% phosphate buffer (0.2 M, pH 7.4), and 5 mM ferrocene carboxylate, signals were recorded at an AC potential of 5 mV, a DC potential of 0.17 V, and in the frequency range of 100,000–0.1 Hz. The equivalent electric circuit compatible with the Nyquist diagrams was shown on the top. Here, Rs is the solution resistance, Rct is the charge transfer resistance, Cdl is double layer capacitance, and ZW is Warburg impedance. (b) Stability of the cyclic voltammetry signals of Ru(NH3)63+ on 1-hexanethiol-passivated thiolated anchor strand-modified Au-SPE during 25 times potential scans at the scan rate of 100 mV·s−1 in a solution containing (v/v) 50% DMEM and 50% phosphate buffer (0.2 M, pH 7.4). (c–e) Square wave voltammetry of the electrochemical c) 12 pN TGT sensor, d) 43 pN TGT sensor, e) 56 pN TGT sensor after adding 100 μL of 1×106 HeLa cells/mL for 0, 15, 30, 45, and 60 min, respectively, followed by washing. The measurement was performed in a solution containing (v/v) 50% DMEM and 50% phosphate buffer (0.2 M, pH 7.4). The step potential was set as 20 mV, the pulse amplitude was at 50 mV, and the frequency was at 20 Hz. (f) Cell incubation time-dependent changes in the peak current value of 12 pN, 43 pN, and 56 pN TGT sensors. Shown are the mean and standard error peak values after subtracting the background signals from four replicated tests.

We next measured the surface densities of the thiolated anchor strand on the Au-SPE electrode using Ru(NH3)63+-based assay (Ge et al., 2007; Vogiazi et al., 2021). Ru(NH3)63+ could quantitatively interact with the negatively charged phosphate backbones of DNAs, and then the amount of surface-adsorbed Ru(NH3)63+ could be determined via the CV. As shown in Fig. S2b and S2a, the CV curves of the thiolated anchor strand-modified Au-SPE indeed exhibited a couple of well-defined redox peaks of Ru(NH3)63+, which were stably shown during 25 times potential scans at the rate of 100 mV·s−1. By further increasing the scan rate (ν) in the range of 25–300 mV·s−1, both anodic and cathodic peak currents (Ip) were observed to follow a linear correlation with the changes in the scan rate, indicating a surface-controlled process with the surface coverage of Ru(NH3)63+calculated to be ~4.72×10−11 mol·cm−2 (Fig. S2b and S2c). Correspondingly, the surface coverage of the thiolated anchor strand was estimated to be ~2.8×1013 per cm2 (supporting information). The surface of Au-SPE is highly packed with DNA probes.

After the further addition of 5’-MB-3’-biotin-modified reporter strand, the surface coverage of these reporter strands can be directly measured based on the redox peaks of MB in the CV curves. The CV curves of MB did not exhibit any obvious change after 25 times of scanning the potential in the range of −0.7 V to −0.1 V (Fig. S3a), indicating the high stability of the TGT sensor signal for potential precise detection. Similar to the above-mentioned Ru(NH3)63+-based assay, we further recorded the CV curves of MB at different scan rates (25–300 mV·s−1) (Fig. S3b). The peak currents were again linearly proportional to the scan rate (Fig. S3c), suggesting a surface-controlled redox process. The surface coverage of MB and consequently the reporter strand were calculated to be ~2.65×1013 per cm2. By comparing this value with that of the thiolated anchor strand (i.e., ~2.8×1013 per cm2), >95% of the surface-attached anchor strand could be used to generate double-stranded TGT sensors.

3.2. Electrochemical detection of cell-generated adhesion forces with the TGT sensor

To test whether our fabricated TGT sensors can indeed be used to detect cell-generated forces, we first added 100 μL of HeLa cells (1×106 cells·mL−1 concentration) respectively onto the Au-SPE electrodes that were modified with the 12 pN, 43 pN, or 56 pN TGT sensors. HeLa cells were chosen here because their membrane-expressed integrin αvβ3 can specifically recognize the cRGDfK ligand on the TGT sensors and facilitate force-mediated cell adhesion. To study the effect of these adhesion forces, we decided to measure the SWV signals of the TGT sensors because of the high sensitivity of SWV (Bard et al., 2022). As shown in Fig. 2c–2e, after a 15–60 min incubation and washing, while almost no change in the peak current signals of the 56 pN TGT sensor was observed, a gradually decreased peak current was shown for both 12 pN and 43 pN TGT sensors.

To further quantify these force-induced SWV signals, we calculated the relative peak current intensity change (I%) as I% =(Ic–I0)/I0×100% (Idili et al., 2019), where I0 and Ic are the peak current signals of the TGT sensors in the absence and presence of HeLa cells, respectively. After a 60-min incubation with 100 μL of 1×106 HeLa cells/mL, the 12 pN, 43 pN, and 56 pN TGT sensors exhibited respectively an ~94%, ~37%, and ~0.4% decrease in the relative peak current signal (Fig. 2f). A decreased SWV peak current suggested a reduced amount of MB-modified reporter strands on the electrode. After incubating with the HeLa cells, the 12 pN and 43 pN TGT sensors could be ruptured and release the reporter strands. In contrast, our results indicate that forces generated by the HeLa cells could not reach 56 pN to rupture the shear-mode TGT sensors.

In the case of the 56 pN TGT sensor, we wondered whether the HeLa cells could indeed adhere to the surface of the electrode, but due to the high threshold force to rupture the TGT probes, the cells will be instead immobilized on the electrode. To test this, we first imaged the surfaces of TGT-modified transparent gold electrodes after a 60-min incubation with HeLa cells and washing. The number of HeLa cells on the surface of the 56 pN TGT electrode was indeed much higher than that of 43 pN and 12 pN sensors (Fig. S4a). The EIS and Nyquist plots of the 56 pN TGT sensor were also used to study this cell adhesion process. After incubating with HeLa cells, the Rct value of the sensor increased dramatically from ~2510 Ω to ~5670 Ω (Fig. S4b), suggesting that less amount of the ferrocene carboxylate redox probe could now reach the surface of the electrode due to mass transfer limitation from the surface-attached HeLa cells. All these data indicated that the HeLa cells could not rapture the 56 pN TGT sensor and thus still attached to the surface of the gold electrode.

We next wanted to test the sensitivity of the electrochemical TGT sensors by detecting adhesion forces from a smaller number of cells. Using the 12 pN TGT sensor as the example, we incubated a total of ~100, 1×103, 1×104, and 1×105 HeLa cells respectively on the surface of the electrode for 60 min, and afterward a ~32%, 57%, 86%, and 94% decrease in the relative SWV peak current intensity was clearly observed (Fig. 3a and 3b). The TGT sensors could indeed be used to measure forces generated by a small number of cells.

Fig. 3.

Fig. 3.

Open in a new tab

(a) The square wave voltammetry of the 12 pN TGT sensor in the absence of HeLa cell (dotted line) or in the presence of 100 μL of 1×103, 1×104, 1×105, or 1×106 HeLa cells/mL for 60 min and washing. (b) The relative peak current intensity change (I%) as a function of adding different concentrations of the HeLa cells. The peak current value obtained from the dotted line, i.e., without adding HeLa cells, was used as the reference, I0. (c) The square wave voltammetry of the 12 pN TGT sensor in the absence of HeLa cell (dotted line) or after adding 100 μL of 1×106 HeLa cells/mL that have been treated with 0 μM, 20 μM, or 40 μM latrunculin B for 60 min and washing. (d) The relative peak current intensity change (I%) as a function of latrunculin B treatment. The peak current value obtained from the dotted line, i.e., without adding HeLa cells, was used as the reference, I0. The measurement was performed in a solution containing (v/v) 50% DMEM and 50% phosphate buffer (0.2 M, pH 7.4). The step potential was set as 20 mV, the pulse amplitude was at 50 mV, and the frequency was at 20 Hz. Shown are representative curves from four replicated tests.

We also wondered whether the TGT sensors could also be applied to study the effect of potential force-inhibiting drugs on the cell-generated forces. For this purpose, we chose to test latrunculin B, an inhibitor of actin polymerization that can bind to cellular G-actin and prevent its transition into F-actin (Mahlum et al, 2000). As shown in Fig. 3c and 3d, in the presence of 20 μM or 40 μM of latrunculin B, after incubating 100 μL of 1×106 HeLa cells/mL with the 12 pN TGT electrode for 60 min, the I% value of the sensor was significantly changed to ~22% and 9%, as compared to that of ~94% in the absence of force-inhibiting drugs. As expected, after the inhibition of actin polymerization, the integrin αvβ3 of the HeLa cells could not apply the forces to rupture the TGT sensors. All these above-shown results demonstrated that our proposed electrochemical TGT sensors can indeed be used as a highly sensitive and easy-to-use tool for detecting cell-generated adhesion forces.

3.3. Construction and characterization of the electrochemical DNA hairpin-based force sensors

Our next goal was to study whether electrochemical DNA hairpin-based sensors could be developed to measure cellular adhesion forces. In our design, the hairpin force sensor comprises a 5’-RGD-3’-thiol-modified anchor hairpin strand and a 5’-MB-modified reporter strand (Table S1). Once the cRGDfK ligands experience sufficient tensions (in this case, ~18 pN) from cellular integrin αvβ3, the unfolding of DNA hairpins will result in an increased distance of MB from the Au-SPE surface, and consequently, a decreased electrochemical signal (Bidinger et al, 2022; Ji et al., 2023; Li et al., 2023). In contrast to the TGT sensors, hairpin-based force sensors can be more reversibly open and close to monitor dynamic changes in the cellular forces. The construction of hairpin force sensors started by annealing a 5’-biotin-3’-thiol-modified anchor strand with a 5’-MB-labeled reporter strand, which duplex was then added and self-immobilized onto surface-cleaned Au-SPE. Like the TGT force sensors, after washing, 1-hexanethiol passivation, and stepwise addition of streptavidin and biotinylated cRGDfK, the DNA hairpin probes were finally fabricated on the surface of the Au-SPE electrode (Fig. 1).

EIS was also used to characterize these construction steps. As shown in Fig. 4a, the unmodified Au-SPE exhibited a low Rct at ~125 Ω, while after adding the conjugates of the thiolated anchor strand and MB-labeled reporter strand, the Rct value was increased to ~1270 Ω, indicating a layer of DNA conjugates was immobilized on the surface of the electrode that could electrostatically repulse the negatively charged ferrocene carboxylate. Based on these Rct values, the surface coverage of the DNA strand θ(t) was found to be ~90%. After additionally assembling streptavidin and biotinylated cRGDfK, the Rct value increased to ~2140 Ω, suggesting a further limited mass transfer of ferrocene carboxylate to the electrode surface.

Fig. 4.

Fig. 4.

Open in a new tab

(a) Nyquist plots of the unmodified Au-SPE (green line), thiolated DNA hairpin probe-modified Au-SPE (orange line), 1-hexanethiol-passivated thiolated DNA hairpin probe-modified Au-SPE (blue line), 1-hexanethiol-passivated thiolated DNA hairpin probe-modified Au-SPE after adding streptavidin and biotinylated cRGDfK, i.e., DNA hairpin force sensor (red line), and that after adding 100 μL of 1×106 HeLa cells/mL onto the DNA hairpin force sensor (purple line). The electrochemical impedance spectroscopy measurement was performed in a solution containing (v/v) 50% DMEM, 50% phosphate buffer (0.2 M, pH 7.4), and 5 mM ferrocene carboxylate, signals were recorded at an AC potential of 5 mV, a DC potential of 0.17 V, and in the frequency range of 100,000–0.1 Hz. The equivalent electric circuit compatible with the Nyquist diagrams were shown on the right. Rs is the solution resistance, Rct is the charge transfer resistance, Cdl is double-layer capacitance, and ZW is Warburg impedance. (b) Stability of the cyclic voltammetry signals of methylene blue on the DNA hairpin force sensor-modified Au-SPE during 25 times potential scans at the scan rate of 400 mV.s−1 in a solution containing (v/v) 50% DMEM and 50% phosphate buffer (0.2 M, pH 7.4). (c) Before (dotted line) and after (solid line) adding HeLa cells, a linear correlation was also observed between both anodic (orange line) and cathodic (blue line) peak currents (Ip) and the changes in the scan rate (ν).

To determine the surface coverage of the hairpin probes, we next measured the CV curves of MB on the Au-SPE electrode at a scan rate of 400 mV·s−1 (Fig. S5a). The paired redox peaks of MB were stably shown during 25 times of potential scans (Fig. 4b), indicating a high stability of the electrochemical DNA hairpin-based sensor. By recording the CV curves at different scan rates in the range of 100–450 mV·s−1, both the oxidation and reduction peaks of MB followed a linear relationship with the scan rate (Fig. 4c and S5b). Similar to the TGT sensors, the surface coverage of DNA hairpin probes can be calculated as a surface-controlled electrochemical system, with ~1.9×1013 probes being modified per cm2 electrode (supporting information).

3.4. Electrochemical detection of cell-generated adhesion forces with the DNA hairpin sensor

We next wanted to study whether the surface attachment of HeLa cells will induce the opening of DNA hairpin probes. For this purpose, we measured the heterogeneous electron transfer rate constant (Ks) of MB, as the Ks value highly depends on the distance of MB from the surface of Au-SPE (Kurnik et al., 2020; Li. et al., 2023; Ji et al., 2023). Upon the unfolding of the DNA hairpin, a dramatically decreased Ks will be observed. After adding 100 μL of HeLa cells (1×106 cells/mL) for 60 min, the CV curves of the hairpin sensor were recorded at different scan rates (100–450 mV·s−1). Based on a Laviron’s formula (Laviron, 1979), the average Ks of MB was calculated to be ~0.21 s−1 (Fig. 4c and S5c). While in the absence of cells, a Ks value of ~7.1 s−1 was shown. This dramatic decrease in the Ks of MB (~33-fold) indicated that the distance between MB and the surface of the electrode was indeed significantly increased after the adhesion of HeLa cells.

To further validate whether the surface attachment of HeLa cells indeed depends on the specific interactions between integrin αvβ3-RGD and cRGDfK, EIS measurements were performed to measure cell adhesion-induced changes in the charge-transfer resistance of the electrode. As expected, after a 60-min incubation with HeLa cells (100 μL of 1×106 cells/mL), the Rct of the hairpin sensor increased dramatically from ~2140 Ω to ~3420 Ω (Fig. 4a), indicating that the surface attachment of HeLa cells hindered ferrocene carboxylate from reaching the surface of the electrode. While in contrast, after the same batch of HeLa cells was incubated with a control Au-SPE electrode that was immobilized with the same DNA hairpin probe just without the attachment of biotinylated cRGDfK, a minimal change in the Rct value (from ~1590 Ω to ~1650 Ω) supported that the electrode surface adhesion of HeLa cells highly relies on the presence of integrin–RGD interactions (Fig. S6a).

Our next goal was to study the sensitivity of the hairpin sensor in measuring cell-generated adhesion forces. A total of ~100, 1×103, 1×104, and 1×105 HeLa cells were added and incubated respectively for 60 min on the surface of the hairpin sensor. As shown in Fig. 5a and 5b, the SWV peak current of the electrode continuously decreased as the concentration of cells increased, suggesting that more and more DNA hairpins were unfolded by surface-attached HeLa cells. We also calculated the relative peak current intensity change of the hairpin sensors. Our results showed that the I% value schanged from ~7%, 25%, 33%, to ~42% as the amount of HeLa cells increased from 100 to 1×105. Like the TGT sensors, hairpin-based electrochemical sensors could also be used to measure forces generated from a small number of cells.

Fig. 5.

Fig. 5.

Open in a new tab

(a) The square wave voltammetry of the DNA hairpin sensor in the absence of HeLa cell (dotted line) or in the presence of 100 μL of 1×103, 1×104, 1×105, or 1×106 HeLa cells/mL for 60 min and washing. (b) The relative peak current intensity change (I%) as a function of adding different concentrations of the HeLa cells. The peak current value obtained from the dotted line, i.e., without adding HeLa cells, was used as the reference, I0. (c) The square wave voltammetry of the DNA hairpin sensor in the absence of HeLa cell (dotted line) or after adding 100 μL of 1×106 HeLa cells/mL that have been treated with 0 μM, 20 μM, or 40 μM latrunculin B for 60 min and washing. (d) The relative peak current intensity change (I%) as a function of latrunculin B treatment. The peak current value obtained from the dotted line, i.e., without adding HeLa cells, was used as the reference, I0. (e) The square wave voltammetry of the DNA hairpin sensor in the absence of HeLa cell (dotted line) or after adding 1×106 cells/mL HeLa cells for different incubation times. (f) Cell incubation time-dependent changes in the peak current value of the DNA hairpin sensor. Shown are the mean and standard error peak values after subtracting the background signals from four replicated tests. The measurement was performed in a solution containing (v/v) 50% DMEM and 50% phosphate buffer (0.2 M, pH 7.4). The step potential was set as 20 mV, the pulse amplitude was at 50 mV, and the frequency was at 20 Hz. Shown are representative curves from four replicated tests.

As another control experiment to test if the observed changes in the SWV peak current could be possibly due to simple cell attachment but without the opening of DNA hairpins, we also immobilized cRGDfK ligands directly onto the surface of Au-SPE, together with cRGDfK-free DNA hairpin probes. In the presence of ferrocene carboxylate, we then measured the SWV signals of both MB and ferrocene carboxylate before and after a 60-min incubation with 1×103 HeLa cells (Fig. S6b). The peak current of ferrocene carboxylate decreased significantly due to mass transfer limitation caused by the attachment of HeLa cells. While in contrast, the MB signals exhibited no change after adding the HeLa cells. These results indicated that without integrin–RGD interaction-mediated unfolding of DNA hairpins, simple attachment of HeLa cells will not affect the hairpin sensor signals.

The hairpin force sensor can be used to detect the force-inhibition function of latrunculin B (Fig. 5c and 5d). After adding 20 μM or 40 μM of latrunculin B to 1×105 HeLa cells and incubating for 60 min, the SWV peak current of the hairpin sensor increased dramatically (with an I% value of ~19% and 8%) as compared to that without the drug treatment (~42% I%). This result was expected, as the inhibition of integrin forces could result in less amount of DNA hairpin probes being unfolded.

Lastly, we wondered whether the hairpin force sensor could be applied to monitor the dynamic changes in the cell-generated adhesion forces. To test this, the SWV response of the sensor was measured every 10 min after adding 1×105 HeLa cells (Fig. 5e). Interestingly, our results indicated that during the first 40-min incubation, a slow decrease in the SWV peak current was observed, while afterward, a much faster and more dramatic change was shown (Fig. 5f). Since the initial HeLa cell surface attachment occurred immediately after adding the cells onto the electrode, we realize that the observed 40-min delay in the SWV signal change is likely due to the time requirement for cell spreading and focal adhesion formation (Zhang et al., 2014). Upon forming focal adhesions, integrins could then apply strong forces to the cRGDfK ligand and open the DNA hairpin probe, which has a force threshold value of ~18 pN. Again, all these data suggested the power and feasibility of electrochemical DNA hairpin-based sensor for measuring cellular adhesion forces.

4. Conclusion

In this project, three types of TGT probes and also a DNA hairpin probe have been fabricated onto a smartphone-based electrochemical device for the sensitive and robust detection of cell-generated adhesion forces. By simply changing the DNA sequence or force application geometry, these electrochemical DNA-based sensors could modularly detect cellular forces at different threshold values. Meanwhile, the in situ dynamic changes in cellular adhesion forces as well as the regulation effect of force inhibitor drugs could also be measured using these novel electrochemical sensors. In addition, by potentially altering the ligand modification of the DNA probes, cellular adhesions mediated by different membrane ligand–receptor pairs could be possibly investigated at the molecular level.

As far as we know, this is the first attempt to measure cellular forces using electrochemical DNA-based molecular sensors. Compared to previously reported fluorescence-based force sensors (Rashid et al., 2023; Liu et al., 2016), these electrochemical tools could be more advantageous in terms of their portability, cost efficiency, and simplicity of use. While to achieve real single-cell or even single-molecule level cellular force detection, future studies are still needed to further improve the sensitivity and spatial resolution of these electrochemical force sensors, such as by combining with signal amplification techniques (Duan et al., 2021; Duan et al., 2023) or electroluminescence. We expect these sensors could be potentially useful for studying fundamental cell signaling pathways and also for developing novel strategies and materials in tissue engineering and pharmaceutics.

Supplementary Material

1

Acknowledgments

The authors gratefully acknowledge the support from NIH R35GM133507,T32GM139789, and Camille Dreyfus Teacher-Scholar Award. The authors also thank other members of the You Lab for useful discussion and valuable comments.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supporting Information Available

Supplementary data associated with this article can be found in the online version.

References

  1. Ahmad HMN, Dutta G, Csoros J, Si B, Yang R, Halpern JM, Seitz WR, Song E, 2021. ACS Appl. Polym. Mater 3(1), 329–341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bard AJ, Faulkner LR, White HS, 2022. Electrochemical methods: fundamentals and applications. John Wiley & Sons. [Google Scholar]
  3. Bidinger SL, Keene ST, Han S, Plaxco KW, Malliaras GG, Hasan T, 2022. Sci. Adv 8(46), eadd4111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blakely BL, Dumelin CE, Trappmann B, McGregor LM, Choi CK, Anthony PC, Duesterberg VK, Baker BM, Block SM, Liu DR, Chen CS, 2014. Nat. Methods 11(12), 1229–1232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brockman JM, Su H, Blanchard AT, Duan Y, Meyer T, Quach ME, Glazier R, Bazrafshan A, Bender RL, Kellner AV, Ogasawara H, Ma R, Schueder F, Petrich BG, Jungmann R, Li R, Mattheyses AL, Ke Y, Salaita K, 2020. Nat. Methods 17(10), 1018–1024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bustamante CJ, Chemla YR, Liu S, Wang MD, 2021. Nat. Rev. Methods Primers, 1, 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dalby MJ, Gadegaard N, Oreffo ROC, 2014. Nat. Mater 13(6), 558–569. [DOI] [PubMed] [Google Scholar]
  8. Dembo M, Wang Y, 1999. Biophys. J 76(4), 2307–2316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Duan Y, Glazier R, Bazrafshan A, Hu Y, Rashid SA, Petrich BG, Ke Y, Salaita K, 2021. Angew. Chem. Int. Ed 60, 19974–19981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Duan Y, Szlam F, Hu Y, Chen W, Li R, Ke Y, Sniecinski R, Salaita K, 2023. Nat. Biomed. Eng 7, 1404–1418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Finklea HO, Snider DA, Fedyk J, Sabatani E, Gafni Y, Rubinstein I, 1993. Langmuir 9(12), 3660–3667. [Google Scholar]
  12. Friedl P, Alexander S, 2011. Cell 147(5), 992–1009. [DOI] [PubMed] [Google Scholar]
  13. Ge B, Huang YC, Sen D, Yu H-Z, 2007. J. Electroanal. Chem 602(2), 156–162. [Google Scholar]
  14. Goktas M, Blank KG, 2017. Adv. Mater. Interfaces 4(1), 1600441. [Google Scholar]
  15. Hinterdorfer P, Dufrêne YF, 2006. Nat. Methods 3, 347–355. [DOI] [PubMed] [Google Scholar]
  16. Idili A, Parolo C, Ortega G, Plaxco KW, 2019. ACS Sens. 4(12), 3227–3233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Janiszewska M, Primi MC, Izard T, 2020. J. Biol. Chem 295(8), 2495–2505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ji X, Lin X, Rivnay J, 2023. Nat. Commun 14(1), 1665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Keshri P, Zhao B, Xie T, Bagheri Y, Chambers J, Sun Y, You M, 2021. Angew. Chem. Int. Ed 60(28), 15548–15555. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kurnik M, Pang EZ, Plaxco KW, 2020. Angew. Chem. Int. Ed 59, 18442–18445. [DOI] [PubMed] [Google Scholar]
  21. Laviron E, 1979. J. Electroanal. Chem. Interf. Electrochem 101(1), 19–28. [Google Scholar]
  22. Lazanas AC, Prodromidis MI, 2023. ACS Meas. Sci. Au 3(3), 162–193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Li Q, Chen Z, Zhang Y, Ding S, Ding H, Wang L, Xie Z, Fu Y, Wei M, Liu S, Chen J, Wang X, Gu Z, 2023. Nat. Commun 14, 7369. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li S, Dai J, Zhu M, Arroyo-Currás N, Li H, Wang Y, Wang Q, Lou X, Kippin TE, Wang S, Plaxco KW, Li H, Xia F, 2023. ACS Nano 17(18), 18525–18538. [DOI] [PubMed] [Google Scholar]
  25. Li S, Dai J, Zhu M, Arroyo-Currás Li. N. H, Wang Y, Wang Q, Lou X, Kippin TE, Wang S, Plaxco KW, Li H, Xia F, (2023). ACS Nano, 17, 18525–18538. [DOI] [PubMed] [Google Scholar]
  26. Liu Y, Blanchfield L, Ma VPY, Andargachew R, Galior K, Liu Z, Evavold B, Salaita K, 2016. Proc. Natl. Acad. Sci. U.S.A 113(20), 5610–5615. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ma VP, Salaita K, 2019. Small 15(26), e1900961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Mahlum JD, Pellitero MA, Arroyo-Currás N, 2021. J. Phys. Chem. C 125(17), 9038–9049. [Google Scholar]
  29. Morton WM, Ayscough KR, McLaughlin PJ, 2000. Nat. Cell Biol, 2, 376–378. [DOI] [PubMed] [Google Scholar]
  30. Rashid SA, Dong Y, Ogasawara H, Vierengel M, Essien ME, Salaita K, 2023. ACS Appl. Mater. Interfaces 15(28), 33362–33372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Robinson C, Juska VB, O’Riordan A, 2023. Environ. Res 237, 116877. [DOI] [PubMed] [Google Scholar]
  32. Schlichthaerle T, Lindner C, Jungmann R, 2021. Nat. Commun 12(1), 2510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Tan JL, Tien J, Pirone DM, Gray DS, Bhadriraju K, Chen CS, 2003. Proc. Natl. Acad. Sci. U.S.A 100(4), 1484–1489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Tian Q, Keshri P, You M, 2022. Chem. Commun 58(30), 4700–4710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Ungai-Salánki R, Peter B, Gerecsei T, Orgovan N, Horvath R, Szabó B, 2019. Adv. Colloid Interface Sci 269, 309–333. [DOI] [PubMed] [Google Scholar]
  36. Vogiazi V, de la Cruz AA, Varughese EA, Heineman WR, White RJ, Dionysiou DD, 2021. ACS EST Engg. 1(11), 1597–1605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Wang X, Ha T, 2013. Science 340(6135), 991–994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Zhang S, Zhou G, Xu X, Cao L, Liang G, Chen H, Liu B, Kong J, 2011. Electrochem. commun 13(9), 928–931. [Google Scholar]
  39. Zhang Y, Ge C, Zhu C, Salaita K, 2014. Nat. Commun 5, 5167. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1974117-supplement-1.docx (929.7KB, docx)

RESOURCES