This is the accepted version of the journal article: Flauzino, J. M. R.; Nguyen, Emily P.; Yang, Qiuyue; [et al.]. «Label-free and reagentless electrochemical genosensor based on graphene acid for meat adulteration detection». Biosensors and Bioelectronics, Vol. 195 (January 2022), art. 113628. DOI 10.1016/j.bios.2021.113628 This version is available at https://ddd.uab.cat/record/266840 under the terms of the license 1 LABEL-FREE AND REAGENTLESS ELECTROCHEMICAL GENOSENSOR BASED ON 2 GRAPHENE ACID FOR MEAT ADULTERATION DETECTION 3 4 José M. R. Flauzino1,2, Emily P. Nguyen2, Qiuyue Yang2, Giulio Rosati2, David Panáček2,3,4, Ana G. 5 Brito-Madurro1, João M. Madurro1,5, Aristeidis Bakandritsos3,6, Michal Otyepka3,7, Arben 6 Merkoçi2* 7 1 8 Institute of Biotechnology, Federal University of Uberlândia, 38405-319, Uberlândia-MG, Brazil. 2 9 Catalan Institute of Nanoscience and Nanotechnology, Autonomous University of Barcelona, 10 08193, Bellaterra, Barcelona, Spain. 3 11 Regional Centre of Advanced Technologies and Materials, Czech Advanced Technology and 12 Research Institute (CATRIN), Šlechtitelů 241/27, 783 71, Olomouc, Palacký University Olomouc, 13 Czech Republic. 4 14 Department of Physical Chemistry, Faculty of Science, Palacký University Olomouc, 17. listopadu 15 1192/12, 771 46 Olomouc, Czech Republic. 5 16 Institute of Chemistry, Federal University of Uberlândia, 38400-902, Uberlândia-MG, Brazil. 17 6 Nanotechnology Centre, Centre of Energy and Environmental Technologies, VŠB–Technical 18 University of Ostrava, 17. listopadu 2172/15, 708 00 Ostrava-Poruba, Czech Republic. 19 7 IT4Innovations, VŠB–Technical University of Ostrava, 17. listopadu 2172/15, 708 00 Ostrava- 20 Poruba, Czech Republic. 21 22 23 24 *Corresponding author. Nanobioelectronics & Biosensors Group, Institut Català de Nanociància i 25 Nanotecnologia (ICN2), CSIC and The Barcelona Institute of Science and Technology (BIST), 26 Campus UAB, 08193, Bellaterra, Barcelona, Spain. E-mail address: arben.merkoci@icn2.cat 1 1 Abstract 2 3 With the increased demand for beef in emerging markets, the development of quality-control 4 diagnostics that are fast, cheap and easy to handle is essential. Especially where beef must be free 5 from pork residues, due to religious, cultural or allergic reasons, the availability of such diagnostic 6 tools is crucial. In this work, we report a label-free impedimetric genosensor for the sensitive 7 detection of pork residues in meat, by leveraging the biosensing capabilities of graphene acid - a 8 densely and selectively functionalized graphene derivative. A single stranded DNA probe, specific 9 for the pork mitochondrial genome, was immobilized onto carbon screen-printed electrodes 10 modified with graphene acid. It was demonstrated that graphene acid improved the charge transport 11 properties of the electrode, following a simple and rapid electrode modification and detection 12 protocol. Using non-faradaic electrochemical impedance spectroscopy, which does not require any 13 electrochemical indicators or redox pairs, the detection of pork residues in beef was achieved in a 14 short-time of 45 min (including sample preparation), with a limit of detection of 9% w/w pork 15 content in beef samples. Importantly, the sensor retained its performance properties unchanged for 16 at least 4 weeks. These set of features places the present pork DNA sensor among the most 17 attractive for further development and commercialization, as well lay the ground for the 18 development of sensitive and selective sensing devices for label-free, fast, simple and reliable 19 monitoring of meat purity. 20 21 Keywords: food adulteration; DNA biosensor; non-faradaic electrochemical impedance 22 spectroscopy; beef; pork. 23 24 25 26 2 1 1 INTRODUCTION 2 The beef industry and the food demand are rapidly growing, mainly in developing countries, 3 due to the rising population (Hansen, 2018). However, food adulteration in the meat industry is a 4 recurring topic, where, in pieces of solid meat, adulteration is challenging, and processed meats, 5 such as hamburgers and meatballs, the mixture of different types of meat is imperceptible (Rahmati 6 et al., 2016). Final consumers, especially from groups with dietary restrictions (e. g. Halal and 7 Kosher diet followers), must be sure of the origin of the meat they consume, especially regarding 8 the total absence of pork, as well as there is a potential danger of serious allergic responses or 9 intolerance to proteins derived from specific animals (Wilson and Platts-Mills, 2018). 10 Meat quality control analysis usually applies physical, sensory, anatomical and histological 11 methods and, in rare cases, molecular techniques such as polymerase chain reaction (Okuma and 12 Hellberg, 2015) and loop-mediated isothermal amplification (Kumar et al., 2017). However, these 13 methods of beef purity analysis are time consuming, require special equipment, reagents and 14 qualified personnel. Biosensors enter this scenario with the proposal to develop devices that are 15 easy to handle, quick to respond, do not require qualified personnel and can be used in the various 16 stages of the meat industry (Montiel et al., 2017; Singh et al., 2016). 17 Nucleic acid-based biosensors, also called genosensors (Ratajczak et al., 2018a, 2018b), are 18 promising tools for meat adulteration analysis, because nucleic acids such DNA are more stable in 19 meat processing steps than proteins (Hellberg et al., 2017). Meat, as a muscle tissue, contains cells 20 including many mitochondria that hold mitochondrial DNA molecules. The mitochondrial DNA 21 differs from the nuclear, because it is smaller, circular and present in larger quantity, making it a 22 more promising target for detection techniques (Irwin et al., 1991). Within the different types of 23 genosensors, those employing electrochemical techniques are fast-responding and easy to 24 miniaturize, however, they in general require the use of DNA intercalants (De Castro et al., 2018), 25 metal complexes (Moço et al., 2019) or other electroactive molecules (Flauzino et al., 2021a) 26 including nanoparticles (De La Escosura-Muñiz et al., 2016; Merkoçi, 2010) for targeted detection, 3 1 which makes the sensing process more laborious, expensive and non-eco-friendly. Therefore, the 2 search of new label-free and reagentless technologies is extremely important. 3 The definition of a label-free electrochemical biosensor is unclear in the literature. However, 4 any molecule that is introduced into the medium, except for the electrolyte (generally a buffer or 5 acid solution), the analyte and the immobilized biomolecule on the surface of the transducer, can be 6 recognized as a label (Chen et al., 2019). Examples of employed labels in electrochemical 7 biosensors are redox probes e. g. potassium ferro/ferricyanide (Oliveira et al., 2018a), electroactive 8 mediators/intercalants such as methylene blue (Idili et al., 2021), ferrocene (Rabti et al., 2016) and 9 ethidium bromide (Oliveira et al., 2018b) and enzymatic reaction indicators e. g. 10 tetramethylbenzidine (Alves-Balvedi et al., 2016). Therefore, genuinely label-free electrochemical 11 biosensors are rare, and their study is still open toward achieving better detection performance. 12 Electrochemical impedance spectroscopy is a very sensitive technique and widely used in 13 biosensors (Strong et al., 2021), and its non-faradaic version only requires a reagentless support 14 electrolyte to perform the detection (Cecchetto et al., 2017; Rosati et al., 2019; Stevenson et al., 15 2018). The detection of small DNA fragments by hybridization in a label-free approach was 16 proposed (Berggren et al., 1999), but it still remains a challenge in bioelectrochemistry. 17 Nevertheless, new nanomaterials and platforms are making this type of detection more and more 18 tangible. 19 The use of screen-printed carbon electrodes (SPCE) offers great advantages in 20 bioelectrochemistry, since they are inexpensive and simple to build, though in general they need an 21 additional matrix modification to be a suitable substrate for biomolecule immobilization 22 (Thiyagarajan et al., 2014). Nanomaterials derived from carbon are excellent candidates for 23 modifying SPCE, as they have an inherent affinity for the surface of the SPCE, can be conductive 24 and may have functional groups that facilitate the immobilization of biomolecules (Quesada- 25 González and Merkoçi, 2018; Ribeiro et al., 2020; Vermisoglou et al., 2020). Recently, new 4 1 fluorographene derivatives have been discovered (Chronopoulos et al., 2017), including graphene 2 acid (Bakandritsos et al., 2017). 3 Graphene acid (GA) is a selectively, homogenously, and densely functionalized carboxylic 4 graphene derivative. Its functionalization degree of 13 at% (corresponding to more than 30 mass% 5 in carboxylic groups), its sheet resistance of 6800 Ω sq-1, which is five orders of magnitude lower 6 than that of graphene oxide, render GA the most conductive graphene derivative with such a high 7 content in carboxyl groups. The titration profile of GA closely resembles that of molecular organic 8 acids, with pKa of 5.2, reflecting its well-defined structure. Such properties shape a great 9 combination for bioelectrochemistry applications, allowing attachment of biomolecules on a 10 conductive support (Seelajaroen et al., 2020). Its use was reported for chemical sensors, catalysis 11 and supercapacitor applications (Blanco et al., 2019; Heng Cheong et al., 2019; Lenarda et al., 12 2019; Šedajová et al., 2020; Seelajaroen et al., 2020), but its potential application in biosensors has 13 not yet been studied. 14 To this end, in this work we developed a new DNA biosensor for the detection of 15 adulteration of beef with pork, utilizing as matrix graphene acid on SPCE via non-faradaic 16 impedance spectroscopy, which is a label-free and reagentless electrochemical technique, measuring 17 the admittance and capacitance of the biosensor surface after the DNA hybridization events. 18 19 2 MATERIAL AND METHODS 20 2.1 Chemicals and apparatus 21 All reagents were of analytical grade, without previous purification. Ultra-high purity water 22 (18.2 MΩ cm-1, Mili-Q, USA) was used for preparation of all solutions. Phosphate Saline Buffer 23 solution (PBS, Sigma-Aldrich, USA, 10 mmol L-1) was used as electrolyte and for DNA dilutions. 24 K4[Fe(CN)6], K3[Fe(CN)6], Tris–HCl, EDTA, NaCl, H2SO4, Sodium dodecyl sulfate (SDS), 1- 25 Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and N-Hydroxysuccinimide (NHS) were 26 purchased from Sigma-Aldrich, USA. Solutions utilized in electrochemical analyzes were 5 1 deoxygenated by ultrapure nitrogen bubbling in time proportional to volume (1 min mL-1) before 2 use. 3 All oligonucleotides were obtained from Biomers.net GmbH, Germany. The immobilized 4 probe consisted on a single-stranded DNA oligonucleotide with the following sequence: 5'- 5 (NH2)AAAAAAAGCTGATAGTAGATTTGTGATGACCGTA-3'. This DNA probe has a 5' end 6 amination for better attachment with the nanomaterial, a spacer consisting of a sequence of 7 7 adenines nucleotides and a 27 base sequence complementary with the pig (Sus scrofa) 8 mitochondrial DNA, specifically with a region of the cytochrome b gene (Matsunaga et al., 1999). 9 To perform hybridization detection tests, complementary (5'- 10 TACGGTCATCACAAATCTACTATCAGC-3') and non-complementary (Bos taurus sequence: 5'- 11 CTAGAAAAGTGTAAGACCCGTAATATA-3') were used. 12 All electrochemical measurements were conducted in an Autolab PGSTAT302N potentiostat 13 (Netherlands) with a FRA2 module by NOVA 2.1.3. software, and data was analyzed and ploted 14 utilizing Origin 2018 software. All data was presented as mean ± standard deviation, with an n = 3. 15 16 2.2 Meat sample preparation 17 Fig. 1 shows a scheme of the meat sample preparation. Knuckle beef and pork loin of low 18 fat were purchased from a butchery in Barcelona, Catalunya, Spain. 19 20
21 22 The sample preparation was carried out according to literature with modifications (Khairil 23 Mokhtar et al., 2020). 5.0 g of each meat samples where blend with 10 mL of ice-cold extraction 24 buffer (0.2 mol L-1 Tris–HCl, 0.01 mol L-1 EDTA, 0.5 mol L-1 NaCl and 1% SDS v/v) with a hand 25 blender for 5 minutes. After that, the samples were sonicated for 15 min at 80 ºC. This step is 26 crucial to disrupt both cell and mitochondrial membranes, as well to facilitate DNA fragmentation. 6 1 Mitochondrial DNA is a large circular molecule of approximately 16.600 base pairs (bp), thus 2 requiring smaller fragments to facilitate its interaction with the 27 bp DNA probe attached on the 3 surface of the biosensor. The resulting solution was centrifuged (12.000 rpm, 15 min) and the upper 4 phase was recovered. To minimize time and cost, no purification step was performed further. The 5 DNA concentration in both beef and pork solutions were measured with a Nanodrop 6 spectrophotometer (Thermo Fischer, USA), normalized with PBS solution to 100 ng mL-1 and 7 stored at -20 ºC until use. 8 9 2.3 SPCE construction and modification with graphene acid 10 Lab-made SPCE were printed on heat-stabilized polyester films (Mac Dermide Enthone, 11 USA) in a DEK 248 (UK) semi-automatic printer, utilizing commercial inks (Gwent Group/Sun 12 Chemical, UK) by layer-by-layer strategy as follows: firstly C2180423D2 Silver Paste was printed 13 to pattern the conductive paths and auxiliary electrodes, then C2130809D5 Silver/Silver Chloride 14 Paste was printed for reference electrode, afterwards C2030519P4 Carbon Paste was printed for 15 work (Ø = 3.1 mm) auxiliary electrodes. D2070423P5 Dielectric Paste was used for contact 16 isolation and delimitation of the electrode areas. After each printing step, the substrate with printed 17 materials was cured in the oven at 110 °C for 30 min. Prior to use, the SPCE were preconditioned 18 by cyclic voltammetry in sulfuric acid solution (0.1 mol L-1) between -0.5 and 1.2 V, 10 cycles, 100 19 mV s-1, to ensure a surface free of impurities and to obtain a common voltammetric profile. Then, 20 they were rinsed with deionized water and dried with N2 gas. Graphene acid (GA) was synthetized 21 as previous works (Bakandritsos et al., 2017). A GA suspension (1.00 mg mL-1) was sonicated for 22 10 minutes then centrifuged (14.000 rpm, 10 min). The concentration of the suspension was 23 mensurated before and after centrifugation with UV-vis spectroscopy (λ = 660 nm). The GA 24 supernatant was drop-casted onto the working area of the SPCE (5 µL, dried at 37 ºC). The number 25 of drop-casting rounds were optimized, with four consecutive rounds being tested, to ensure a 7 1 complete cover of the SPCE working area. Then, the SPCE was washed by immersion in PBS for 2 10 s and dried with N2. 3 4 2.4 Genosensor construction 5 6
7 8 Fig. 2 shows a scheme of the electrode modification and the genosensor construction. In the 9 modified SPCE working area, 10 µL of EDC/NHS solution (50 and 30 mmol L-1 respectively, 10 freshly prepared in PBS) was added and kept in room temperature (25 ºC) for 1 hour, aiming the 11 activation of the carboxyl groups of the GA and consequently attachment of the DNA probe. The 12 electrodes were then washed by immersion in PBS for 10 s and dried with N2. 10 µL of the DNA 13 probe solution (10 µmol L-1, PBS) was dropped onto the working area of the SPCE, and it was kept 14 at 4 ºC for 12 h for the formation of the amide bonds between the nanomaterial and the 5’ amino- 15 terminal group of the DNA probe. Then, the electrodes were washed by immersion in PBS for 10 s 16 and dried with N2. To block possible active EDC/NHS sites, 10 µL of ethanolamine solution (1 mol 17 L-1 in water) was added onto the working area of the SPCE and kept in room temperature for 10 18 minutes. Finally, the electrode was washed by immersion in PBS for 10 s and dried with N2. This 19 was the ready-to-use biosensor. To determine whether the immobilization process of the DNA probe 20 was effective, Fourier-transform infrared spectroscopy (FTIR) analyzes were performed on the 21 surface of the working electrode utilizing a PerkinElmer spectrophotometer in ATR (Attenuated 22 Total Reflectance) mode, with and without the use of the EDC/NHS protocol, to verify the 23 formation of the amide bonds. 24 25 2.5 Electrochemical characterization of the electrodes 8 1 Bare SPCE, SPCE modified with GA and the ready-to-use biosensor were characterized by 2 cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) employing the well- 3 known ferrocyanide/ferricyanide redox pair (K4[Fe(CN)6] and K3[Fe(CN)6], 1 mmol L-1, in PBS 4 solution 10 mmol L-1). For the CV measurements 5 potential cycles from -0.5 to 0.8 V, 50 mV s-1 5 were applied, in which the anodic peak current (Ipa) and potential (Epa) values where monitored. For 6 EIS measurements 0.12 V (half-wave potential of the redox pair) was applied, in a frequency range 7 from 10000 Hz to 0.1 Hz, amplitude 10 mV. The resistance to charge transfer (R CT) values were 8 calculated utilizing NOVA 2.1.3 software, fitting the experimental data of the Nyquist plots in the 9 equivalent circuit, with a χ2 < 10-2. 10 11 2.6 Morphological characterization of the electrodes 12 The surface morphology of the electrodes was assessed by scanning probe microscopy 13 (Model 5100 N, Hitachi), performed in atomic force microscopy (AFM) mode, and by scanning 14 electron microscopy (SEM) on an EVO MA10 Zeiss microscope. AFM microscopies were analyzed 15 with Gwyddion 2.58 software. 16 17 2.7 Faradaic detections of complementary target and real samples 18 Square wave voltammetry (SWV) and faradaic EIS detections were performed onto the 19 electrode to verify successful DNA hybridization, utilizing complementary and non-complementary 20 oligonucleotide sequences as target. 10 µL of the target solution (10 µmol L-1, PBS) was applied 21 onto the SPCE working area and kept at 57 ºC (the annealing temperature of the probe DNA with 22 the complementary target) for 20 min. The electrode was washed by immersion in PBS for 10 s and 23 dried with N2. The values of peak current (Ip) were monitored by SWV, applying a potential from - 24 0.1 V to +0.4 V, amplitude 25 mV, increment 4 mV, frequency 15 Hz, in ferrocyanide/ferricyanide 25 redox pair solution (K4[Fe(CN)6] and K3[Fe(CN)6], 1 mmol L-1, in PBS 10 mmol L-1. Faradaic EIS 26 was carried out as described in item 2.5. The prepared pure beef and pork samples were also used 9 1 for faradaic detection to ensure that its components do not interfere with the DNA hybridization 2 onto the biosensor surface. 3 4 2.8 Non-faradaic detections of real samples 5 For the construction of a calibration curve, the normalized meat samples were mixed in 6 different proportions (v/v) as follows: 100% pork (total positive sample), 100% beef (total negative 7 sample), a mixture of 50% each, 90% beef with 10% pork and 99% beef with 1% pork. 10 µL of the 8 samples were applied onto the SPCE working area and kept at 57 ºC for 20 min. The electrode was 9 washed by immersion in PBS for 10 s and dried with N2. A blank sample (PBS only) was also used. 10 Electrochemical detection of the target was performed by non-faradaic EIS (0 V, frequency range 11 from 10000 Hz to 0.1 Hz, amplitude 10 mV) in PBS solution. The admittance and impedance 12 values were calculated utilizing the NOVA 2.1.3 software, fitting the experimental data of the 13 Nyquist and Bode plots in the equivalent circuit, with a χ2 < 10-2. 14 15 2.9 Stability study 16 To determine the stability of the genosensor, the prepared bioelectrodes were maintained in a 17 sealed container, with atmospheric air at room temperature for 4 weeks. Analysis by non-faradaic 18 EIS using pure pork samples were conducted each week, in order to evaluate whether the biosensor 19 was still functional and could recognize its target. 20 21 3 RESULTS AND DISCUSSION 22 3.1 Electrochemical characterizations 23 It was observed that from the initial 1.00 mg mL-1 of suspended graphene acid, an average 24 of 0.45 mg mL-1 was obtained in the final supernatant after centrifugation. The sonication and 25 centrifugation steps of the graphene acid suspension aim to discard large clumps of nanomaterial, in 26 order to obtain the lightest and loosest sheets to enhance the drop-cast procedure. The repetition of 10 1 the drop-casting for two consecutive times showed the lowest resistance values and the highest 2 current (Fig. S1), so it was maintained for further analysis. That can be explained because one drop- 3 casting round is not enough to cover the entire working electrode surface with the nanomaterial, 4 while in three repetitions or more there is an accumulation of nanomaterial, and its electronic 5 properties diminished. 6 The electrochemical properties of the graphene acid-coated SPCE were probed with cyclic 7 voltammetry in ferro/ferricyanide aqueous electrolyte (Fig. 3A, data in Table S1). For the case of 8 the bare SPCE, the oxidation potential of ferrocyanide was the highest measured (0.301 ± 0.021 V), 9 while the anodic peak current was the lowest (14.39 ± 1.56 µA). After the modification of SPCE 10 with graphene acid, a decrease in the oxidation potential (0.176 ± 0.015 V) and an increase in 11 anodic peak current (23.48 ± 2.24 µA) were recorded, indicating a more facile oxidation of 12 ferrocyanide, and improved charge-transport properties, respectively. These data support the 13 capability of graphene acid to promote the electrocatalytic redox reactions on the electrode, due to 14 its higher conductivity, as shown by the 14 times higher charge transfer resistance of the bare SPCE 15 (12.54 ± 1.19 kΩ, Fig. 3B) in comparison to the graphene acid modified one (0.85 ± 0.07 kΩ). The 16 improved performance may also be affected by the higher surface area of the electrode after 17 deposition of graphene acid (as later discussed). 18
19 20 The modification of the electrode with the DNA probe decreased the oxidation current 21 (17.83 ± 1.64 µA) and shifted the potential to a slightly anodic value (0.204 ± 0.018 V) in 22 comparison with the graphene acid-modified SPCE. This can be attribute to two main causes: 1) the 23 inherent steric hindrance of the DNA molecules that prevents the ferro/ferricyanide ions from 24 reaching the electrode surface, and 2) the negative net charge of the biomolecule, due to the 25 presence of negatively charged phosphate groups in a neutral medium (PBS, pH = 7.4), thus also 26 repelling the negatively charged ferro/ferricyanide ions. Likewise, the resistance to charge transfer 11 1 mean increased in 2 kΩ after the DNA probe incorporation, indicating that the biomolecule is well 2 incorporated into the working electrode surface by the EDC/NHS mechanism which links the 3 carboxyl groups of the nanomaterial to the amino group of the biomolecule. The FTIR analysis 4 (Fig. S2) also indicates the bond formation between the DNA probe and the electrode surface with 5 the EDC/NHS protocol, as there is an increase in the band region between 1510 to 1700 cm-1, which 6 involves the 1665 cm-1 characteristic amide I band, as well a small band in 1700 cm-1 corresponding 7 to the amide carbonyl band. There was also an increase in the amide A band (around 3500 cm-1), 8 another characteristic band, showing the formation of the amide bond between the carboxylic group 9 of GA and the amino group of the DNA 5' end, mediated by the EDC/NHS chemical mechanism. 10 11 3.2 Morphological characterizations 12 The analysis of the electrode morphology can provide interesting data in relation to the 13 behavior of the nanomaterial on the solid surface and can also corroborate with the electrochemical 14 analysis. Fig. 4 shows SEM and AFM images of the working electrode in the genosensor 15 construction steps and Table S2 summarizes some surface parameters extracted from AFM 16 microscopy. Printed carbon electrodes usually have rough texture due to ink constituents, as can be 17 observed in the morphology in SEM images and the roughness coefficient of the bare SPCE (120.15 18 ± 9.78 nm, surface area of 33.19 ± 4.91 µm²). The addition of GA promotes an increase in this 19 inherent roughness (Rq = 209.84 ± 21.67 nm), in addition to an increase of more than three times in 20 the surface area of the electrode (106.54 ± 11.45 µm²). The surface area increase can be associated 21 with the electrochemical results which showed an increase in the current and a decrease in the 22 charge transfer resistance after graphene acid deposition. The incorporation of the DNA probe 23 caused a small drop in the roughness and surface area values (183.67 ± 15.45 nm and 90.47 ± 10.56 24 µm², respectively), that can be attributed to the formation of cross-links between the nanomaterial 25 and the biomolecule. 26 12 1
2 3 3.3 Faradaic detections of complementary target and meat samples 4 To ensure that the genosensor is working properly and the real sample is recognized by the 5 immobilized DNA probe, complementary targets and pure pork and beef samples were detected by 6 SWV (Fig. 3C) and by faradaic EIS (Fig. 3D). Without any target, the genosensor presented with 7 the highest peak current and lowest resistance to charge transfer values (IP = 12.63 ± 1.43 µA and 8 RCT = 2.91 ± 0.21 kΩ, Table S3), since electron transfer is less impeded in absence of the 9 complementary target. Similar values were recorded in presence of non-complementary target (IP = 10 11.35 ± 1.09 µA, RCT = 3.14 ± 0.29 kΩ) and the beef sample (IP = 11.07 ± 1.04 µA, RCT = 3.75 ± 11 0.36 kΩ), highlighting the absence of any binding events on the graphene acid modified SPCE and 12 the selectivity of the genosensor. On the other hand, the application of the complementary DNA 13 solution decreased the Ip and raised the RCT values (7.26 ± 0.85 µA, 6.23 ± 0.57 kΩ). Due to 14 hybridization of the complementary DNA strands, the quantity of biomolecules on the surface 15 increases and leads to greater steric hindrance, which obstructs the ferrocyanide ions from reaching 16 the electrode surface. The presence of pork sample resulted in the lowest peak current and highest 17 resistance to charge transfer values (6.89 ± 0.71 µA, 7.64 ± 0.71 kΩ), because it is a more complex 18 sample and contains larger fragments of mitochondrial DNA, thus creating a greater electron 19 transfer barrier during the redox events. The DNA probe sequence was selected to be 20 complementary to the pork cytochrome b gene because that is a barcode gene, as it is used to 21 discern different species as it is highly specific. In addition, the use of a mitochondrial DNA 22 sequence has the advantage that this genetic material is present in greater amounts in muscle tissue 23 than nuclear DNA, facilitating its detection. 24 25 3.4 Non-faradaic detection of the meat samples 13 1 Pure pork and beef samples were analyzed by non-faradaic EIS, and the Nyquist and Bode 2 plots where extracted (Fig. 5A and 5B, respectively). Non-faradaic impedance was employed 3 because it is the only operative electrochemical technique in the absence of any redox label in the 4 electrolyte solution. Thus, the charge transfer-related parameters (RCT and Warburg diffusional 5 impedance) are negligible, and the equivalent Randles circuit can be simulated only by considering 6 the solution resistance (RS) and a capacitor element. Due to the use of an imperfect rough carbon 7 electrode, the capacitor element is substituted for a constant phase element (CPE). In this case, the 8 Nyquist plot assumes a straight-line (Fig. 5A) that forms an angle to the real impedance axis. This 9 angle is proportional to how similar the CPE is in relation to a real capacitor, as a perfect capacitor 10 assumes a vertical straight-line in the Nyquist Plot. The Impedance Bode plot (Fig. 5B, black lines) 11 forms two regions: a horizontal region at high frequencies, corresponding to the resistance of the 12 electrolytic solution, and a sloped region at low frequencies, corresponding to the CPE. In the Bode 13 plot from Figure 5B, at frequencies of 10 Hz and below, the impedance response is driven by the 14 capacitive reactance, indicating the non-faradaic charging of the double layer (Stevenson et al., 15 2018). 16 17
18 19 The CPE admittance values (Y0) can be obtained by fitting the experimental data with a 20 simple equivalent circuit. Admittance can be defined as how easily a circuit or device will allow a 21 current to flow, and it varies according to the electrode surface coverage (Lee and Shim, 2001). In a 22 non-faradaic experiment, the presence of biomolecules on the electrode surface will disturb the 23 delicate electrical double layer formation due to electrostatic interactions with the electrolyte 24 solution ions. Therefore, the higher the concentration of biomolecules on the electrode surface, the 25 lower the admittance values of the formed double electric layer against the sinusoidal perturbation. 26 Considering the greater retention of pork DNA on the electrode surface, due to its hybridization 14 1 with the immobilized probe, a major disruption in the double electrical layer is observed, as 2 indicated by a decrease of 22% in its CPE admittance values (38.23 ± 0.76 µS sN in pork versus 3 49.44 ± 0.98µS sN in beef). 4 The impedance values also demonstrated differences between the samples: in low frequency 5 regions (starting in 10 Hz, constant after that) there was an 20% increase in the total impedance 6 values (Z) with the addition of the pork sample (50.76 ± 1.54 kΩ in f = 1 Hz) in comparison to the 7 beef sample (40.69 ± 0.85 kΩ in f = 1 Hz). This observation corroborates with the decrease in the 8 admittance because they are inverse parameters: the presence of more biomolecules on the electrode 9 surface results in a greater steric hindrance, leading to a greater impedance at lower frequencies to 10 the sinusoidal disturbance. 11 12 3.5 Calibration curves and genosensor stability 13 Two calibration curves were constructed, based on the admittance values (Fig. 5C) and on 14 the total impedance values at low frequency (Fig. 5D). A linear behavior was observed for both 15 variables, and the experimental calibration curves were fitted by linear equations: Y0 = 49.123 - 16 0.109*(pork content percentage), r = 0.997 (n = 3) and Z = 40.95 + 0.096*(pork content 17 percentage), r = 0.991 (n = 3). We verified that the admittance was a more adequate parameter for 18 measurement, since it presented a slightly higher slope value, which indicates a greater sensitivity, 19 and a higher r coefficient, implying better linearity, both desired analytical parameters for 20 biosensors. The calculated limit of detection (LOD = 3*blank standard deviation/slope) was 9.35% 21 of pork content in beef. As this LOD was obtained with a non-purified and non-amplified sample, it 22 is possible that trace levels detection could be achieved with sample purification and amplification 23 steps. 24 The stability study (Fig. 5E) showed that the biosensor remained functional even after 4 25 weeks of storage at room temperature, as the electrochemical parameters of the detection of pork 15 1 residues in beef remained practically intact: after 28 days the signals varied by 3.2% for admittance 2 and 3.6% for impedance. 3 Biosensors based in nucleic acids for pork detection in meat samples are scarce in the 4 literature (Table 1). Most reported genosensors require long sample preparation and assay time, 5 while platform stability studies of such sensors remained vastly unexplored, although it is an 6 important parameter for defining the shelf-life of the biosensor, dictating its future potential for 7 commercialization. Although this work reports a higher detection limit than some previously 8 developed biosensors, it is necessary to emphasize that it did not require any DNA purification or 9 amplification technique, nor was any electrochemical indicator or redox molecule required for 10 detection, and it presented a simple drop-casting protocol for electrode surface modification. These 11 features render the present pork-residue detecting genosensor as the most sensitive one among the 12 label-free and reagentless sensors in the field. 13 14 4 CONCLUSIONS 15 The combination of a cost effective lab-made screen-printed carbon electrode, with a 16 selective and densely functionalized graphene derivative with carboxyl groups, and further modified 17 with a single stranded DNA probe, resulted in an impedimetric diagnostic device that can detect 18 adulteration of beef by pork with a limit of detection of 9% w/w. Importantly, the overall protocol 19 and the impedimetric technique are simple, without requiring expensive and/or hazardous reagents, 20 neither extensive purification steps or amplification of the target genetic material in the sample. 21 Furthermore, the method has a sample preparation and assay time of 45 minutes, and the sensor 22 retains its full functional performance for at least 4 weeks of storage. 23 These set of features places the present pork DNA sensor among the most attractive for 24 further development and commercialization. Electrochemical biosensors that use non-faradic 25 techniques are scarce in the scientific literature, especially employing DNA as a biorecognition 26 element. Therefore, this work lays the ground for the development of new biosensing devices that 16 1 do not require specific reagents for detection and that are truly label-free, powered by the recent 2 achievements in fluorographene’s chemistry. In addition, the reported platform has the potential for 3 immobilization of other DNA sequences specific to a multitude of targets, thus tremendously 4 expanding its application potential in the biomedical, crop and food analysis fields. 5 6 Acknowledgments 7 We acknowledge the financial support from the EU Graphene Flagship Core 2 Project (No. 8 785219). This article reflects only the author’s view, and the European Commission is not 9 responsible for any use that may be made of the information it contains. ICN2 is funded by the 10 CERCA programme, Generalitat de Catalunya. The ICN2 is supported by the Severo Ochoa Centres 11 of Excellence programme, funded by the Spanish Research Agency (AEI, grant no. SEV-2017- 12 0706). J. M. R. Flauzino is grateful for the Coordenação de Aperfeiçoamento de Pessoal de Nível 13 Superior scholarship (CAPES-PRINT, Brazil, Grant number: 88887.371591/2019-00). D. Panáček 14 acknowledges the Internal Student Grant Agency of the Palacký University in Olomouc, 15 (IGA_PrF_2021_031). A. G. Brito-Madurro acknowledges the funding from Conselho Nacional de 16 Desenvolvimento Científico e Tecnológico - CNPq (310782/2018-0). J. M. Madurro acknowledges 17 the funding from Fundação de Amparo à Pesquisa do Estado de Minas Gerais – FAPEMIG (CEX- 18 APQ-02902-17) and Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq 19 (311737/2018-8). A. Bakandritsos acknowledges the funding from the Czech Science Foundation, 20 (project GA CR – EXPRO, 19-27454X). M. Otyepka acknowledges the ERC grand 2D-CHEM (No 21 683024 from H202). The work was supported also by the ERDF/ESF project "Nano4Future" 22 (No.CZ.02.1.01/0.0/0.0/16_019/0000754). 23 24 25 26 17 1 References 2 3 Ahmed, M.U., Hasan, Q., Mosharraf, M.H., Saito, M., Tamiya, E., 2010. Food Control 21, 599– 4 605. 5 Ali, M.E., Hashim, U., Mustafa, S., Che Man, Y.B., Yusop, M.H.M., Kashif, M., Dhahi, T.S., Bari, 6 M.F., Hakim, M.A., Latif, M.A., 2011. J. Nanomater. 2011, 1–11. 7 Ali, M.E., Hashim, U., Mustafa, S., Che Man, Y.B., Islam, K.N., 2012. J. Nanomater. 2012. 8 Alves-Balvedi, R.P.P., Caetano, L.P.P., Madurro, J.M.M., Brito-Madurro, A.G.G., 2016. Biosens. 9 Bioelectron. 85, 226–231. 10 Bakandritsos, A., Pykal, M., Błoński, P., Jakubec, P., Chronopoulos, D.D., Poláková, K., 11 Georgakilas, V., Čépe, K., Tomanec, O., Ranc, V., Bourlinos, A.B., Zbořil, R., Otyepka, M., 12 2017. ACS Nano 11, 2982–2991. 13 Berggren, C., Stålhandske, P., Brundell, J., Johansson, G., 1999. Electroanalysis 11, 156–160. 14 Blanco, M., Mosconi, D., Otyepka, M., Medveď, M., Bakandritsos, A., Agnoli, S., Granozzi, G., 15 2019. Chem. Sci. 10, 9438–9445. 16 Cecchetto, J., Fernandes, F.C.B., Lopes, R., Bueno, P.R., 2017. Biosens. Bioelectron. 87, 949-956. 17 Chen, H.J., Chen, R.L.C., Hsieh, B.C., Hsiao, H.Y., Kung, Y., Hou, Y. Te, Cheng, T.J., 2019. 18 Biosens. Bioelectron. 131, 53–59. 19 Chronopoulos, D.D., Bakandritsos, A., Pykal, M., Zbořil, R., Otyepka, M., 2017. Appl. Mater. 20 Today. 9, 60–70. 21 De Castro, A.C.H., Kochi, L.T., Moço, A.C.R., Coimbra, R.S., Oliveira, G.C., Cuadros-Orellana, S., 22 Madurro, J.M., Brito-Madurro, A.G., 2018. J. Solid State Electrochem. 22, 2339–2346. 23 De La Escosura-Muñiz, A., Baptista-Pires, L., Serrano, L., Altet, L., Francino, O., Sánchez, A., 24 Merkoçi, A., 2016. Small 12, 205–213. 25 El Sheikha, A.F., 2019 Trends Food. Sci. Technol. 86, 544–552. 26 Flauzino, J.M.R., Peres, R.C.S., Alves, L.M., Vieira, J.G., dos Santos, J.G., Brito-Madurro, A.G., 18 1 Madurro, J.M., 2021a. Bioelectrochemistry 140, 107801. 2 Flauzino, J.M.R., Pimentel, E.L., Alves, L.M., Madurro, J.M., Brito-Madurro, A.G., 2021b. 3 Electroanalysis 33, 296-303. 4 Hansen, A., 2018. Geoforum 93, 57–68. 5 Hellberg, R.S., Hernandez, B.C., Hernandez, E.L., 2017. Food Control 80, 23–28. 6 Heng Cheong, Y., Nasir, M.Z.M., Bakandritsos, A., Pykal, M., Jakubec, P., Zbořil, R., Otyepka, M., 7 Pumera, M., 2019. ChemElectroChem 6, 229–234. 8 Idili, A., Parolo, C., Alvarez-Diduk, R., Merkoçi, A., 2021. ACS Sensors 9 Irwin, D.M., Kocher, T.D., Wilson, A.C., 1991. J. Mol. Evol. 32, 128–144. 10 Khairil Mokhtar, N.F., El Sheikha, A.F., Azmi, N.I., Mustafa, S., 2020. J. Sci. Food Agric. 100, 11 1687–1693. 12 Kumar, Y., Bansal, S., Jaiswal, P., 2017. Compr. Rev. Food Sci. Food Saf. 16, 1359–1378. 13 Lee, T.Y., Shim, Y.B., 2001. Anal. Chem. 73, 5629–5632 14 Lenarda, A., Bakandritsos, A., Bevilacqua, M., Tavagnacco, C., Melchionna, M., Naldoni, A., 15 Steklý, T., Otyepka, M., Zbořil, R., Fornasiero, P., 2019. ACS Omega 4, 19944–19952. 16 Matsunaga, T., Chikuni, K., Tanabe, R., Muroya, S., Shibata, K., Yamada, J., Shinmura, Y., 1999.. 17 Meat Sci. 51, 143–148. 18 Merkoçi, A., 2010. Biosens. Bioelectron. 26 (4), 1164-1177 19 Moço, A.C.R., Guedes, P.H., Flauzino, J.M.R., da Silva, H.S., Vieira, J.G., Castro, A.C.H., Gomes, 20 É.V.R., Tolentino, F.M., Soares, M.M.C.N., Madurro, J.M., Brito‐Madurro, A.G., 2019. 21 Electroanalysis 31, 1580–1587. 22 Montiel, V.R.V., Gutiérrez, M.L., Torrente-Rodríguez, R.M., Povedano, E., Vargas, E., Reviejo, 23 Á.J., Linacero, R., Gallego, F.J., Campuzano, S., Pingarrón, J.M., 2017. Anal. Chem. 89, 24 9474–9482. 25 Okuma, T.A., Hellberg, R.S., 2015. Food Control 50, 9–17. 26 Oliveira, D.A., de Rezende Rodovalho, V., Flauzino, J.M.R., da Silva, H.S., Araujo, G.R., Vaz, 19 1 E.R., Vieira, C.U., Madurro, J.M., Madurro, A.G.B., 2018a. Protein Pept. Lett. 25, 878–883. 2 Oliveira, D.A., Silva, J. V., Flauzino, J.M.R., Castro, A.C.H., Moço, A.C.R., Soares, M.M.C.N., 3 Madurro, J.M., Brito-Madurro, A.G., 2018b. Anal. Biochem. 549, 157–163. 4 Quesada-González, D., Merkoçi, A., 2018. Chem. Soc. Ver. 47, 4697–4709. 5 Rabti, A., Raouafi, N., Merkoçi, A., 2016. Carbon 108, 481–514. 6 Rahmati, S., Julkapli, N.M., Yehye, W.A., Basirun, W.J., 2016. Food Control 68, 379–390. 7 Ratajczak, K., Krazinski, B.E., Kowalczyk, A.E., Dworakowska, B., Jakiela, S., Stobiecka, M., 8 2018a. ACS Appl. Mater. Interfaces 10, 17028–17039. 9 Ratajczak, K., Krazinski, B.E., Kowalczyk, A.E., Dworakowska, B., Jakiela, S., Stobiecka, M., 10 2018b. Nanomaterials 8, 510. 11 Ribeiro, S.H.D., Alves, L.M., Flauzino, J.M.R., Moço, A.C.R., Segatto, M.S., Silva, J.P., Borges, 12 L.F.A., Madurro, J.M., Madurro, A.G.B., 2020. Electroanalysis 32 (10), 2316-2322 13 Rosati, G., Ravarotto, M., Sanavia, M., Scaramuzza, M., De Toni, A., Paccagnella, A., 2019. Sens 14 Bio-Sensing Res. 26, 100308. 15 Šedajová, V., Jakubec, P., Bakandritsos, A., Ranc, V., Otyepka, M., 2020. Nanomaterials 10, 1–14. 16 Seelajaroen, H., Bakandritsos, A., Otyepka, M., Zbořil, R., Sariciftci, N.S., 2020. ACS Appl. Mater. 17 Interfaces 12, 250–259. 18 Singh, P.K., Jairath, G., Ahlawat, S.S., Pathera, A., Singh, P., 2016. J. Food Sci. Technol. 53, 1759– 19 1765. 20 Stevenson, H., Radha Shanmugam, N., Paneer Selvam, A., Prasad, S., 2018. SLAS Technol. 23, 5– 21 15. 22 Strong, M.E., Richards, J.R., Torres, M., Beck, C.M., La Belle, J.T., 2021. Biosens. Bioelectron. 23 177, 112949. 24 Thiyagarajan, N., Chang, J.L., Senthilkumar, K., Zen, J.M., 2014. Electrochem. Commun. 38, 86– 25 90. 26 Torelli, E., Manzano, M., Marks, R.S., 2017. Sens. Actuators B Chem. 247, 868–874. 20 1 Vermisoglou, E., Panáček, D., Jayaramulu, K., Pykal, M., Frébort, I., Kolář, M., Hajdúch, M., 2 Zbořil, R., Otyepka, M., 2020. Biosens. Bioelectron. 166. 3 Wilson, J.M., Platts-Mills, T.A.E., 2018. Mol. Immunol. 100, 107–112. 4 5 FIGURES 6 7 Figure 1 – Scheme of the meat sample preparation steps. 8 9 Figure 2 – Scheme of the SPCE modification and genosensor construction. 21 1 2 Figure 3 – Electrochemical characterization of the genosensor construction steps: (A) cyclic 3 voltammograms of the bare screen-printed carbon electrode (SPCE), SPCE modified with graphene 4 acid (SPCE/GA) and SPCE/GA after the immobilization of the DNA probe (SPCE/GA/DNA 5 Probe). (B) Nyquist plots obtained by EIS analysis in each genosensor construction step; inset 1: 6 equivalent Randles circuit, where RS, RCT, W, and CPE represent the solution resistance, the 7 charge-transfer resistance, the Warburg diffusion resistance, and the constant phase element, 8 respectively; inset 2: high-frequency region zoom. Faradaic detections of complementary target and 9 meat samples: (C) square wave voltammograms and (D) Nyquist plots obtained by EIS; inset: 10 equivalent Randles circuit. Electrolyte: PBS (10 mmol L-1), containing K4[Fe(CN)6] and 11 K3[Fe(CN)6] (1 mmol L-1). CV parameters: potential range from -0.5 to 0.8 V, 50 mV s-1. EIS 12 parameters: frequency range from 10000 Hz to 0.1 Hz, potential applied 0.12V, amplitude 10 mV. 22 1 SWV parameters: potential range from -0.1 V to +0.4 V, amplitude 25 mV, increment 4 mV, 2 frequency 15 Hz. 3 4 Figure 4 – Scanning Electron and Atomic Force micrographs of the working electrode of the bare 5 SPCE (A, D), SPCE modified with graphene acid (B, E) and SPCE modified with graphene acid 6 and the DNA probe (C, F). 7 8 23 1 2 Figure 5 – Detection of pork and beef samples with non-faradaic electrochemical impedance 3 spectroscopy: (A) Nyquist plots. Inset: equivalent Randles circuit where RS is the electrolyte 4 solution resistance and CPE is the constant phase element. (B) Bode plots with total impedance (Z) 5 and Phase values in function of the frequency. Calibration curves of the developed biosensor based 6 on (C) admittance values and (D) total impedance values (f = 1 Hz). (E) Stability of the genosensor 7 response over time with a pure pork sample in function of time. Electrolyte: PBS (10 mmol L-1), 8 frequency range from 10000 Hz to 0.1 Hz, potential applied 0 V, amplitude 10 mV. Data is 9 presented as mean ± standard deviation, n = 3. 24 1 Table 1 – Reported genosensors for detection of pork DNA in beef samples. DNA purification Sample Detection Label-free and Detection Platform and/or preparation and Stability Ref method/technique Reagentless limit amplification assay time (Torelli et Optical fiber Chemiluminescent No 1% w/w Yes 150 min n. a. al., 2017) (Ali et al., Gold Nanoparticles Spectrophotometry Yes 10% w/w Yes n. a. n. a. 2012) (El Sheikha, Lateral Flow Naked eye No 0.1% w/w No 45 min n. a. 2019) Disposable Linear Sweep (Ahmed et electrochemical printed No 20% w/w Yes 65 min n. a. Voltammetry al., 2010) chips (Ali et al., Gold Nanoparticles Spectrophotometry No 1% w/w Yes 150 min n. a. 2011) Graphite disk modified SWV and Faradaic (Flauzino et with polymeric No 1% w/w No 60 min 4 weeks EIS al., 2021b) nanocomposite SPCE modified with Non-faradaic EIS Yes 9% w/w No 45 min 4 weeks This work Graphene Acid 2 n. a.: not available 25 1 LABEL-FREE AND REAGENTLESS ELECTROCHEMICAL GENOSENSOR BASED ON 2 GRAPHENE ACID FOR MEAT ADULTERATION DETECTION 3 4 José M. R. Flauzino, Emily P. Nguyen, Qiuyue Yang, Giulio Rosati, David Panáček, Ana G. Brito- 5 Madurro, João M. Madurro, Aristeidis Bakandritsos, Michal Otyepka, Arben Merkoçi 6 7 Supplementary Material 8 9 Figure S1 – Optimization of the drop-casting procedure of the graphene acid solution: electrochemical characterization 10 of the electrode surface. Electrolyte: PBS (10 mmol L-1), containing K4[Fe(CN)6] and K3[Fe(CN)6] (1 mmol L-1). CV 11 parameters: potential range from -0.5 to 0.8 V, 50 mV s-1. EIS parameters: frequency range from 10000 Hz to 0.1 Hz, 12 potential applied 0.12V, amplitude 10 mV. 13 26 1 2 Figure S2 – FTIR spectra of the SPCE with and without the application of the linker EDC/NHS. 3 4 Table S1 – Electrochemical data of the biosensor characterization steps (mean ± standard deviation, n=3) Electrode ipa (µA) Epa (V) RCT (kΩ) SPCE 14.39 ± 1.56 0.301 ± 0.021 12.54 ± 1.19 SPCE/GA 23.48 ± 2.24 0.176 ± 0.015 0.85 ± 0.07 SPCE/GA/DNA probe 17.83 ± 1.64 0.204 ± 0.018 2.91 ± 0.217 5 6 7 Table S2 – Surface parameters extracted from AFM microscopies. Electrode Surface Root Mean Square Roughness (nm) Surface Area (µm²) SPCE 120.15 ± 9.78 33.19 ± 4.91 SPCE/GA 209.84 ± 21.67 106.54 ± 11.45 SPCE/GA/DNA 183.67 ± 15.45 90.47 ± 10.56 8 27 1 2 3 4 Table S3 – Electrochemical data of faradaic detections of targets (mean ± standard deviation, n = 3) Target sample ip (µA) RCT (kΩ) None (probe only) 12.63 ± 1.43 2.91 ± 0.21 Non-Complementary target 11.35 ± 1.09 3.14 ± 0.29 Complementary target 7.26 ± 0.85 6.23 ± 0.57 Pure Beef 11.07 ± 1.04 3.75 ± 0.36 Pure Pork 6.89 ± 0.71 7.64 ± 0.71 5 6 7 Table S4 - Electrochemical data obtained from non-faradaic EIS detections (mean ± standard deviation, n = 3) Beef:Pork sample ratio (v/v) Y0 (µS sN) Z at 1 hz (kΩ) 100:0 (Pure beef) 49.44 ± 0.98 40.69 ± 0.85 99:1 48.96 ± 1.06 41,34 ± 0.95 90:10 47.78 ± 0.95 42.12 ± 0.98 50:50 43.56 ± 0.87 45.32 ± 1.16 0:100 (Pure Pork) 38.23 ± 0.76 50.76 ± 1.54 Blank (PBS) 52.68 ± 0.34 38.78 ± 0.76 8 9 10 28