Semiautomated Electrochemical Melting Curve Analysis Device for the Detection of an Osteoporosis Associated Single Nucleotide Polymorphism in Blood

The detection of single nucleotide polymorphisms (SNPs) is of increasing importance in many areas including clinical diagnostics, patient stratification for pharmacogenomics, and advanced forensic analysis. In the work reported, we apply a semiautomated system for solid-phase electrochemical melting curve analysis (éMCA) for the identification of the allele present at a specific SNP site associated with an increased risk of bone fracture and predisposition to osteoporosis. Asymmetric isothermal recombinase polymerase amplification using ferrocene labeled forward primers was employed to generate single stranded redox labeled amplicons. In a first approach to demonstrate the proof of concept of combining asymmetric RPA with solid-phase éMCA, a simplified system housing a multielectrode array within a polymeric microsystem, sandwiched between two aluminum plates of a heater device, was used. Sample manipulation through the microfluidic channel was controlled by a syringe pump, and an external Ag/AgCl reference electrode was employed. Individual electrodes of the array were functionalized with four different oligonucleotide probes, each probe equivalent in design with the exception of the middle nucleotide. The isothermally generated amplicons were allowed to hybridize to the surface-tethered probes and subsequently subjected to a controlled temperature ramp, and the melting of the duplex was monitored electrochemically. A clear difference between the fully complementary and a single mismatch was observed. Having demonstrated the proof-of-concept, a device for automated éMCA with increased flexibility to house diverse electrode arrays with internal quasi-gold reference electrodes, higher resolution, and broader melting temperature range was developed and exploited for the detection of SNP hetero/homozygosity. Using the optimized conditions, the system was applied to the identification of the allele present at an osteoporosis associated SNP site, rs2741856, in 10 real fingerprick/venous blood samples, with results validated using Sanger sequencing.

Electrode array fabrication.Nine-gold working electrode configuration: Briefly, a positive photoresist AZ1505 (MicroChemicals GmbH, Germany) was deposited by spin coating at 4000 rpm for 30 s on a pre-cleaned and dried glass slide.The photoresist was then exposed to UV light for 4 s using a chromium mask in contact mode (LED Paffrath GmbH, Rose FotoMasken, Germany) and the transferred pattern was developed using the commercial developer AZ726.Following development, the glass slide was introduced into the sputtering chamber (ATC Orion 8-HV, AJA International Inc., USA) and was subjected to an oxygen plasma etching using AC O 2 /Ar (5 cm 3 s −1 of Ar, 5 cm 3 s −1 of O 2 , 50 W) for 5 minutes.A layer of 30 nm of Ti/TiO 2 was then sputtered (oxygen flow rate: 5 cm 3 s −1 of O2 for the first 10 nm, then increased to 20 cm 3 s −1 for the last 5 nm.Ar flow rate: constant 5 cm 3 s−1).This was followed by the deposition of 100 nm of Au by AC sputtering (5 cm 3 s −1 of Ar, 5 cm 3 s −1 of O 2 , 50 W).Lift-off was achieved using sonication in acetone for 5 minutes, then 5 minutes in isopropanol and finally rinsing with Milli-Q water.
Electrochemical cleaning of electrode array and probe immobilisation.The array of 9 working gold electrodes and a common gold counter electrode were used with a classical reference electrode Ag/AgCl.Working electrodes were cleaned by sequential sonication in isopropanol, acetone and water for 5 minutes each, followed by electrochemical cleaning by cycling 10 times, firstly in 0.5 M KOH between 0.0 and -1.2 V vs Ag/AgCl, and secondly in 0.5 M sulfuric acid between 0.0 and 1.6 V vs Ag/AgCl at a scan rate of 100 mVs -1 .Washing steps in Milli-Q water were included in between both cleaning procedures and again at the end.After drying with nitrogen, the electrodes were immediately used for surface functionalization.
In the case of the screen-printed 64 electrode array, the cleaning procedure was significantly simpler based on sequential washing steps with 50 % v/v isopropanol in water, and water.After drying with nitrogen, the electrodes were immediately used for surface functionalisation.Thiolated capture probes (Table S1) (probe-A, probe-T, probe-C and probe-G) were selfassembled on each of the individual electrodes in duplicate, by spotting 1 µL of 1 µM thiolated capture probe solution + DT1 as backfiller with a 1:100 Probe:DT1 molar ratio freshly prepared in 1 M KH 2 PO 4 .This functionalization solution was left to assemble at 37°C for 3 h in humidity chamber to avoid evaporation, followed by a thorough wash with Milli-Q water and finally dried with nitrogen.
Sequences design.dbSNP database from National Center for Biotechnology Information (NCBI) was used to obtain the osteoporosis sequence related with SNP rs2741856.For the melting curve analysis a 138-mer synthetic target containing the SNP site and four probes sequences in such a way to capture the SNP in the target at the middle according to previous work 1 were designed (Table S1).Moreover, to mimic the genomic DNA, dsDNA should be produced and, specific primers to the synthetic 138-mer sequence were designed using Primer Blast software (Table S2).For electrochemical detection the redox label ferrocene (Fc) was introduced in the amplified product using a Fc-labelled forward primer in an RPA reaction.
Preparation of the dsDNA from synthetic 138-mer ssDNA by PCR, for further RPA amplification.A dsDNA was produced to mimic genomic DNA.This dsDNA was produced by PCR from a synthetic 138-mer sequence (Table S2).The PCR was carried out using a T100 thermal cycler (Biorad) following the protocol: 95C for 2 minutes, followed by 25 cycles at 95C for 30 seconds, 60C for 30 seconds, and 72C for 30 seconds, with a final elongation step at 72C for 5 minutes.One hundred microliters of PCR reaction mixture contained 1 unit of DreamTaq and DreamTaq buffer 1X, both forward and reverse primers without modifications at 200 nM, dNTPs at 200 µM and 100 pM synthetic 138-mer DNA as final concentrations.A DNA Clean & Concentrator™ kit was used to purify the products from the reaction mixture and the amplification products were visualized using agarose gel electrophoresis as described below.The concentration of the amplicon was measured using a SimpliNano spectrophotometer (Fisher Scientific, Spain).

Gel electrophoresis Analysis.
For gel electrophoresis analysis, a gel was made with certified molecular biology agarose gel powder (2.6% w/v) in 1× Tris-Borate-EDTA buffer (TBE) and stained with GelRed™ nucleic acid stain.A mixture of 5 µL of asymmetric-RPA product with 4 µL of loading buffer 6x was loaded per gel well and the electrophoresis was performed at 100 V, for 30 min and gels were visualized in a UV transilluminator.Finally, the samples were used immediately or stored at -20 °C until use.For better understanding of the gel, Image J software was used to calculate the intensity of the band and the obtained values were normalised by percentage and ploted in a graph.
Assymmetric RPA reaction.Two optimised conditions were carried out: 1) Reaction time optimisation for asymmetric RPA for single stranded DNA generation.A common master mix containing Fc-FwP (1000 nM) and reverse primer (200 nM) were prepared and aliquoted in five vials and the RPA reaction carried out for different times (5, 10, 15, 20 and 30 min).
2) Primer ratio optimisation for Asymmetric RPA for single stranded DNA generation.Using 15 min amplification, different ferrocene labelled forward primer:reverse primer (Fc-FwP:RevP) molar ratios were evaluated to find the optimum conditions for amplification and subsequent electrochemical detection.The reverse primer concentration was kept constant at 200 nM and the Fc-FwP concentrations (200, 500, 1000, 2000 and 5000 nM) was consecutively increased, resulting in Fc-FwP:RevP molar ratios of 1:1, 2:1, 5:1, 10:1 and 25:1 (Table S3).Table S1.Sequences of target and capture probes related with SNP used for éMCA (each one has a SNP in the middle).The regions of the target sequence complementary to the capture probes are underlined) and also the fully complementary capture probe: Probe-C.SNP-GG related to osteoporosis.

5′-pho-GAAGGTATGTAGGGGGTGGAGAAG-3′
Table S3.Ratio between forward and reverse primers used for optimisation of the asymmetric-RPA Table S4.Correlation between the SNP detected by the éMCA method and results obtained using Sanger sequencing 40 (used as a reference method).Each of the DPV curves in Figures S9 to S11 represents the DPV taken at different temperatures, as indicated in the inset on the right handside, with decreasing peaks, and peak currents due to the thermally induced dissociation of the ferrocene labelled target ssDNA.The raw data taken using the DPV measurement at the different temperatures are then used to create a malting curve for the determination of the melting temperature.Firstly, the baseline correction of DPV peaks was done using the GPES program.After that, the average of the maximum height (maximum current) of each replicate of DPV curves at each temperature was calculated.Finally, taking the DPV carried out at the starting temperature as 100%, the percentage of the DPV signal at each temperature was calculated and the melting profiles were constructed by plotting these percentages vs temperature.For fitting the curves, the mathematic Biphasic Dose Response Function was applied using the Levenberg Marquardt iteration algorithm and the melting temperatures were obtained from the extrema (maximum) of their first derivative.References:

The Process:
Initial Dispense: To communicate with the device, you must first establish a serial connection.
Install and run the Arduino IDE, then plug the device into your computer via USB.
Navigate to "Tools" and then select whichever port the device has connected to on your PC (this differs from PC to PC).
It is recommended to use the Arduino serial terminal, but if you are confident with another terminal or wish to integrate it into your own software, feel free to do so.
Next, select Arduino Uno as the Board.
Finally, navigate to the "Tools" menu and select "Serial Monitor".
In the bottom right corner of the serial monitor, set the baud rate to 115200 baud.Set the carriage return/new line drop down box to "No line ending".

Initial Dispense:
Upon establishing a serial connection, the user will be presented with the initial dispense menu.
In the first screenshot, it can be seen that "YES" has been typed into the top of the serial monitor.Upon pressing send or the enter key, this text will be sent to the device.
The second picture shows that the device has registered the "YES" command, as it has repeated it back and progressed to the next prompt.
The user will be asked whether they would like to dispense liquid for 1ms -10000ms.This allows the liquid being tested to be progressed through the piping, to the chip.By following the on-screen prompts, the user can run this initial dispense as many times as they desire before starting the main process.Once the initial dispense menu has been exited, the device must be restarted to regain access to this feature.

Ramp and Detection Time Setup:
The initial dispense menu has been excited by sending "NO" to "would you like to dispense?".
The user is prompted to enter a ramp time from 0 to 120 seconds.This is the time allowed between temperature intervals for the heaters to reach and settle at the new temperature

FiguresFigure
Figures Figure S1.(a) First generation of the set-up; (b) Break-out box developed by Labman for this project.

Figure S2 .
Figure S2.Electrode arrays used in the present work (a) 9 x gold working electrode array and (b) 64 x gold working array.

Figure S3 .
Figure S3.Real picture of the box containing a small pump and Arduino board with the internal view in the inset.The assembly of the 64 x working electrode array for the melting procedure for the second generation device.

Figure S4 .
Figure S4.(a) Top and front views of the second generation of the set-up developed for the present work.(b)Electrochemical Melting Curve System diagram.(c) Heater and channel detail.

Figure S5 .
Figure S5.Screenshot of the software specifically developed to run the electrochemical melting curve analysis prototype showing how the flow rate of the washing buffer and the rate of the temperature ramps are controlled, and the temperature continuously measured throughout.

Figure S6 .
Figure S6.Screenshot of the continuous DPV measurement throughout the temperature ramp.

Figure S7 .
Figure S7.Preparation of ssDNA with a higher excess of Fc-Forward primer than the used in Figure3of the manuscript (2.6 % w/v agarose gel after electrophoresis of asymmetric-RPA products obtained using PCR generated dsDNA as target)

Figure S9 .
Figure S9.Differential pulse voltammograms (DPV) recorded for each of the four probes hybridised to the Fc-labelled synthetic target after the sensor had been exposed to each temperature starting from 25 until 40 °C.The parameters employed in the DPV experiments were: potential window between 0 and 0.6 V (vs.Ag/AgCl 3M KCl), step potential 10 mV, modulation amplitude 10 mV, modulation time 0.015 s and interval time 0.1 s.

Figure S10 .
Figure S10.Differential pulse voltammograms (DPV) recorded for each of the four probes hybridised to the Fc-labelled ssDNA generated from blood (Sample 1) after the sensor had been exposed to each temperature starting from 25 until 40 °C.The parameters employed in the DPV experiments were: potential window between 0 and 0.6 V (vs.Ag/AgCl 3M KCl), step potential 10 mV, modulation amplitude 10 mV, modulation time 0.015 s and interval time 0.1 s.

Figure S11 .
Figure S11.Differential pulse voltammograms (DPV) recorded for each of the four probes hybridised to the Fc-labelled ssDNA generated from blood (Sample 1) after the sensor had been exposed to each temperature starting from 25 until 40°C.The parameters employed in the DPV experiments were: potential window between 0.1 and 0.3 V (vs.Au), step potential 10 mV, modulation amplitude 10 mV, modulation time 0.015 s and interval time 0.1 s.

Figure S12 .
Figure S12.Comparison of the reproducibility of the (a) 9-electrode array vs the (b) 64 electrode array configurations by recording the DPVs for identically decorated electrodes with the thiolated PC and hybridised the fully complementary Fc-ssDNA (SNP G).

Figure S13 .
Figure S13.Study of inter-array reproducibility by comparison of the éMCA profiles obtained by plotting the maxima of the DPV recorded with T ramping (1°C /step) for a fingerprick blood sample (Sample 1) using each of the devices.Three electrodes were used for constructing the melting profiles in each case.

Figure S15 .
Figure S15.a) Analytical data of the éMCA profiles obtained by recording DPVs with T ramping (1°C /step) and the corresponding first derivatives.Triplicates were carried out for each probe and sample used.

Figure S16 .
Figure S16.éMCA profiles (as % DPV signal) obtained using the second-generation device by recording DPV with T ramping (1°C /step) of 10 representative blood samples and the corresponding first derivatives.Triplicates were carried out for each probe and sample used.