Multi-parameter surface plasmon resonance instrument for multiple nucleic acid quantitative detection

Multiplex nucleic acid assays can simultaneously detect the characteristics of different target nucleic acids in complex mixtures and are used in disease diagnosis, environmental monitoring, and food safety. However, traditional nucleic acid amplification assays have limitations such as complicated operation, long detection time, unstable fluorescent labeling, and mutual interference of multiplex nucleic acids. We developed a real-time, rapid, and label-free surface plasmon resonance (SPR) instrument for multiplex nucleic acid detection. The multiparametric optical system based on total internal reflection solves the multiplex detection problem by cooperating with linear light source, prism, photodetector, and mechanical transmission system. An adaptive threshold consistency correction algorithm is proposed to solve the problem of inconsistent responsiveness of different detection channels and the inability of quantitative comparison. The instrument achieves label-free and amplification-free rapid detection of these biomarkers for miRNA-21 and miRNA-141, which are widely expressed in breast cancer and prostate cancer. The multiplex nucleic acid detection takes 30 min and the biosensor has good repeatability and specificity. The instrument has a limit of detection (LODs) of 50 nM for target oligonucleotides, and the smallest absolute amount of sample that can be detected is about 4 pmol. It provides a simple and efficient point-of-care testing (POCT) detection platform for small molecules such as DNA and miRNA.


Introduction
Multiplex nucleic acid detection refers to the detection of the characteristics of different target nucleic acids in complex mixtures at the same time in one reaction.We can screen multiple pathogenic factors at the same time when testing a single sample, which greatly reduces the time and cost of nucleic acid detection (Zhang et al. 2020).Therefore, multiple nucleic acid detection is commonly used in the fields of disease diagnosis, environmental monitoring, and food safety.MicroRNAs (miRNAs) are small non-coding endogenous RNA molecules that play a major role in cellular gene expression by affecting target mRNAs.Relevant studies have shown that the abnormal expression of miRNAs is closely linked to human cardiovascular diseases, immune diseases, and cancer (Liu et al. 2017).miRNA has become an important emerging biomarker for early clinical diagnosis, and it is of great practical significance to realize the rapid and accurate detection of miRNA.For example, in association with epithelial tissues, miRNA-141 is widely expressed in common human cancers such as breast, lung, colon, and prostate cancer (Yin et al. 2012).As well as being reported as an anti-apoptotic factor of cancer cells, miRNA-21, which is often overexpressed in breast cancer, ovarian cancer, prostate cancer and myocardial diseases (Wang et al. 2016).Clinical diagnosis requires simultaneous quantitative analysis of multiple miRNAs.Conventional multiple nucleic acid analytical methods include nucleic acid amplification, nucleic acid hybridization, etc. Specific techniques include multiplex PCR (Gaňová et al. 2021), DNA microarray (Sterpone 2009), capillary electrophoresis (Dangerfield et al. 2021), etc.All of these techniques can accurately detect multiple nucleic acids but suffer from intrinsic limitations of 24 Page 2 of 12 sophisticated operation, high instrument costs, strict experimental environment requirements, and low sensitivity.In recent years, modern techniques including electrochemical methods (Xu et al. 2020), chemiluminescence (Hu et al. 2021), and surface plasmon resonance (Jebelli et al. 2020;da Silva et al. 2021) have been developed for multiplex nucleic acid detection.
Surface plasmon resonance (SPR) is a label-free technique compared with other methods.By measuring the changes in the surface refractive index during the binding or dissociation of complexes, it can be used to detect various proteins, nucleic acids, and other biomolecules (Sarkar et al. 2022).It exhibits the characteristics of simple operation, high sensitivity, and good specificity.Zhang et al. proposed a reflective SPR sensor that can simultaneously detect temperature and nitrate concentration using the dualwavelength matrix method (Zhang et al. 2019).Wang et al. proposed a U-shaped cascade dual-channel SPR optical fiber sensor (Wang et al. 2021).The two channels of the sensor are covered by the gold film, and two independent SPR signals can be obtained by controlling the U-shaped bending radius.Rampazzi et al. designed and tested a novel low-cost multiparametric stand-alone LSPR imaging instrument for biosensing applications (Rampazzi et al. 2016).
The monitoring methods of SPR for biochemical reactions mainly include angle modulation, intensity modulation, and wavelength modulation.The angle modulation method has higher sensitivity and versatility, but the mechanical structure design of the detection system is more complicated and the cost is higher.The system based on the intensity modulation method is simpler and more compact.Changes in reflected light intensity at a fixed angle are used to monitor biochemical reactions.A CCD camera is generally chosen as the imaging device to monitor multiple interactions on the sensor chip surface in real-time by capturing the reflected light from a region of interest (ROI) (Singh 2016).Silicon photodiode arrays can also serve as photodetectors to observe SPR reactions.Although it is not possible to visualize the surface of the sensor, the photodiode has a lower noise level and faster signal conversion, and the detection system can achieve better sensitivity and signal-to-noise ratio (Zybin et al. 2007).And it is easier to integrate into low-cost portable electronic devices, enabling POCT deployment.For example, the integrated SPR sensor platform Spreeta 2000 developed by Texas Instruments uses a linear diode array detector (Chinowsky et al. 2003), and the size of the detection system is only 1.5 cm × 0.7 cm × 3 cm.Due to its angular analysis method, multiple Spreeta devices are required to achieve multiplex detection.When the photodiode array work in the intensity modulation mode, if the photosensitive pixel and the sensor chip reaction area correspond perfectly, the real-time SPR monitoring of multiple biochemical reactions on the sensor can be realized by reading the reflected light intensity value of the pixels.However, the reflected illuminance of the light source passing through the prism is uneven, which leads to the different responsiveness of the photodetector array, and it is difficult to ensure the consistency of the response channels.In the qualitative analysis of multiple nucleic acids, it is usually necessary to compare the differences in different reaction areas, and in the quantitative analysis, the concentration of the reactants in the reaction areas should be calculated.All these require the channel consistency of the sensing platform.
In this paper, we describe how we designed and developed a real-time, rapid, and label-free SPR instrument for multiplex nucleic acid detection.The multiparametric optical system based on total internal reflection solves the problem of simultaneous detection of multiple nucleic acids by cooperating with linear light source, prism, photodetector, and mechanical transmission system.The SPR sensing platform is equipped with a functional biosensor, which is suitable for multiparametric analysis of complex chemical and biological samples.We selected miRNA-21, miRNA-141, and a 30 bp DNA single-stranded oligonucleotide as detection targets, demonstrating the multiple nucleic acid detection capabilities of the SPR biosensing platform.The DNA sequence was used for comparison with miRNA and does not correspond to a particular cancer.A multi-channel consistency correction algorithm based on the adaptive threshold is proposed to solve the problem of inconsistent responsivity of different detection channels and the inability of quantitative comparison.We evaluated the sensitivity of the instrument by measuring the refractive index changes of glycerol solutions at different concentrations.The case introduced in this paper demonstrates the sensing ability of the device in miRNA detection and the sensitivity and specificity of the SPR biosensor in nucleic acid molecular hybridization.The results show that the detection platform has great application potential in the field of biomedicine and early disease diagnosis.

Materials and reagents
Phosphate buffered saline (PBS), glycerol, ethanolamine hydrochloride, Glycine-HCL (10 mM), Sodium Acetate (pH 4.5), N-hydroxy succinimide (NHS), and N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC) were purchased from GE Healthcare.Refractive index matching liquid was purchased from Shanghai Specimen mode.PAGE-purified miRNAs (miRNA-21 and miRNA-141), and HPLC-purified DNA capture probes were synthesized by BGI (Beijing, China). 1 mg/ml streptavidin (SA), 30 bp DNA capture probe (with biotin at the 5' end), and DNA oligonucleotides were donated by the Institute of Biophysics, Chinese Academy of Sciences.The sequences of DNA and miRNA were listed in Table S1.All oligonucleotides were dissolved in PBS buffer to give a 100 μM solution, which was diluted with buffer to the desired concentration before use.All of the solutions were treated with DEPC water and autoclaved to protect them from RNase degradation.The aqueous solutions used in the experiments were all prepared using ultrapure water.

Basic SPR theory
As shown in Fig. 1a, a p-polarized light is coupled to the metal/dielectric interface through the prism under the conditions of total internal reflection.At a specific angle of incidence, resonant coupling occurs when the wave vector of the incident light matches the wave vector of the surface plasmons (SPs) at the metal/dielectric interface.The energy of the incident light is absorbed by the surface plasmons and the intensity of the reflected light is greatly reduced.This phenomenon is called SPR.Equation ( 1) denotes the wave vector K sp of a surface plasmon propagating along the surface of the metal film.Equation (2) denotes the parallel component of the wave vector of a light ray incident to the metal film with an angle of incidence θ (Oliveira et al. 2013).When K sp equals k x, the resonance is excited.
where ε 0 denotes the dielectric constant of the optical prism, ε 2 is the dielectric constant of the sensing layer on the surface of the metal film, ε 1Re is the real part of the complex dielectric constant of the metal film, and λ is the wavelength of the incident light in vacuum.When K sp = k x , the surface plasmon resonance is excited and the minimum reflectivity can be observed in the reflectivity curve.θ is the SPR resonance angle, which is related to the position of the minimum reflectivity.The refractive index n 2 of the sensing (1) layer can be obtained by solving the equation.μ r refers to relative permittivity.
A four-layer Kretschmann coupling model was used to describe the SPR sensing system.As shown in Fig. 1a, the different media layers are described as follows: Medium 1 is a prism.Medium 2 is a thin gold film.Medium 3 is the sensing layer where the chemical interactions occur, and Medium 4 is usually an aqueous solution containing the substance to be analyzed.The dielectric constant of each layer is j (j = 0,…3) and the refractive index is n j (j = 0,…3).The wave vector of incident light in each layer is k j (j = 0,…3) and the thickness of each layer is d j (j = 0,…3).
The Fresnel's equations were applied to the theoretical description of Kretschmann-type SPR model.According to Fresnel's equations, the reflectivity R is a function of θ, n 2, and λ (Raether 1977;Thirstrup et al. 2004), and the details of the relevant equations and simulations are uploaded as supplementary files.

Multi-parameter SPR instrument
The multi-parameter SPR instrument we developed works in the intensity modulation mode.The SPR detection system is mainly composed of biological sensor chips, optical systems, data acquisition, control circuits, microfluidic systems, and human-computer interaction equipment.The fabrication process of the SPR sensor chip is consistent with our previous work (Wang et al. 2022a, b), including two steps of goldplated chip fabrication and surface BSA modification.The sensing chip was fabricated by depositing a stack of 2 nm chromium and 48 nm gold on the microscope slide substrate by magnetron sputtering.Then modify the surface of the gold-plated sensor chip with bovine serum albumin (BSA).
As shown in Fig. 1d, the optical structure is based on the classic Kretschmann configuration.After the linear semiconductor laser (632 nm) enters the prism from one side, it is focused at the bottom to form a rectangular light spot with a size of 1 mm × 15 mm.The position of the flow cell is fixed by the position limit device so that the four microchannels flow horizontally through the central area of the light spot, as shown in Fig. 1d.The photodiode array is used to detect the intensity change of the reflected light on the other side of the prism.Adjust the array position so that the photosensitive area of the photodiode and the outgoing light coincide. (3) According to the relative positions of the light beam and the microchannels, the number of photosensitive pixels corresponding to the four microchannels is determined to be pixel 5, 7, 9, and 11 respectively.Collecting the light intensity values of these photosensitive pixels can realize real-time multi-parameter SPR monitoring of biochemical reactions occurring on the chip surface in the microchannel.
As shown in Fig. 1b, the functions of the circuit system of the multi-parameter SPR detection platform mainly include the drive of the stepping motor, photoelectric detection, multi-channel data acquisition, and data processing, communication, and control based on the microcontroller.First, the stepping motor is controlled to rotate.Then the symmetrical link mechanism is driven by the belt to move, and the incident angle of the laser and the position of the photodetector for receiving the reflected light are changed at the same time.The SPR angular scanning curve can be obtained by collecting the reflected light intensity signal of the photodetector as the angle changes (Fig. S3b).There is an encoder on the stepping motor, which can record and feedback the position information of the motor in real-time, and realize high-precision closed-loop motion control.After acquiring the angular scanning data, the stepper motor is controlled to drive the connecting rod to the steepest position of the SPR reflection spectrum curve (the point with the largest absolute value of the slope).This position serves as the working point of the detection platform.The working point is the most sensitive position of the refractive index change, and the strongest signal change caused by the refractive index change can be observed (Zybin et al. 2007).
The silicon photodiode array S4111-16 from Hamamatsu converts the reflected light signal into a current signal, which is processed by TIA amplification and filtering circuit and then transmitted to the A/D acquisition card for analog-todigital conversion.The microprocessor performs digital filtering on the collected data and transmits it to the host computer.A GUI is developed by Qt to complete the realtime display and storage of experimental data.Relying on the above design, we can monitor the change of the refractive index of the sample in the flow cell in real-time by detecting the change of the reflected light intensity, and realize the intensity modulation of the surface plasmon resonance.

Microfluidic system
The main function of the microfluidic system is to inject and discharge experimental reagents.We designed a flow cell with four microchannels as the reaction area.The PDMS oligomer and a cross-linking agent (Sylgard 184) were mixed in a 10:1 ratio under vacuum, and the mixture was poured into a siliconized glass mold.The flow cell was prepared by heating in an oven at 80 °C for 2 h (Wang et al. 2022a, b).The size of the flow cell matches the sensor chip and the four microchannels with a width of 0.5 mm and a depth of 0.8 mm.During the experiment, the sensor chip was in contact with the surface of the prism through the refractive index matching liquid, and then the flow cell was tightly attached to the sensor chip through the pressure structure.Finally, the sensor chip was fixed on the optical platform with an adjustable incident angle.The peristaltic pump of GE Company was used as the liquid feeding device.

Process of the SPR assay
In order to achieve multiplexed nucleic acid detection, the experimental preparation and detection process can be divided into five stages.The principle of SPR detection of nucleic acid and a typical sensorgram are shown in Fig. 2. First, relying on the Au-S interaction, BSA is immobilized on the surface of the gold sensor chip.Then inject the activation reagent to activate the binding site on BSA, and streptavidin (SA) is injected later.After the binding of SA and BSA is complete, the binding site needs to be blocked to prevent non-specific binding during nucleic acid hybridization.The DNA capture probe was bound to SA via biotin and attached to the sensor chip surface.Finally, inject the target oligonucleotides for nucleic acid hybridization.After hybridization, the chip surface was completely regenerated with the regeneration solution.The hybridized duplex is unwound, and the chip can be reused.

Sensor sensitivity test
To test the refractive index resolution of the SPR platform, we prepared glycerol solutions with a concentration gradient of 1-10% and injected them into the microchannels of the flow cell.The refractive index of glycerol solution is related to its concentration and can be used to test the sensitivity of SPR biosensors.We adjust the position of the microchannels of the flow cell so that one of the microchannels coincides with the rectangular light spot.Therefore, all the pixels of the photodiode array reflect the change of the refractive index in the same microchannel.1%-10% glycerol was injected in sequence.After the baseline was stabilized, 500 sampling points were taken for each glycerol concentration, and the average light intensity of the sampling points corresponded to the glycerol refractive index signal.The experiment was repeated three times.The corresponding refractive index of glycerol at different concentrations was measured with an Abbe refractometer (10 -4 RIU resolution).The refractive index-light intensity fitting curve obtained by least-squares was shown in Fig. 3a.The glycerol refractive index response curve of each channel after baseline alignment was shown in Fig. 3b.
The noise of the silicon photodiode array detection platform is obtained by measuring deionized water.Collect 5 sets of data, each with 500 sampling points of reflected light intensity.The average of the root mean square errors of these groups is taken as the system noise.The calculated noise is 55, and the resolution of the refractive index of the system is 6.75E-6RIU.To compare the sensitivity and signal-to-noise ratio of the photodiode and CCD, we used a Basler 12-bit area CCD camera as the light intensity detector under the same experimental conditions.The refractive index-light intensity fitting curve of the CCD is shown in Fig. 4a.The ordinate is the average gray value of the ROI, and the fitting line equation is y = 6.643E4 × x -877757.826.The noise of the CCD platform is also obtained by measuring deionized water, and the calculated noise value is 0.3564.As far as the glycerol refractive index measurement experiment is concerned, the signal-to-noise ratio (S/N) of the photodiode array platform is 4672.5.However, the signal-to-noise ratio of the CCD is 2000.5.As shown in Fig. 4a, it is obvious that the sensitivity of the photodiode is also higher than the CCD.

Consistency correction algorithm for multiple detection channels
Relying on the symmetrical optical and mechanical structure, the SPR resonance angles of each microchannel in the flow cell are approximately equal.The photosensitive pixels of the photodetector correspond to the refractive index change of the microchannels.In theory, the sensitivity of each channel is equal.However, the light reflected by the line laser after passing through the prism is not uniform, so the responsiveness of the photodiode is not consistent.
To solve the problem of inconsistent responsiveness of different detection channels and the inability of quantitative comparison, we propose an adaptive threshold algorithm that introduces a sensitivity correction coefficient to correct for uneven illumination results.The specific algorithm is as follows: Firstly, the light intensity data collected continuously from each channel are acquired, and the average value is taken as the response value corresponding to the glycerol concentration.The refractive index-light intensity response curve of the sensor at 1%-10% concentration is approximately linear.We establish a simple linear regression model: y = wx + b.The refractive index-light intensity data were extracted as x, y, and the squared loss error was defined as the loss function.The values of w and b are constantly updated by the gradient descent algorithm, and the optimal solution is finally calculated.w 1 , w 2 , …, w n is used as the responsiveness weight ratio of each channel, and the sensitivity correction coefficient λ is obtained by calculating the weight ratio of the channel.To draw the response curve, we divide the signal value of each channel minus the baseline by the calculated correction coefficient λ as the response signal.The range of refractive index variation corresponding to the sensitivity correction coefficient was limited, and the threshold was determined by the linear regression equation.If the sensor chip was replaced, the threshold value and correction coefficient could be recalculated through the glycerol experiment.Figure 4b showed the glycerol refractive index response curve calibrated by the correction coefficient.In order to achieve the expected uniformity of the light source and to solve the problem of inconsistent responsiveness of different detection channels for quantitative comparison, we introduced the correction coefficient and the results exhibited better consistency.As shown in Fig. 5c, we finally immobilized DNA capture probes with (5) = w i ∕w n , i = 1, 2, … , N complementary sequences of target oligonucleotides on the three channels respectively.Figure 5a is the response curve in the preparation stage.Figure 5b, d are the response curves calibrated by the sensitivity correction coefficient.It is obvious that the response signal changes due to activation and blocking reagent injection are approximately equal after correction.
There are some differences in the refractive index change after binding of SA, which is related to the non-uniform sensitivity of the protein film on the surface of the biosensor.Since the relative molecular mass of miRNA-141, miRNA-21, and 30 bp DNA was similar, there was little difference in the change level of response curve signal values caused by the capture probe immobilization, which was also confirmed in Fig. 5d.Therefore, the sensitivity correction coefficient makes the response curve have better consistency, which is convenient for us to compare the differences between channels and perform quantitative and qualitative analysis.

Analytical performance of nucleic acid detection
To verify the consistency of the experimental method we developed in terms of interday and intraday measurements, intraday 10 repeat tests and interday (three consecutive days) repeat tests were performed.The results of the supplementary experimental tests are shown in Table S2.The independent-sample t-test results of the two experimental groups were 0.095, which less than the standard t-test value of 2.201 at a confidence level of 95% (α = 0.05).The statistical results show that the results of the experimental method developed in this paper are consistent between the interday and intraday measurements.
To investigate the nucleic acid detection performance of the SPR biosensor, we immobilized miRNA-141, miRNA-21, and DNA capture probes on the microchannels of the sensor surface, respectively.In the concentration range of 0-1 μM, different concentrations of the target oligonucleotides were selected to evaluate the analytical performance of the biosensor.The measurements were repeated three times on the same chip, and the error bars represent the standard error of the mean of the three measurements.The surface would be regenerated after each hybridization, and the next detection was performed after the baseline was stabilized.The regeneration protocol had been proven to maintain the activity of the capture probe for at least 50 detection/regeneration cycles.
The calibration curves obtained by miRNA-141, miRNA-21, and DNA hybridization detection are shown in Fig. 6a-c, respectively.Figure S5 shows the original sensing spectra of the SPR response for nucleic acid hybridization with different concentrations of target oligonucleotides.Tables 1  and 2 show the recovery and RSD of mir-141 and mir-21 series.The recovery rate and RSD respectively reflect the accuracy and repeatability of the detection method.The recovery based on SPR related experiments is generally in the range of 85% ~ 115%.The recovery evaluation standards for different experimental samples also differ.Meanwhile, at low concentrations, the recovery was below the accepted    et al. 2017).Therefore, it can be seen from Tables 1 and  2 and the calibration curve that the SPR sensor has good repeatability and accuracy.
The results showed that the response signal increases with the concentration of target oligonucleotides within a certain concentration range.Figure 6a, b showed that the response signal and concentration of miRNA-141 and miRNA-21 were linear in the range of 50-500 nM. Figure 6c showed that the response signal and concentration of DNA were also roughly linearly distributed in the range of 50-300 nM.The limits of detection (LODs) of the target oligonucleotide were estimated to be 50 nM.Based on the standard deviation of the response value and the standard curve slope method (LOD = 3σ/S), the calculated limits of detection of miRNA-141, miRNA-21, and DNA were estimated to be 41.2 nM, 33 nM, 52.3 nM.
The difference between the calculated and actual detection limits may be related to the experimental environment and the condition of the instrument.Considering that the sample volume required for each analysis was 80 μl, the smallest detectable amount of target oligonucleotide was less than 4 pmol.Table 3 compares the performance of our proposed method with other methods reported for miRNA detection, including the limit of detection, assay time, and linear range.The results show that the sensitivity and limit of detection of the current measurement methods are equivalent to those of other works, such as electrochemistry (Kilic et al. 2013), fluorescence (Lavaee et al. 2019), and colorimetry (Agahi and Rahaie 2022).

Specificity of SPR biosensor
To evaluate the specificity of the SPR sensor, we detected different oligonucleotide sequences, including miRNA-141, miRNA-21 and DNA.The specificity detection results are shown in Fig. 7.The response signal of the capture probe to the target oligonucleotides was significantly increased, and there was almost no cross-reaction with non-target.It proved that capture probes only specifically hybridize with  IgG(Mouse IgG).The specificity assay results are shown in Fig. 8, and the sensor showed little cross-reactivity for non-targeted proteins.These results indicated that the SPR biosensor displayed excellent specificity for the detection of nucleic acid.

Real-time quantitative detection of multiple nucleic acids
To achieve real-time quantitative detection of multiplexed nucleic acids, miRNA-21, miRNA-141 and DNA oligonucleotide solutions with a concentration of 1 μM were mixed and injected into the flow cell connected with tubes.
As shown in Fig. 9, with the injection of the mixed solution, miRNA-141, miRNA-21, and DNA were specifically bound with the capture probe respectively.The corresponding refractive index changed, and the signal values of the three channels increased sequentially and reached relative stability.In order to unwind the double strand of hybridization between the probe and miRNA and make the chip regenerate successfully, 10 mM glycine-HCl was injected.With the injection of glycine-HCl, the change The results showed that there was no non-specific binding between the capture probe and the non-complementary oligonucleotide.Real-time quantitative detection of various oligonucleotides could be achieved through the response curve.These results indicated that the SPR instrument we designed exhibits good specificity and multi-channel multiplexing ability in the detection of oligonucleotide molecules.

Conclusion
SPR is a label-free detection method, which is widely used in biomolecular detection and its interaction.However, most commercial SPR instruments, such as Biacore, are expensive and bulky, which is not suitable for point-of-care testing (POCT).We developed a real-time, rapid, label-free SPR instrument suitable for multiparametric analysis of complex chemical and biological samples with a functional biosensor.The multiparametric optical system based on total internal reflection solves the problem of simultaneous detection of multiple nucleic acids by cooperating with linear light source, prism, photodetector and mechanical transmission system.The detection system has better sensitivity and signal-to-noise ratio, which is in accordance with the development trend of high throughput, array, and miniaturization of biomedical detection instruments.Table 4 compares our equipment with commercial instruments, including instrument cost, size, weight, resolution, and number of detection channels.The cost of the equipment is estimated at less than $1000, and it can also be used by untrained personnel.
The sensitivity of the instrument to the change of bulk refractive index has been measured.The resolution of glycerol solution with different concentrations is about 6.75E-6RIU.In the previous section, the sensing capability of the device in multiplex nucleic acid detection was demonstrated through experiments, and the target oligonucleotide analysis with a minimum detectable amount of 4 pmol can be achieved within 30 min.The sensor shows good specificity and repeatability, and has great application potential in biomedicine and early disease diagnosis.Higher sensitivity is required for the detection of nucleic acid or protein molecules, especially miRNAs with properties of small size and low abundance.We can use various amplification strategies in the future, such as nanoparticle enhancement (Tamada et al. 2007), enzyme signal amplification (Li et al. 2014a, b), and hybridization chain reaction (HCR) (Li et al. 2014a, b), to improve the detection sensitivity and get a lower detection limit.

Fig. 1 a
Fig. 1 a Four-layer Kretschmann coupling model used for SPR biosensing and optical boundary conditions used for the Fresnel calculations.b Schematic illustration of the integrated opto-mechatronic SPR sensing platform, including the optical system, electronic ele-

Fig. 2
Fig. 2 Schematic representation of nucleic acid detection assay using SPR biosensor and typical SPR response curves

Fig. 5 a
Fig. 5 a Raw response curves of the preparation stage for channels 1-3.b The response curve of preparation stage calibrated by the sensitivity correction coefficient.c Raw response curves of immobilized

Fig. 6
Fig. 6 Calibration curve and scaling of the linear regions respectively corresponding to various oligonucleotides a mir-141, b mir-21, and c DNA

Fig. 7
Fig. 7 Cross-reaction of the SPR response signal obtained by injecting mir-141, mir-21, and DNA into the channel on which the capture probe was immobilized, respectively

Fig. 8
Fig. 8 Cross-reaction of the SPR response signal obtained by injecting Kappa, HSA, and IgG into the channel on which the capture probe was immobilized, respectively

Table 1
Percentage of recovery and RSD (n = 3) for determination of mir-141 Series

Table 2
Percentage of recovery and RSD (n = 3) for determination of mir-21 Series

Table 3
Comparison of our proposed method with other methods for miRNA detection