Fluorescence polarization-based rapid detection system for salivary biomarkers using modified DNA aptamers containing base-appended bases

The field of care testing toward the analysis of blood and saliva, lacks nowadays simple test techniques for biomarkers. In this study, we have developed a novel nucleobase analog, Ugu, which is an uracil derivative bearing a guanine base at the 5-position. Moreover, we attempted the development of aptamers that can bind to secretory immunoglobulin A (SIgA), which has been examined as a stress marker in human saliva. It was observed that the acquired aptamer binds strongly and selec-tively to the SIgA dimer (Kd = 13.6 nM) without binding to the IgG and IgA monomers of human serum. Reduction of the aptamer length (41 mer) successfully improved fourfold the binding affinity (Kd = 3.7 nM), compared to the original, longer aptamer (78 mer). Furthermore, the development of a simple detection system for human saliva samples by fluorescence polarization was investigated, using the reported human salivary α-amylase (sAA) and the SIgA-binding aptamer. Compari-son of the present method with conventional enzyme-linked immunosorbent assay techniques highlighted a significant Pearson's correlation of 0.94 and 0.83 when targeting sAA and SIgA, respectively. It is thus strongly suggested that a new simple test of stress markers in human saliva can be quantified quickly without bound/free (B/F) separation.

S aliva is a biological sample that can be noninvasively collected, and its biomarkers can be used to detect changes in biological stress systems, in order to easily investigate the effect of acute and chronic stress to physical health and illness. 1−3 Current research focuses on the stress status estimation and the timely diagnosis of various systemic and local diseases, by measuring salivary biomarkers. 4 Among them, salivary α-amylase (sAA) is secreted due to the direct nerve action of the sympathetic nervous system and the regulation of norepinephrine, and its concentration rises due to acute stress. Therefore, sAA is a sensitive biomarker affected by physical stress-related changes that reflect the nervous system activity. 5,6 Furthermore, salivary secretory immunoglobulin A (SIgA) has been investigated as a decline biomarker of the immune system function, 7,8 while its association with stress and upper respiratory tract infection established SIgA as a reliable biomarker for identifying infection risk in athletes. 9,10 Although the ELISA assay is mainly applied to measure saliva markers, its utilization at the point of care is difficult, because the operation is complicated and time-consuming. Furthermore, the development of an evaluation system that could simultaneously measure multiple markers, at low cost, is strongly desired in large-scale cohort studies. 11 A simple method for measuring the sAA concentration based on the amylase activity has been developed and commercialized, 12 but it is widely influenced by pH and salt concentration, whereas it cannot be measured along with other markers. Therefore, this method cannot be used as a general-purpose platform. Additionally, there is no other method available for measuring SIgA except for ELISA, reinforcing the need for novel, simple techniques to determine the salivary SIgA concentration.
Aptamers are single strand DNA or RNA molecules, 13,14 and a large number has already been reported in the literature. 15−17 Chemical modifications or additions/extensions of additional functions are more feasible in aptamers compared to antibodies, while they can be employed as switching sensors through structural changes, because of their stability at room temperature and resistance to thermal denaturation. Therefore, they have attracted attention as diagnostic agents and molecular elements of simple sensors, also due to their high detection sensitivity in combination with the polymerase chain reaction (PCR). 18,19 Despite numerous reports of successful aptamer acquisition and the production of aptamer sensors for the detection of various targets, not many cases of successful detection of targets in actual samples have been reported so far. 20 Aptamers with affinity for a desired target are selected from a library of random sequences with a primer region, using the systematic evolution of ligands by exponential enrichment (SELEX) method. 13,14 Disclosed aptamers with high binding affinity and slow off-rate use, apart from natural nucleic acids, modified nucleic acids that introduce highly hydrophobic amino acids to the bases. 21, 22 In our previous study, we have developed a DNA aptamer with a base-appended base (BAB) modification and adenine derivatives at the 5-position of uracil, which displayed high affinity for small molecules and proteins. 23,24 Moreover, we have successfully developed recently an aptamer with a high binding affinity for sAA, using a library containing the nucleobase analog U ad . 25 It is known that this sAA-binding aptamer exhibits a more diverse complex structure than the same oligomer consisting of natural bases. The uniqueness of the three-dimensional structure is considered to contribute to its high binding affinity. The binding of sAA-binding aptamers to salivary α-amylase has been confirmed by pull-down experiments, using aptamer beads, and by lateral flow devices using gold colloid that could be a simple detection method for salivary sAA. Furthermore, we examined aptamers as simpler detection systems for SIgA, which is a stress biomarker in saliva. Similar to sAA-binding aptamer, the use of a library containing the nucleobase analog U ad disclosed some clones showing binding to SIgA. However, the clone with the strongest binding affinity exhibited crossreactivity to IgG.
Hence, in this study, we developed and applied a novel analog, U gu , which features a new BAB modification and bears a guanine base at the 5-position. Thereby, the interaction between the nucleic acid bases is strengthened compared to analog U ad , enhancing thus its structural specificity. As a result, it was possible to obtain a clone, which can bind specifically to SIgA, without cross-reactivity to IgG, and develop a detection system for stress biomarkers in saliva based on a fluorescence polarization (FP) assay, using the SIgA-binding aptamer in addition to the sAA-binding aptamer. 26 ■ EXPERIMENTAL SECTION Materials. The selection target, SIgA, was purchased from MP Biomedicals (Santa Ana, CA, USA). Dynabeads MyOne Carboxylic Acid magnetic beads for the target immobilization and Dynabeads MyOne SA C1 magnetic beads for the recovery of biotinylated DNA were purchased from Invitrogen (Carlsbad, CA, USA). KOD dash polymerase for PCR and the incorporation of base-modified oligonucleotide was obtained from Toyobo Co., Ltd. (Osaka, Japan). Primer, 5′-fluorescent (TYE665) labeled primer, random pool, and template for aptamer clones were purchased from Integrated DNA Technologies MBL K. K. (Tokyo, Japan). The ELISA kit for the measurement of SIgA in human saliva was purchased from Cloud-Clone Corp. Human Saliva. Methods were conducted according to the relevant guidelines. Informed consent was obtained from all the participants. All experiments on the use of human saliva were approved by the Institutional Biosafety Committee of NEC Solution Innovators, Ltd.
Synthesis of dU gu TP. A novel modified nucleoside triphosphate, dU gu TP, was synthesized from 2,6-dichloropurine via a 7-step reaction sequence (Scheme S1). The detailed synthetic procedure and the product characterization are included in the Supporting Information.
Systematic Evolution of Ligands by Exponential Enrichment (SELEX) Method. The applied SELEX methodology was an improvement of previously reported methods ( Figure S1). 25 For the preparation of the magnetic beads with SIgA as the target, SIgA was immobilized on Dynabeads MyOne Carboxylic Acid magnetic beads according to the manufacturer's instructions and washed with a selection buffer (SB; 40 mM HEPES, pH 7.5, 125 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 0.01% Tween 20). 5′-Biotinylated DNA template (5′-GAT AAG CCC GCC CAT TGT AGT TTC-N 30 -GAT TAG GGA GAT TGC GTC CAA ACC-3′), forward (Fw) primer (5′-GGT TTG GAC GCA ATC TCC CTA ATC-3′), and KOD Dash polymerase were used for the preparation of dsDNA containing U ad or U gu . After binding dsDNA to Dynabeads MyOne SA C1 magnetic beads, ssDNA was eluted with 0.02 M NaOH (aq), and the elution was neutralized with 0.08 M HCl (aq) to yield the ssDNA library containing U ad or U gu . Afterward, 50 pmol of the library was mixed with 250 μg of the target beads at 25°C for 15 min, and the beads were washed with SB and eluted with 7 M urea to yield ssDNA. The elution was then amplified by PCR, using the Fw primer and the 5′-biotinylated reverse (Rv) primer. The amplified dsDNA was incubated with Dynabeads MyOne SA C1 magnetic beads, treated with 0.02 M NaOH (aq) to elute the Fw primerelongated products, and then washed with SB. Furthermore, to prepare ssDNA, dsDNA containing U ad or U gu was prepared as described above using the Rv primer-elongated products, which were immobilized on the magnetic beads, Fw primer, and dU ad TP or dU gu TP, and eluted with 0.02 M NaOH (aq). The elution was used also for the next rounds. After eight rounds of selection, the enriched library was amplified by PCR using the Fw primer and the nonlabeled Rv primer and then sequenced with a GS junior sequencer (Roche, Indianapolis, IN, USA).
Surface Plasmon Resonance (SPR) Assay. The SPR analyses were performed at 25°C using ProteON XPR360 (Bio-Rad Laboratories, Inc., Hercules, CA, USA). 25 A 20-mer deoxyadenylic homooligonucleotide (poly(A 20 )) was attached to the 3′-end of an aptamer clone, which was hybridized with a 5′-biotinylated 20-mer thymidylic homooligonucleotide (dT 20 ), immobilized onto an NLC sensor chip, and used as a ligand. 25 The target proteins, sAA and SIgA, were used as the analyte, and the above-mentioned SB was used as the running buffer. We regarded the molecular mass of sAA and SIgA as 57 kDa and 160 kDa, respectively. The dissociation constant (K d ) was calculated according to the manufacturer's instructions.

Analytical Chemistry
Article template, and 5biotinylated Fw primer were mixed with 3 mg of Dynabeads MyOne SA C1 magnetic beads in SB at room temperature for 30 min, the magnetic beads ware washed in triplicate with SB.
The DNA template was then eluted with 0.02 M NaOH (aq), and the beads ware washed with SB three times and suspended in SB. Clone beads (200 μg), i.e., SIgA capture beads or primer-immobilizing magnetic beads, i.e., control beads, were mixed with 5 μg of SIgA or 90% human saliva in SB at 25°C for 60 min. After removing the supernatant and washing with SB in triplicate, the bound proteins were eluted after treatment with 0.1% SDS at 95°C for 10 min. The eluted samples were electrophoresed on PAGEL C520L (ATTO, Japan), according to the manual.
FP Assay. The FP assay was conducted according to the improved methodology of Gokulrangan et al. 27 Clones of the fluorescent-modified aptamer were prepared using a 5′fluorescent (TYE665) labeled primer. TYE665 is an alternative fluorescent dye for Cy5, which has an excitation/emission wavelength at 645 nm/665 nm, and other information is not publicized. Fluorescent-labeled clones (10 nM) were denatured at 95°C for 5 min, folded at 4°C for 5 min, and then incubated with the target samples in a selection buffer 2 (SB2; 40 mM HEPES pH 7.5, 125 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , and 0.1 mM dextran) at 25°C for 5 min. After incubation, samples were measured at 25°C using FP spectroscopy on TECAN infinite M1000PRO (TECAN, S Switzerland), according to the manual (N = 2) with 635 nm excitation and monitoring at 665 nm. The reference value of an unbound state aptamer was set to 20 mP, as a G-Factor reference. SIgA, sAA, and human saliva were used as target samples. Human saliva was collected using Salivette propylene (PP)−polyethylene (PE) polymer swabs (Art. No. 51.1534.901J, Sarstedt K. K., Tokyo, Japan). The volumes of sAA and SIgA in human saliva were recorded using a human amylase AssayMax ELISA kit (AssayPro, USA) and an ELISA kit for secretory immunoglobulin A (Cloud-Clone Corp., USA), respectively, based on the manual.

■ RESULTS AND DISCUSSION
The aptamer selection was realized using a library of known, modified nucleic acids (U ad ), to acquire SIgA-binding aptamers ( Figure S1). 24,25 After eight rounds of SIgA aptamer selection, five sequences, each of which occupies more than 5% of the enriched library, were chosen as aptamer candidates (Table  S1).
Measurement of the binding affinities of five sequences for SIgA on a SPR instrument, indicated that all five clones were binding to SIgA ( Figure S2). Furthermore, we measured the binding affinities for several types of IgG and IgA from human serum by SPR to point out IgA ad4 clones with the highest affinity for SIgA. The IgA human serum exists predominantly in the form of monomers, whereas the majority of IgA (SIgA) in external secretions, e.g., saliva, is present in the form of dimers. 28 As a result, IgA ad4 showed cross-reactivity to full-size IgG and IgA from human serum ( Figure S3).
Therefore, we prepared a fresh library containing modified nucleobase U gu (Figure 1) and conducted aptamer selection. After eight rounds of SIgA aptamer selection, three sequences, each of which occupies more than 5% of the enriched library, as shown in Table 1, were chosen as aptamer candidates.
The evaluation of the binding affinities of these three sequences to U gu clone by SPR indicated that all candidates bind to SIgA. Among the tested clones, IgA gu1 had the highest binding capacity to SIgA, based on the K d values and the value of the SPR resonance units for SIgA at 200 nM ( Figure 2). The binding of IgA gu1 clone to human serum IgG and IgA was measured by SPR, revealing that the IgA gu1 clone did not bind to IgG or serum IgA, whereas it showed high binding specificity to SIgA (Figure 3). SIgA has not been reported to bind nucleic acids except for hemagglutinin (HA) protein of influenza A virus. 29 Thus, these results suggested that IgA gu1 clone binds to SIgA specifically.
IgA ad1 , IgA ad2 , IgA ad4 , IgA gu1 , and IgA gu3 clones were used to generate magnetic beads, able to bind to SIgA. A pull-down assay was then conducted for SIgA from human saliva in SB, utilizing the generated beads. It was observed that all the clones bind to SIgA in SB (Figure 4a), but binding to SIgA in human saliva could be confirmed only for 2 clones, IgA gu1 and IgA ad4 (Figure 4b).
In general, IgA ad4 showed broad binding specificity for antibodies, whereas IgA gu1 showed high specificity only for the dimer IgA. IgA ad4 strongly bound to SIgA (K d = 1.14 nM, Figure S2), and binding to SIgA in saliva was also confirmed, as displayed in Figure 4. Considering the binding specificity of the examined clones to SIgA, IgA gu1 was selected as a SIgA-binding aptamer candidate. In view of the salivary SIgA concentration range (0.6−1.2 μM) that has been reported so far, 8 the high affinity of IgA gu1 for SIgA (K d = 13.6 nM) can be sufficiently used as a detection element for SIgA in saliva.
Generally, long sequences are likely to form various secondary structures that destabilize the conformation of the target aptamer's binding site. 30 Therefore, we considered that the binding specificity could be improved by suppressing the  2.3% ttACAtAAGtGCCAACGttAtCAACAtACt a Sequence ratio was defined as the ratio of the sequence to the total number of sequences, which was generated by a next generation sequencer. b "t" indicates U gu . structural instability, which could be achieved by minimizing the sequences. 31 In this case, the sequence minimization was examined using IgA gu1 as a SIgA-binding aptamer candidate and the SPR methodology as an index for binding recordings (Figure 5a,b). As a result, we achieved the minimization by 41 bases and improved the K d value to 3.7 nM (Figure 5c). Our next goal was to develop a simple detection system of sAA and SIgA by the FP assay. This system has the great advantage of simple detection, which does not require B/F separation, while the measurement of the sample concentrations requires only mixing of the fluorescent-labeled aptamer with the examined sample. 32 It is generally considered that the sensitivity of FP is affected by the difference in molecular weight between the target molecule and the molecule bearing the fluorophore and by the local fluorophore movement. 27,33 In order to enhance the method's detection sensitivity, it is important to minimize the size of the fluorescent-labeled aptamer, i.e., increase the molecular weight ratio with the target. Therefore, we analyzed in detail the minimized sequences of the sAA-binding (AMY ad1-2 ) and SIgA-binding (IgA gu1-3 ) aptamers 25 and then sought for sequences that increase the fluorescence anisotropy (mP) during the target binding. Specifically, various sequences, labeled with a fluorescent dye (TYE665) at the 5′-end, were prepared.

Analytical Chemistry
Then, a sequence with substantial changes in the mP values was mixed with each target molecule, while AMY ad1-2-3 and IgA gu1-3-3 aptamers were selected as detection clones for sAA and SIgA, respectively ( Figures S4 and S5). The binding of AMY ad1-2-3 and IgA gu1-3-3 to the targets was also confirmed by SPR ( Figure S6). AMY ad1-2-3 had almost the same K d value in SPR with the original minimized clone (AMY ad1-2 ), 25 but IgA gu1-3-3 exhibited an about 3-fold higher K d value in SPR than the original minimized clone (IgA gu1-3 ) (Figures 5c and S6b). Because of the minor alteration of the fluorescence anisotropy, when IgA gu1-3 binds to 1 μM SIgA (Figures S5b), IgA gu1-3 cannot be used for the detection of SIgA by the FP method. Instead, considering the actual concentration of SIgA in human saliva (0.6−1.2 μM), 8 we assessed that IgA gu1-3-3 could be sufficiently used as a candidate aptamer for FP assays.
Our study confirmed that the mP values changed in a target concentration-dependent way, when the binding of AMY ad1-2-3 and IgA gu1-3-3 to each target (sAA and SIgA) was examined in SB2 ( Figure 6). We also confirmed target specificity of AMY ad1-2-3 and IgA gu1-3-3 between sAA and SIgA in FP measurement ( Figure 6). Thus, those aptamers enable to measure sAA and SIgA with high specificity using FP measurement.
A good regression line could be obtained by fitting the change in fluorescence anisotropy using a 4-parameter logistic nonlinear regression model, and the inflection point values were estimated at 3.1 nM for sAA ( Figure 6a) and 287 nM for SIgA (Figure 6b). These values are about 7-fold greater for sAA and about 27-fold greater for SIgA than the K d values calculated from the SPR method (Figures 6 and S6). The difference in binding activity between different methodologies  (a) SPR response curve of the interaction between the SIgA or other types of IgG and the aptamer IgA gu1 . SIgA, IgG1 kappa, IgG, or IgG-Fc (each of 400 nM) were injected over the respective aptamerimmobilizing sensor chips for 120 s at a flow rate of 50 μL/min. The red line represents the SIgA measured curve, and the black and blue lines represent the IgG1 kappa and IgG measured curves, respectively. The green and the orange lines represent the IgG-Fc1 and IgG-Fc2 measured curves, respectively. (b) SPR response curves of the interaction between the SIgA or serum IgA and the aptamer IgA gu1 . SIgA or serum IgA (each of 50 μg/mL) was injected over the respective aptamer-immobilizing sensor chips for 120 s at a flow rate of 50 μL/min. The red line represents the SIgA measured curve, and the gray lines represent the serum IgA measured curves, respectively.

Analytical Chemistry
Article was reported previously, 34 and the difference in the binding activity between FP and SPR in this study could be related to the immobilization status of the binding assay. The 3′ tail of aptamers and free aptamers were used for SPR and FP, respectively.
In addition, we prepared a DNA oligo in which all the U gu in IgA gu1-3-3 were substituted with natural T (designated as IgA gu1-3-3N ) and checked the binding affinity using the SPR and FP assay (Figure S7a,b). The result clearly showed that IgA gu1-3-3N was unable to bind to SIgA, suggesting that the BAB modification was important for the recognition of SIgA.
To compare the difference in target specificity between IgA ad4 and IgA gu1 , we predicted the secondary structure of the IgA ad4 and IgA gu1 sequence that substituted U ad or U gu to natural base (T) using the VALFold program 35 with the general DNA parameters 36 ( Figure S8). Though we could not explain the reason for the high specificity of IgA gu1 judging from the comparison of the second structure between IgA gu1 and IgA ad4 (Figure S8), we speculated that guanine bases of IgA gu1 are likely to contribute various conformations through hydrogen bonds and stacking interactions, as noticed in guanine quadruplex structures.
Though the maximal fluorescence anisotropy (FA) variation observed between unbound state aptamer and protein binding state aptamer is supposed commonly ranging from 50 to 100 mP, 37 a smaller FA variation (20 to 40 mP) in our study is significant to monitor the concentration change of sAA or SIgA ( Figure 6). Because the magnitude of FA variation was

Analytical Chemistry
Article reported to be influenced by selecting fluorophore, 38 changing TYE665 to other fluorophore, for example, tetramethylrhodamine, could probably increase the range of FA values. Furthermore, using bivalent aptamer could be effective to enhance the target detection sensitivity. 39 The human saliva samples were measured by the FP method, as well. Particularly, both sAA and SIgA exhibited significant correlation between the FP and the ELISA method, when the measured mP values of the same saliva sample were compared with the absorbance (monitoring at 450 nm) recordings of the ELISA method (Figure 7). Especially in the case of sAA, Pearson's correlation coefficient showed a high correlation of 0.94 between the FP and the ELISA method (Figure 7a).
In general, ELISA is a time-consuming method and needs two kinds of antibodies to detect the target molecules on the basis of the sandwich method. For example, sAA measurement using a human amylase AssayMax ELISA kit required about 4 h and eight experimental steps, while the FP aptamer assay needs only one kind of aptamer and could measure sAA concentration within about 5 min with two experimental steps. Therefore, the FP aptamer assay takes great advantage for the practical application of monitoring sAA or SIgA in real time. 33 In the FP method, the molecular weight ratio of the target molecule to the fluorescent-labeled molecule greatly affected the change in FP upon binding. Generally, when a protein is used as the target, an antibody cannot be used as the fluorescent label molecule. This is because the molecular weight ratio to the target molecule is low, and only a very small change in FP can be observed. 37 On the other hand, fluorescent-labeled aptamers may be able to detect the target protein molecules by FP, if the sequence can be optimized and the molecular size can be minimized.
Furthermore, the FP assay does not require a B/F separation operation, reducing thus the measurement time compared to  . sAA-ELISA: the saliva was diluted to 1/30,000; sAA-FP: the saliva was diluted to 1/1,000. (b) The outcome of the FP using the fluorescent-labeled SIgA-binding aptamer (IgA gu1-3-3 ). SIgA-ELISA: the saliva was diluted to 1/1,000; SIgA-FP: the saliva was diluted to 1/100. Because of the difference of detection sensitivity between FP and ELISA, the dilution factor of saliva was optimized for FP and ELISA, respectively. The mP values of the sAA and SIgA-binding aptamers using the FP method were compared with the absorbance recording (monitoring at 450 nm) from the ELISA kit. Sixteen saliva samples were collected twice from eight people (30−50 years old) on different days.

Analytical Chemistry
Article the ELISA method. However, in this case aptamers should feature high binding affinity and specificity.

■ CONCLUSIONS
In the current study, we developed a new BAB modification by introducing a guanine base on uracil, and acquired a SIgAbinding aptamer that can selectively bind to SIgA, a highly specific salivary marker for immune stress. To our knowledge, this is the first report of an aptamer that binds to SIgA in human saliva. The sequences of sAA and SIgA-binding aptamers were also minimized and optimized, while a new simple detection system, capable of quantifying sAA and SIgA in saliva through the FP method, without B/F separation, was developed. The recorded values of sAA and SIgA in human saliva samples showed significant correlation with the results of sAA and SIgA recordings of the predominant ELISA method. These results strongly suggest that a new simple test system for salivary stress markers can emerge, with anticipated applications on future large-scale cohort studies that require sAA and SIgA recordings. Future plans aim also to the development of novel aptamers of high affinity and specificity that could selectively target various salivary biomarkers, through diverse BAB modifications.
Materials and additional text showing detailed experimental procedures, characterization of all compounds, and Supplementary Scheme S1,