Direct Detection of Thrombin Binding to 8-Bromodeoxyguanosine-Modified Aptamer: Effects of Modification on Affinity and Kinetics

The affinity of an 8-bromodeoxyguanosine- (8-BrdG-) substituted thrombin-binding aptamer (TBA-Br), which has the 1st and 10th guanosine residues replaced with 8-BrdG, was estimated using reflectometric interference spectroscopy (RIfS). When comparing TBA-Br with unmodified TBA (TBA-H), it was demonstrated that the modification effectively improved the affinity of TBA; dissociation constants (KD) of TBA-H and TBA-Br were 45.4 nM and 1.99 nM, respectively. These values, which were obtained by direct observation of thrombin binding using RIfS, have the same order of magnitude as those obtained in our previous study utilizing conformational changes in TBA to detect thrombin binding, thus confirming the validity of the obtained KD values. RIfS measurements also revealed that the 8-BrdG modification resulted in a lower dissociation rate constant (kd), which suggests that the enhancement of affinity can be attributed to the stabilization of the G-quadruplex structure on introduction of 8-BrdG.


Introduction
Chemical modification of nucleic acids is a useful approach to improve the stability of higher-order structures [1][2][3][4][5]. Thrombin-binding aptamer (TBA), d(GGTTGGTGTGGT-TGG), is a nucleic acid for which modification has attracted considerable attention, as the improved stability may lead to enhanced affinity for the target species, thrombin, which is potentially useful for the diagnosis and treatment of various conditions. Because the relationship between higher-order structure and affinity is well known, its modification has been widely studied. TBA is known to fold into a Gquadruplex structure consisting of two G-planes connected via a TGT loop and two TT loops, as shown in Figure 1(a) [6][7][8][9][10][11].
Strategies for modifying TBA can be classified into two categories: modification of the nucleotide backbone such as Locked Nucleic Acid (LNA) [12,13] or 2 -fluoro-arabinonucleic acid (2 -F-ANA) [14] and modification of nucleobases, such as alkylation or phenylation [15]. Both strategies have been shown to be effective for stabilizing higher-order structure by stabilizing specific glycosidic bond conformations in the G-plane.
We recently demonstrated that the higher-order structure of TBA can be stabilized by introducing 8-bromodeoxyguanosine (8-BrdG, Figure 1(b)), which stabilizes a syn conformation of a glycosidic bond by steric hindrance between the bromo group at the C8 position and the deoxyribose moiety [16][17][18]. When two 8-BrdG residues were introduced in place of the 1st and 10th guanosine residues, which have a syn conformation, the higher-order structure of TBA was significantly stabilized, with an increase in melting point of 15.8 • C, and the affinity for thrombin showed a 12.5-fold greater binding constant [19]. In a previous study, however, the binding of thrombin to TBA was detected indirectly; conformational transition of TBA from a single strand to a G-quadruplex, which is induced by the binding of thrombin, was monitored using a fluorescent dye. Therefore, direct measurement of the thrombin-binding phenomenon has been required in order to verify the estimated binding constant values and the effects of 8-BrdG on enhancing the affinity of TBA for thrombin. Furthermore, the practical usefulness of the 8-BrdG-modified aptamer has not been investigated.
In this study, we preformed direct observation of thrombin binding to an 8-BrdG-modified TBA (TBA-Br) using reflectometric interference spectroscopy (RIfS), which can measure changes in optical thickness of a molecular layer on a sensor chip and is, therefore, expected to detect thrombin binding to the modified TBA immobilized on a chip [20][21][22][23] in order to assess the binding constants of TBA-Br and the effectiveness of TBA-Br for developing a sensitive TBA sensor.

Chip Preparation.
Sensor chips were prepared in situ using RIfS (Figure 2). A hetero bifunctional cross-linker, Sulfo-EMCS, which bears a maleimide group and succinimide group at each end, was used to immobilize a thiol-terminated aptamer onto an aminated SiN chip. HTT was used to cap the unreacted maleimide termini of the immobilized Sulfo-EMCS moieties. After injection of aptamer for immobilization, an approximately 1.2-nm shift in wavelength was observed with both the unmodified (TBA-H) and modified aptamers (TBA-Br), thus confirming the same degree of immobilization for both aptamers. TBA-Br possesses two 8-BrdG residues at the 1st and 10th guanosine sites, and based on our previous study, is optimized in terms of the number and position of the 8-BrdG groups [19].

Kinetic Measurement.
The sensor chips immobilized with TBA-H and TBA-Br were assessed by injection of 100 nM thrombin. The response was reproducible with variations of less than 8.7%. When compared to the TBA-Himmobilized chip (Figure 3(a)), the TBA-Br-immobilized chip showed a larger response (Figure 3(b)). These results suggest that TBA-Br has a higher affinity for thrombin than the TBA-H and accordingly captured a larger number of thrombin molecules. Using the sensorgrams obtained on injection of 100 nM thrombin, kinetic analysis was performed (Table 1). TBA-Br showed a larger association rate constant (k a ) (1.58 × 10 5 M −1 s −1 ) than TBA-H (7.45 × 10 4 M −1 s −1 ). However, the effect on dissociation rate constant was more significant; TBA-Br showed a dissociation rate constant an order of magnitude lower (k d = 3.15 × 10 −4 s −1 ) than TBA-H (k d = 3.38 × 10 −3 s −1 ). Because thrombin binding with TBA is accompanied by G-quadruplex formation of TBA, the decrease in dissociation rate constant may be attributed to the stabilization of G-quadruplex structure of TBA (TBA-Br) due to the introduction of 8-BrdG; the Tm value for TBA-H was 50.7 • C while that for TBA-Br was 66.5 • C [19].
Dissociation constants (K D ) for TBA-H and TBA-Br were 45.4 nM and 1.99 nM, respectively. These values are within the same order of magnitude as those obtained by our previous study using a fluorescent dye [19]. It is also notable that the dissociation constant for TBA-H reported elsewhere is between 20 and 102.6 nM, which was obtained by other detection methods, such as SPR [24,25], QCM [26], ITC [27], and an electrochemical sensor [28], thus supporting the suitability of the applied conditions and estimation protocol. Direct detection of binding between TBA and thrombin appeared to be successful using RIfS, and the enhancement of affinity by introducing 8-BrdG to TBA was confirmed. Interestingly, the dissociation constant (K D ) for TBA-Br estimated by RIfS was slightly lower than that obtained using the fluorescent dye [19], whereas the dissociation constant (K D ) for TBA-H estimated by RIfS was larger than that obtained using the fluorescent dye. Although the cause of these discrepancies is not currently understood, it is possible that even in the absence of thrombin, TBA-Br forms a Gquadruplex in the binding buffer at the temperature used for RIfS measurements due to the improved stability, which would be favorable for binding with thrombin as a result of preorganization effects.

Effects of Aptamer Modification on Sensitivity.
Sensors that sensitively detect thrombin are potentially useful for diagnosis of thrombosis. The TBA-Br-immobilized chip is expected to offer improved sensitivity due to its high affinity for thrombin. Thus, we investigated the sensitivity of the RIfS  sensor with an TBA-Br-immobilized chip. Responses to various concentrations of thrombin (from 0.01 nM to 500 nM) are plotted in Figure 4. The TBA-Br-immobilized chip showed a significant response to 1 nM thrombin with variations of less than 5%, whereas the detection limit of the TBA-H-immobilized chip was higher than 5 nM. Although the degree was moderate, it confirms that modification of TBA to improve affinity can contribute to enhancing thrombinsensor sensitivity.

Conclusions
In this study, we directly observed the association between TBA-Br and thrombin by RIfS in order to estimate the kinetic rate and affinity constants. The estimated dissociation constants had the same order of magnitude as those obtained in a previous study using a fluorescent dye, thus suggesting that modification with 8-BrdG at appropriate residues in the Gplane increases the affinity of TBA. In addition, the decrease observed in dissociation rate constant by introduction of 8-BrdG suggests that affinity enhancement is due to the effect of 8-BrdG in stabilizing the G-quadruplex structure of TBA. A preliminary test on the sensitivity of TBA-immobilized sensor chips also demonstrated that modification to improve affinity resulted in a better detection limit although the effect is currently moderate.