Facile and Cost-Effective Detection of Saxitoxin Exploiting Aptamer Structural Switching

Aptamers are short single-stranded nucleic acids that bind to a defi ned target with high affi nity and specifi city (1,2). They are selected from a random library of 1012–1024 oligonucleotides, through a process called SELEX (Systematic Evolution of Ligands by Exponential Enrichment) where target-bound strands are isolated and PCR amplifi ed. The combination of an aptamer specifi city and affi nity, together with the inherent avoidance of an animal host makes aptamers an extremely att ractive alternative to the use of antibodies as analyte recognition elements (3,4). Aptamers are highly cost-eff ective, highly stable and are routinely produced by easier and more consistent manufacturing methods than antibodies.


Introduction
Aptamers are short single-stranded nucleic acids that bind to a defi ned target with high affi nity and specifi city (1,2).They are selected from a random library of 10 12 -10 24 oligonucleotides, through a process called SELEX (Systematic Evolution of Ligands by Exponential Enrichment) where target-bound strands are isolated and PCR amplifi ed.The combination of an aptamer specifi city and affi nity, together with the inherent avoidance of an animal host makes aptamers an extremely att ractive alternative to the use of antibodies as analyte recognition elements (3,4).Aptamers are highly cost-eff ective, highly stable and are routinely produced by easier and more consistent manufacturing methods than antibodies.
Aptamers have been developed for a wide range of diff erent molecules (5)(6)(7)(8), and considerable eff ort has been placed on engineering and formatt ing selected aptamers into analytical devices.Since aptamers are a class of nucleic acid molecules, they are inherently much more fl exible than antibodies and have been used in various formats not amenable to antibodies, including displace-ment assays (9,10), apta-PCR (11,12) and molecular aptamer beacons (13,14).The arrangement of an aptamer as a sensing platform is oft en dependent on its interaction with the target (15)(16)(17)(18)(19).Some assays are designed similarly to those with antibodies, based on strategic labelling/immobilization of the participant molecules.Other common formats measure changes in surface properties when an att ached molecule binds a free counterpart and recent developments rely on the structural switch that the aptamer itself undergoes when binding its target.In particular, the tertiary structural switch of aptamers that bind small molecules is oft en signifi cant, and can be measured by a number of strategies (20)(21)(22)(23).Molecular aptamer beacons are of particular interest as they allow assays to be performed in a single experimental step that does not require washing or further developing, with the only required end user intervention being sample addition.Successful examples are fl uorophore-quencher, dual-pyrene and quantum dot interactions (24)(25)(26)(27)(28).
Here, we exploit a previously discovered aptamer (29), which we adapted as a label-free fl uorescent sensor for the detection of saxitoxin (STX), a main component of the paralytic shellfi sh poison (PSP), concentrated by bivalve molluscs when they consume toxic dinofl agellates from a harmful algal bloom (HAB).Contaminated mollusc consumption can be fatal and HABs are thus not only a problem for public health, but also for mollusc fi sheries, as well as tourism since they aff ect living resources in coastal systems (30)(31)(32).PSP toxicity detection is typically performed through a mouse bioassay, with obvious disadvantages, and several bioanalytical systems are currently being proposed to replace it (33)(34)(35).Herein, we detail the use of the reported STX-binding aptamer as a facile and economically viable molecular analytical tool for the sensitive and specifi c detection of PSP toxin, exploiting the fl uorescence detection of the conformational change of aptamer structure upon binding with its cognate target.

Materials and Methods
Single-stranded DNA corresponding to the aptamer reported to have been selected against STX (29) (STXH aptamer: GGT ATT GAG GGT CGC ATC CCG TGG AAA CAT GTT CAT TGG GCG CAC TCC GCT TTC TGT AGA TGG CTC TAA CTC TCC TCT) as well as the random DNA strand used by the same authors as a negative control (RAN: GGT ATT GAG GGT CGC ATC TAG TAG AAA AGT GCT GAG TAG TTT TAC CTG GTA GAT ATG CGA TGG CTC TAA CTC TCC TCT) were synthesized by and purchased from Integrated DNA Technologies (Hayward, CA, USA).
STX and gonyautoxin 2/3 (GTX 2/3) preserved in 0.001 M HCl were lyophilized and later resuspended in RNAse-and DNase-free purifi ed water to constitute 80 μg/mL of stock concentration.
The high quantum yield fl uorescent dye Evagreen was purchased from Biotium, Inc. (Coralville, KY, USA).Evagreen uses a 'release on demand' mechanism, which greatly diminishes background noise, emitt ing fl uorescence upon switching confi guration when binding to double-stranded DNA.
Saxitoxin detection was performed in a 10-μL reaction volume containing 10 μM of the STX aptamer STXH in purifi ed water and Evagreen dye at 5× fi nal concentration.The fl uorescence was measured by performing high resolution melting (HRM) analyses in a real time thermocycler Eco Illumina (Illumina, San Diego, CA, USA) with excitation/emission of 500/530 nm.The HRM denaturation graphics were monitored in order to verify the change from dsDNA to ssDNAs with increasing temperature, while the fl uorescence value taken at 30 °C was used as a baseline value to establish the relation between STX concentration and fl uorescence.The HRM protocol includ ed a denaturation step of 5 s at 95 °C, followed by renaturation at 30 °C and a melting step from 30 to 95 °C at a ramp rate set at 0.08 °C/s (36).All measures were carried out in triplicate and the melting analysis was based on the derivative of the fl uorescence signal normalized to a control that did not contain STX.
Goodness-of-fi t for correlations (R 2 values) and their signifi cance (p-values calculated from an F-test) were per-formed using GraphPad Prism v. 5 for Windows (San Diego, CA, USA).

Results
The normalized derivative melting profi le of the aptamer bound to STX can be seen in Fig. 1, where the confi guration change of the aptamer with 5 μM STX (full line) is clear when normalized to the control signal (dotted line), suggesting a change in the secondary structure stability upon target binding.This melting profi le is consistent in all experiments performed with a range of toxin concentrations.
Since the mouse bioassay has a detection limit of approx. 1 μM STX (34), two independent assays were performed around that value.The fl uorescence signal of Evagreen bound to the STXH aptamer was tested for ten diff erent concentrations of the toxin between 50 nM and 1 μM and between 1 and 10 μM, yielding a signifi cant linear correlation of 0.821 (p=0.01) and 0.885 (p<0.001)respectively, while the same assays applied on the randomly generated ssDNA sequence RAN yielded correlations of 0.040 and 0.198 respectively, which were not signifi cant (p>0.1).
The lowest amount of STX that infl uenced the melting profi le of STXH aptamer was observed to be 25 nM.At that concentration, the profi les are signifi cantly distinguishable from the melting profi le of the unbound aptamer.However, the linearity of the relation concentration/ fl uorescence greatly increases only when points over 50 nM are considered.
The mouse bioassay is meant to quantify saxitoxin within a rough shellfi sh extract (37).Therefore, the detection system developed here was tested using this matrix.The linear correlation of Evagreen fl uorescence and the toxin in concentrations between 1 and 10 μM, while signifi cant (p<0.01), was considerably aff ected by the compounds of the extract (R 2 =0.618).
Potential cross-reactivity of the STXH aptamer was then tested with a range of concentrations of GTX 2/3, a similar toxin to that of STX, but with an OSO 3 -group instead of H group at one of the radical groups (38).No signifi cant change in fl uorescence of the Evagreen bound to the aptamer was observed between 1 and 10 μM (R=0.074;p>0.1), again highlighting the specifi city of this fl uorescence detection method.

Discussion
Most advances on label-free aptasensor development have been based on strategies where the target analyte sequesters the aptamer from a competitive binding molecule, oft en a complementary DNA strand (39)(40)(41).Since in these cases fl uorescence is emitt ed by a dsDNA-binding dye, it is inversely proportional to analyte concentration.Double-stranded DNA is formed with fl uorescently labelled DNA that is complementary to the aptamer, and a fl uorescent signal is observed, with a quantitative decrease in fl uorescence observed upon displacement of the aptamer from its labelled complementary sequence.The sensitivity of this strategy depends on the affi nity of the aptamer for its target, which competes with the Watson--Crick interaction between the aptamer and its complementary strand.
Here, a novel strategy was developed by directly measuring the change in fl uorescence of a dsDNA-binding fl uorophore, Evagreen, due to a modifi cation in aptamer structure upon target binding, a phenomenon that has commonly been described for aptamers that bind to small molecules (5,19,42).With the STX-binding aptamer, the analyte contributes to secondary structure stability, resulting in a diff erent double-stranded confi guration compared to that of the free aptamer.Even though the unbound aptamer has a distinct denaturation profi le, which indicates that without the target, other double--stranded confi gurations might be formed (43)(44)(45), the amount of dsDNA area increases with target concentration.The increased stability of the aptamer bound to the toxin was verifi ed by the distinct melting profi les observed in the HRM assays.This is the fi rst report of a fl uorescent aptasensor for the detection of saxitoxin.However, while signifi cant correlation was obtained under simulated fi eld conditions, a weaker correlation compared to that of the laboratory conditions suggests an aptasensor of low robustness.Apparently, STXH aptamer performance might be very sensitive to external conditions, which might explain the poor results of surface plasmon resonance (SPR) analysis published previously (29) and reviewed in our facilities (∆RU<50).Possibly, small steric eff ects caused by STXH aptamers att ached to the SPR chip surface might impede proper STX binding (46).Moreover, when STXH aptamer was tested with STX for competition with its complementary strand, inverse correlation between STX concentration and fl uorescence was not present at 30 °C (data not shown).Thus, STXH binding affi nity towards its target, not reported in the original publication (29), might not be high enough to overcome Watson-Crick interaction, as opposed to what has been described for other aptamers (39)(40)(41).
In this work, the successful detection of STX was performed by taking advantage of STXH confi guration change upon binding its target, which increased the amount of dsDNA area.Indeed, aptamer-based sensors are designed specifi cally depending on the 3D structure and folding of each particular aptamer, which is usually modifi ed when the aptamer binds its target (47)(48)(49).Usually, reporter molecules are added so they transduce the confi guration change as a measurable signal.Hence, extensive eff orts have been directed towards understanding the folding confi guration of every developed aptamer.This can be estimated roughly through folding algorithms (50), but they typically have to be refi ned in order to obtain an accurate tertiary structure.Characterization of folding/unfolding confi gurations has been performed mainly through nuclear magnetic resonance analysis (51,52), but also through mutational analysis (43,53) and recently through the analysis of stretching force by optical tweezers (54,55).
Here, for the fi rst time, we characterized an aptamer and its interaction with the target by its unfolding speed in relation to small increments of temperature in HRM assays.HRM has been used so far to distinguish among different PCR amplicons in population studies, where a given DNA fragment can be identifi ed by a particular denaturation profi le, even when its sequence diff ers by a single base pair, applying constant temperature increments homogeneously among diff erent samples (56,57).The technique is performed by most nowadays real-time thermocyclers and it is considered one of the most precise, inexpensive and fastest methods to distinguish base--pair binding diff erentials (58,59).The STXH aptamer showed a distinct HRM profi le compared to the RAN sequence (data not shown) and additionally, dramatically changed its HRM profi le upon interacting with STX.HRM assays could then be used as a simple and quick system to characterize aptamer confi guration and to evaluate its response to its target, as well as to structural modifi cations or changes in the reaction media.

Conclusions
Here the feasibility of quantifying saxitoxin (STX) with specifi c aptamers was shown through the use of Evagreen, a dsDNA-binding fl uorophore.When the aptamers are in a solution containing STX, they present a distinguishable melting profi le, which is probably given by specifi c dsD-NA areas that are formed when the target stabilizes the structure.Additionally, at a given temperature, fl uorescence of the Evagreen bound to the dsDNA can be correlated with STX concentration.The method developed here is inexpensive, easy to prepare, fast and sensitive.However, correlation falls dramatically when trying to quantify STX in a rough shellfi sh extract.This is probably caused by the relatively low affi nity of the aptamer for STX, thus, future developments towards high affi nity aptamers could lead to a robust aptamer-mediated method to quantify STX in fi eld samples.