General Approach for Engineering Small-Molecule-Binding DNA

Copyright © 2013 American Chemical Society. *E-mail: [email protected]. ... route for the engineering of additional split aptamers, dramatically...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/ac

General Approach for Engineering Small-Molecule-Binding DNA Split Aptamers Alexandra D. Kent,§ Nicholas G. Spiropulos,§ and Jennifer M. Heemstra* Department of Chemistry and the Center for Cell and Genome Science, University of Utah, Salt Lake City, Utah 84112, United States S Supporting Information *

ABSTRACT: Here we report a general method for engineering three-way junction DNA aptamers into split aptamers. Split aptamers show significant potential for use as recognition elements in biosensing applications, but reliable methods for generating these sequences are currently lacking. We hypothesize that the three-way junction is a “privileged architecture” for the elaboration of aptamers into split aptamers, as it provides two potential splitting sites that are distal from the target binding pocket. We propose a general method for split aptamer engineering that involves removing one loop region, then systematically modifying the number of base pairs in the remaining stem regions in order to achieve selective assembly only in the presence of the target small molecule. We screen putative split aptamer sequence pairs using split aptamer proximity ligation (StAPL) technology developed by our laboratory, but we validate that the results obtained using StAPL translate directly to systems in which the aptamer fragments are assembling noncovalently. We introduce four new split aptamer sequences, which triples the number of small-molecule-binding DNA split aptamers reported to date, and the methods described herein provide a reliable route for the engineering of additional split aptamers, dramatically advancing the potential substrate scope of DNA assembly based biosensors.

S

The research presented here explores this hypothesis and demonstrates that three-way junction DNA aptamers can be reliably engineered into new split aptamers. Stojanovic and coworkers have recently described a method for generating threeway junction aptamers using structurally biased DNA libraries.12 Starting from these sequences, we report a general approach to aptamer splitting that consists of sequence division, systematic truncation of complementary stem regions, and optimization of buffer conditions to achieve selective assembly in the presence of the desired small-molecule target (Figure 1b). To evaluate the assembly properties of each putative split aptamer sequence, we utilize split aptamer proximity ligation (StAPL) technology developed by our laboratory, as this translates DNA assembly events into the output of DNA ligation, which can be easily quantified by polyacrylamide gel electrophoresis (PAGE).13 We observe slight changes in target binding preference upon aptamer splitting, but the overall target selectivity of our new split aptamers is very similar to that of their respective parent aptamer sequences. We also demonstrate that the small-molecule-dependent DNA assembly properties observed by StAPL are analogous to those observed when no ligation step is used. Thus, the split aptamers reported here are well-suited for use in a wide variety of biosensing

plit aptamers are composed of two nucleic acid strands that assemble selectively in the presence of a small-molecule or protein target.1−3 Small-molecule-binding split aptamers have found wide use as recognition elements for biosensing applications.4,5 However, despite the broad utility of these nucleic acids, reliable methods for the development of new split aptamer sequences are lacking. In principle, a split aptamer can be generated via strategic division of an aptamer sequence. However, approximately 100 small-molecule-binding aptamers have been reported to date,6 and only two of these have been successfully engineered into split aptamers.1,2 The difficulty of engineering aptamers into split aptamers likely arises from the fact that many aptamers have a hairpin or hairpin-like architecture,7 and the small-molecule target often binds to nucleobases in the loop region of the hairpin.8 While the reported ATP split aptamer does have a hairpin-like architecture with target binding in the stem region,1 we have found that in general, hairpin-like aptamers are challenging to divide without perturbing the target binding pocket.9 In contrast, the widely used cocaine split aptamer is derived from a parent aptamer having a three-way junction architecture.10 We hypothesize that this three-way junction is a “privileged architecture” for the engineering of aptamers into split aptamers, as the small-molecule target typically binds in the central junction region,11 providing two loop regions where the sequence can be divided without perturbing the target binding pocket (Figure 1a). © 2013 American Chemical Society

Received: August 8, 2013 Accepted: September 13, 2013 Published: September 13, 2013 9916

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

Article

Agilent Technologies 1260 Infinity series HPLC. Mass spectra were collected using the University of Utah Mass Spectrometry Core Facility. Absorbance values for DNA were measured using a Take3 plate on a Synergy MX multimode microplate reader (BioTek). Molar extinction coefficients were calculated using OligoAnalyzer 3.1 from Integrated DNA Technologies and were used to calculate DNA concentration. Fluorescence quenching was measured using the Synergy MX microplate reader. PAGE gels were analyzed for Cy3-fluorescence using a Typhoon FLA9500 scanner (GE Healthcare Life Sciences) with a 532 nm excitation laser and a BPG1 (570DF20) filter. Fluorescence volumes were quantified using the ImageQuant TL 1D gel analysis program, with manual integration, automatic background subtraction, and signal detection. Modifiers Used for DNA Synthesis. All modified phosphoramidites and CPG cartridges were purchased from Glen Research. Amine functional groups were installed using a 3′-PT-amino-modifier C6 CPG (20-2956). Cy3 fluorophores were installed using Cy3 phosphoramidite (10-5913). Azide functional groups were installed using bromohexyl phosphoramidite (10-1946), which was converted to azide according to manufacturer’s protocol. Black Hole Quencher 2 (BHQ2) was installed using 3′-BHQ2 CPG (20-5932). Cyclooctyne Functionalization of Amine-Modified DNA. A mixture of ALO cyclooctyne carboxylic acid (10 mg, 55 μmol), 1-ethyl-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) (10 mg, 52 μmol), and N-hydroxysuccinimide (10 mg, 87 μmol) was dissolved in anhydrous DMF (200 μL) in a 2.0 mL microcentrifuge tube. The reaction mixture was shaken at room temperature for 1 h. Amine-modified DNA (20 nmol) dissolved in 200 μL phosphate buffer (25 mM, pH 8.2) was then added, and the reaction mixture was allowed to shake at room temperature for an additional 1 h. The reaction mixture was desalted using a NAP 5 column (GE Healthcare), and the ALO-functionalized DNA was purified using reversephase HPLC (Agilent ZORBAX Eclipse XDB-C18, 5 μm particle size, 4.6 × 150 mm) with a binary mixture of 0.1 M aqueous triethylammonium acetate (TEAA)/MeCN. The DNA was eluted at a flow rate of 2 mL/min starting at 95:5 0.1 M TEAA/MeCN and ramping up to 60:40 0.1 M TEAA/MeCN over a period of 16 min. The eluted fraction was monitored at 570 and 260 nm to detect Cy3 and DNA, respectively. DNA 1ALO, 2-ALO, and 3-ALO were collected at retention times of 12.1, 11.7, and 11.1 min, respectively. Purified DNA was frozen and lyophilized to afford 4−6 nmol of pink solid (25−35%). Purified DNA 1-ALO, 2-ALO, and 3-ALO were analyzed by MALDI-TOF mass spectrometry in linear positive mode (see Supporting Information). Purified DNA was dissolved in MilliQ water, divided into aliquots, and stored at −80 °C until needed. General Protocol for Split Aptamer Ligation. Stock solutions of DIS, DOG, DCA, and BE were prepared in 10− 80% DMSO in H2O and sonicated at 37 °C for 20 min to allow all solids to dissolve. To a 0.2 mL PCR tube was added steroid solution, Tris (pH 8.2), NaCl (1 M), azide-modified DNA (30 μM), ALO-modified DNA (10 μM), and water to bring the final volume to 20 μL. The PCR tubes were capped, centrifuged, and incubated at room temperature for the desired reaction time. The reactions were quenched using 20 μL 2× PAGE loading buffer containing 7 M urea, and then they were analyzed by denaturing PAGE on a 10% Tris/Borate/EDTA (TBE)/polyacrylamide gel. Denaturing polyacrylamide gels

Figure 1. (a) Splitting of aptamers having hairpin or three-way junction architecture. (b) General approach to engineering a split aptamer from a three-way junction aptamer.

formats. Together, this research provides a general approach for the engineering of new DNA split aptamers, and this work introduces four new split aptamer sequences. Introduction of these new sequences triples the number of small-moleculebinding DNA split aptamers reported to date, and the methods described herein provide a reliable route for the engineering of additional split aptamers, dramatically advancing the potential substrate scope of DNA assembly based biosensors.

■

EXPERIMENTAL SECTION General Techniques. Unless otherwise noted, all starting materials were obtained from commercial suppliers and were used without further purification. β-Estradiol (BE), sodium deoxycholate (DCA), dehydroisoandrosterone-3-sulfate sodium salt dihydrate (DIS), and deoxycorticosterone 21-glucoside (DOG) were purchased as solids (Sigma-Aldrich). DOG was stored at −20 °C until needed; all other steroids were stored at room temperature. DOG is slightly insoluble in water, and BE is almost completely insoluble in water. Thus, a binary mixture of water/DMSO was used to increase solubility of DOG and BE. Human blood serum was purchased from Sigma-Aldrich. Artificial urine media was prepared as reported in the literature.14 All reactions were performed under an inert atmosphere using N2. NMR spectra were recorded on a Varian 300 MHz NMR instrument. Thin-layer chromatography was performed on glass plates (EMD Chemicals) and visualized using iodine staining. Cyclooct-1-yn-3-glycolic acid (ALO) was synthesized following procedures adapted from those described in the literature.15 DNA was purchased from the University of Utah DNA/Peptide Core Facility. DNA was purified using an 9917

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

Article

were imaged as described above. The yield for ligation reactions was calculated according to eq [1]: %yield = 100 × [Vp/(Vp + VR )]

([1])

in which VR is the fluorescence volume of the band for reactants and Vp is the fluorescence volume of the band for ligated products. For split aptamer ligation in biological fluids, ligation reactions were performed as described above, but using 50% human blood serum or 50% artificial urine media in place of Tris buffer. General Protocol for Fluorescence Measurements. All measurements were carried out in 25 mM Tris, pH 8.2, and 115 mM NaCl in 0.2 mL PCR tubes. Fmax was measured using a solution containing only DNA 4-Cy3. To observe liganddependent split aptamer assembly, we prepared solutions containing DNA 4-Cy3 and DNA 1-BHQ2, along with the specified concentration of steroid. The total volume for all solutions was 100 μL. Each mixture was incubated for 80 min at room temperature and transferred to a Costar 96-well nontreated, opaque flat-bottom, black polystyrene plate (Corning, Inc.). Fluorescence was measured using excitation at 550 nm and emission at 570 nm.

■

RESULTS AND DISCUSSION Aptamer SELEX is typically carried out using a random nucleic acid library.16 This maximizes sequence diversity, but it often provides aptamers having a hairpin or hairpin-like architecture,7 as this is the simplest and most common nucleic acid structure element. This does not pose a problem for the aptamers themselves, but it does pose a significant barrier to elaboration of the aptamers into split aptamers. In contrast, SELEX can be carried out using a structurally biased library in which some portions of the sequence are held constant to enforce specific structure elements.17 Using the structurally biased DNA library shown in Figure 2a, Stojanovic and co-workers have generated a series of four three-way junction aptamers that bind with varying affinity and selectivity to the steroid targets DIS, DOG, DCA, and BE (Figure 2b).12 We envisioned that these DNA aptamers could be converted into split aptamers by removing one of the loop regions and reducing the number of base pairs in the bottom stem region. The underlying principle behind split aptamer assembly is that the sequences have some complementarity to one another, but the enthalpic gain from base pairing and base stacking is not sufficient to overcome the entropic cost of assembly. Upon addition of the target ligand, enthalpy is gained from favorable interactions between the ligand and the aptamer fragments, which tips the thermodynamic balance and drives the equilibrium toward assembly.13,18 Thus, a key aspect of split aptamer engineering is fine-tuning the number of base pairs formed between the sequences, such as to thermodynamically poise the system at the brink of assembly. In choosing how many nucleotides to remove from the three-way junction aptamers in Figure 2a, we were initially guided by mimicking the number of base pairs present in the widely used cocaine split aptamer.5 However, for each split aptamer, we investigated multiple sequence pairs in order to find the sequences that demonstrate the optimal thermodynamic balance for targetdriven assembly. To quantitatively evaluate the assembly properties of each putative split aptamer sequence pair, we turned to StAPL technology previously developed by our laboratory. In StAPL,

Figure 2. (a) Three-way junction aptamers generated via structurally biased SELEX. (b) Chemical structure of steroid targets.

reactive groups are appended to the termini of the split aptamer fragments, and upon aptamer assembly, these reactive groups are brought into close proximity, thus promoting a ligation reaction (Figure 3a).13 One of the aptamer fragments is

Figure 3. (a) Split aptamer proximity ligation (StAPL). (b) Putative DOGS.1 split aptamer sequence pairs having varying degrees of base pairing. 9918

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

Article

synthesized having a Cy3 fluorophore, which enables fast and accurate quantification of ligation yield using denaturing PAGE. Our first attempt to apply our split aptamer engineering method focused on aptamer DOGS.1. We removed one of the three-nucleotide loop regions to divide the aptamer into two fragments, and we then screened putative split aptamer sequences having 4−6 base pairs in the lower stem region (Figure 3b). We chose to use strain-promoted azide−alkyne cycloaddition as our ligation chemistry,19 as this reaction is bioorthogonal and had provided excellent results in our earlier StAPL studies. Thus, one aptamer fragment was synthesized having a 3′ amine modification, then coupled with ALO cyclooctyne carboxylic acid15 to give DNA 1-ALO, and the second aptamer fragment was synthesized having a 5′ azide (4a-c-N3). To vary the number of base pairs in the stem region, the length of the ALO strand was held constant, and the length of the azide strand was varied systematically. In addition to varying the extent of base pairing between the split aptamer fragments, the equilibrium for split aptamer assembly can also be shifted by varying the ionic strength of the medium. At high ionic strength, electrostatic repulsion between the DNA strands is reduced via Debye screening, effectively increasing the enthalpic driving force for assembly.20 Thus, we investigated the assembly properties of our putative DOGS.1 split aptamers by carrying out StAPL with a reaction time of 80 min in pH 8.2 Tris buffer having 15, 115, or 415 mM NaCl. We performed each reaction with either no ligand or 1 mM DOG, as this ligand was shown to bind the tightest to the DOGS.1 parent aptamer. The reactions having no ligand serve as a control to determine the amount of unwanted background assembly, and the reactions with 1 mM DOG show the efficiency of the desired target-driven assembly. As shown in Figure 4, split aptamer assembly is dependent upon both base pairing and ionic strength, as increasing either of these variables leads to increased ligation yields in both the presence and absence of target. For the 1-ALO:4a-N3 sequence pair having the shortest stem region, we observe only a moderate ligation yield of 44% in the presence of 415 mM NaCl and 1 mM DOG. However, results with 1-ALO4:b-N3 and 1-ALO:4c-N3 having slightly longer stem regions were more encouraging. In 115 mM NaCl, 1-ALO:4b-N3 demonstrated excellent targetdependent assembly and signal-to-background ratio with ligation yields of 90% and 1% for 1 mM DOG and control, respectively. For 1-ALO:4c-N3 having the longest stem region, we found that similar ligation yields of 85% and 1% could be achieved at the lower ionic strength of 15 mM NaCl. These results provided initial validation for our hypothesis that threeway junction aptamers could be reliably elaborated into split aptamers and demonstrated that assembly of the split aptamer fragments can be easily fine-tuned by varying either the number of base pairs or the ionic strength of the buffer. Typically, split aptamer assembly is a noncovalent process where the extent of assembly remains constant once equilibrium is reached. However, in StAPL, assembly events are covalently trapped, enabling signal to accrue over time. Thus, we were curious to observe the effect of ligation time on reaction yield and signal-to-background for our DOGS.1 split aptamer pairs. As anticipated, the data in Figure 5 show that increasing either base pairing or ligation time leads to increased ligation yields in both the presence and absence of DOG. Sequence pair 1-ALO:4b-N3 reacting for 80 min still showed the highest signal-to-background ratio. Nevertheless, moderate signal-to-background with yields of 72 and 12% for 1 mM

Figure 4. (a) Yield of ligated product and (b) signal-to-background ratio as a function of NaCl concentration for DOGS.1 split aptamer sequence pairs. Conditions: 10% DMSO, 25 mM Tris, pH 8.2, 0.5 μM 1-ALO, 2.0 μM 4a-c-N3, 80 min. Data shown reflect a single trial.

DOG and control, respectively, could be achieved with 1ALO:4c-N3 at a reduced ligation time of 20 min. While these results have minimal bearing on traditional noncovalent splitaptamer-based assays, they demonstrate that incubation time in our StAPL-based assays can be reduced by using split aptamers having slightly increased levels of base pairing. Having optimized the sequence and ligation conditions for our DOGS.1 split aptamer, we next investigated the selectivity of the split aptamer for the steroid targets shown in Figure 2b. The parent DOGS.1 aptamer binds strongest to DOG and DIS, binds more weakly to DCA, and does not show observable binding with BE.12 Using the 1-ALO:4b-N3 sequence pair and our optimized ligation conditions, we measured dose-dependent ligation for each of the four ligands. The data in Figure 6 show that our split aptamer parallels the substrate selectivity of 9919

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

Article

Figure 6. (a) Denaturing PAGE of ligation reactions. Lower bands represent unreacted 1-ALO. Upper bands represent 1-ALO + 4b-N3 ligated product. (b) Yield of ligated product as a function of ligand concentration for DOGS.1 split aptamer sequence pair 1-ALO:4b-N3. Conditions: 10% DMSO (DIS, DOG, DCA), 40% DMSO (BE), 25 mM Tris, pH 8.2, 115 mM NaCl, 0.5 μM 1-ALO, 2.0 μM 4b-N3, 80 min. Error bars represent standard deviation of three independent trials.

Thus, we wanted to validate that our new split aptamer is also capable of target-dependent assembly in the absence of a ligation step. To evaluate noncovalent assembly, we modified our DOGS.1 split aptamer sequences to give 1-BHQ2 and 4bCy3 (Figure 7a). If these DNA strands are not assembled, we observe maximum fluorescence emission from the Cy3 fluorophore. However, upon assembly, the Cy3 fluorescence is quenched by the Black Hole Quencher 2 (BHQ2), enabling spectroscopic monitoring of split aptamer assembly. We were encouraged to observe that the noncovalent target-dependent assembly almost perfectly mirrors that observed using StAPL. As seen in Figure 7b, 1-BHQ2:4b-Cy3 begins to show assembly at 1 μM DIS or DOG, but it requires 100 μM DCA for observable assembly and shows no assembly with up to 1 mM BE. This demonstrates that StAPL provides an accurate evaluation of the target-dependent assembly properties of our putative split aptamer pairs and that the split aptamers developed using our method are well-suited for use in a wide variety of assay formats. Most split-aptamer-based assays are designed for use in clinical diagnostics applications. Thus, we wanted to explore the ability of our new DOGS.1 split aptamer to undergo targetdependent assembly in complex biological fluids such as serum or urine. The additional salts and biomolecules present in these matrices can interfere with fluorescence measurements, thus we

Figure 5. (a) Yield of ligated product and (b) signal-to-background ratio as a function of ligation time for DOGS.1 split aptamer sequence pairs. Conditions: 10% DMSO, 25 mM Tris, pH 8.2, 115 mM NaCl, 0.5 μM 1-ALO, 2.0 μM 4a-c-N3. Data shown reflect a single trial.

the parent aptamer as the ligation yields for DIS and DOG were above background at concentrations as low as 1 μM and rose to 74 and 93% in the presence of 1 mM DIS and DOG, respectively. In contrast, DCA did not show ligation above background until 100 μM, and yields for BE rose only slightly above background even in the presence of 1 mM target. StAPL provides a convenient, accurate method for evaluating split aptamer assembly, and the results obtained from the ligation reactions can be directly applied toward the development of our StAPL-based small-molecule-detection assays. However, the primary goal of the research reported here is to develop methods for generating new split aptamers that can be used in multiple assay formats, and many of these splitaptamer-based assays rely solely on noncovalent assembly. 9920

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

Article

Figure 8. Ligation yield as a function of ligand concentration for DOGS.1 split aptamer in biological fluids. Conditions: 10% DMSO, 50% human blood serum or artificial urine media, 25 mM Tris, pH 8.2, 215 mM NaCl (buffer), 100 mM NaCl (serum or urine), 0.5 μM 1ALO, 2.0 μM 4b-c-N3, 20 min. Error bars represent standard deviation of three independent trials.

Having established our methodology for engineering of three-way junction aptamers into split aptamers, we next sought to explore the generality of this method by repeating the process with the three other aptamers reported by Stojanovic and co-workers.12 We again divided the sequences by removal of one of the loop regions, then truncated the bottom stem region to achieve the thermodynamic balance needed for selective target-dependent assembly. This provided sequence pairs 1-ALO:5-N3 from splitting of aptamer DISS.1, 2-ALO:6N3 from splitting of aptamer BES.1, and 3-ALO:7-N3 from splitting of aptamer DCAS.1 (Figure 9a). We evaluated the target-dependent assembly properties of each of these split aptamers using StAPL (Figure 9b), and we found that they each show excellent signal-to-background and ligation efficiency with their optimal targets. We observed that the overall target binding preferences of the split aptamers were very similar to those of their respective parent aptamer sequences. However, in some cases, subtle changes to substrate preference were noted. For example, the DISS.1 parent aptamer binds more strongly to DIS than DOG, but the corresponding split aptamer (1ALO:5-N3) shows a slight preference for DOG over DIS. These subtle differences are not entirely unexpected given the difference in methods and conditions used to test the aptamers versus the split aptamers. First, target binding for the parent aptamers was tested at high ionic strength (40 mM Tris, pH 7.4, 2 M NaCl, 20 mM MgCl2), whereas the assembly properties of the split aptamers require that they be utilized at moderate ionic strength (25 mM Tris, pH 8.2, 65−415 mM NaCl). Second, the fluorescence-based assay used to monitor target binding to the parent aptamers provides a different maximum signal for each target. In contrast, StAPL provides a consistent reaction end point for all targets. Thus, even if the overall binding preferences do not change significantly, target binding data for the split aptamers is unlikely to perfectly mirror that observed for the parent aptamers.

Figure 7. (a) Measuring noncovalent split aptamer assembly using Cy3-BHQ2 FRET. (b) F/Fmax as a function of ligand concentration for DOGS.1 split aptamer sequence pair 1-BHQ2:4-Cy3. Conditions: 10% DMSO (DIS, DOG, DCA), 40% DMSO (BE), 25 mM Tris, pH 8.2, 115 mM NaCl, 600 nM 1-BHQ2, 600 nM 4-Cy3. Error bars represent standard deviation of three independent trials.

turned again to StAPL to evaluate split aptamer assembly. We found that DOG is degraded over time in biological fluids, likely via hydrolysis of the glycosidic bond. Therefore, we aimed to reduce our ligation time to 20 min in order to outpace target degradation. To accomplish this, we increased the overall ionic strength of the reaction mixtures. Both artificial urine media and human blood serum have significant salt content,14,21 thus we added only 100 mM NaCl to reactions containing either of these media, and we added 215 mM NaCl to our reaction in buffer. Additionally, for the ligation in 50% human blood serum, we utilized the 1-ALO:4c-N3 sequence pair, having the extended stem region. The 1-ALO:4b-N3 sequence pair was utilized for our reactions in buffer and 50% artificial urine media. As shown in Figure 8, the presence of complex biological fluids does not significantly interfere with aptamer assembly, as we observe similar dose-dependent ligation in buffer, serum, and urine media. The shorter ligation time provides a slightly reduced signal-to-background ratio (as also demonstrated in Figure 5), thus a higher concentration of target (approximately 10 μM) is required to achieve signal over background. However, the fact that our DOGS.1 split aptamer can undergo target-dependent assembly in complex biological fluids with only a slight decrease in efficiency and sensitivity is very encouraging for the use of our new split aptamers in clinical diagnostics assays. 9921

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

Article

Figure 9. (a) Sequences of 1-ALO:5-N3 (DISS.1), 2-ALO:6-N3 (BES.1), and 3-ALO:7-N3 (DCAS.1) split aptamer sequence pairs. (b−d) Yield of ligated product as a function of ligand concentration for (b) 1-ALO:5-N3, (c) 2-ALO:6-N3, and (d) 3-ALO:7-N3. Conditions: 10% DMSO (DIS, DOG, DCA), 40% DMSO (BE), 25 mM Tris, pH 8.2, 0.5 μM 1−3-ALO, 2.0 μM 5−7-N3. See Supporting Information for [NaCl] and ligation time for each sequence pair. Error bars represent standard deviation of three independent trials.

into split aptamers without compromising this selectivity. The ability to systematically and reliably convert aptamers into split aptamers is anticipated to significantly advance the scope of small-molecule targets that can be detected using split-aptamerbased DNA biosensors.

The biased library used for selection of the three-way junction aptamers was intended to result in an identical secondary structure for all of the aptamers. However, due to mutations during the SELEX process, the aptamers each have slightly different sequences and lengths in their stem-loop regions, as well as different numbers of unpaired nucleobases in their putative target binding pockets. Thus, our ability to successfully engineer each of the three-way junction aptamers into a split aptamer despite their differences in secondary structure demonstrates the generality and robustness of our methodology.

■

ASSOCIATED CONTENT

S Supporting Information *

General experimental protocols, synthesis of ALO cyclooctyne, MALDI-TOF mass spectra of ALO-functionalized DNA, denaturing PAGE, and tabular data for optimization and dose-dependent ligation of split aptamers. This material is available free of charge via the Internet at http://pubs.acs.org.

■

CONCLUSIONS The research reported here introduces the first general method for engineering of DNA split aptamer sequence pairs. We demonstrate that the three-way junction is a “privileged architecture” for the conversion of aptamers into split aptamers, and we successfully generate four new split aptamers from previously reported three-way junction aptamer sequences. Using StAPL, we systematically explore the effect of sequence length and ionic strength on split aptamer assembly and find that either of these factors can be varied in order to fine-tune the assembly properties of the aptamer fragments. We show that the assembly properties using StAPL mirror those that are observed in the absence of a ligation step, making our new split aptamers well-suited for use in assay formats that utilize noncovalent split aptamer assembly. We also show that our model split aptamer can be used in complex biological fluids without significantly sacrificing performance. Importantly, we find that the split aptamers have substrate binding selectivity that is similar to that of their respective parent aptamer sequences. Thus, aptamers that are generated to bind selectively to specific small-molecule targets can be converted

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions §

These authors contributed equally.

Notes

The authors declare no competing financial interests.

■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (CHE 1308364 to J.M.H.) and the Army Research Office (59482CH to J.M.H.). N.G.S. gratefully acknowledges support from the National Science Foundation (Graduate Research Fellowship 1256065). A.D.K. gratefully acknowledges support from the University of Utah Undergraduate Research Opportunities Program. 9922

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923

Analytical Chemistry

■

Article

REFERENCES

(1) Huizenga, D. E.; Szostak, J. W. Biochemistry 1995, 34, 656−665. (2) Stojanovic, M. N.; de, P. P.; Landry, D. W. J. Am. Chem. Soc. 2000, 122, 11547−11548. (3) Yamamoto-Fujita, R.; Kumar, P. K. R. Anal. Chem. 2005, 77, 5460−5466. Chen, J.; Zhang, J.; Li, J.; Yang, H.-H.; Fu, F.; Chen, G. Biosens. Bioelectron. 2010, 25, 996−1000. (4) Zuo, X.; Xiao, Y.; Plaxco, K. W. J. Am. Chem. Soc. 2009, 131, 6944−6945. Golub, E.; Pelossof, G.; Freeman, R.; Zhang, H.; Willner, I. Anal. Chem. 2009, 81, 9291−9298. Freeman, R.; Sharon, E.; TelVered, R.; Willner, I. J. Am. Chem. Soc. 2009, 131, 5028−5029. Wen, Y.; Pei, H.; Wan, Y.; Su, Y.; Huang, Q.; Song, S.; Fan, C. Anal. Chem. 2011, 83, 7418−7423. Sharma, A. K.; Kent, A. D.; Heemstra, J. M. Anal. Chem. 2012, 84, 6104−6109. (5) Zhang, J.; Wang, L.; Pan, D.; Song, S.; Boey, F. Y. C.; Zhang, H.; Fan, C. Small 2008, 4, 1196−1200. (6) McKeague, M.; DeRosa, M. C. J. Nucleic Acids 2012, 2012, 748913. (7) Uphoff, K. W.; Bell, S. D.; Ellington, A. D. Curr. Opin. Struct. Biol. 1996, 6, 281−288. (8) Patel, D. J. Curr. Opin. Chem. Biol. 1997, 1, 32−46. (9) Spiropulos, N. G.; Heemstra, J. M. Unpublished results. (10) Stojanovic, M. N.; de, P. P.; Landry, D. W. J. Am. Chem. Soc. 2001, 123, 4928−4931. (11) Neves, M. A. D.; Reinstein, O.; Johnson, P. E. Biochemistry 2010, 49, 8478−8487. (12) Yang, K.-A.; Pei, R.; Stefanovic, D.; Stojanovic, M. N. J. Am. Chem. Soc. 2012, 134, 1642−1647. (13) Sharma, A. K.; Heemstra, J. M. J. Am. Chem. Soc. 2011, 133, 12426−12429. (14) Brooks, T.; Keevil, C. W. Lett. Appl. Microbiol. 1997, 24, 203− 206. (15) Agard, N. J.; Baskin, J. M.; Prescher, J. A.; Lo, A.; Bertozzi, C. R. ACS Chem. Biol. 2006, 1, 644−648. (16) Stoltenburg, R.; Reinemann, C.; Strehlitz, B. Biomol. Eng. 2007, 24, 381−403. (17) Davis, J. H.; Szostak, J. W. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 11616−11621. Ruff, K. M.; Snyder, T. M.; Liu, D. R. J. Am. Chem. Soc. 2010, 132, 9453−9464. (18) Neves, M. A.; Reinstein, O.; Saad, M.; Johnson, P. E. Biophys. Chem. 2010, 153, 9−16. (19) Sletten, E. M.; Bertozzi, C. R. Angew. Chem., Int. Ed. 2009, 48, 6974−6998. Debets, M. F.; van, B. S. S.; Dommerholt, J.; Dirks, A. J.; Rutjes, F. P. J. T.; van, D. F. L. Acc. Chem. Res. 2011, 44, 805−815. (20) Schildkraut, C.; Lifson, S. Biopolymers 1965, 3, 195−208. Gong, H.; Freed, K. F. Phys. Rev. Lett. 2009, 102, 057603−1−4. (21) Payne, R. B.; Levell, M. J. Clin. Chem. 1968, 14, 172−178.

9923

dx.doi.org/10.1021/ac402500n | Anal. Chem. 2013, 85, 9916−9923