DNA split aptamers as a biosensing platform for the detection of small drug molecules

Prescription drug overdose and abuse is a leading cause of death in the United States. It is a serious issue and has become increasingly problematic as opioids are being prescribed with a higher frequency. For this reason, fast, accurate detection of small drug molecules is crucial. The current standard for use in clinical drug detection is an enzyme‐linked immunosorbent assay (ELISA) that uses a series of antibodies to bind to the target drug and enable quantification via a colorimetric output. However, the antibodies used in an ELISA often cannot distinguish between similar molecules. They are generated in vivo, causing them to have a limited potential target scope as well as being costly. Deoxyribonucleic acids (DNA) can have a wide variety of functions outside of simply encoding genetic information. Aptamers are short sequences of DNA that are capable of binding target small‐molecules. They have emerged as a promising alternative to antibodies, as they are generated in vitro, where negative selections can be used to increase target selectivity. These aptamers can be cleaved to make split aptamers that only assemble in the presence of the target small molecule. In Chapter 1, we report a method of detection analogous to that of an ELISA. The cocaine split aptamer is used in conjunction with Split Aptamer Ligation (StAPL) technology. We attach one split aptamer fragment to a microplate, and in the presence of the target small molecule and the other aptamer fragment, the two aptamer fragments assemble. The reactive groups located on their ends undergo a cycloaddition reaction, covalently attaching the full sequence to the microplate. An attached biotin/streptavidin‐horseradish peroxidase complex allows for a colorimetric output upon addition of TMB substrate. We successfully used this system to detect varying concentrations of cocaine in buffer and biological fluids. In Chapter 2, we investigate the problem of limited numbers of split aptamers for small‐molecule targets. While there are many known aptamers, there are very few known split aptamers that bind small‐molecules. This research generates four new steroid binding split aptamers from their three‐way junction counterparts. We explore optimization, sequence changes, and selectivity of these new split aptamers. We successfully demonstrate a reliable method of separating aptamers with a privileged structure to generate new split aptamers for more targets. Finally, in Chapter 3, we report an effort to make our split aptamer ligation system compatible with qPCR. This would allow for a semi automatable system of detection, which would be more useful for potential clinical applications. In order to accomplish this, we alter the covalent linkage from a cycloaddition adduct to a morpholino linkage. This, as well as a primer binding region on the 5' end of one of the split aptamer fragments, allows the ligated split aptamer to be PCR readable. This is exciting progress in the effort toward creating a semi automatable system of detection for small drug molecules using split aptamers.

DNA SPLIT APTAMERS AS A BIOSENSING PLATFORM FOR THE DETECTION OF SMALL DRUG MOLECULES by Alexandra Kent A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science In Chemistry Approved: _______________________________________ _______________________________________ Dr. Jennifer Heemstra Dr. Cynthia Burrows Research Advisor Chair, Department of Chemistry ______________________________________ ______________________________________ Dr. Thomas Richmond Dr. Sylvia D. Torti Department Honors Advisor Dean, Honors College 2 ABSTRACT Prescription drug overdose and abuse is a leading cause of death in the Unit‐ed States. It is a serious issue and has become increasingly problematic as opioids are being prescribed with a higher frequency. For this reason, fast, accurate detec‐tion of small drug molecules is crucial. The current standard for use in clinical drug detection is an enzyme‐linked immunosorbent assay (ELISA) that uses a series of antibodies to bind to the target drug and enable quantification via a colorimetric output. However, the antibodies used in an ELISA often cannot distinguish between similar molecules. They are generated in vivo, causing them to have a limited poten‐tial target scope as well as being costly. Deoxyribonucleic acids (DNA) can have a wide variety of functions outside of simply encoding genetic information. Aptamers are short sequences of DNA that are capable of binding target small‐molecules. They have emerged as a promising alternative to antibodies, as they are generated in vitro, where negative selections can be used to increase target selectivity. These ap‐tamers can be cleaved to make split aptamers that only assemble in the presence of the target small molecule. In Chapter 1, we report a method of detection analogous to that of an ELISA. The cocaine split aptamer is used in conjunction with Split Aptamer Ligation (StAPL) technology. We attach one split aptamer fragment to a microplate, and in the presence of the target small molecule and the other aptamer fragment, the two aptamer fragments assemble. The reactive groups located on their ends undergo a cycloaddition reaction, covalently attaching the full sequence to the microplate. An attached biotin/streptavidin‐horseradish peroxidase complex allows for a colori‐ 3 metric output upon addition of TMB substrate. We successfully used this system to detect varying concentrations of cocaine in buffer and biological fluids. In Chapter 2, we investigate the problem of limited numbers of split ap‐tamers for small‐molecule targets. While there are many known aptamers, there are very few known split aptamers that bind small‐molecules. This research generates four new steroid binding split aptamers from their three‐way junction counterparts. We explore optimization, sequence changes, and selectivity of these new split ap‐tamers. We successfully demonstrate a reliable method of separating aptamers with a privileged structure to generate new split aptamers for more targets. Finally, in Chapter 3, we report an effort to make our split aptamer ligation system compatible with qPCR. This would allow for a semi automatable system of detection, which would be more useful for potential clinical applications. In order to accomplish this, we alter the covalent linkage from a cycloaddition adduct to a mor‐pholino linkage. This, as well as a primer binding region on the 5' end of one of the split aptamer fragments, allows the ligated split aptamer to be PCR readable. This is exciting progress in the effort toward creating a semi automatable system of detec‐tion for small drug molecules using split aptamers. 4 TABLE OF CONTENTS ABSTRACT 2 CHAPTER 1 5 INTRODUCTION 5 METHODS 8 RESULTS AND DISCUSSION 14 CONCLUSIONS 18 REFERENCES 19 CHAPTER 2 20 INTRODUCTION 20 METHODS 22 RESULTS AND DISCUSSION 26 CONCLUSIONS 32 REFERENCES 33 CHAPTER 3 34 INTRODUCTION 34 METHODS 36 RESULTS AND DISCUSSION 39 CONCLUSIONS 41 REFERENCES 43 FINAL REMARKS 44 ACKNOWLEDGMENTS 44 5 CHAPTER 1: Enzyme‐Linked Small‐Molecule Detection Using Split Aptamer Ligation INTRODUCTION Aptamers are short sequences of DNA that are capable of selectively binding a small‐molecule target. Upon binding, the aptamer undergoes a conformational change into a secondary structure around the small‐molecule. Such sequences can be used for a variety of applications and targets as they provide the aforementioned advantages over their antibody counterparts. Split aptamers are simply aptamers that have been split into two fragments, but re‐tain specificity for their targets1. This confers the benefit of adding two places on the aptamer fragments that are now available for modifica‐tions (Figure 1.1). Split aptamers also have the potential to detect a wider range of targets be‐cause the surface area of the target can be much smaller compared to aptamers. The Heemstra group has used this to pioneer Split Aptamer Proximity Ligation (StAPL) technology in order to better detect small‐molecules2. In this technology, reactive groups are covalently attached to the two free ends of the aptamer fragments. Upon binding the target and the as‐sembly of the split aptamer from the free fragments, these reactive groups are brought into close proximity (Figure Figure 1.1: a) Aptamer fragment folding around a target mole‐cule. b) Split aptamer compris‐ing of two strands folding around a target small molecule. Figure 1.2: Assembly of the split aptamer around its target and subsequent ligation. Depiction of slow ligation in an untemplated reaction. 6 1.2) and covalently react. The proximity allows for a very significant increase in re‐action rate over that of the untemplated reaction. Previously, we investigated some of the applications of such a system in regards to the cocaine split aptamer. The re‐active groups used were a cyclooctyne and an azide. Upon binding cocaine, a cycloaddition reaction occurred creating an essentially irreversible, covalent linkage between the two fragments. This, along with the attachment of a fluorophore to the 5' end of the aptamer, allowed for detection of cocaine using gel electrophoresis. However, this method proved to be somewhat slow and using gel electrophoresis was non‐ideal for clinical applications. In order to remedy some of these limitations, we turned to an assay that could have a wider application. Herein, we report a split aptamer based analog of a sandwich enzyme‐linked immunosorbent assay (ELISA). An ELISA utilizes antibodies to detect a target protein or small molecule (Figure 1.3)3. One antibody, which binds to a target, is immobilized on a surface. Up‐on addition of the detec‐tion antibody and necessary reagents, the presence of the target can be determined. We ap‐plied this same concept to StAPL technology. This research represents the first attempt to produce a split aptamer based assay that utilizes the reporter enzyme‐chromogenic substrate format inherent in ELISAs and the current standard in clinical diagnostics. This assay provides a colorimetric out‐ Figure 1.3: Depiction of a typical ELISA where capture and de‐tection antibodies work in tandem to detect a target. 7 put that is easy to read and amenable to multiplexing for high‐throughput analysis4. StAPL technology has several unique features that make this method uniquely adept at small‐molecule detection. In this approach, the first aptamer fragment comprising of an amine on the 3' terminus and an azide on the opposite terminus is added to an N‐hydroxysuccinimide (NHS) functionalized microplate. The aptamer fragment is co‐valently attached to the microplate via amide bond formation. The remaining NHS groups are chemically blocked using bovine serum albumin (BSA). Next, a test sam‐ple and the second DNA strand are added to the microplate. The second sequence is biotinylated on the 5' terminus and has a cyclooctyne on the opposite terminus. The presence of cocaine in the sample causes the second aptamer fragment to bind to the first around the cocaine molecule. This brings the cyclooctyne and azide into close proximity allowing them to undergo a cycloaddition reaction5 and covalently attach the biotin to the microplate6,7. The concentration of the target is directly cor‐related to the ligation yield, and thus, to the amount of biotin now on the plate. Streptavidin, functionalized with horseradish peroxidase (HRP), binds tightly to bio‐ Figure 1.4: StAPL applied to an enzyme‐linked assay in order to detect cocaine. 8 tin. Once added, the chromogenic substrate, tetramethylbenzidine (TMB), is con‐verted by HRP to an optically observable product (Figure 1.4). StAPL technology is a key aspect of the assay, allowing it to have the format of an ELISA. In principle, split aptamers could be generated for a wide variety of targets, potentially extending the scope of the assay to be broadly applicable to small‐molecule detection. METHODS All chemicals were used without further purification. Cocaine was purchased from Sigma‐Aldrich as a 1 mg/mL solution in acetonitrile. It was then diluted with water, lyophilized, and dissolved again in water at a concentration of 1 mg/mL. DIBAC (DBCO)‐NHS ester was purchased from Click Chemistry Tools, Inc. Clear NHS‐ functionalized DNA‐BIND microplates were purchased from Fisher. Streptavi‐din− HRP conjugate was purchased from Thermo Scientific. TMB substrate was purchased from SurModics. DNA was purchased from the University of Utah DNA/Peptide Synthesis Core Facility. High‐Performance Liquid Chromatography (HPLC) was conducted using a 1260 Infinity HPLC system from Agilent Technolo‐gies. Human blood serum was purchased from Sigma‐Aldrich. Mass spectra were obtained through the Mass Spectrometry Core Facility at the University of Utah. PAGE gels were analyzed for Cy3 fluorescence using a Typhoon 9400 scanner from Amersham Biosciences using a 532 nm excitation laser and 580 BP 30 emission fil‐ter. Fluorescence volumes were corrected for background by subtracting the fluorescence volume of an identically sized area of the gel in which no fluorescence was observed. Absorbance values were obtained using a Biotek Synergy MX micro‐plate reader. NMR spectra were recorded on a Varian 300 MHz NMR instrument. 9 Synthesis of 8,8‐dibromobicyclo[5.0.1]octane A solution of cycloheptene (2.21 mL, 19mM) and tert‐butoxide (4.26 g, 38 mmol) in anhydrous pentane (8 mL) was vigorously stirred at ‐10°C. To this solu‐tion, bromoform (2.45 mL, 28.5 mmol) was added drop wise. The solution was then allowed to warm to room temperature and left to stir for 24 hours. Water (30 mL) was added to the solution, and then neutralized with HCl (1M). The aqueous layer was extracted with hexanes (3 x 30 mL). The organic layers were washed with wa‐ter (3 x 30 mL), dried over MgSO4, and concentrated in vacuo. The product was then purified with silica gel chromatography (5% EtOAc:hexanes), yielding the purified product as a pale yellow oil (4.30g, 84%). 1H NMR (300 MHz, CDCl3) δ 2.26‐2.20 (m, 2H), 1.80‐1.96 (m, 3H), 1.65‐1.78 (m, 2H), 1.30‐1.45 (m, 2H), 1.08‐1.24 (m, 3H) [lit., (400 MHz, CDCl3) δ 2.25‐2.28 (m, 2H), 1.76‐1.92 (m, 3H), 1.62‐1.72 (m 2H), 1.28‐ 1.40 (m, 2H), 1.05‐1.22 (m, 3H)]. 13C NMR (300 MHz, CDCl3) δ 40.9, 34.8, 32.5, 29.1, 28.2 [lit., (100 MHz, CDCl3) δ 40.7, 34.8, 32.2, 28.9, 27.9] GCMS (m/z) calculated for [C8H12Br2] 268, found 268. Synthesis of methyl 2‐bromocyclooct‐1‐en‐3‐glycolate 8,8‐dibromobicyclo[5.0.1]octane (2.5 g, 9.3 mmol) was added to a stirring so‐lution of methyl glycolate (6.35 mL, 83.9 mmol) in anhydrous toluene (3 mL). This was followed by the portion‐wise addition of silver perchlorate (3.85 g, 18.6 mmol). The reaction was protected from light by aluminum foil and allowed to stir for 3 hours at room temperature. The resulting brown precipitate was filtered through celite, washed with EtOAc, dried on MgSO4, and concentrated in vacuo. The crude 10 product was purified by alumina gel column chromatography (5% EtOAc:hexanes, then 15% EtOAc:hexanes) to yield the product as a light yellow oil (1.20 g, 60%). 1H NMR (300 MHz, CDCl3) δ 6.20 (dd, 1H), 4.23 (d, 1H), 4.10 (dd, 1H), 3.95 (d, 1H), 3.72 (s, 3H), 2.63‐2.84 (m, 1H), 2.24‐2.25 (m, 1H), 1.2‐2.2 (m, 7H), 0.75‐0.90 (m, 1H) [lit., (200 MHz, CDCl3) δ 6.35 (dd, 1H), 4.23 (d, 1H), 4.10 (dd, 1H), 3.94 (d, 1H), 3.72 (s, 3H), 2.24‐2.35 (m, 2H), 1.2‐1.9 (m, 7H), 0.7‐0.9 (m, 1H)]. 13C NMR (300 MHz, CDCl3) δ 170.7, 133.5, 131.7, 84.9, 65.8, 53.8, 52.2, 39.8, 36.4, 33.8, 29.1, 26.4 [lit., (50 MHz, CDCl3) δ 170.7, 133.0, 131.4, 84.8, 65.4, 53.48, 51.8, 39.3, 36.5, 33.4, 28.0, 26.2]. LRMS (ESI, pos) calculated for [C11H17BrO3Na]+ 299.1, found: 299.4. Synthesis of cyclooct‐1‐yn‐3‐glycolic acid (ALO) Methyl 2‐bromocyclooct‐1‐en‐3‐glycolate (200 mg, 0.72 mmol) was added to a solution of NaOMe/MeOH (0.5 M, 12.2 mL, 6.1 mmol) and anhydrous DMSO (1 mL). The solution was stirred for 24 h at room temperature. The solvent was re‐moved under reduced pressure to afford a crude mixture, which was dissolved in DCM and acidified using HCl (1 M) to a pH of 2. The aqueous layer was extracted with DCM (3 x 20 mL), dried on MgSO4, and concentrated in vacuo. The crude prod‐uct was purified by silica gel column chromatography (2% MeOH:CH2Cl2, then 5% MeOH:CH2Cl2) to yield cycloocty‐1‐yn‐3‐glycolic acid (ALO) as a light yellow oil (80 mg, 61%). 1H NMR (300 MHz, CDCl3) δ 4.45 (m, 1H), 4.20 (d, 1H), 4.03 (d, 1H), 1.20‐ 2.30 (m, 10H) [lit., (400 MHz, CDCl3) δ 8.12 (s, 1H), 4.58 (d, 1H), 4.32‐4.45 (m, 1H), 4.45 (d, 1H), 1.3‐2.3 (m, 10H)]. 13C NMR (300 MHz, CDCl3) δ 172.2, 101.0, 91.7, 72.7, 66.1, 42.1, 34.1, 29.9, 26.1, 20.7 [lit., (100 MHz, CDCl3) δ 174.1, 101.3, 90.7, 72.6, 65.3, 41.8, 33.9, 29.3, 25.9, 20.3]. LRMS (ESI, neg) calculated for [C10H13O3]‐ 181.2, 11 found 181.3. DNA Sequences The following sequences (Table 1.1) were the modified DNA aptamer frag‐ments used in this chapter. Capture strand 1a and detection strand 2a were used together for the initial assay tests with ALO. Sequence 1b is the mutation of se‐quence 1a for use with DIBAC. Sequences 1c and 1d are also mutations of 1a, but with A10 linkers. Detection strand 2b has the same sequence as 2a, but replacing ALO with DIBAC. Strands 2a and 2b also have Biotin on the 5' end. Sequences 3a‐3h resemble sequence 1a, but modified to be usable in solution phase reactions and containing a varying number of mutations. Sequences 4a and 4b are also for use in solution phase reactions and have either ALO or DIBAC respectably. Strands 4a and 4b also have the fluorophore, Cy3, on the 5' end of the sequence. Combinations of these sequences are used for all the experiments in this chapter. 12 Label Sequences (5'‐3') 1a N3‐GTT CTT CAA TGA AGT GGG ACG ACA‐NH2 1b N3‐CTT CTT CAA CGA AGT GGG ACG ACA‐NH2 1c N3‐CTT CTT CAA CGA AGT GGG ACG ACA‐A10‐NH2 1d N3‐GTC CTT CAA CGA AGT GGG ACG ACA‐A10‐NH2 2a Biotin‐GGG AGT CAA GAA C‐NH‐ALO 2b Biotin‐GGG AGT CAA GAA C‐NH‐DIBAC 3a N3‐GTT CTT CAA TGA AGT GGG ACG ACA 3b N3‐CTT CTT CAA CGA AGT GGG ACG ACA 3c N3‐GCT CTT CAA TGA AGT GGG ACG ACA 3d N3‐GCT CCT CAA CGA AGT GGG ACG ACA 3e N3‐GTC CCT CAA TGA AGT GGG ACG ACA 3f N3‐GTC CCT CAA CGA AGT GGG ACG ACA 3g N3‐GCC CCT CAA TGA AGT GGG ACG ACA 3h N3‐GCC CTT CAA CGA AGT GGG ACG ACA 4a Cy3‐GGG AGT CAA GAA C‐NH‐ALO 4b Cy3‐GGG AGT CAA GAA C‐NH‐DIBAC Table 1.1: DNA Sequences Used in Chapter 1 Coupling ALO to DNA sequence 2a: To a mixture of 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) in a 2.0 mL microcentrifuge tube was added anhydrous DMF (200 μL). The reaction mixture was allowed to shake at room tem‐perature for 1 h. A 200 μL solution of amine modified DNA (20 nmol) in phosphate buffer (25 mM, pH 8.2) was then added. The reaction mixture was allowed to shake at room temperature for an additional 1 h before it was desalted using a NAP 5 col‐umn (GE Healthcare). DNA was purified using reverse phase HPLC (Agilent ZORBAX Eclipse XDB‐C18 5 um particle size, 4.6 x 150 mm) with a binary mixture of 0.1 M triethylammonium acetate (TEAA) in MeCN. MALDI‐TOF (linear positive mode) calcd [M+H]+ 4949.1, found 4950.2. Coupling DIBAC to DNA sequence 2b: 13 A solution of DIBAC‐NHS ester (100 μL, 200 mM in DMF) was added to a so‐lution of DNA (350 μL, 85 μM, in sodium phosphate buffer, pH 7.2). The reaction was allowed to proceed for 2 hours at room temperature, then desalted using a NAP‐5 column. DNA was purified using reverse phase HPLC (Agilent ZORBAX Eclipse XDB‐C18 5 um particle size, 4.6 x 150 mm) with a binary mixture of 0.1 M triethylammo‐nium acetate (TEAA) in MeCN. MALDI‐TOF (linear positive mode) calcd [M+H]+ 5158.5, found 5159.2. Microplate functionalization for use in enzyme‐linked assay: Each well of a 96‐well NHS functionalized DNA‐BIND plate was reacted with 100 μL of DNA capture strand 1 (1μM) in binding buffer (0.5 mM NaPi, pH 8.5). The plate was then shaken for 24 hours and each well was washed first with 3 x 100 μL of wash buffer (10 mM NaPi, pH 7.4, 150 mM NaCl, 0.05% Tween 20), then 3 x 100 μL of water. The remaining unreacted NHS groups were blocked by adding 100 μL of 3% BSA in binding buffer blocking solution, which was followed by incubation at 37 °C for 18 hours. A series of washes were then performed: 3 x 100 μL of water, 3 x 100 μL of wash buffer, and 3 x 100 μL of water. The plate was stored dry and wrapped in foil at 4 °C until use. General method of cocaine detection using the Enzyme‐Linked Assay: A solution (100 μL total volume) containing 100 pmol of detection strand 2, cocaine, and buffer or serum was added to the microplate well. The plate was shak‐en for the incubation time, and then washed (3 x 100 μL water, 3 x 100 μL wash buffer, and 3 x 100 μL water). Then, 100 μL of TMB substrate was added and the ab‐ 14 sorbance measured after 10 min (ALO) or 20 min (DIBAC) using an absorbance plate reader. Net absorbance was calculated by subtracting the absorbance of a con‐trol without cocaine from the values gathered for the test solution. Solution‐phase ligation reactions: Solutions containing cocaine, DNA 3, and DNA 4 were diluted into 50 mM TRIS buffer to give final concentrations specified for each reaction. The reaction was allowed to remain at room temperature for a specified amount of time. Then, 2x PAGE loading buffer was added and the reaction mixtures were separated on 12% TBE/urea polyacrylamide gels. Gel images were analyzed and ligation yields were calculated using Equation 1.1 where VR is the band for DNA 4 and VP is the band for DNA 3 + DNA 4. %Yield=100[ Vp Vp+VR ] Equation 1.1: Calculation for ligation yields RESULTS AND DISCUSSION ALO enzyme‐linked assay reaction time dependence: In initial attempts to develop an Enzyme‐Linked Assay that utilized StAPL technology, we employed DNA sequences 1a and 2a as the capture and detection strands respectively. Sequence 2a had been functionalized with ALO (Figure 1.5). In previous studies, ALO showed dose‐dependent ligation after four hours. However, in 15 the microplate format, almost no reaction was seen after four hours even at high concentrations of cocaine. Thus, the reac‐tion time was increased to 20 hours in order to see similar dose dependence (Figure 1.6). We hypothesized that the slower reaction kinetics in the micro‐plate format arises from the reduced accessibility of the azide‐functionalized strand. However, a reaction time of 20 hours is not competitive with similar ELISAs. In order to combat this problem, we turned to a cyclooctyne with much faster reaction kinetics. Solution phase studies of DIBAC reactivity: The cyclooctyne chosen for its increased reaction rate was DIBAC (Figure 1.5). This cyclooctyne reacts approximately 240‐fold faster than its ALO counter‐part. In order to investigate its reactivity, we turned to solution phase ligation reactions as used in the original publication. We began by using the same sequences as used with the ALO Enzyme‐Linked Assay. However, these sequences showed sig‐nificant background ligation after only 30 minutes of reaction time (Figure 1.7, Lane 2). High salt concentrations can increase the affinity of DNA strands for one another Figure 1.5: Structures and relative rate con‐stants for cyclooctynes, ALO and DIBAC. Figure 1.6: Enzyme‐linked assay using strands 1a and 2a. Reaction times were 4 hours (red squares) and 20 hours (blue circles). Conditions are 100 pmol 2a, 25 mM Tris, pH 8.2, and 5 mM NaCl. 16 causing the subsequent liga‐tion reaction even without the presence of the target small molecule. We believe that the increased lipo‐philiciy of DIBAC relative to ALO may also be acting simi‐larly to promote annealing and ligation via hydrophobic interactions between reac‐tive groups. In order to decrease the background ligation, mutations were added to give DNA strand 3b. A GC base pair was mu‐tated to a CC mis‐match at the reactive end of the sequence, and a GT wobble pair was mutated to a GC base pair (Figure 1.8). This overall mutation of less than one base pair was enough to reduce the background ligation sufficiently to regain dose dependent ligation (Figure 1.7, Lanes 3‐7). DIBAC enzyme‐linked assay studies: After designing the sequences that were feasible for ligation using DIBAC, we began applying them to the enzyme‐linked assay. Using capture strand 1b im‐ Figure 1.7: Sequence 3a shows high background with DIBAC, but mutated sequence 3b reduces background ligation while maintaining dose dependent ligation. Conditions are 0.5 μM 4b, 2.0 μM 3a or 3b, 25 mM Tris, pH 8.2, and 30 minute reac‐tion time. Figure 1.8: Original sequence (3a:4b), mutation to reduce back‐ground ligation due to faster reaction kinetics (3b:4b), and mutation to reduce background ligation due to higher ionic strength of serum (3f:4b). 17 mobilized on the plate and detec‐tion strand 2b, modest dose de‐pendence was observed after a 30 minute ligation time. Howev‐er, inserting an A10 linker be‐tween the microplate and the strand (1c) resulted in much im‐proved signal (Figure 1.9), the linker allowing for better access to the capture strand. The detec‐tion limit of this system is 100 nM, an order of magnitude improved from solution phase detection and is comparable to some of the most sensitive aptamer based as‐says. The total detection time for the assay is two hours, which is comparable to that of an ELISA. Compatibility in human blood serum: The final step in creat‐ing a functioning assay for the detection of cocaine is devel‐oping it to have dose depend‐ence in complex biological fluids such as human blood Figure 1.9: Ligation reaction showing dose dependence with either capture strand 1b (red squares) or 1c (blue circles). Conditions are 1 μM 2b, 25 mM Tris, pH 8.2, 5 mM NaCl, and 30 minute reaction time. Figure 1.10: Screening of capture sequences 3b‐h with detec‐tion sequence 4b for the lowest background option in serum. Conditions are 0.5 μM 4b, 2.0 μM 3b‐h, 50% human blood se‐rum in water, and 30 minute reaction time. 18 serum. However, serum has a high salt concentration, which as noted above, can cause large increases in background ligation. As predicted, the strands 3b and 4b in solution phase ligation showed significant background ligation in 50% blood serum. Therefore, a series of mutations of various severities (3c‐h) were tested for their abilities to lower background ligation (Figure 1.10). Sequence 3f showed the most promise in lowering background ligation. In this sequence, an AT base pair is mutat‐ed to an AC mismatch causing a bulge in the middle of the duplex region and thus having a pronounced affect on binding affinity between aptamer fragments (Figure 1.8). Sequence 3f was then converted to sequence 1d to be compatible with the en‐zyme‐ linked assay, and an A10 linker was added. These se‐quences showed good dose de‐pendence in the range of 1 μM‐ 100 μM with the enzyme‐linked assay in 50% serum showing the successful development of a sen‐sitive assay for cocaine detection (Figure 1.11). CONCLUSIONS In this chapter, we have developed the first DNA based analog of a sandwich ELISA that is capable of cocaine detection via a standard colorimetric output. This is an important foundation in developing improved assays for small molecule detec‐ Figure 1.11: Enzyme‐linked assay showing compat‐ibility in 50% human blood serum using capture strand 1d. Conditions are 1 μM 2b, 75 mM NaCl, 50% human blood serum in water, and 30 minute reaction time. 19 tion. Reaction time was then improved from 20 hours to 30 minutes by switching ALO for the more reactive DIBAC giving an overall assay time of 2 hours. This assay was successfully tested with biological samples to demonstrate its usefulness in clinical applications. We also showed significant improvement in sensitivity by using StAPL to covalently trap the complex. For cocaine, the detection limit in serum is an order of magnitude better than other comparable split aptamer based assays. Antibody based assay have a bet‐ter limit of detection due to the inherent lower binding affinity of the cocaine split aptamer. However, the split aptamer based system has many distinct advantages over antibodies. Importantly, split aptamers have the capabilities to be much more selective for their targets. This becomes especially relevant for the long‐term goals of detecting and differentiating between opioid targets. The positive screening cut‐off for opioids when using antibodies is well within the sensitivity range of the DNA based analog, thus, the selectivity of split aptamers should allow for a marked im‐provement over antibodies in accurately detecting opioid targets. REFERENCES 1) Huizenga, D. E.; Szostak, J. W. Biochemistry 1995, 34, 656−665 2) Sharma, A. K.; Heemstra, J. M. J. Am. Chem. Soc. 2011, 133, 12426−12429 3) Kerrigan, S.; Phillips, W. H., Jr. Clin. Chem. 2001, 47, 540−547 4) Engvall, E.; Perlmann, P. Immunochemistry 1971, 8, 871−874 5) Agard, N. J.; Baskin, J. M.; Prescher, J. A.; Lo, A.; Bertozzi, C. R. ACS Chem. Biol. 2006, 1, 644−648 6) Stojanovic, M. N.; de Prada, P.; Landry, D. W. J. Am. Chem. Soc. 2000, 122, 11547−11548 7) Zhang, J.; Wang, L.; Pan, D.; Song, S.; Boey, F. Y. C.; Zhang, H.; Fan, C. Small 2008, 4, 1196−1200 20 Chapter 2: General Approach for Engineering Small‐Molecule‐Binding DNA Split Aptamers INTRODUCTION The application of StAPL technology to the detection of small molecule tar‐gets is a very promising avenue in the quest for the improvement of drug testing methods1. However, one major limitation of this system is its requirement for ap‐tamers that are capable of being split while still maintaining their ability to bind their targets2,3. There were only two known split aptamers at the start of this re‐search. Therefore, the generation of new split aptamers was a very important step in continuing to broaden their utility, specifically in detecting more targets. In order to pursue this, we began looking to generate new split aptamers from small molecule binding aptamers. In principle, split aptamers can be generated from the division of their ap‐tamer counterparts. Unfortunately, this proves problematic due to the structure the aptamer conforms to upon binding the small molecule. Many aptamers have sec‐ondary structures that resemble a hairpin4 where the small molecule binds to the nucleobases in the loop of the hairpin. This architecture is not very amenable to splitting because changes in the aptamer are more likely to disturb the binding pocket. We hypothesize that the three‐way junction ar‐chitecture adopted by the cocaine aptamer is a privileged structure for generating split ap‐tamers5. In this secondary structure the Figure 2.1: Splitting of hairpin or three‐way junction architectures. 21 small molecule binds some distance from the splitting site and is additionally stabi‐lized by the third stem loop6 (Figure 2.1). Herein we present a method to reliably generate DNA split aptamers from three‐way junction parent aptamers. Recently, Stojanovic and coworkers have presented a method of generating three‐way junction aptamers by beginning with a structurally biased DNA library during selections (Figure 2.2). This was specifically demonstrated by generating steroid binding aptamers for the targets DIS, DOG, DCA, and BE (Figure 2.3). Beginning from these sequences, we present a systematic method of splitting, truncating, and optimizing (Fig‐ure 2.4) in order to produce working split aptamers from those original aptamers. StAPL technology and polyacrylamide gel electrophoresis were then em‐ployed to investigate their assembly properties. In order to ensure that StAPL was Figure 2.2: Steroid binding aptamers generated by Stoja‐novic and associates using structurally biased SELEX. Figure 2.3: Steroid targets bound by aptamers in Figure 2. 22 not interfering with the binding properties of the split aptamer, studies were also done with a fluorophore‐quencher system. The selectivity for the targets was ob‐served to be very similar to the parent aptamers. Thus, the new split aptamers generated have the potential to be used for a wide variety of biosensing applica‐tions. Furthermore, other three‐way junction aptamers that may be generated would be able to be split for use in StAPL assays. METHODS All starting materials were ob‐tained from commercial suppliers and were used without further purification. Β‐Estradiol (BE), sodium deoxycholate (DCA), dehydroisoandrosterone‐3‐sulfate sodium salt dihydrate (DIS), and deoxy‐coticosterone 21‐glucoside (DOG) were purchased from Sigma Aldrich. Because DOG and BE were somewhat insoluble in water, a mixture of water/DMSO was used to increase solubility. Human blood se‐rum was purchased from Sigma Aldrich. Cyclooct‐1‐yn‐3‐glycolic acid (ALO) was synthesized according to the procedures detailed in Chapter 1. DNA was purchased from the University of Utah DNA/Peptide synthesis Core Facility. DNA was modified with ALO according to the procedure detailed in Chapter 1. Heating and cooling cy‐cles for fluorescence quenching studies were performed using a Bio Rad T100 Thermal Cycler. Fluorescence quenching was measured using a Biotek Synergy MX Figure 2.4: General scheme for generating split aptamers from their three‐way junction aptamer counterparts. 23 microplate reader. PAGE gels were analyzed Cy3 for fluorescence using the Typhoon FLA 9500 scanner from GE Healthcare Life Sciences with a 532 nm excitation laser and a BPG1 (570DF20) filter. High‐Performance Liquid Chromatography (HPLC) was conducted using a 1260 Infinity HPLC system from Agilent Technologies. Mass spectra were collected through the Mass Spectrometry Core Facility at the Universi‐ty of Utah. DNA Sequences The following sequences (Table 2.1) are the split aptamer fragments used in this chapter. Sequence 1‐ALO corresponds to the cyclooctyne functionalized strand for the DOGS.1 and DISS.1 split aptamers. 2‐ALO corresponds to the cyclooctyne functionalized sequence for the BES.1 split aptamer, and 3‐ALO corresponds to the cyclooctyne functionalized sequence for the DCAS.1 split aptamer. Sequences 4a‐c N3 correspond to azide functionalized strands with varying base changes associated with the DOGS.1 split aptamer. Sequences 5‐N3, 6‐N3, and 7‐N3 correspond to azide functionalized strands associated with DISS.1, BES.1, and DCAS.1 split aptamers re‐spectively. Sequence 1‐BHQ2 corresponds to a sequence with the dye, Black Hole Quencher‐2, attached to the 3' end of one fragment of the DOGS.1 split aptamer. 4‐ Cy3 is the other split aptamer fragment associated with DOGS.1 with the dye Cy3 attached to the 5' end of the sequence. 24 Label Sequences (5'‐3') 1‐ALO Cy3‐CGG GAC GTG GA‐ALO 2‐ALO Cy3‐CGG GAC GAC ATG GA‐ALO 3‐ALO Cy3‐CGG GAC GCT GGG‐ALO 4a‐N3 N3‐TCC ACA AAC CAG AAT GGT GTC C 4b‐N3 N3‐TCC ACA AAC CAG AAT GGT GTC CC 4c‐N3 N3‐TCC ACA AAC CAG AAT GGT GTC CCG 5‐N3 N3‐TCC GCA TAC GAA GTT GTC CC 6‐N3 N3‐TCC ATC AAC GAA GTG CGT CCG TCC C 7‐N3 N3‐CCC AGG ACG AAG TCC GTC CC 1‐BHQ2 CGG GAC GTG GA‐BHQ2 4‐Cy3 Cy3‐TCC ACA AAC CAG AAT GGT GTC CC Table 2.1: DNA Sequences Used in This Chapter General method for split aptamer ligation in buffer: First, a steroid solution at a specified concentration was added. This was fol‐lowed by the addition of Tris buffer (pH 8.2), NaCl (1 M), azide modified DNA (30 μM), and then ALO modified DNA (10 μM). The volume was adjusted to 20 μL with ultrapure water. The reactions were then incubated at room temperature for a spec‐ified amount of time, and quenched using 20 μL 2x PAGE loading buffer containing 7 M urea. They were then analyzed by denaturing PAGE on 8% TBE/urea polyacryla‐mide gels. Gel images were analyzed and ligation yields were calculated using Equation 2.1 where VR is the band for the reactants and VP is the band for the ligated products. %Yield=100[ Vp Vp+VR ] Equation 2.1: Calculation for ligation yields 25 General method for split aptamer ligation in biological media: Ligation reactions were carried out in either 50% serum or 50% urine. A steroid solution at a specified concentration was added. This was followed by the addition of serum or urine, NaCl (1 M), azide modified DNA (30 μM), and then ALO modified DNA (10 μM). The volume was adjusted to 20 μL with ultrapure water. The reactions were then incubated at room temperature for a specified amount of time, and quenched using 20 μL 2x PAGE loading buffer containing 7 M urea. They were then analyzed by denaturing PAGE on 8% TBE/urea polyacrylamide gels. Gel images were analyzed and ligation yields were calculated using Equation 1. General method for fluorescence measurements: Measurements were taken in 25 mM Tris, pH 8.2, 115 mM NaCl. The steroid solution was added first, and then the maximum fluorescence in the absence of BHQ2 was determined by adding 2 uL of 30 μM DNA 4‐Cy3 and 2 μL of ultrapure wa‐ter (in place of BHQ2 strand). Fluorescence quenching measurements were carried out by adding 2 uL of 30 μM DNA 4‐Cy3 and 2 μL of 30 μM DNA 1‐BHQ2. The total volume was 100 μL after the addition of steroid and DNA strands. Each reaction so‐lution and control was prepared in triplicate for estimation of error. The mixtures were incubated for 80 minutes and transferred into a Costar 96‐well non‐treated, opaque flat bottom, black polystyrene assay plate (Corning Incorporation). The flu‐orescence was measured with a 550 nm excitation filter and a 570 nm emission filter. The average relative fluorescence intensities were divided by the fluores‐cence intensity maximum of solutions without BHQ2. 26 RESULTS AND DISCUSSION Generating and optimizing a split aptamer for DOGS.1: As was shown with the cocaine split aptamer, one loop region of the aptamer was not present allowing for the addition of reactive groups. We envisioned a simi‐lar system with the steroid binding aptamers. Beginning with the DOGS.1 aptamer, we removed a loop from the end of the aptamer. One important enthalpic aspect of split aptamer assembly is generating a split aptamer that will have enough base complementarity to come together in the presence of the target, but not enough that it will have significant assembly without the target. This enthalpic balance can be fine‐tuned by variations in the length of the stem region. Ionic strength also impacts the ligation due to the reduced electrostatic repulsion between DNA strands at higher salt concentrations. Therefore, we began generating stem region varia‐tions for the DOGS.1 split ap‐tamer (Figure 2.5). Those stem variations were then tested with a range of salt concentrations and at 0 or 1000 μM DOG (Figure 2.6a). As can be seen, 1‐ALO with azide sequence 4b‐N3 shows the best ligation while main‐taining low off target assembly. Sequence 4a‐N3 showed very little ligation demonstrating that the enthalpic barrier for assembly with the target is too high. Conversely, sequence 4c‐N3 showed significant assembly without the target at mid‐dle to high salt concentrations. Next, the signal to background ratio for the three Figure 2.5: Stem length iterations of the DOGS.1 split aptamer. 27 sequences was investigated at different salt concentrations for continued optimization. A higher signal to background ratio is an important measure of the poten‐tial sensitivity of the split aptamer indicating a better de‐tection limit. The signal to background ratio chart was gen‐erated from the previous experiment. At a middle range salt concentration 4b‐N3 showed very good signal to background ratio. 4c‐N3 also showed a good signal to background ratio at very low salt concentrations (Figure 2.6b). Because in StAPL, assem‐bly of the aptamer generates a covalent linkage, more signal ac‐cumulates over time. In order to determine the effect of this phe‐nomenon on the newly generated DOGS.1 split aptamer, we began testing the time Figure 2.6: Conditions are 10% DMSO, 25 mM Tris, pH 8.2, 0.5 μM 1‐ALO, 2.0 μM 4a‐c N3, 80 minute ligation time. a.) Ligation yield at 0 and 1000 μM DOG. b.) Sig‐nal to background ratio of DOGS.1 split aptamer. 28 dependence of the ligation (Figure 2.7a). Again, a signal to background chart was generated (Figure 2.7b). As expected, more signal was ob‐served as the time of the ligation increased. Sequence 4b‐N3 again showed the highest signal to back‐ground ratio at 80 minutes and 115 mM NaCl. However, it should be not‐ed that sequence 4c‐N3 could also be used at lower salt concentrations and shorter times. The combination of the data obtained from this series of ex‐periments allowed us to determine the optimum split aptamer sequence and conditions for DOGS.1. Dose dependence and selectivity of split aptamers: Figure 2.7: Conditions are 10% DMSO, 25 mM Tris, pH 8.2, 115 mM NaCl, 0.5 μM 1‐ALO, 2.0 μM 4a‐c N3. a.) Yield of ligated product as a function of in‐creasing time. b.) Signal to background ratio of DOGS.1 as a function of time. 29 Having split and opti‐mized the DOGS.1 aptamer, we next wanted to determine the selectivity for varying tar‐get steroids. The parent aptamer binds to the ligands DOG and DIS and binds DCA and BE weakly. Thus in order to determine the selectivity of the newly generated split ap‐tamer; we performed a dose dependent ligation with each of the ligands (Figure 2.8). We found that the split ap‐tamer bound the steroid targets with very similar se‐lectivity as the parent aptamer, confirming our suc‐cessful generation of a working split aptamer. We then tested to see if this method was generalizable over the other three aptamers generated by Stojanovic. After op‐timization and dose dependent ligation with the steroid targets, we showed the Figure 2.8: Conditions are 10% DMS (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 minutes. a.) PAGE gel of reactions. Lower bands are 1‐ALO, and upper bands are 1‐ALO:4b‐N3. b.) Yield of ligated product for 1‐ALO:4b‐N3. 30 successful generation of split aptamers from the three other aptamers (Figure 2.9). Non‐covalent assembly of DOGS.1: Next, we wanted to confirm that the ligation step was not affecting the selec‐tivity or binding properties of the newly generated split aptamers. In order to test this, we used a version of the DOGS.1 split aptamer that did not include the azide or cyclooctyne (1‐BHQ2:4b‐Cy3). To measure a binding event, a quencher was attached to one aptamer fragment (Figure 2.10a). When the DNA strands are not assembled in the absence of the target, a large fluorescence signal is measured from Cy3. How‐ever, upon addition of the target and assembly of the split aptamer, Black Hole Quencher 2 (BHQ2) quenches Cy3 fluorescence. Using this system to measure dose dependent ligand binding, we saw that it mirrored the trends in selectivity and Figure 2.9: Conditions are 10% DMSO (DIS, DOG, DCA), 40% DMSO (BE), 25 mM Tris, pH 8.2, 0.5 μM 1‐ALO, and 2.0 μM 5‐7 N3. a.) Sequences of split aptamers DISS.1, BES.1, and DCAS.1 split ap‐tamers. b‐d.) Yield of ligated product as a function of ligand concentration. 31 binding seen using StAPL (Figure 2.10b). Therefore, we concluded that StAPL is an accurate way of assessing assembly properties of the split aptamers. Detection in biological fluids: Because the eventual goal for generating split aptamers is for applications in clinical diagnostics, we wanted to test the relevance of this assay in biological fluids such as serum and urine. Unfortunately, DOG is degraded over time in bio‐logical fluids as it undergoes hy‐drolysis of the glycosidic bond. Because of this fact, it was neces‐sary to lower the ligation time in order to have minimal target degradation. To do this, we first increased the ionic strength of the reaction to drive the ligation to occur faster. In serum, we used the sequences 1‐ALO:4c‐N3 for its extended stem region while continuing to use 1‐ ALO:4b‐N3 for urine experiments. As shown, the presence of 50% serum or 50% urine still allows for dose dependent ligation (Figure 2.11). Due to the shorter liga‐tion time, the signal to background ratio is decreased, raising the detection limit to Figure 2.10: Conditions are 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. a.) Depiction of aptamer assembly with fluorophore quencher pair Cy3:BHQ2. b.) Measured fluorescence over maximum fluorescence as a function of ligand con‐centration. 32 10 μM. This is only a slight decrease in sensitivity, encouraging our view that these newly generated split aptamers could be clinically relevant. CONCLUSIONS A wide range of factors was considered in the quest to generate new split ap‐tamers. After the initial splitting, the first factor explored was truncation. We found that the length of the stem region had a significant effect on the assembly of the split aptamer. This alteration is very sensitive, where a single base pairing interaction can cause a substantial change in the ease of assembly. Ionic strength also af‐fected split aptamer assembly. A higher ionic strength correspond‐ed to a higher entropic favorability for the assembly of the split aptamers. This allowed us to use salt concentration as tunable factor for split aptamer assembly. Finally, we explored the time dependence of the reac‐tion. Because we are generating a covalent linkage, there is a natural accumulation of signal over time. This allowed us to determine optimum reaction times for different stem regions and salt concentra‐ Figure 2.11: Conditions are 10% DMSO, 50% human blood serum or artificial urine, 25 mM Tris, pH 8.2, 215 mM NaCl (buffer), 100 mM NaCl (serum or urine), 0.5μM 1‐ALO, 2.0 μM 4b‐cN3, 20 minutes. Ligation yield as a function of DOG concentration in biological fluids. 33 tions to maximize the signal to background ratio. After the optimization, we con‐firmed that the ligation step was not causing a major perturbation in aptamer assembly by using a fluorophore‐quencher system. Lastly, we demonstrated the usefulness of the newly generated split aptamers in biological fluids. This could al‐low them to have clinical applications in the future. We have successfully generated four new steroid binding split aptamers from their parent aptamers. The three‐way junction architecture is a preferred structure for splitting over its more common counterpart, the hairpin. This allowed us to stra‐tegically divide, truncate, and optimize the split aptamers, establishing a generalized method for generating split aptamers from three‐way junction aptamers. These mir‐ror the selectivity and functionality of their parent aptamers. Because there were only two known split aptamers at the start of this research, we have tripled the number of known split aptamers. We anticipate that the ability to generate split ap‐tamers will continue to advance the scope of small‐molecule detection by DNA based biosensors. REFERENCES 1.) Sharma, A. K.; Kent, A. D.; Heemstra, J. M. Anal. Chem. 2012, 84(14), 6104‐ 6109 2.) Huizenga, D. E.; Szostak, J. W. Biochemistry 1995, 34, 656−665 3.) Stojanovic, M. N.; de Prada, P.; Landry, D. W. J. Am. Chem. Soc. 2000, 122, 11547−11548 4.) Uphoff, K. W.; Bell, S. D.; Ellington, A. D. Curr. Opin. Struct. Biol. 1996, 6, 281−288 5.) Stojanovic, M. N.; de, P. P.; Landry, D. W. J. Am. Chem. Soc. 2001, 123, 4928−4931 6.) Neves, M. A. D.; Reinstein, O.; Johnson, P. E. Biochemistry 2010, 49, 8478−8487 7.) Yang, K. A.; Pei, R.; Stefanovic, D.; Stojanovic, M. N. J. Am. Chem. Soc. 2012, 134, 1642−1647 34 CHAPTER 3: Coupling Split Aptamer Ligation with qPCR for the Semi‐Automatable Detection of Steroid Targets INTRODUCTION As we have demonstrated, StAPL technology using an azide and a cy‐clooctyne as reactive groups is a promising alternative to antibodies. StAPL has been combined with both colorimetric techniques and gel electrophoresis to detect small molecule targets1,2. However, this technique does not allow for real time detection and quantification. It can also be labor intensive, making it non ideal for a clinical setting. To remedy this, herein we propose a method of detection analogous to a Proximity Ligation Assay (PLA) by coupling split aptamer ligation to quantitative polymerase chain reaction (qPCR). PLA is an antibody based method of detection for antigens. It primarily uti‐lizes PLA probes, which are conjugated to short oligonucleotides that are then joined via enzymatic ligation and can undergo rolling circle amplification. Comple‐mentary oligonucleotide probes are then added and the product can be detected through fluorescence intensity3. However, as previously stated, antibodies are ex‐pensive and have various limitations due to their in vivo generation. Therefore, replacement with nucleic acids would be advantageous. In qPCR, primers bind the DNA and polymerases are used to amplify it through a series of heating and cooling steps. This results in exponential amplifica‐tion of the sequences. The progress of this amplification can be detected by using SYBR Green; a double stranded DNA intercalating fluorescent dye. The resulting flu‐orescent signal is directly proportional to sample concentration during the 35 exponential phase of amplification4. Because this method can be semi automatable, we wanted to apply it to our split aptamer ligation technology. In this iteration, the sequence would only be amplified if the full aptamer product is formed from the two fragments after ligation in the presence of the small molecule. However, the current cycloaddition adduct cannot be read through by polymerases. In order to develop a linkage that was PCR readable, we looked toward link‐ages that mimicked the phosphodiester bonds of native DNA. Nicholas Hud and associates report that a morpholino linkage has this quality and is readable with specific thermophilic DNA polymerases5 (Figure 3.1). Therefore, we began looking to adapt this linkage to split aptamer ligation. In this system, an amine is added to Figure 3.1: Reaction scheme for the formation of a morpholino linkage. Each sugar is attached to a DNA sequence. 36 the 5' end of one aptamer fragment and a ribonucleotide is added to the 3' end of the other aptamer fragment. Upon assembly in the presence of the small molecule target and the addition of sodium periodate, the hydroxyl groups on the ribonucleotide are oxidized to aldehydes that react with the amine. With the addition of sodium cyanoborohydride, the newly formed dihydroxymorpholine is reduced forming a covalent link between the aptamer frag‐ments. We chose to use the BES.1 split aptamer (Figure 3.2) and the target molecule β‐estradiol (BE) due to the good signal to background ratio. The BES.1 ligated split aptamer does not have quite enough nucleobases to be amplified on its own. To remedy this, a primer binding region was added to the 5' end of the ribonucleotide modified strand. With this modi‐fication, the sequence can be amplified with PCR. We found that there was amplification of the full length sequence in the presence of BE making this method another potential system for small molecule detection. METHODS All starting materials were obtained from commercial suppliers and were used without further purification. Β‐Estradiol (BE), sodium periodate, and sodium cyanoborohydride were purchased from Sigma Aldrich. Because BE was somewhat insoluble in water, it was first dissolved in DMSO before it was diluted to 10% DMSO when added to the reaction mixture. Reagents necessary for PCR as well as DNA polymerases were purchased from New England BioLabs. DNA with all necessary modifications was purchased from the University of Utah DNA/Peptide synthesis Figure 3.2: Split aptamer sequence used in this chap‐ter. X represents an amine and Y represents a ribonu‐ 37 Core Facility. PAGE gels were analyzed Cy3 for fluorescence using the Typhoon FLA 9500 scanner from GE Healthcare Life Sciences with a 532 nm excitation laser and a BPG1 (570DF20) filter. PCR reactions were heated and cooled using a Bio Rad T100 Thermal Cycler. DNA Sequences: The following sequences (Table 3.1) are the DNA strands used in this chap‐ter. Sequences Cy3‐S1 and NH2‐S1 are the split aptamer fragments corresponding to the BES.1 split aptamer with one strand modified with an amine and the other modi‐fied with a uracil ribonucleotide. Cy3‐P‐S1 and NH2‐P‐S1 correspond to the first iterations of the BES.1 split aptamer with inserted primer binding sites. Sequences NH2‐P‐S2‐(1‐6) correspond to six different primer binding regions inserted into the original NH2‐S1 amine sequence. Sequences FP and RP‐6 correspond to the forward primer and reverse primer for use in PCR with sequences Cy3‐S1:NH2‐P‐S2‐6 and PCR‐FS‐C. Finally, PCR‐FS‐C is the full sequence of Cy3‐S1:NH2‐P‐S2‐6 with the mor‐pholino linkage replaced with a thymine to be used as a control in PCR. 38 Label Sequences (5'‐3') Cy3‐S1 Cy3‐CGG GAC GAC ATG GAT‐U(ribo) Cy3‐P‐S1 Cy3‐GGA GGC ACC ACG GCT GGA TCC CGG GAC GAG ATG GAT‐U(ribo) NH2‐S1 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC NH2‐P‐S1 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC CCT TGG TCA TTA GGA TCG NH2‐P‐S2‐1 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC GAG AAT TCC GAC CAG AAG NH2‐P‐S2‐2 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC AGC AGC ACA GAG GTC AGA TG NH2‐P‐S2‐3 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC GAC GGA ATA TAA GCT GGT GG NH2‐P‐S2‐4 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC CTC CTC TGA CTG TAA CCA CG NH2‐P‐S2‐5 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC CGA AGT CGC CAT CTC TTC NH2‐P‐S2‐6 NH2‐TTT CCA TCA ACG AAG TGC GTC CGT CCC ATA CCA GCT TAT TCA ATT FP CGG GAC GTG GAT TTT TCC RP‐6 AAT TGA ATA AGC TGG TAT GGG PCR‐FS‐C CGG GAC GTG GAT TTT TCC ATC AAC GAA TGC CGT GCG TCC CAT ACC AGC TTA TTC AAT T Table 3.1: DNA Sequences Used in This Chapter General method for morpholino forming ligation reaction: First, TAPS buffer (30 mM, pH 7.6) was added to a PCR tube followed by NaCl (1 M), Cy3 labeled DNA (5μM), and amine DNA (20 μM). Then BE dissolved in DMSO was added to equal a specified concentration. After the addition of sodium periodate (60 mM), the mixture was allowed to react at room temperature for 30 minutes. Then sodium cyanoborohydride (0.5M) was added and the reaction was allowed to proceed at room temperature for an hour and a half. The reaction was quenched with 1.25 μL of 200 mM ethylene diamine. The reaction mixtures were then ana‐lyzed by denaturing PAGE on 12% TBE/urea polyacrylamide gels. Gel images were analyzed and ligation yields were calculated using Equation 3.1 where VR is the band for the reactants and VP is the band for the ligated products. 39 %Yield=100[ Vp Vp+VR ] Equation 3.1: Calculation for ligation yields General method for PCR of morpholino linkage: After quenching of the reaction with ethylene diamine, the reaction mixture is diluted 100 fold before being carried on to PCR. For each reaction, 10x reaction buffer, dNTPs (10 mM), sequence FP (10 μM), sequence RP‐6 (10 μM), and diluted reaction mixture is added to a PCR tube. Lastly, DeepVentR (DVR) polymerase is added. A series of heating and cooling reactions are then carried out (Table 3.2). Temperature (°C) # Of Cycles Time (s) 85 1 30 95 30 45 30 45 68 30 68 1 300 4 Hold Infinite Table 3.2: Details of PCR heat/cool cycles. Bolded rows represent three steps that are repeated. Then 20 μL 2x PAGE loading buffer containing 7 M urea is added. The reaction mix‐tures were analyzed by denaturing PAGE on 12% TBE/urea polyacrylamide gels. The gels were then shaken with diluted ethidium bromide solution for 15 minutes. The resulting bands were then visualized with UV light. RESULTS AND DISCUSSION Ligation using a morpholino linkage with primer binding regions The first challenge in developing this method was to switch the linkage from a cycloaddition reaction to a morpholino linkage. Not only was the linkage being changed, but we also needed to add a primer binding region to split aptamer frag‐ 40 ments. The ligated split aptamer on its own is not long enough to be amplified using PCR. Therefore, we first began using a known primer binding region, and we modified both split aptamer fragments with a primer binding sequence. We per‐formed the ligation reactions with var‐ying combinations of the first four se‐quences in Table 3.1 (Figure 3.3). Sequences Cy3‐S1 and NH2‐S1 without primer binding re‐gions showed high background ligation without a large increase in the presence of BE. While sequences Cy3‐P‐S1 and NH2‐P‐S1 showed very little ligation reaction. We hypothesize this is due to the extra bases on the end of both sequences are interacting with the split aptamer to form secondary structures that do not bind the target BE as well. How‐ever, allowing only one split aptamer fragment to have a primer binding region shows the best ligation yield. Specifically, having the primer binding region on the amine labeled strand gave the most promising result (Cy3‐S1:NH2‐P‐S1). Next, we investigated other primer binding sequences to couple with the amine strand. We chose six other sequences as potential candidates to continue im‐proving ligation yields while maintaining low background (Figure 3.4). Of those Figure 3.3: Ligation yields for morpholino linkage demonstrating the effect of primer binding regions. Conditions are 20 μL total volume, 0.5 μM Cy3‐DNA, 2.0 μM NH2‐DNA, 30 mM TAPS, pH 7.6, 80 mM NaCl, 10% DMSO, 3 mM NaIO4 for 30 minutes, and 100 mM NaCNBH3 for 1.5 hours. 41 candidates, pair Cy3‐1:NH2‐P‐S2‐6 showed the best signal to background ratio. Therefore, we chose to continue using it through further studies. PCR of full length sequence with morpholino linkage: Nicholas Hud and associates reported that certain thermophilic DNA poly‐merases are capable of reading through the non native morpholino linkage. The most characterized of these is DVR. Therefore, we chose to use it as the polymerase for these reactions. Upon staining the PAGE gel containing the PCR reactions with ethidium bromide, we were able to see amplification of the full length sequence when the target was present, but no amplification of the full length sequence in the absence of the target (Figure 3.5). Thus, we can successfully translate StAPL to a PCR amplifiable system. CONCLUSIONS In this chapter, we demonstrated a method that adapts a split aptamer sys‐tem for use with PCR. We utilized the BES.1 split aptamer and successfully detected BE using PCR. In order to do this, we explored the effect of primer binding region on Figure 3.4: The effect of changing the sequence of the primer binding region of the amine functionalized sequence. Conditions are 20 μL total volume, 0.5 μM Cy3‐DNA, 2.0 μM NH2‐DNA, 30 mM TAPS, pH 7.6, 80 mM NaCl, 10% DMSO, 3 mM NaIO4 for 30 minutes, and 100 mM NaCNBH3 for 1.5 hours. 42 ligation. When neither split aptamer sequence con‐tains a primer binding region, the background liga‐tion is high. However, when both contain primer binding regions, there is very little ligation seen, even at high concentrations of ligand. We hypothe‐sized that this was due to secondary structures formed by the aptamer fragments, which were in‐hibiting ligand binding. Having only one primer binding region on the amine modified split aptamer fragment showed the best signal to background ratio. Therefore, we chose to use this iteration of poten‐tial primer binding combinations. Then, the effect of dif‐ferent sequences that comprised the primer binding sequence was tested allowing us to find a better primer binding region. We then optimized the chosen sequence for use with PCR. Upon completing PCR, we showed that the full length ligated split aptamer sequence was amplifiable. This is the first step in creating a semi automatable system of detection for small molecule targets. Further research focuses on performing qPCR and de‐termining the dose dependence of the system. Figure 3.5: Ligation reactions im‐aged with Cy3 (above) and corresponding PCR reactions im‐aged with ethidium bromide (below). C corresponds to the control sequence, PCR‐FS‐C. Liga‐tion conditions are 20 μL total volume, 0.5 μM Cy3‐DNA, 2.0 μM NH2‐DNA, 30 mM TAPS, pH 7.6, 80 mM NaCl, 10% DMSO, 3 mM NaIO4 for 30 minutes, and 100 mM NaCNBH3 for 1.5 hours. PCR conditions are detailed in meth‐ods. 43 REFERENCES 1.) Sharma, A.; Heemstra, J. J. Am. Chem. Soc. 2011, 133 (32), 12426‐12429 2.) Sharma, A. K.; Kent, D. A.; Heemstra, J. M. Anal. Chem. 2012, 84, 6104‐6109 3.) Gullberg, M. et al. PNAS, 2004, 22, 8420‐8424 4.) Sigma Aldrich. (2008). qPCR Technical Guide. Retrieved from: 􀀼<http://www.sigmaaldrich.com/etc/medialib/docs/Sigma/General_Inform ation/qpcr_t 􀁥echnical_guide.Par.0001.File.tmp/qpcr_technical_guide.pdf> 5.) Engelhart, A. E.; Cafferty, B. J.; Okafor, D. C.; Chen, M. C.; Williams, L. D.; Lynn, D. G.; Hud, N. V. Chem. Bio. Chem. 2012, 13, 1121‐124 44 FINAL REMARKS This research demonstrates a contribution towards the development of split aptamers as DNA based biosensors. Split aptamers have numerous advantages over their antibody counterparts. They are more selective, can be generated for more targets, and are less expensive to produce. We have explored some aspects of split aptamers and how they can be adapted for use in a clinical setting. The development of a microplate assay was the first step in this endeavor. With its creation, we opened the doors to even more possibilities for detection. To broaden the scope of use, we then worked to generate split aptamers for more targets. Developing a method of easily generating split aptamers from three‐way junction aptamers not only tripled the number of known split aptamers, but established the tools to pro‐duce even more. Finally, we made strides toward building a semi automatable system of detection utilizing a morpholino linkage and qPCR. Further research seeks to continue developing this system and exploring its range of potential applications. In conclusion, we have made progress toward improving current drug detection methods through StAPL technology. ACKNOWLEDGEMENTS Thanks are given to Dr. Jennifer Heemstra and the Heemstra research group for their support throughout the years during which this research was completed. Spe‐cial acknowledgement is given to Dr. Ashwani Sharma and Nicholas Spiropulos for their assistance with the experiments detailed in Chapters 1 and 2 respectively. This research was supported through the National Science Foundation, the Army Re‐search Office, the University of Utah, and the Undergraduate Research Program.