Communication pubs.acs.org/JACS
In Vitro Selection of Diversely Functionalized Aptamers Dehui Kong,† Wayland Yeung,† and Ryan Hili*,†,‡ †
Department of Chemistry, University of Georgia, 140 Cedar Street, Athens, Georgia 30602, United States Department Chemistry, York University, 4700 Keele Street, Toronto, ON M3J 1P3, Canada
‡
S Supporting Information *
function, including improved binding properties13a and divalent metal-independent nuclease activity.13b,e The most widely used method for incorporating modifications throughout a nucleic acid polymer is to perform templated primer extension using a polymerase and modified dNTPs. This essentially limits the number of unique modifications to the number of nucleobases used in the genetic code. Thus, if a standard four letter genetic code is used, a maximum of four modifications could be incorporated. Further increasing the diversity of nucleic acid polymers should provide a basis for Darwinian evolution of expanded molecular function. Our group recently developed the Ligase-catalyzed OligOnucleotide PolymERization (LOOPER) method to access this class of sequence-defined synthetic nucleic acids (Figure 1).14
ABSTRACT: We describe the application of T4 DNA ligase-catalyzed DNA templated oligonucleotide polymerization toward the evolution of a diversely functionalized nucleic acid aptamer for human α-thrombin. Using a 256membered ANNNN comonomer library comprising 16 sublibraries modified with different functional groups, a highly functionalized aptamer for thrombin was raised with a dissociation constant of 1.6 nM. The aptamer was found to be selective for thrombin and required the modifications for binding affinity. This study demonstrates the most differentially functionalized nucleic acid aptamer discovered by in vitro selection and should enable the future exploration of functional group dependence during the evolution of nucleic acid polymer activity.
C
hemical diversity increases molecular function.1 Thus, it is not surprising that nature places proteins at the apogee of cellular control. With 20 standard amino acids, and numerous post-translational modifications,2 molecular evolution can finely tune the functions of proteins for a vast array of important biological roles within the realm of structure, molecular recognition, and catalysis. Conversely, nucleic acid polymers have only four canonical base pairs consisting of comparatively limited functional groups. This dearth of chemical diversity has minimized its role beyond information storage and transfer. Nucleic acid polymers do have a specific advantage over their proteinogenic counterparts. Since they simultaneously represent both phenotype and genotype, this enables their rapid evolution in the laboratory through iterative cycles of in vitro selection, also termed SELEX.3 Furthermore, they have key structural advantages, including a high propensity to fold via intramolecular hybridization and the ability to reproducibly renature, making them more suitable for long-term storage and enable reproducible binding studies.4 These physical properties have captured the imagination of chemists who have sought out ways of expanding the chemical diversity of nucleic acid polymers in an attempt to enable access to higher molecular function.5−9 Indeed, expanding the chemical space of nucleic acid polymers by including the homomultivalent display of a hydrophobic functional group has increased the binding affinity and lowered kinetic off-rates of nucleic acid aptamers raised against various proteins10a,b and cell targets.10c Furthermore, the homomultivalent display of various functional groups has increased the catalytic potential of nucleic acid polymers for ribonuclease activity11 and protease activity.12 Expanding DNA into a heteromultivalent polymer comprising two or three functional groups has enabled the evolution of higher molecular © 2017 American Chemical Society
Figure 1. General LOOPER process for the generation of modified DNA.
The method relies upon the T4 DNA ligase-catalyzed DNAtemplated copolymerization of a library of modified 5′phosphorylated pentanucleotides. As the method employs codons, rather than single nucleotides, the number of modifications that can be incorporated throughout a nucleic acid polymer depends on the codon set size. We previously optimized LOOPER with a 256-membered codon set defined as NNNNT, where N = A, C, G, or T. The corresponding ANNNN comonomer library was divided into 16 sublibraries, each comprising 16 co-monomers (Figure 2). This enabled the sequence-defined incorporation of 16 different functional groups throughout a nucleic acid polymer in a library format. The process proceeded with high efficiency and with 94% fidelity.14d The 16-membered library comprised desired functional groups for molecular recognition, including aliphatic and aromatic groups, as well as Brønsted acids and bases. With a method to equip nucleic acids with protein-like chemical diversity, we sought to determine if these diverse modifications would increase binding properties or open access to different binding motifs. To explore these concepts, we Received: July 11, 2017 Published: September 22, 2017 13977
DOI: 10.1021/jacs.7b07241 J. Am. Chem. Soc. 2017, 139, 13977−13980
Communication
Journal of the American Chemical Society
mM KCl, 1 mM MgCl2, and 1 mM CaCl2. The library was then subjected to positive selection pressure, which involved incubation with sepharose-immobilized human α-thrombin beads for 15 min at room temperature. After the beads were washed three times with selection buffer, binders were eluted by denaturation at 90 °C for 4 min using a buffer comprising 40 mM Tris−HCl pH8, 10 mM EDTA, 3.5 M urea, and 0.02% TWEEN-20. Eluted aptamers were desalted and subjected to amplification. PCR was performed using KOD polymerase, a family B polymerase. These polymerases have been shown to accommodate bulky modifications in the major groove,6 which are present in our system. Templates for subsequent rounds of LOOPER/SELEX were generated by streptavidin-mediated strand separation and gel purification. Iterative rounds of LOOPER/SELEX were performed until library diversity was sufficiently decreased (six rounds), determined using qPCR thermal melting.15 The surviving library members were subjected to highthroughput paired-end DNA sequencing and analyzed for frequency and consensus sequences (Tables S1 and S3 and Figures S2−S8). A clear consensus sequence was revealed, which consisted of a C-rich motif. The top five sequences (Table S1), which together represent 3% of the sequence reads, show the motif comprising four instances of cytosine pairs interspaced by two or three nucleotides. All lead sequences also exhibited a stem-loop motif within the reading frame, whereby the conserved sequence was displayed within the loop region and the stem region was largely variable (Figure S2). For the purposes of this study, we moved forward with the most highly represented aptamer sequence from the library (Figure 4a). We
Figure 2. 256-membered ANNNN library was subdivided into a 16 × 16 library for the generation of heteromultivalent DNA by LOOPER. Inset: structure of modified dA.
sought to evolve a modified aptamer against human αthrombin, which has generated many well-studied nucleic acid aptamers. To evolve these modified aptamers, we ported our LOOPER system into a SELEX cycle for molecular recognition (Figure 3). To synthesize a modified aptamer
Figure 3. In vitro selection cycle of highly functionalized nucleic acid polymers for molecular recognition of human α-thrombin.
Figure 4. Isolated modified thrombin aptamer and SPR analysis. (a) Predicted secondary structure of thrombin aptamer with modifications at dA shown. (b) Single-cycle kinetics SPR of modified thrombin aptamer for thrombin binding. (c) Single-cycle kinetics SPR of unmodified reverse complement for thrombin binding.
library, we used 120 pmol of a 5′-biotinylated ssDNA template library, comprising two 18-nt primer regions flanking an eightcodon encoding region derived from the NNNNT codon set. With our previously validated14d comonomer library (Figure 2), we used LOOPER to generate the functionalized nucleic acid polymer library. Strand separation and isolation of the functionalized ssDNA library from its biotinylated templates was achieved by using streptavidin-coated magnetic particles. The modified ssDNA library was then renatured in selection buffer comprising 20 mM Tris−HCl pH 7, 140 mM NaCl, 5
readily identified that the consensus sequence highly represented throughout the postselection library was the reverse complement of a G-quadruplex. G-quadruplexes are known to bind thrombin and are the primary sequence motif isolated from in vitro selections of ssDNA libraries for thrombin.16,17 Concerned that trace amounts of the Gquadruplex-containing template was contaminating the LOOPER strand library following strand separation, which could also result in convergence on the observed consensus sequence, we sought to study both strands for their binding affinity for thrombin. 13978
DOI: 10.1021/jacs.7b07241 J. Am. Chem. Soc. 2017, 139, 13977−13980
Communication
Journal of the American Chemical Society
dissociation constant of 1.6 nM. The modifications were required for binding and did not result in the nucleic acid polymer having nonspecific binding to proteins. Moreover, the secondary structure of the modified aptamer was beyond the traditional G-quadruplex motifs seen with known unmodified thrombin DNA aptamers. Remarkably, the reverse complement of the aptamer also bound thrombin; further studies are needed to determine if there is any synergistic role in cosurvival. Nonetheless, to the best of our knowledge, this is the first case where both the genotype and phenotype have the same function. Given our limited knowledge concerning the influence of chemical modifications on the co-operative network of weak interactions that comprise a functional aptamer, we anticipate that this platform will enable the discovery of new molecular function for nucleic acid polymers. Furthermore, increasing the chemical diversity of nucleic acid polymers may assist in developing novel aptamers against previously refractory targets or enabling higher selectivity of molecular recognition.
The highest frequency modified DNA sequence (Figure 4a) was synthesized by LOOPER and characterized by MALDI MS (Figure S26). Preliminary binding analysis by label-free microscale thermophoresis showed a promising dissociation constant Kd = 4 nM for thrombin, while nonspecific binding to bovine serum albumin (BSA) was not observed (Figure S3). Using single-cycle kinetics surface plasmon resonance (SPR),18,19 the LOOPER aptamer exhibited a Kd of 1.6 nM (Figure 4b), with an association rate constant kon = 1.93 × 106 M−1 s−1 and a dissociation rate constant koff = 3.03 × 10−3 s−1. Curious to determine if the modifications played a critical role in binding, we prepared the unmodified version of the aptamer and analyzed thrombin binding by single-cycle kinetics SPR. Interestingly, the aptamer exhibited no observable affinity (Figure S5), highlighting that the nucleic acid component was not purely responsible for binding, but rather these modifications play an essential role in molecular recognition of thrombin. Importantly, the LOOPER aptamer exhibited no observable binding to BSA by SPR (Figure S5), which suggests that the modifications did not result in undesired nonspecific binding to proteins. Consistent with other G-quadruplexes, the reverse complement of the LOOPER strand (GATTCGCCTGCCGTCGCAAGCATCGAATAGGATTGGATAGGTTGGAATTAGATGCAGTCACGTGGAGCTCGGATCC) also exhibited binding to thrombin, albeit weaker, with a Kd = 307 nM. Taken together, these data confirm that LOOPER can be ported into a SELEX cycle for the molecular evolution of modified nucleic acid polymers for molecular recognition. Indeed, four of the five top postselection sequences demonstrate detectable levels of thrombin binding (Table S4 and Figure S6). Since reported DNA aptamers for thrombin contain a Gquadruplex motif, the isolated modified DNA aptamer from this study represents a novel binding motif for thrombin recognition. Examining the conserved loop region, which spans codons 3−7 within the aptamer sequences, reveals that acidic modifications (ANNCT: COOH) as well as hydrophobic groups (ANNCA: phenyl and ANNCC: cyclopentyl) comprise approximately 50% of the modifications present in the postselection sequences (Table S3). Interestingly, the Cterminal tail of hirudin, DFEEIPEEYL, is known to have a strong interaction with the fibrinogen-binding exosite of thrombin. The abundant acidic residues in this region of hirudin are known to engage in electrostatic interactions with basic residues residing within the long groove that extends off the active site cleft of thrombin.20 Furthermore, the C-terminus also engages in hydrophobic interactions within the same groove. In light of these facts, the high frequency members of our postselection library may exhibit similar interactions with thrombin. It is important to note that, despite having a codon enrichment bias of 0.70,21 the ANNCT pentanucleotide building block, which is modified with COOH, is highly enriched in the postselection library. This important finding suggests that selections using LOOPER will not favor only those codons that are enriched during the LOOPER process. Future studies surrounding the mutational analysis of the modified aptamer and the location of the binding site on thrombin are required to further illuminate the molecular interactions between this modified aptamer and thrombin. In summary, we have developed an in vitro selection system for the generation of aptamers with diverse chemical functionality. We demonstrated the utility of this system by evolving a modified aptamer against human α-thrombin with a
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b07241. Supporting data, experimental methods, Tables S1−S4, and Figures S1−S30 (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*
[email protected] ORCID
Dehui Kong: 0000-0002-5906-8814 Ryan Hili: 0000-0002-8933-970X Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the National Institutes of Health (R21CA207711) and the Office for the Vice President of Research, University of Georgia. We thank the Georgia Genomics Facility for DNA sequencing services and Dr. Dennis Philips from the Proteomics and Mass Spectrometry core facility at the University of Georgia for their help in the characterization of oligonucleotides.
■
REFERENCES
(1) Lutz, J. F.; Ouchi, M.; Liu, D. R.; Sawamoto, M. Science 2013, 341, 1238149. (2) Prabakaran, S.; Lippens, G.; Steen, H.; Gunawardena, J. WIREs Syst, Biol, Med. 2012, 4, 565−583. (3) (a) Tuerk, C.; Gold, L. Science 1990, 249, 505−510. (b) Ellington, A. D.; Szostak, J. W. Nature 1990, 346, 818−822. (4) Keefe, A. D.; Pai, S.; Ellington, A. Nat. Rev. Drug Discovery 2010, 9, 537−550. (5) Malyshev, D. A.; Romesberg, F. E. Angew. Chem., Int. Ed. 2015, 54, 11930−11944. (6) (a) Hottin, A.; Marx, A. Acc. Chem. Res. 2016, 49, 418−427. (b) Sawai, H.; Ozaki, A. N.; Satoh, F.; Ohbayashi, T.; Masud, M. M.; Ozaki, H. Chem. Commun. 2001, 2604−2605. (c) Kuwahara, M.; Nagashima, J.; Hasegawa, M.; Tamura, T.; Kitagata, R.; Hanawa, K.; Hososhima, S.; Kasamatsu, T.; Ozaki, H.; Sawai, H. Nucleic Acids Res. 2006, 34, 5383−5394. (7) Hocek, M. J. Org. Chem. 2014, 79, 9914−9921. 13979
DOI: 10.1021/jacs.7b07241 J. Am. Chem. Soc. 2017, 139, 13977−13980
Communication
Journal of the American Chemical Society (8) Kong, D.; Yeung, W.; Hili, R. ACS Comb. Sci. 2016, 18, 355−370. (9) Meek, K. N.; Rangel, A. E.; Heemstra, J. M. Methods 2016, 106, 29−36. (10) (a) Rohloff, J. C.; Gelinas, A. D.; Jarvis, T. C.; Ochsner, U. A.; Schneider, D. J.; Gold, L.; Janjic, N. Mol. Ther.–Nucleic Acids 2014, 3, No. e201. (b) Davies, D. R.; Gelinas, A. D.; Zhang, C.; Rohloff, J. C.; Carter, J. D.; O’Connell, D.; Waugh, S. M.; Wolk, S. K.; Mayfield, W. S.; Burgin, A. B.; Edwards, T. E.; Stewart, L. J.; Gold, L.; Janjic, N.; Jarvis, T. C. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 19971−19976. (c) Renders, M.; Miller, E.; Lam, C. H.; Perrin, D. M. Org. Biomol. Chem. 2017, 15, 1980−1989. (11) Santoro, S. W.; Joyce, G. F.; Sakthivel, K.; Gramatikova, S.; Barbas, C. F. J. J. Am. Chem. Soc. 2000, 122, 2433−2439. (12) Zhou, C.; Avins, J. L.; Klauser, P. C.; Brandsen, B. M.; Lee, Y.; Silverman, S. K. J. J. Am. Chem. Soc. 2016, 138, 2106−2109. (13) (a) Gawande, B. N.; Rohloff, J. C.; Carter, J. D.; von Carlowitz, I.; Zhang, C.; Schneider, D. J.; Janjic, N. Proc. Natl. Acad. Sci. U. S. A. 2017, 114, 2898−2903. (b) Perrin, D. M.; Garestier, T.; Helene, C. J. J. Am. Chem. Soc. 2001, 123, 1556−1563. (c) Hollenstein, M.; Hipolito, C. J.; Lam, C. H.; Perrin, D. M. Nucleic Acids Res. 2009, 37, 1638−1649. (d) Hollenstein, M.; Hipolito, C. J.; Lam, C. H.; Perrin, D. M. ACS Comb. Sci. 2013, 15, 174−182. (e) Thomas, J. M.; Yoon, J. K.; Perrin, D. M. J. J. Am. Chem. Soc. 2009, 131, 5648−5658. (14) (a) Hili, R.; Niu, J.; Liu, D. R. J. J. Am. Chem. Soc. 2013, 135, 98−101. (b) Guo, C.; Watkins, C. P.; Hili, R. J. J. Am. Chem. Soc. 2015, 137, 11191−11196. (c) Lei, Y.; Kong, D.; Hili, R. ACS Comb. Sci. 2015, 17, 716−721. (d) Kong, D.; Lei, Y.; Yeung, W.; Hili, R. Angew. Chem., Int. Ed. 2016, 55, 13164−13168. (e) Guo, C.; Hili, R. Bioconjugate Chem. 2017, 28, 314−318. (f) Lei, Y.; Hili, R. Org. Biomol. Chem. 2017, 15, 2349−2352. (15) Vanbrabant, J.; Leirs, K.; Vanschoenbeek, K.; Lammertyn, J.; Michiels, L. Analyst 2014, 139, 589−595. (16) Bock, L. C.; Griffin, L. C.; Latham, J. A.; Vermaas, E. H.; Toole, J. J. Nature 1992, 355, 564−566. (17) Tasset, D. M.; Kubik, M. F.; Steiner, W. J. J. Mol. Biol. 1997, 272, 688−698. (18) Handbook of Surface Plasmon Resonance; Schasfoort, R. B., Ed.; Royal Society of Chemistry, 2017. (19) Scapin, G.; Yang, X.; Prosise, W. W.; McCoy, M.; Reichert, P.; Johnston, J. M.; Kashi, R. S.; Strickland, C. Nat. Struct. Mol. Biol. 2015, 22, 953−958. (20) Rydel, T. J.; Ravichandran, K. G.; Tulinsky, A.; Bode, W.; Huber, R.; Roitsch, C.; Fenton, J. W., II Science 1990, 249, 277−280. (21) Codon bias is equal to the standard deviation of the frequency of the codon in the product divided by the frequency of the codon in the template. A codon enrichment bias of 1 represents no enrichment, while less or greater than 1 represents de-enrichment or enrichment of the codon, respectively. See ref 14d for study.
13980
DOI: 10.1021/jacs.7b07241 J. Am. Chem. Soc. 2017, 139, 13977−13980