Discovery at the VICB
In Quest of Better Aptamers
By: Carol A. Rouzer, VICB Communications
Published: September 15, 2016
A simple substitution in the phosphate backbone of RNA aptamers increases binding affinity to target proteins, promising to improve their therapeutic potential.
Aptamers are short RNA or RNA-based polynucleotides that are designed to bind tightly but noncovalently to a protein target. They can provide selectivity and affinity similar to those of antibodies and have been used as affinity reagents and bioimaging probes. They also hold significant therapeutic promise as in the case of pegaptanib, a vascular endothelial growth factor-targeting aptamer that is currently being used to treat age-related macular degeneration. One process often used for the discovery of aptamers is SELEX (systematic evolution of ligands by exponential enrichment). This process can yield aptamers with impressive binding affinity to the desired target, but often further modification is necessary to improve stability and further augment affinity. In the past, efforts to increase binding affinity of aptamers have focused on the ribose moieties, an approach that usually succeeds in achieving improved affinity but has little effect on stability. This led Vanderbilt Institute of Chemical Biology members Martin Egli and Terry Lybrand, along with their collaborator Xianbin Yang at AM Biotechnologies to explore modification of the phosphate groups as a better way to achieve enhanced aptamer performance [N. D. Abeydeera, et al. (2016) Nuc. Acids Res., published online August 26, DOI:10.1092/nar/gkw725].
The phosphate backbone of RNA is usually well exposed, and the phosphate oxygen atoms frequently interact via salt bridges or hydrogen bonds with amino acid residues of protein binding partners. Prior work had shown that replacement of the two non-bridging atoms of a phosphate group with sulfur atoms, yielding a phosphorodithioate (PS2) linkage, could significantly increase the potency and stability of siRNA. Evidence suggests that the affinity effects associated with PS2 substitution were the result of hydrophobic interactions that cannot occur with phosphate linkages. Another possibility is that the higher polarizability of the P=S bond as compared to the P=O bond could lead to tighter binding interactions. To systematically explore the effects of PS2 substitution on aptamer affinity, and to better assess the mechanism(s) of those effects, the investigators conducted a "PS2 walk" using two known aptamers. The PS2 walk entails creating a library of aptamers, each of which carries a PS2 substitution at a single phosphate group. The complete library covers every phosphate group across the entire aptamer chain (Figure 1). The aptamers the researchers chose for their study were AF83-1, which binds to the VEGF165 variant of vascular endothelial growth factor (KD 2.1 nM) and AF113-1, which binds to α-thrombin (KD = 2.8 nM) (Figure 2).
FIGURE 1. Diagrammatic representation of the PS2-walk process. Starting with a known aptamer, a library of new aptamers is synthesized. Each of these contains a PS2 substitution at a single phosphate group, and every phosphate group is altered. The library of aptamers is then screened for changes in affinity and/or other properties. Figure reproduced under the Creative Commons Attribution-NonCommercial 4.0 International License from N. D. Abeydeera, et al. (2016) Nuc. Acids Res., published online August 26, DOI:10.1092/nar/gkw725.
FIGURE 2. Structures of the AF81-1 (A) and AF113-1 (B) aptamers and the sites of PS2 modification that resulted in increased affinity (highlighted in blue). Figure reproduced under the Creative Commons Attribution-NonCommercial 4.0 International License from N. D. Abeydeera, et al. (2016) Nuc. Acids Res., published online August 26, DOI:10.1092/nar/gkw725.
As soon as the PS2 walk libraries were complete, the researchers evaluated the binding affinity of each aptamer using biolayer interferometry. They found that in the vast majority of cases PS2 substitution had little effect on aptamer binding affinity. However, substitution of AF83-1 at positions 7 (AF83-7, KD = 1.0 pM) or 19 (AF83-19, KD = 1.0 pM) or AF113-1 at position 18 (AF113-18, KD = 1.8 pM) increased affinity by ~1000-fold.
To study the binding affinity of AF83-7 in intact cells, the researchers labeled the aptamer with fluorescein, and incubated it with HT-29 colorectal cancer cells that express VEGF165 or with MRC-5 fibroblasts that do not express the target protein. They then used flow cytometry and fluorescence confocal microscopy (Figure 3) to show that AF83-7 bound only to the HT-29 cells. These results, along with additional studies showing that AF83-7 exhibited increased stability relative to that of AF83-1 upon incubation with human serum, suggested that the modified aptamer could be of value as an affinity probe to detect pathological VEGF165 expression.
FIGURE 3. As seen using confocal microscopy, fluorescein- (FITC)-labeled AF83-7 (green color) binds to HT-29 cells that express VEGF165, whereas an FITC-labeled control aptamer with a scrambled sequence does not. Neither aptamer binds to MRC-5 cells, which do not express VEGF165. Figure reproduced under the Creative Commons Attribution-NonCommercial 4.0 International License from N. D. Abeydeera, et al. (2016) Nuc. Acids Res., published online August 26, DOI:10.1092/nar/gkw725.
Further characterization by CD spectroscopy and thermodynamic parameter analysis demonstrated that the PS2 substitution had minimal effect on the overall structure of the AF83-7 and AF113-18 aptamers as compared to their native RNA counterparts. In addition, biolayer interferometry demonstrated that the PS2 substitutions did not alter the binding specificity of the aptamers.
A crystal structure of the AF113-18:thrombin complex was particularly revealing (Figure 4). Comparison to the previously published structure of the AF113-1:thrombin complex showed little difference in the overall conformations of either the protein or the aptamer. However, in the region of the PS2 substitution, notable differences were apparent. The uracil base at position 17 (U17) and its associated ribose were displaced by about 2 Å, while the adjacent PS2 linkage was rotated by 90o in the AF113-18:thrombin complex relative to the positions of the corresponding structures in the AF113-1:thrombin complex. As a result, the PS2 group established interactions in the AF113-18:thrombin complex that were not possible for the corresponding phosphate group in the AF113-1:thrombin complex. Specifically, the sulfur atoms of the PS2 group of AF113-18 formed intramolecular hydrogen bonds with the N6-amino group of the adenine at position 7 and a hydrophobic contact with C8 of the guanine at position 16 of the aptamer. New contacts between the aptamer and the protein included hydrophobic interactions between the PS2 sulfur atoms and carbons in the side chain of Arg126 and the phenyl ring of Phe232.
FIGURE 4. Comparison of the structures of an AF113-18:thrombin and native AF113-1:thrombin complex. The aptamer and protein in the AF113-18:thrombin complex are colored by atom, with RNA carbons shown in magenta and protein carbons shown in green. The corresponding structures in the AF113-1 complex are colored gray. Highlighted are new interactions between the PS2 group of AF113-18 and residues A7 and G16 of AF113-18 in addition to interactions between the PS2 group of AF113-18 and residues Arg126, and Phe232 of α-thrombin. The inset highlights the region around the corresponding phosphate group of AF113-1 in the AF113-1:thrombin complex. Figure reproduced under the Creative Commons Attribution-NonCommercial 4.0 International License from N. D. Abeydeera, et al. (2016) Nuc. Acids Res., published online August 26, DOI:10.1092/nar/gkw725.
The results provide strong support for use of the PS2 walk to systematically explore modifications that can improve both the affinity and stability of aptamers. They also demonstrate that increased affinity associated with PS2 substitution can result from the ability of the sulfur atoms to engage in hydrophobic contacts with protein residues that are not possible with oxygen atoms. In addition, the authors note that, although the PS2 group of the AS113-18 aptamer contacts a predominantly hydrophobic pocket on the α-thrombin surface, polar residues are in close proximity to this pocket. Thus, the greater polarizability of the P=S bond relative to the P=O bond likely also contributes to the greater binding affinity of AS113-18 as compared to AS113-1 for the target protein. Regardless of the relative contribution of these various mechanisms, the PS walk approach has great promise for future design of high affinity aptamers for diagnosis and treatment.
View Nucleic Acids Research article: Evoking picomolar binding in RNA by a single phosphorodithioate linkage