In vitro evolution of chemically-modified nucleic acid aptamers: Pros and cons, and comprehensive selection strategies

ABSTRACT

Nucleic acid aptamers are single-stranded DNA or RNA oligonucleotide sequences that bind to a specific target molecule with high affinity and specificity through their ability to adopt 3-dimensional structure in solution. Aptamers have huge potential as targeted therapeutics, diagnostics, delivery agents and as biosensors. However, aptamers composed of natural nucleotide monomers are quickly degraded in vivo and show poor pharmacodynamic properties. To overcome this, chemically-modified nucleic acid aptamers are developed by incorporating modified nucleotides after or during the selection process by Systematic Evolution of Ligands by EXponential enrichment (SELEX). This review will discuss the development of chemically-modified aptamers and provide the pros and cons, and new insights on in vitro aptamer selection strategies by using chemically-modified nucleic acid libraries.

Abbreviations

SELEX

Systematic Evolution of Ligands by EXponential enrichment

US FDA

United States Food and Drug Administration

AMD

Age-related macular degeneration

VEGF

Vascular endothelial growth factor protein, 2′-NH₂, 2′-Amino

2′-OH

2′-Hydroxyl

K_d

Equilibrium dissociation constant

2′-OMe

2′-O-Methyl

Bfgf

Basic fibroblast growth factor

2′-F

2′-Fluoro

PSMA

Prostate specific membrane antigen

IFN-γ

Interferon-gamma

KGF

Keratinocyte growth factor

4′-S

4′-Thio

2′-FANA

2′-Fluroarabino nucleic acid

HNA

1,5-Anhydro hexitol nucleic acid

TAR

transactivation responsive element

TNA

Threose nucleic acid

LNA

Locked nucleic acid

SOMAmers

Slow Off-rate Modified Aptamers

Introduction

Nucleic acid aptamer technology has attracted considerable attention in recent years in light of their widespread applications in therapeutic development, targeted drug delivery, bio-sensing and accurate molecular imaging. Aptamers are short single-stranded DNA or RNA oligonucleotides with unique 3-dimensional shape that can bind to their specific target with very high affinity and specificity.^1-5 Aptamers are generally developed from a large pool of oligonucleotide libraries containing approximately 10¹⁴ members by a reiterative process referred to as SELEX which involves selection, separation and enrichment steps (Fig. 1).^6,7 Till now, antibodies have been widely used for target specific molecular recognition.⁸ However, compared to antibody-based technologies, aptamers may possess a number of advantages including easy laboratory production in vitro effectively eliminating the use of live animals, no batch to batch variation, low or no immunogenicity, freedom to introduce multiple chemistries during synthesis without losing the affinity and specificity, small size that allows faster tissue penetration, ability to reverse target binding interactions using its complementary antidote sequence, significantly longer shelf-life and low cost. In 2004, an aptamer drug Macugen (Pegaptanib Sodium) was approved by United States Food and Drug Administration (US FDA) for the treatment of neovascular age-related macular degeneration (AMD) by targeting vascular endothelial growth factor protein 165 (VEGF₁₆₅).^9,10 Currently, a number of aptamer-based therapeutic candidates are in preclinical development and in different stages of clinical trials.¹¹

Figure 1.

Schematic illustration of the SELEX and post-SELEX methods for developing aptamers.

Typically, aptamers are developed with naturally occurring nucleotides. However, aptamers composed of natural nucleotide monomers are not suitable for theranostic applications as they possess very poor resistance to enzymatic degradation and show decreased binding affinity, rendering poor pharmacokinetic properties. To circumvent these shortcomings, aptamers containing chemically-modified nucleotide analogs with high stability against nucleases are normally used. Early examples of modified aptamers were primarily produced by post-SELEX-based approach. In this process, the aptamers were first isolated using natural RNA or DNA random sequences by SELEX method and then modified as per demand on affinity, stability and functionality. For this purpose, appropriate chemically-modified nucleotides are systematically incorporated into an existing DNA/RNA aptamer during solid-phase oligonucleotide synthesis. Generally, a web-based secondary structure prediction algorithm (e.g. mfold,¹² RNAfold¹³) is used as a tool to assist with the positioning of chemically-modified nucleotides and to truncate the overall size of the selected aptamers during chemical synthesis. Such chemically-fabricated aptamer variants are then tested for binding affinity, and the best candidates are used for further analysis and down-stream applications in vitro and in vivo. Evolution of chemically-modified aptamers would be a far more powerful approach to develop modified aptamers as this could generate aptamers with unique structures and with even shorter libraries.

Evolution of chemically-modified aptamers

A large repertoire of chemically-modified nucleotide analogs with remarkable biophysical properties have been developed in recent years and a few papers have reviewed the use of these chemically-modified nucleotides in the generation of aptamers and nucleic acids with enzymatic activity including DNAzymes and ribozymes.^14-18 But, their applicability in de novo evolution of aptamers via SELEX methodologies are rather impeded by poor or lack of enzymatic recognition capabilities. Conventional selection methodologies involve multiple enzymatic steps that are required to amplify and regenerate chemically-modified nucleotide-containing libraries. Some sugar-modified nucleotides are reported to tolerate few commercially available DNA or RNA polymerases,¹⁹ making them promising candidates for aptamer selection. However, it is worth mentioning that the level of enzymatic recognition capabilities of the reported modified nucleotides varies depending on the specific chemical modification. Still, the substrate properties of a number of other promising analogs have not been reported which might be due to the lack of enzymatic recognition. One option is to evolve an enzyme specific to the modified nucleotide, and there are reports of successful aptamer selection using engineered enzymes for chemically-modified nucleotides.^20,21 An overview of possible strategies that can be applied for selecting chemically-modified aptamers are outlined below (Fig. 2).

Figure 2.

The construction of oligonucleotide libraries for developing chemically modified libraries.

Generation of chemically-modified libraries and selection strategies

Starting oligonucleotide libraries containing the desired chemically-modified nucleotide or a combination of different modifications can be chemically synthesized using an oligonucleotide synthesizer via standard phosphoramidite chemistry. Another approach could be to synthesize normal DNA library which subsequently can be converted to a chemically-modified library by following enzymatic protocols involving PCR and/or in vitro transcription reactions with modified nucleotide triphosphates using the specific polymerase which could also be used in subsequent amplification/regeneration steps of multiple selection rounds. Synthetic chemically-modified libraries can be constructed by incorporating the modified nucleotides in the randomized region or in the primer-binding region.

Library with modified random region

Modified nucleotides can be incorporated as mixmers together with natural nucleotide monomers or as fully modified in the random region to furnish a chemically-modified oligonucleotide library for aptamer selection. However, the libraries containing fully modified random region with all four modified nucleotide bases may be limited to use in one-step aptamer selection-based methodologies. This is mainly because the selected aptamer candidates are regenerated using triphosphate derivatives of the corresponding modified nucleotides, but during an enzymatic synthesis process, the primer-binding region will also be modified with the same modified nucleotides contrary to the starting library design. In SELEX to allow maximum base variations of the chemically-modified nucleotides (good substrates of specific enzymes) in the random region, the nucleotide bases can be mixed with one, two or three modified nucleotide bases of the same or different nucleotide analogs in combination with the natural counterparts. But, depending on the number of modified nucleotide bases in the randomized region, the primer-binding region needs to be adjusted with the base composition of unmodified nucleotides. In this approach, where nucleotide mixmers are randomized, it is difficult to locate the positioning of modified nucleotide analogs, and if the library is constructed with successive stretch of modified nucleotides it may be hard for the polymerases to regenerate these particular sequences for subsequent selection rounds.

Another design approach could be to fix the positioning of modified nucleotides in the random region, which will assist to keep track of the positioning of modified nucleotides after sequencing. But, this could limit the pool diversity to some extent, however, it is worthwhile to perform selection since some chemical modifications could generate unique shapes and enhance target binding affinity. It is best to limit up to two fixed modifications that could be incorporated individually at different positions or as a mixture at the same positions allowing two nucleotide variations. Again, the primer-binding region needs to be constructed only with the unmodified natural nucleotide bases (avoiding similar base counterparts of the chemically-modified nucleotides) to regenerate the selected aptamer candidates for use in the subsequent selection rounds. During the sequence alignment after sequencing, only those candidates that maintain the initial library design need to be considered for further analysis and chemical synthesis.

Library with modified primer-binding region

For some modified nucleotide analogs that offer potential biophysical properties, enzymatic recognition may be difficult limiting their application in aptamer selection using conventional SELEX method. But, such modifications can still be used in aptamer selection using a library with one of the primer-binding regions containing the desired chemically-modified nucleotides, and the other region with all four natural nucleotide monomers. This will essentially eliminate the need of an enzymatic recognition of modified nucleotides as a synthetic primer sequence with the chemical modifications can be used during the enrichment steps. But, the protocol may be limited to DNA-based oligonucleotides. This approach may not fit very well in line to the concept of aptamer selection where it is speculated that the randomized region would promote the formation of target-binding motifs due to high sequence variability. But, one cannot ignore the fact that a part of the primer-binding region also plays a role in the folding pattern of the binding aptamers in majority of the cases. On this ground, aptamer selection using a library with chemically-modified primer-binding regions can be well justified, and can be used as an alternate approach. This method has been successfully applied for generating LNA-modified aptamers.^22,23

∞-aptamer library

Aptamer selections can also be performed using a library containing infinite range (∞) of modified nucleotides with non-natural structural chemistries of the nucleobase, sugar and phosphate backbone. Libraries containing infinite nucleotides could come with diverse range of chemistries bearing positive charges, hydrophobic groups, phosphorothioates, amino acids etc. that could improve and enhance the target binding interactions and nuclease resistance. Such libraries are mainly suitable for one-step selection protocols.^24-27 ∞-nucleotides can be best positioned in the randomized regions. During synthesis, ∞-nucleotides of all four bases can also be mixed with their natural counterparts.

Aptamer selection using sugar-modified nucleotides

Modifications of sugars will result in libraries with a greater functional diversity and can form stable aptamers with unique shapes. Sugar modifications could allow screening of ligands that bind with greater affinity to their targets than their unmodified counterparts. These unnatural modifications are less likely to be recognized by nucleases, making them more stable in serum. Most frequently, chemical modifications are introduced at the 2′-position of the nucleotide for increased nuclease resistance, and binding affinity. The SELEX evolution of various sugar- modified nucleotide (Fig. 3) containing aptamers against different targets, their binding affinities and stabilities are described below. Table 1 shows an overview of sugar-modified aptamer development in recent years.

Figure 3.

Structures of various sugar-modified nucleotides used in aptamer selection by SELEX methodologies.

Table 1.

Aptamers selected from sugar-modified DNA/RNA libraries and the binding affinity of aptamers to their targets.

Modified nucleotide	Target	Selection Method	Polymerases Used	Binding Affinity (Kd)	Ref.
2′-NH2	Human neutrophil elastase	Nitrocellulose membrane filter binding	T7 RNA polymerase	7–25 nM	28
	VPF/VEGF	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	2.4 nM	29
	mRNAs	Affinity column	T7 RNA polymerase	No Kd (Kcat = 0.04 min−1)	30
	bFGF	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	Apparent Kd = 0.35 nM	31
	KGF	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	0.4 nM	41
	IFN-γ	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	1.8 nM	40
	MiniTAR	Magnetic bead immobilisation	DNA polymerase	26–47 nM	32
2′-F	VEGF165	Nitrocellulose membrane filter binding	T7 RNA Polymerase	2 pM	33
	PSMA	Magnetic bead immobilisation	Y639F T7 RNA Polymerase	No Kd (IC50 = 27 nM)Ki = 11.9 nM	34, 35
	Human complement C5 component	Nitrocellulose membrane filter binding	T7 RNA polymerase	2–5 nM	36
	KGF	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	14F Kd = 0.3–3 pM	41
	Factor IXFactor IXa	Nitrocellulose membrane filter binding	AMV Reverse Transcriptase	Kd = 0.64 nM (9.3)Kd = 364 pM (Clone 9D-6)	37,38
	IFN-γ	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	106 nM	40
	IFN-γ	Nitrocellulose membrane filter binding	Y639F T7 RNA Polymerase	6.8 nM	40
	Cancer Antigen 125	Magnetic bead immobilisation	Y639F T7 RNA Polymerase	CA125.1 Kd = 4.13 nM	39
2′-OMe	VEGF	Electrophoretic mobility shift assay	Y639F/H784A/K378R T7 RNA Polymerase	2 nM	42
	Interleukin-23 and thrombin	Nitrocellulose membrane filter binding	Y639F/H784A/K378R T7 RNA Polymerase	Clone A5 Kd = 8.4 nMClone B4 Kd = 26 nM	43
	Tissue factor pathway inhibitor	Nitrocellulose membrane filter binding	Not Specified	ARC17480 Kd = 2.8 nMARC19499 IC50 = 17.9 nM	44
4′-S	Human thrombin	Nitrocellulose membrane filter binding	T7 RNA polymerase	4.7 nM	50
2′-FANA	HIV-1 reverse transcriptase	Electrophoretic mobility shift assay	Taq polymerase	4 pM	51
HNA	TAR	Magnetic bead immobilisation	Pol6G12	28–67 nM	52
	HIV Hen egg lysozyme	Magnetic bead immobilisation	Pol6G12	HNA11 Kd = 107 nMHNA19 Kd = 141 nM	52
TNA	Human thrombin	Capillary electrophoresis	Therminator DNA polymerase	200–900 nM	53
BNA/LNA	Human thrombin	Capillary electrophoresis	an enzyme mix of KOD Dash and KOD mutant DNA polymerases	0.26–27 nM	23,70
	CD73	Affinity column	KOD XL	3.7 nM	22
Spiegelmers	D-adenosine	Affinity column	T7 polymerase	1.7 µM	71
	L-arginine	Affinity column	T7 polymerase	Bound to L-arginine with Kd = 129 µM	72
	HIV-1 Tat protein			Bound to HIV-1 tat protein with Kd=26 µM
	D-vasopressin	Electrophoretic mobility shift assay	T7 polymerase	0.9 µM	73
	D-staphylococcal enterotoxin B	Electrophoretic mobility shift assay	Taq DNA polymerase	200 nM	74
	Nociceptin/ orphanin FQ	Affinity column	T7 polymerase	L-NOX2149 Kd = 0.3 µML-NOX2137 Kd = 0.7 µM	77
	D-gonadotropin-releasing hormone	Affinity column	T7 polymerase	L-S42 Kd = 45 nML-A10 Kd not specified (IC50 = 200 nM)	78
	Ghrelin	Electrophoretic mobility shift assay	T7 polymerase	44.4 nM	79,80
	Calcitonin gene-related peptide	Affinity column	T7 polymerase	Apparent Kd = 2.5 nM	75
	Substance P	Affinity column	T7 polymerase	40 nM	76

2′-Amino (2′-NH₂) modified aptamers

Aptamers have been selected from libraries where the 2′-hydroxyl (2′-OH) group of the pyrimidine is replaced by a NH₂ group. A modified aptamer was screened from a 2′-NH₂ -modified RNA library involving 2′-NH₂-UTP and 2′-NH₂-CTP replacing unmodified UTP and CTP during enrichment steps, specific to the human neutrophil elastase using nitrocellulose filter binding assay.²⁸ The selected aptamers showed high affinity to the target with the equilibrium dissociation constant (K_d) in the range of 7–30 nM which is much lower than that of the unmodified RNA aptamer (K_d > 1 µM).²⁸ These aptamers were found to be very selective to human neutrophil elastase.²⁸ These modified RNA aptamers also improved nuclease resistance, with an extended half-life (20 h and 9 h in serum and urine respectively) than the unmodified aptamer (degraded in less than 8 and 5 mins in serum and urine respectively).²⁸ Another 2′-NH₂-pyrimidine-modified RNA aptamer was developed against VPF/VEGF by SELEX process.²⁹ The affinity was measured through nitrocellulose filter binding assay and one aptamer NX-178 had a K_d of 2.4 nM, with an increased half-life in urine (17 h).²⁹ Interestingly the introduction of a 2′-O-Methyl (2′-OMe) nucleotide post-SELEX further increased the binding affinity to K_d = 0.14 nM and half-life in urine to 31 h.²⁹ This study suggests that the affinity and nuclease resistance of the RNA modified aptamers developed by SELEX methodology can be further improved by post-SELEX modification. In another study, Beaudry et al. developed a 2′-NH₂ pyrimidine-modified ribozyme which was identified by affinity column-based SELEX.³⁰ The selected aptamer had a half-life of 16 h in human serum, whereas the unmodified counterpart degraded in serum in just 5 mins.³⁰ Jellinek et al. developed a different 2′-NH₂ pyrimidine-modified RNA aptamer, selected through nitrocellulose membrane filter binding assay specific to basic fibroblast growth factor (bFGF).³¹ The identified aptamer m21A bound to bFGF with high binding affinity (K_d = 0.35 nM).³¹ The stability of this aptamer in serum is 100 fold higher than the natural RNA aptamer.³¹ In another study, Bugaut et al. used an interesting methodology whereby SELEX and dynamic combinatorial chemistry were combined to select conjugated RNA aptamers against MiniTAR by magnetic bead immobilisation using a 2′-NH_2- modified RNA library. The resulting conjugated aptamers developed bound to MiniTAR with high affinity (K_d = 26–47 nM).³²

2′-fluoro (2′-F) modified aptamers

Ruckman et al. developed an RNA aptamer specific to VEGF₁₆₅ isoform by nitrocellulose filter binding assay using 2′-F modified RNA libraries.³³ Most of the selected 2′-F RNA aptamers had affinities for VEGF₁₆₅ in the pM range (aptamer with the lowest affinities had a K_d value of 2 pM).³³ This candidate aptamer after further chemical modifications has been approved by the US FDA for the treatment of AMD. Lupold et al. developed another 2′-F modified RNA aptamer A10, targeting prostate specific membrane antigen (PSMA), cell surface receptors expressed on prostate cancer cells by SELEX using magnetic bead separation-based method.³⁴ The aptamer A10 bound to PSMA with a K_d of 11.9 nM in LNCaP cells.³⁴ This was later used to link siRNA to form aptamer-siRNA chimeras that allowed targeted delivery of siRNA to cells expressing PSMA wherein the siRNA reduced the expression of two survival genes, Polo-like kinase 1 and B-cell lymphoma 2.^34,35 In another study, Biesecker et al. developed another aptamer against human complement C5 component using 2′-F pyrimidine-modified RNA libraries by nitrocellulose membrane filter binding assay.³⁶ The aptamers generated had K_d of between 20–40 nM and this was further improved by a second biased SELEX experiment where the generated aptamers had a K_d of 2–5 nM.³⁶ The aptamers generated were able to bind to human complement C5 component with high affinity and were able to inhibit its activity in human serum with high stability.³⁶ Rusconi et al. developed 2′-F modified aptamers for Factor II, Factor VII, Factor IX and Factor X through nitrocellulose membrane filter binding assay.^37,38 By this approach a RNA aptamer was identified against Factor IXa with the highest affinity (K_d of 0.64 nM), and was a potent anticoagulant.³⁸ Another aptamer generated against Factor IX had a K_d of 364 pM.³⁷ Lamberti et al. developed an aptamer against Cancer Antigen 125 by magnetic bead immobilisation using a 2′-F modified RNA library.³⁹ Of the two RNA aptamers isolated, CA125.1 had a K_d of 4.15 nM and can be developed further for use as a diagnostic tool.³⁹

Comparison of 2′-F and 2′-NH₂ modifications

Various studies indicate that both 2′-F and 2′-NH₂ modifications increase the nuclease resistance and binding affinity of aptamers toward their targets. There are a few studies comparing the two modifications to validate which modification was more powerful. Kubik et al. reported an aptamer that bound to interferon-gamma (IFN-γ) which was screened from RNA libraries modified at the 2′-position of pyrimidine nucleotides with F, NH₂, or a mixture of F and NH₂ (2′-F/NH₂) groups using nitrocellulose filter binding assay.⁴⁰ The binding affinity of the modified aptamers to IFN-γ differed depending on the sugar modifications.⁴⁰ 2′-NH₂ modified aptamers had the highest binding affinity to IFN-γ with a K_d of 1.8 nM, whereas the 2′-F modified aptamers with the highest binding affinity had a K_d of 6.8 nM and the aptamers containing 2′-F/NH₂ mixture had a K_d of 106 nM.⁴⁰ The half-lives of these aptamers were measured through their stability in human serum, where 2′-NH₂ modified aptamer had a half-life of 80 h, 2′-F modified aptamers had a half-life of 6 h and the mixed 2′F/NH₂ modified aptamers had a half-life of 48 h.⁴⁰ However, all three were more nuclease-resistant than the unmodified aptamer which had a half-life of 20 s in human serum.⁴⁰ 2′-NH₂ modified aptamers seemed to confer the highest binding affinity to IFN-γ and nuclease resistance in human serum.⁴⁰ 2′-NH₂ aptamer also inhibited IFN-γ binding to its receptor on A549 human lung carcinoma cells with an ID₅₀ of 10 nM.⁴⁰ Another study by Pagratis et al. suggested that 2′-F modified aptamers improved the binding affinity compared to the 2′-NH₂ modified aptamers.⁴¹ In this study, aptamers have been selected successfully from libraries containing 2′-NH₂ and 2′-F modified pyrimidines by nitrocellulose filter binding assay and compared the affinities of the two different aptamer variants to keratinocyte growth factor (KGF).⁴¹ The best 2′-F modified aptamer 14F bound to KGF with a higher affinity (K_d = 0.3–3 pM) than the best 2′-NH₂ modified aptamer (K_d = 0.4 nM).⁴¹ 2′-F modified aptamers (6F and 14F) inhibited KGF binding to their receptor in PC3 cells with a K_i of 100 and 200 pM respectively and had a much higher inhibitory activity than the 2′-NH₂ modified variants, 14N and 29N (K_i = 1.4 nM).⁴¹ Although if 2′-F or 2′-NH₂ modification is more powerful is arguable the use of 2′-NH₂ modifications result in problems in solid-phase synthesis and the ribose sugar adopting an unfavorable conformation which has resulted in the reduced use of 2′-NH₂.

2′-OMe modifications:

2′-OMe modification involves the replacement of the 2′-OH group of the nucleotide with a methoxy group. 2′-OMe is a common sugar modification that has been used post-SELEX to generate aptamers with high nuclease resistance. Burmeister et al. developed an aptamer against VEGF from oligonucleotide libraries modified with 2′-OMe adenine and guanines by electrophoretic mobility shift assay.⁴² The selected aptamer candidate, ARC245 had a K_d of 2 nM, however this is lower than Macugen, the FDA approved aptamer drug targeting VEGF (K_d = 50 pM).⁴² ARC245 showed inhibition of VEGF binding to VEGF receptor in 293 cells at 10 nM concentrations and was also highly stable in serum for up to 96 h.⁴² Burmeister et al. also screened aptamers comprising 2′-OMe pyrimidines by the SELEX process to multiple protein targets including thrombin and interleukin-23, using proteins immobilized on a 96-well plate and nitrocellulose filter binding assay.⁴³ The identified aptamers targeting interleukin-23 had a K_d of 8.4 nM while the aptamer targeting thrombin had a K_d of 26 nM.⁴³ Waters et al. developed 2′-OMe modified aptamers, ARC19499 and ARC17480 against tissue factor pathway inhibitor by nitrocellulose filter binding using a 2′-OMe modified library.⁴⁴ ARC17480 bound to the tissue factor pathway inhibitor with a K_d of 2.8 nM, however as ARC19499 was not viable in experiments requiring radiolabelling a competition-binding experiment showed that ARC19499 competed with radiolabelled ARC17480 with an IC₅₀ of 17.9 nM.⁴⁴ 2′-OMe incorporation into libraries and using these modified libraries for SELEX is hard due to lack of enzymes that are capable of recognizing these 2′-OMe modified bases. However, recently polymersases have been evolved that accept 2′-OMe modified triphosphates which would allow generation of additional 2′-OMe modified aptamers.^21,45-49

4′-Thio (4′-S) modifications

4′-S modifications had a sulfur atom at the 4′-position of the sugar moiety. Kato et al. developed a 4′-S modified RNA aptamer that bound to human α-thrombin by nitrocellulose filter binding assay.⁵⁰ In the presence of RNase, the stability of the aptamer candidate, thioRNA59, had a half-life of 1174 mins and was stable even after 12 h of incubation, which was 50 times greater than that of the corresponding natural RNA (completely degraded in the presence of RNase A in 3 h), and it also showed high binding affinity to thrombin with a K_d of 4.7 nM.⁵⁰

2′-fluroarabino nucleic acid (2′-FANA) modified aptamers

Very recently, Alves Ferreira-Bravo et al. reported a 2′-FANA aptamer to human immunodeficiency virus-1 (HIV-1) reverse transcriptase.⁵¹ This 2′-FANA aptamer was isolated from a 2′-FANA modified DNA pool through SELEX using electrophoretic mobility shift assay.⁵¹ The developed 2′-FANA modified DNA aptamer, FA₁ had a K_d value of 4 pM and showed greater resistance to nucleases.⁵¹

1,5-anhydro hexitol nucleic acid (HNA) modified aptamers

Pinheiro et al. developed HNA aptamers against HIV-1 transactivation responsive element (TAR) and hen egg lysozyme through SELEX using magnetic bead-based separation.⁵² HNA aptamer, T5-S8-7 bound to TAR with high specificity and had a K_d of 28–67 nM.⁵² HNA aptamers that bound specifically to hen egg lysozyme had a K_d of 107 to 141 nM as determined by surface plasmon resonance.⁵²

Threose nucleic acid (TNA) modified aptamers

Yu et al. selected a TNA modified aptamer against thrombin using affinity column and capillary electrophoresis.^53,54 TNA backbone is nuclease-resistant as TNA remained undigested even after 72 h of incubation with a pure nuclease unlike DNA and RNA which exhibit half-lives of 30 mins and less than 10s respectively.⁵⁴ The developed TNA aptamers to thrombin had a K_d in the range of 200–900 nM.⁵³ A new manganese-independent TNA polymerase was evolved using droplet-based optical polymerase sorting which can be used in TNA aptamer selection protocol.⁵⁵

Locked nucleic acid (LNA) modified aptamers

In LNA, the sugar ring is locked by a 2′O, 4′C methylene linkage and is conformationally restricted adopting a C3′-endo conformation.^56-61 Toward the development of LNA-modified aptamers, Veedu et al. and others have extensively investigated the enzymatic recognition capabilities of LNA-nucleotides.^58-68 DNA aptamers containing LNA (BNA/LNA) nucleotides were developed against human thrombin using capillary electrophoresis SELEX. BNA/LNA-modified nucleotides were introduced in the primer-binding region of the library, and after selection the selected candidates were enriched by PCR using the LNA-modified primer sequence to regenerate the selected aptamer candidates using KOD polymerase.^23,69 Many aptamers showed K_d values in the low nanomolar range.^23,70 In another study, Elle et al. generated LNA-modified aptamers against CD73 immobilized on anti-His tag plates.²² A LNA-modified DNA aptamer NAC6772 showed high binding affinity to CD73 through surface plasmon resonance experiment with a K_d of 3.54 nM and inhibited the CD73 activity by 85%.²²

Spiegelmers

Spiegelmers are mirror images of the natural aptamers in which the D-ribose (the natural ribose) are replaced with the unnatural L-ribose. Spiegelmers prevent enzymatic degradation as the chiral form of the nucleic acid (the L-form) is unnatural and is not recognized by nucleases. However, the L-form of the nucleic acid like their D-form natural counter parts can bind to the target with high affinity and will not trigger the immune response. Similar to the other sugar modifications described above, spiegelmers can be enriched by T7 RNA polymerases. Klussmann et al. generated one of the first spiegelmers through SELEX against the naturally occurring D-adenosine.⁷¹ The high affinity L-RNA aptamers against D-adenosine were isolated through affinity column and had a K_d of 1.7 µM which was comparable to the affinity displayed by the D-RNA aptamers.⁷¹ Nolte et al. reported another spiegelmer that bound to L-arginine that could also bind to a short peptide that contains an arginine-rich region in HIV-1 Tat protein.⁷² The aptamer, L-R16c was isolated through affinity column based SELEX.⁷² L-R16c bound to L-arginine with a K_d of 129 µM and to the Tat protein with a K_d of 26 µM.⁷² L-R16c was very stable in serum and did not degrade even after 60 h of incubation.⁷² In another study, spiegelmers were selected against D-vasopressin and D-staphylococcal enterotoxin B through SELEX using electrophoretic mobility shift assay.⁷³ The L-DNA aptamers targeting D-vasopressin showed no degradation with exo- and endonucleases and showed high binding affinity for D-vasopressin with a K_d of 0.9 µM.⁷³ The corresponding D-aptamer degraded in 10 s while the L-aptamer showed no degradation even after 10 days incubation with exo- and endonucleases and 7 days in serum.⁷³ A DNA spiegelmer targeting D-staphylococcal enterotoxin B, B12b10 had a K_d of 200 nM.⁷⁴ Vater et al. isolated a RNA spiegelmer against migraine associated calcitonin gene-related peptide through SELEX using affinity columns.⁷⁵ The developed RNA spiegelmer STAR-F12 bound to calcitonin gene-related peptide with a K_d of 2.5 nM and inhibits its activity.⁷⁵ In another study, a RNA spiegelmer targeting Substance P was generated through SELEX using affinity column.⁷⁶ The developed L-RNA aptamer, SUP-A-004, which was further truncated and the resulting spiegelmer bound to Substance P with a K_d of 40 nM.⁷⁶ SUP-A-004 inhibited L-substance P mediated Ca²⁺ release in AR42J pancreatic cells.⁷⁶ Faulhammer et al., in another study reported the development of L-RNA aptamers NOX2149 and NOX2137 against Nociceptin/orphanin FQ through SELEX using affinity column.⁷⁷ The L-NOX2149 and L-NOX2137 bound to the nociception/orphanin FQ with a K_d of 0.3 µM and 0.7 µM respectively which were similar to the binding affinity to their D-aptamer counterparts.⁷⁷ However L-aptamers had the advantage of being more stable in serum.⁷⁷ In another study, L-DNA/RNA aptamers that bind to D-gonadotropin-releasing hormone with high affinity and specificity were isolated out of L-RNA and DNA libraries through SELEX using affinity column.⁷⁸ The L-DNA aptamer, S42 bound to gonadotropin-releasing hormone with a K_d of 45 nM which is similar to that of the D-DNA aptamer and the IC₅₀ was 50 nM.⁷⁸ The D-RNA aptamer, A10 bound with K_d of 55 nM, however the K_d of L-RNA aptamer could not be obtained.⁷⁸ Both the DNA and RNA spiegelmers bound to gonadotropin-releasing hormones and inhibited their binding to their receptor in Chinese hamster ovary cells and there was no immune response produced in zimmermann rabbits.⁷⁸ Helming et al. reported L-RNA aptamers targeting ghrelin through SELEX using electrophoretic mobility shift assay.^79,80 L-NOX-B11, a L-RNA aptamer was isolated and bound to ghrelin with high affinity with a K_d of 44.4 nM.⁸⁰ L-NOX-B11 inhibited ghrelin from binding to its receptor with an IC₅₀ value of 5 nM in Chinese hamster ovary cells.^79,80 NOXXON Pharma have developed many spiegelmers against various targets and a few of the spiegelmers are in various stages of the preclinical and clinical trials, two of which are described in a recent review by Sundaram et al. NOX-A12 is a spiegelmer that bound to and inhibits stroma cell-derived factor-1.⁸¹ After completion of Phase I and Phase IIa trials of NOX-A12, Phase 3 trials have been planned for multiple myeloma.⁸¹ Also, Phase 2b/3 trials have been planned for the same candidate against glioblastoma.⁸¹ Phase 2a clinical trials have been completed for another spiegelmer, NOX-E36, targeting monocyte chemoattractant- protein 1.⁸¹ Several other candidates are currently in different stages of clinical investigation.⁸² NOX-D20 is another spiegelmer that binds to anaphylatoxin C5a with picomolar affinity and is being considered for preclinical and clinical development.⁸³ Wang et al. has described the generation of a chemically-synthesized D-amino acid polymerase that is capable of recognizing and catalyzing transcription and polymerisation of an L-DNA template.⁸⁴ This polymerase could result in generation of new speigelmers in the future.⁸⁴

Aptamer selection using modified nucleotide bases

Incorporating chemical modification to the nucleotide bases could increase the stability and target-binding affinity of aptamers. A number of researchers have succeeded in generating aptamers with base-modified nucleotides (Fig. 4), some of which are in preclinical stages. Various base modifications are incorporated into the nucleic acid library and in many cases have been able to be amplified using family B-DNA polymerases.⁸⁵ Table 2 provides a summary of base-modified aptamer development. There is also a new method called click-SELEX which is used to generate base-modified libraries by replacing the thymidines with C5-ethynyl-2′-deoxyuridine which is then further modified through click reaction with copper(I)-catalyzed alkyne-azide cycloaddition.^86-91 The library is then used in selection experiments to evolve modified aptamers.^86-91 This is a powerful method that allows generation of modified aptamers with new chemical modifications without problems caused by lack of enzymatic recognition as seen with the addition of a cluster of oligomannose glycans.^86-91 This will allow targeting of many molecules and epitopes that is currently not possible with standard SELEX.^86-91 Addition of new nucleotides and/or unnatural base pairs to the libraries and using these libraries for aptamer selection allows generation of modified aptamers with greater sequence and functional diversity.^92-95 Zhang et al. has shown that addition of new nucleotides Z and P allowed generation of sequences that had better binding affinity to the target molecule indicating that new nucleotides are viable in the library and can show higher binding affinity than natural bases in some cases.⁹²

Figure 4.

Structures of various base-modified nucleotides used in aptamer selection by SELEX methodologies.

Table 2.

Aptamers selected from base-modified DNA/RNA libraries and the binding affinity of aptamers to their targets.

Modified nucleotide	Target	Selection Method	Polymerases Used	Binding affinity (Kd)	Ref.
5-pentynyl-dU	Human thrombin	Affinity column	Vent DNA Polymerase	400 nM	96
5-N-(6-aminohexyl) carbamoylmethyl-dU	Thalidomide (T5N and T5-IB)	Streptavidin-sepharose gel	KOD Dash DNA polymerase	T5N Kd= 113 µMT5-1B Kd= 133 µM	100
5-boronic acid-dT	Fibrinogen	Magnetic bead immobilisation	Taq DNA Polymerase	3–30 nM	102
5-iodouridine	HIV-1 Rev protein.	Nitrocellulose membrane filter binding	T7 RNA polymerase	0.8 nM	97
5-(3′-aminopropynyl)-2′-dU	ATP (Sequence 409)	Capillary electrophoresis	Vent DNA polymerase	6 nM	98
5-benzylaminocarbonyl dU	Plasminogen Activator Inhibitor	Magnetic bead immobilisation	KOD DNA Polymerase	≤30 nM	107
5-isobutylaminocarbonyl –dU	Human mobility group -1			≤30 nM
5-tryptaminocarbonyl -dU	Fractalkine			≤30 nM
5-naphthylmethylaminocarbonyl-dU	Human protein targets			≤30 nM
5-(3-aminoallyl)-dC5-guanidinoallyl-dU, 5-imidazolyl-dA	RNA	Magnetic bead immobilisation	Vent (exo -) DNA Polymerase	Kd not specified (DNAzyme12–91 Kobs = 0.06 min−1DNAzyme 16.2–11 Kcat= 1.4–1.5 min−1)	106
7-(2-thienyl)imidazo[4,5-b]pyridine	VEGF165	Magnetic bead immobilisation	Accu Prime Pfx DNA Polymerase	0.65 pM	105
	IFN-γ			0.038 nM
5-tyrosyl-dU	Clostridium difficile binary toxins A and B	Magnetic bead immobilisation	KOD EX DNA polymerase	1.4 nM (Toxin A)0.31 nM (Toxin B)	109
5-(2-naphthylmethylaminocarbonyl)-dU				9.2 nM (Toxin A)0.25 nM (Toxin B)
5-phenylethyl-dU				13.1 nM (Toxin A)0.27 nM (Toxin B)
5-benzylaminocarbonyl-dU				5 nM (Toxin B)
5-naphthylmethylaminocarbonyl-dU				0.43 nM (Toxin B)
5-tryptaminocarbonyl-dU				1.7 nM (Toxin A)0.45 nM (Toxin B)
5-carboxamide-modified-dU	Tumor necrosis factor receptor super family member 9 (6a and 6d)	Magnetic bead immobilisation	Deep Vent and KOD XL DNA Polymerases	6a Kd= 6 nM 6d Kd= 4 nM	103
5-imidazole-dU analog	RNA	Affinity column	DNA Polymerase	Kd not specified (DNAzyme 16.2–11 Kcat= 1.4–1.5 min−1	104

Uridine modifications

Latham et al. developed a modified aptamer containing 5-(1-pentynyl)-2′-deoxyuridine against thrombin by affinity column.⁹⁶ The affinity of the selected aptamer pool was 400 nM.⁹⁶ This was one of the first examples of using a base-modified nucleic acid library for screening aptamers. In another study, Jensen et al. developed RNA aptamers from a random RNA library containing photoreactive chromophore 5-iodouridine using crosslinking SELEX.⁹⁷ The selected RNA aptamers bound and also cross-linked to the target with UV irradiation.⁹⁷ These modified aptamers had high affinity with a K_d of 0.8 nM.⁹⁷ Battersby et al. isolated another base-modified aptamer containing 5-aminopropynyl-deoxyuridine against ATP by electrophoretic mobility shift assay.^98,99 One of the selected aptamer candidate, 409 was found to bind ATP with higher affinity (K_d = 6 µM) than their unmodified counterparts.^98,99 Shoji et al. reported another base-modified aptamer with 5′-N-6-NH₂-hexyl carbamoylmethyl dU.^100,101 The modified library was used to screen aptamers against thalidomide using streptavidin-sepharose gel.¹⁰⁰ Surface plasmon resonance showed that the selected base-modified aptamer T5N and T51B had a K_d of 113 µM and 133 µM respectively.¹⁰⁰ Li et al. selected an aptamer containing boronic acid-modified thymidine against fibrinogen using magnetic bead immobilization-based SELEX.¹⁰² The resulting modified aptamers bound to fibrinogen with high affinity compared to that of the unmodified aptamers (K_d = 3–30 nM compared to K_d = 450 nM).¹⁰² Vaught et al. selected a DNA aptamer against tumor necrosis factor receptor super family 9 from a library possessing 5-amide-modified deoxyuridine via a magnetic bead-based SELEX method.¹⁰³ Six 5-position modified 2′-dUTP derivatives were used.¹⁰³ Two of the modified DNA aptamers developed showed high affinity (K_d = 6 nM and 4 nM respectively) compared to an unmodified RNA aptamer that has been previously reported (K_d = 40 nM).¹⁰³

Imidazole modifications

Aptamers can also act as DNA enzymes which are catalytic nucleic acids and modified libraries have been used to select such aptamers. Sontoro et al. developed a modified DNAzyme aptamer via affinity column from a modified library containing C5-imidazole-functionalised dU.¹⁰⁴ The isolated DNAzyme was one of the smallest nucleic acid enzymes, which showed multiple turnovers in the presence of millimolar concentrations of Zn²⁺, and a catalytic rate of 1.4–1.5 min⁻¹ for cleavage.¹⁰⁴ Kimoto et al. reported the selection of 7-(2-thienyl)imidazole[4,5-b]pyridine modified DNA aptamers against VEGF₁₆₅ and IFN-γ by magnetic bead-based SELEX.¹⁰⁵ The isolated DNA aptamers bound to VEGF₁₆₅ and IFN-γ with 100 fold higher affinity than their natural counterparts with K_d values of 0.64 pM and 0.038 nM respectively.¹⁰⁵ Hollenstein et al. reported the development of Dz12–91, a modified DNAzyme containing imidazole, ammonium and guanidinium groups. Dz12–91 was isolated through SELEX using magnetic streptavidin particles and cleaves all RNA sequence independently of M²⁺ with a K_obs of 0.06 min⁻¹.¹⁰⁶

Slow off-rate modified aptamers (SOMAmers)

Another class of aptamers with improved binding properties through base-modifications that result in slow dissociation rates called SOMAmers. There are many SOMAmer modifications that are used to generate aptamers to difficult protein targets.¹⁰⁷ The affinities of the SOMAmers are consistently in the nanomolar range to their targets and show high nuclease resistance.¹⁰⁷ Various proteins have been used as targets to generate SOMAmers.^107,108 In many SOMAmers, the modifications are introduced at the 5-position of the uridine nucleotide.¹⁰⁷ Gold et al. described the development of SOMAmers against more than 800 targets using multiplex SOMAmer affinity assay system from base-modified libraries including 5-benzylaminocarbonyl, 5-naphthylmethylaminocarbonyl, 5-tryptaminocarbonyl, 5-isobutylaminocarbonyl, 5-tyrosyl, 5-phenylethyl, and 5-guanidinoallyl.^107,109 Ochsner et al. reported development of SOMAmers against toxins A, B and binary toxins using 5-benzylaminocarbonyl, 5-naphthylmethylaminocarbonyl, 5-tryptaminocarbonyl, 5-phenylethyl-1-aminocarbonyl, 5-tyrosylaminocarbonyl or 5-(2-naphthylmethyl)aminocarbonyl-modified uridine containing library using magnetic bead-based SELEX.¹⁰⁹ Most of the modified aptamers bound to the toxin A and B with high binding affinity with the K_d in the nanomolar range.¹⁰⁹ For example, the selected 5-tryptaminocarbonyl dU modified aptamer showed a K_d of 1.7 nM to Toxin A and 0.45 nM for Toxin B.¹⁰⁹

Aptamer selection using phosphate modified nucleotides

A few phosphate-modified aptamers (Fig. 5) were reported as substrates of T7 RNA polymerase that could accept the triphosphate analogs of the modified phosphate moiety, mainly phosphorothioate and boranophosphates. Table 3 provides a summary of base-modified aptamer development.

Figure 5.

Structures of various phosphate-modified nucleotides used in aptamer selection by SELEX methodologies.

Table 3.

Aptamers selected from phosphate-modified DNA/RNA libraries and the binding affinity of aptamers to their targets.

Modified nucleotide	Target	Selection Method	Polymerases Used	Binding affinity (Kd)	Ref.
Boranophosphate-5′-(α-P-borano)-G 5′-(α-P-borano)-U	ATP	Affinity column	T7 RNA Polymerase	Not Specified	113
Phosphorothioate-modified DNA	bFGF	Nitrocellulose membrane filter binding	Taq polymerase	1.8 nM	110
Phosphorothioate-dA	Nuclear factor for human Interleukin6	Nitrocellulose membrane filter binding	Taq DNA Polymerase	Kd not specified (Kobs = 2 nM)	112
Phosphorothioate-dA	Human Nuclear factor-kappa B RelA (p65)	Nitrocellulose membrane filter binding	Taq DNA Polymerase	4.8 nM	111
	Human Nuclear Factor-kappa B RelA (p50)			0.8 nM

Phosphorothioate modifications

Through nitrocellulose membrane filter binding, phosphothioate-modifed RNA aptamer 11–20 was isolated that bound to bFGF with a K_d of 1.8 nM which was similar to the K_d of all RNA aptamers (K_d = 0.19 nM and 0.49 nM).¹¹⁰ K_d of another bFGF aptamer containing 2′-NH₂ pyrimidines was 0.35 nM, which was also comparable to the unmodified RNA aptamers.¹¹⁰ These aptamers also bound to other bFGF members in the family, but not to the unrelated targets and were therefore “semi-specific” for the FGF family members.¹¹⁰ Another phosphorothioate DNA aptamer was selected against nuclear factor for human interleukin 6 which was nuclease-resistant to DNase 1 enzyme at the point of modifications.^111,112 This modified DNA aptamer was selected through nitrocellulose filter binding from a library with thiophosphate backbone substitution at the thymidine positions and the modified aptamer was found to bind with stoichemistry of two protein dimers/duplex unlike the unmodified aptamer which bound with one protein dimer/duplex stoichemistry.¹¹² The binding constant (K_obs) measured by fluorescence anisotropy was 2 nM for a 66-mer.¹¹² Another phosphorothioate aptamer was selected against Nuclear Factor kappa B proteins Rel A (p65) and p50 and the selected aptamer had a K_d value of 4.8 nM and 0.8 nM respectively, which was comparable to the unmodified aptamer (K_d = 4.77 nM).¹¹¹

Boranophosphate modification

Boronated nucleotide analog modifications of guanosine (5′(α-P-borano)triphosphate) for GTP and uridine (5′(α-P-borano)triphosphate) for UTP were also introduced into general libraries to select modified aptamers that bind ATP through affinity matrix containing C8-linked ATP agarose.¹¹³ Each of the 2 modifications were tolerated in the backbone alone, however both modifications in one backbone were not tolerated.¹¹³

Conclusions and future perspectives

Aptamers are nucleic acid ligands that are generally developed through SELEX and bind to their targets with high affinity and specificity. Aptamers can be used therapeutically, diagnostically, as molecular beacons or as DNAzymes in various diseases. Macugen is a therapeutic aptamer approved by the US FDA for the treatment of AMD. One of the major disadvantages of aptamers composed of natural nucleotides is their rapid degradation in vivo. To overcome this obstacle, chemically-modified aptamers have been developed through SELEX using modified-oligonucleotide libraries. There are many potential chemically-modified nucleotides with excellent properties, however many of them cannot be used in the SELEX processes to develop aptamers mainly because of their very limited enzymatic recognition capabilities which are required for enrichment steps. As reviewed here, it is very promising to see the development of chemically-modified aptamers using several sugar, base and phosphate-modified nucleotides. It is also very encouraging to note the development of novel engineered polymerases capable of recognizing a specific modified nucleotide. Recent developments in one-step-based selection methodologies further offer tremendous hope for developing aptamers containing modified aptamers with limited enzymatic recognition capabilities.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgment

RNV acknowledges the funding from McCusker Charitable Foundation, and Western Australian Neuroscience Research Institute. MC thanks the funding from Greg and Dale Higham.