Current Protocols in Nucleic Acid Chemistry || In Vitro Selection Using Modified or Unnatural Nucleotides

UNIT 9.6In Vitro Selection Using Modified orUnnatural Nucleotides

Gwendolyn M. Stovall,1 Robert S. Bedenbaugh,1 Shruti Singh,1

Adam J. Meyer,1 Paul J. Hatala,2 Andrew D. Ellington,1 and Bradley Hall2

1The University of Texas at Austin, Austin, Texas2Altermune Technologies LLC, Austin, Texas

ABSTRACT

Incorporation of modified nucleotides into in vitro RNA or DNA selections offers manypotential advantages, such as the increased stability of selected nucleic acids against nu-clease degradation, improved affinities, expanded chemical functionality, and increasedlibrary diversity. This unit provides useful information and protocols for in vitro selectionusing modified nucleotides. It includes a discussion of when to use modified nucleotides;protocols for evaluating and optimizing transcription reactions, as well as confirming theincorporation of the modified nucleotides; protocols for evaluating modified nucleotidetranscripts as template in reverse transcription reactions; protocols for the evaluation ofthe fidelity of modified nucleotides in the replication and the regeneration of the pool; anda protocol to compare modified nucleotide pools and selection conditions. Curr. Protoc.Nucleic Acid Chem. 56:9.6.1-9.6.33. C© 2014 by John Wiley & Sons, Inc.

Keywords: in vitro selection � aptamer � ribozymes � deoxyribozymes � modifiednucleotides � unnatural nucleotides

INTRODUCTION

In vitro selection is the process by which a pool of nucleic acids is enriched via iterativeselection and amplification for those species that are capable of performing a particulartask. Nucleic acids have been selected that bind to particular targets (aptamers), catalyzereactions (ribozymes or deoxyribozymes), or act as molecular switches (aptazymes).Similarly, nucleic acids have been found in nature to control gene expression uponbinding an analyte (riboswitches).

Instructions for carrying out in vitro selection experiments have been detailed elsewherein this chapter (i.e., UNITS 9.3, 9.4, & 9.5). This unit augments these other units by describinghow modified nucleotides can potentially be incorporated into a selection. It is stronglyrecommended that the researcher be conversant with a “normal” in vitro selection experi-ment prior to attempting selections with modified nucleotides. A normal in vitro selectionexperiment is already fraught with problems and pitfalls, and the addition of modifiednucleotides adds an extra level of difficulty. For simplicity, this unit focuses on in vitroselection using RNA pools; however, similar procedures can be used for DNA pools.

CAUTION: When working with radioactivity, take appropriate precautions to avoidcontamination of the experimenter and surroundings. Carry out the experiment anddispose of wastes in an appropriately designated area, following the guidelines providedby the local radiation safety officer.

NOTE: Experiments involving RNA require careful precautions to prevent contaminationand degradation by RNases (see APPENDIX 2A). The water used to prepare all solutions andbuffers should be RNase-free or treated with diethylpyrocarbonate (DEPC; APPENDIX 2A).Use sterile, disposable plasticware where possible.

Current Protocols in Nucleic Acid Chemistry 9.6.1-9.6.33, March 2014Published online March 2014 in Wiley Online Library (wileyonlinelibrary.com).DOI: 10.1002/0471142700.nc0906s56Copyright C© 2014 John Wiley & Sons, Inc.

CombinatorialMethods inNucleic AcidChemistry

9.6.1

Supplement 56

STRATEGIC PLANNING

Advantages of Using Modified Nucleotides in In Vitro Selections

As discussed in other units (UNITS 9.3, 9.4, & 9.5), the desired outcome of a selection ex-periment should be determined in advance and the selection strategy should be designedaccordingly. The decision to include modified nucleotides in a selection experiment andthe choice of which nucleotide modifications to use should be based on how the nu-cleotides might benefit the selection (Fig. 9.6.1). There are various advantages to usingmodified nucleotides, such as the increased stability of the selected nucleic acid againstnuclease degradation, expanded chemical functionality, and/or improved aptamer bind-ing affinity (recently reviewed in Wilson and Keefe, 2006; Keefe and Cload, 2008; Belland Micklefield, 2009; and Mayer, 2009).

When nucleic acids are utilized as therapeutics or diagnostics, nuclease degradationis a common problem. Whereas unmodified nucleic acids are extremely susceptibleto nuclease degradation, modification of the phosphate backbone or ribose moiety canrender nucleic acids much more impervious to cleavage. For example, the incorporationof pyrimidine nucleosides modified at the 2′-position on the ribose moiety with aminoor fluoro functional groups has been shown to drastically increase the stabilities oftranscribed RNA molecules, in large measure because most ribonucleases polarize the2′-hydroxyl group to attack the phosphodiester linkage. Numerous examples of nuclease-resistant aptamers and ribozymes have been published (Table 9.6.1), including fully 2′-modified aptamers selected against vascular endothelial growth factor (Burmeister et al.,2005) and tissue factor pathway inhibitor (Waters et al., 2011).

nucleasestability

less susceptible tonucleophilic attack

• phosphorothioate• locked nucleic acids

Exa

mp

les

of

mo

dif

ied

nu

cleo

tid

esTy

pe

of

mo

dif

icat

ion

Go

alE

nh

ance

• 2� O-methyl NTPs• 2� fluoro NTPs• 2� amine NTPs• 4� thiol NTPs

• 5-position on base• backbone composition• 2� amine NTPs

• 2� amine NTPs• 5-position on base

expanded alphabetexpanded chemistries

enhanced catalytic activityoffer conjugation moieties

improved hit rateimproved specificity

chemicalfunctionality

bindingaffinity

O

O

O

N O

O

O

O OP

R

OH

OH

NH43

21

1�

2�3�

4�

5�

6

5

O

O OP

O–

O–

U

O

O

O

N O

O

O

O OP

R

OH

OH

NH43

21

1�

2�3�

4�

5�

6

5

O

O OP

O–

O–

U

O

O

O

N O

O

O

O OP

R

OH

OH

NH43

21

1�

2�3�

4�

5�

6

5

O

O OP

O–

O–

U

Figure 9.6.1 Overview of the properties, advantages, and examples of modified nucleotides used in in vitro selections.For specific examples, refer to Table 9.6.1.

9.6.2

Supplement 56 Current Protocols in Nucleic Acid Chemistry

Table 9.6.1 Modified Nucleotides and Examples of Successful In Vitro Aptamer or Ribozyme/Deoxyribozyme Selections

Modifiednucleotidea Polymerasea Target/function (reference)b Notes

Modified sugars

2′-Aminopyrimidinesa

Y639F T7 RNApolymerasea

VEGF/VPF (Green et al.,1995), Human neutrophilelastase (Lin et al., 1994),bFGF (Jellinek et al., 1995),hKGF (Pagratis et al., 1997),IFN-γ (Kubik et al., 1997),and Ribozyme–transcleavage of RNA (Beaudryet al., 2000)

Several 2′-amino pyrimidine modifiedRNA aptamers have encounteredsynthesis difficulties and thereforehave been abandoned as therapeuticcandidates (reviewed in Keefe andCload, 2008).

2′-Fluoropyrimidinesa

Y639F T7 RNApolymerasea

VEGF 165 (Ruckman et al.,1998; Chakravarthy et al.,2006), hKGF (Pagratis et al.,1997), and IFN-γ (Kubiket al., 1997)

VEGF 165 aptamer is the first humanaptamer therapeutic (Macugen). Theaptamer was post modified afterselection to include 2′ O-methylnucleotides in a process called “backfilling.”

2′-O-Methylnucleotidesa

Y693F/H784A (Briebaand Sousa, 2000),“RGVG”(Chelliserrykattil andEllington, 2004),“2P16” (Siegmundet al., 2012), or R425CT7 RNA polymerase(Ibach et al., 2013)

VEGF 165 (Burmeister et al.,2005)

This nucleotide is less costly tosynthesize than 2′-amino or 2′-fluoronucleotides. Also, 2′-O-methyl is acommon post-transcriptionalmodification; for example, there are>100 2′-O-methyl nucleotides in eachribosome (Maden, 1990). Thus, thefact that these nucleotides arenaturally occurring may make FDAapproval more easily attainable. All 23nucleotides in anti-VEGF aptamer are2′-O-methyl nucleotides (Burmeisteret al., 2005).

2′ Compositions,mixed

Y639F/H784A/K378RT7 RNA polymerase

Thrombin and IL-23(Burmeister et al., 2006)

dRmY (dA/dG/mU/mC) was thechosen composition for in vitroselection, although others wereconsidered.

TFPI (Waters et al., 2011) dCmD (mA/mG/mU/dC) was thepublished composition for in vitroselection, although fRmY (fA/fG/mU/mC) was used in a patentedvariant against the same target.

2′-O,4′-C-Methylene-bridgedATP, GTP, TTP(locked nucleicacids, LNAs) and5-methyl-CTPa

KOD Dash DNApolymerasea and KODmutant (KOD 2) DNApolymerase (Kuwaharaet al., 2000) (19:1 ratioof polymerases)(Kasahara et al., 2013)

Thrombin (Kasahara et al.,2013)

KOD Dash DNA polymerase is amixture of a KOD polymerase and anarchaeal DNA polymerase withproofreading activity (available fromToyobo Co.). In the selection,2′-O,4′-C-methylene bridged (B/L)TTP was incorporated enzymatically,whereas the B/L ATP, B/L GTP, and5-methyl-CTP were present only inthe pool (forward) primer region(Kasahara et al., 2013).

continued

9.6.3

Current Protocols in Nucleic Acid Chemistry Supplement 56

Table 9.6.1 Modified Nucleotides and Examples of Successful In Vitro Aptamer or Ribozyme/Deoxyribozyme Selections,continued


4′-Thiopyrimidinesa

T7 RNApolymerasea

Thrombin (Kato et al., 2005) The thio-modified aptamer is 50 timesmore stable in the presence of RNaseA and has an increase in thrombininhibition compared to thecorresponding unmodified RNAaptamer (Kato et al., 2005).

Modified bases

5-(3-Aminoally)deoxycytidinea

Vent (exo-) DNApolymerasea

RNase DNAzyme(Hollenstein et al., 2013)

This modified DNAzyme containsthree modified nucleotides(5-guanidinoallyl-dUTP,5-aminoallyl-dCTP, and5-imidazolyl-dATP) and cleavesall-RNA targets independent of M2+

(Hollenstein et al., 2013).

5-(3-Aminoallyl)deoxyuridinea

Sequenase 2.0 DNApolymerasea

RNase DNAzyme, sequencedirected (Perrin et al., 2001)

This RNase DNAzyme contains boththe 8-[2-(4-Imidazolyl)ethylamino]deoxyadenosine and the5-(3-aminoallyl) deoxyuridinemodified nucleotides and self-cleavedthe internal rC. This is the firstexample of a metal-independentDNAzyme (Perrin et al., 2001).

5-N-(6-Aminohexyl)carbamoylmethyldeoxyuridine

KOD DNApolymerasea

Thalidomide (Shoji et al.,2007)

This aptamer is highly specific for theR-enantiomer of thalidomide (Shojiet al., 2007).

5-(3-Aminopropynyl)deoxyuridine

Vent DNApolymerasea

ATP, ADP, and AMP(Battersby et al., 1999)

This is the first example of theincorporation of a positively chargedfunctional group (Battersby et al.,1999).

5-Benzylamino-carbonyldeoxyuridine(BndU) (Vaughtet al., 2010)

KOD DNApolymerasea

Plasminogen ActivatorInhibitor-1 (PAI-1), as wellas multiple other targets(Gold et al., 2010)

Modified nucleotides were used toselect against thirteen human proteins(“difficult targets”) for whichunmodified RNA/DNA in vitroselection did not yield an aptamer. Theextent to which a given nucleotidemodification was beneficial for aselection was highly dependent on theprotein target (Gold et al., 2010).

5-Boronicacid-modifiedthymidine

Taq DNApolymerasea

Fibrinogen (Li et al., 2008)

5-Bromodeoxyuridinea

Taq DNApolymerasea (Brodyet al., 1999; Goldenet al., 2000) or E.coli DNApolymerase Ia

(Smith et al., 2003)

bFGF (Brody et al., 1999;Golden et al., 2000) andGP120MN (Smith et al.,2003)

Photocrosslinking aptamers weregenerated with this modifiednucleotide.

continued

9.6.4




5-Carboxamide-modifieddeoxyuridine(Vaught et al., 2010)

Deep Vent and KODXL DNApolymerasesa

Tumor necrosis factorreceptor super familymember 9 (TNFRSF9)(Vaught et al., 2010)

PCR amplification could not be performedusing any of the carboxide derivatives ofdUTP, only primer extensions. Therefore,an intermediate PCR step usingunmodified nucleotides was required toexponentially amplify, and thensubsequent primer extension reactionsreincorporated the modified nucleotides.Additionally, amide linkages increase thepossibility of hydrogen bonding with thetarget (Vaught et al., 2010).

5-Guanidinoallyldeoxyuridinea



See Notes for 4-(3-aminoallyl)deoxycytidine.

5-Isobutylamino-carbonyldeoxyuridine(iBudU) (Vaughtet al., 2010)

KOD DNApolymerasea

Human mobility group-1(HMG-1), as well as multipleother targets (Gold et al.,2010)

See Notes for 5-benzylaminocarbonyldeoxyuridine.

5-Imidazolyldeoxyadenosinea



See notes for 4-(3-aminoallyl)deoxycytidine.

5-Imidizole uridine T7 RNApolymerasea

Ribozyme with amidesynthase activity (Wiegandet al., 1997); ribozyme withurea synthase activity(Nieuwlandt et al., 2003)

A side-by-side comparison of a5-imidazol uridine RNA pool versus anunmodified RNA pool used in a selectionfor urea bond catalysts resulted withsignificant catalytic activity of themodified pool after nine rounds ofselection, while no significant increase incatalytic activity was observed overbackground with the unmodified RNAselection after fourteens rounds ofselection (Nieuwlandt et al., 2003)

5-Imidizole-uridineanalogue (unnamed)

Thermostable DNApolymerasesa

(possibly Taq, Vent,Pfu, and rTh DNApolymerases)

RNase DNAzyme, sequencedirected (Santoro et al., 2000)

5-Naphtylmethy-laminocarbonyldeoxyuridine(NapdU)(Vaught et al., 2010)

KOD DNApolymerasea

Human protein targets (Goldet al., 2010)

See notes for 5-benzylaminocarbonyldeoxyuridine.

5-(2-Naphtylmethy-laminocarbonyl)deoxyuridine(2NapdU) (Vaughtet al., 2010 andOchsner et al., 2013)

KOD DNApolymerasea

C. difficile toxins (Ochsneret al., 2013)

continued

9.6.5




5-(1-Pentynyl)deoxyuridine

Vent DNApolymerasea

Thrombin (Latham et al.,1994)

5-Phenylethylduoxyuridine(PEdU) (Vaughtet al., 2010 andOchsner et al., 2013)

KOD DNApolymerasea


5-Pyridylmethyl-carboxamid uridine

Not noted in paper Ribozyme withDiels-Alderase activity(Tarasow et al., 1997)

5-Tyrosyldeoxyuridine(TyrdU) (Vaughtet al., 2010 andOchsner et al., 2013)

KOD DNApolymerasea


5-Tryptaminocarbonyldeoxyuridine(TrpdU) (Vaughtet al., 2010)

KOD DNApolymerasea

Fractalkine(CX3CL-1), aswell as multiple other targets(Gold et al., 2010)

See Notes for 5-benzylaminocarbonyldeoxyuridine.

6-Aminohexyladenosine

T7 RNApolymerasea

Ribozyme with ligase activity(Teramoto et al., 2000)

The ribozyme catalyzed the ligation to its5′ end (Teramoto et al., 2000).

7-(2-Thienyl)imidazo[4,5-b]pyridinea

AccuPrime Pfx DNApolymerasea

VEGF 165 and IFN-γ(Kimoto et al., 2013)

Modified nucleotide exclusively pairs withdiol-modified 2-nitro-4-propynlpyrrole, inessence creating a third base pair. Themodified aptamers have >100-foldbinding affinity over the unmodifiedaptamers (Kimoto et al., 2013).

8-[2-(4-Imidazolyl)ethylamino]deoxyadenosine

Sequenase 2.0 DNApolymerasea

RNase DNAzyme, sequencedirected (Perrin et al., 2001)

See Notes for 5-(3-aminoallyl)deoxyuridine.

Modified phosphate backbone

Boranophosphatelinkages

T7 RNApolymerasea

ATP (Lato et al., 2002)

Phosphorothioate-linked DNAa

(S-linked dNTPs)

Taq DNApolymerasea

NF-kB RelA (p65) (Kinget al., 2002), p50 (King et al.,2002), and NF-IL6 (Kinget al., 1998)

These modifications are nuclease resistantand efficiently internalized by cells (Kinget al., 2002).

Phosphorothioate-linked RNAa

(S-linked NTPs)

T7 RNApolymerasea

bFGF (Jhaveri et al., 1998)

aDenotes the modified nucleotide or polymerase is commercially available.bAbbreviations: bFGF, basic fibroblast growth factor; hKGF, human keratinocyte growth factor; IFN-γ , interferon γ ; NF-IL6, nuclear factor for humaninterleukin 6; VEGF, vascular endothelial growth factor; VPF, vascular permeability factor.

9.6.6


Similarly, substitutions on the nucleic acid backbone, such as replacing the phosphatewith a phosphorothioate (mono- or di-; Zon and Geiser, 1991) or linking the 2′- and4′-positions of the ribose (reviewed in Veedu and Wengel, 2010), have also been shownto increase oligonucleotide stability in the presence of nucleases. An additional benefit isthat phosphorothioate nucleotides have been shown to be incorporated into an elongatingtranscript by T7 RNA polymerase with little or no increase in KM (Griffiths et al., 1987).While DNA is not as vulnerable to hydrolysis as RNA, it is nonetheless susceptible tocleavage by a variety of deoxyribonucleases and phosphodiesterases. The stability ofDNA can also be increased by the incorporation of phosphorothioate nucleotides, andthese can be readily incorporated by Taq DNA polymerase (Nakamaye et al., 1988) andused for selection (King et al., 1998).

In addition to enhancing nuclease resistance, modified nucleotides potentially expandthe chemical functionality of nucleic acids. Modified nucleotides have been includedin selections for catalytic nucleic acids. The resultant catalysts have been shown to behighly dependent upon the modifications for activity (Tarasow et al., 1997; Wiegandet al., 1997; Beaudry et al., 2000; Santoro et al., 2000) and, in some examples, poten-tially improve the catalytic activity. For example, Nieuwlandt et al. (2003) selected amodified (5-imidazolyluridine) ribozyme that catalyzed the conjugation of the 3′-aminogroup of 3′-amino-3′-deoxycytidine to the N-terminus of a tripeptide substrate, form-ing a urea linkage. In the same work, a parallel unmodified selection was performed.Significant catalytic activity was observed from the modified pool after nine rounds ofselection, but no significant increase in catalytic activity was seen in the unmodified RNAselection even after fourteen rounds of selection. In another example, Hollenstein et al.(2013) selected a metal-independent RNase-active deoxyribozyme containing modifiedcytidine, uridine, and adenine nucleosides (5-guanidinoallyl-dU, 5-aminoallyl-dC, and8–2-(4-imidazolyl)ethylamino-dA) from a doped sequence pool that was in turn basedon a previously identified unmodified deoxyribozyme. According to the first-order rateconstants, this modified deoxyribonuclease had improved reaction kinetics over similar,unmodified ribozymes (Faulhammer and Famulok, 1997; Geyer and Sen, 1997). Theidentification of magnesium (Mg2+)-independent catalysts may have been due to theinclusion of modified nucleotides or may have just been due to further selection; parallelexperiments with an unmodified pool were not performed. However, when a similarselection was performed using a deep, random pool (i.e., not doped) as a starting point,no RNase activity was found after nine rounds of selection.

In line with the hypothesis that modified nucleotides expand the functional repertoireof nucleic acids, there are multiple examples of modification-dependent aptamers ordeoxyribozymes (i.e., binding/activity is lost when modified nucleotides are replacedwith unmodified nucleotides). For instance, the first RNA capable of catalyzing theformation of carbon-carbon bonds utilized 5-pyridylmethylcarboxamid-UTP in placeof UTP (Tarasow et al., 1997). When the most active isolate from this selection wastranscribed with unmodified UTP, it was inactive. Similar results were obtained for aribozyme that catalyzed amide bond formation (Wiegand et al., 1997), whose activitywas dependent on the incorporation of 5-imidazolyl-UTP. Additionally, two sequence-specific RNase deoxyribozymes were dependent on the incorporation of a 5-imidazolyl-dUTP (Santoro et al., 2000) and both 8–2-(4-imidazolyl)ethylamino-2′-dATP and 5-(3-aminoallyl)-2′-dUTP (Perrin et al., 2001). However, none of the sequences, motifs, oractivities found in these selection experiments was directly compared with ribozymesthat contained canonical nucleotides and that were sieved from the same pool using thesame selection conditions.

Unexpectedly, there are at least a few conflicting examples and counterexamples suggest-ing that modified nucleotides do not greatly contribute to binding or catalysis relative to


9.6.7


unmodified nucleotides. Santoro and colleagues (Santoro and Joyce, 1997; Santoro et al.,2000) selected deoxyribozymes with RNA hydrolysis activity from different aliquots ofthe same, unamplified random sequence pool. According to the second-order rate con-stants, the selection performed with unmodified nucleotides produced a much fastercatalyst (Santoro and Joyce, 1997) than that identified through the modified selection(Santoro et al., 2000). Ultimately, it is unknown whether this indicates the superiorityof unmodified nucleotides for this pool and this function, or whether the fraction of theoriginal pool used for the selection of the unmodified catalyst just contained a “jackpot”sequence. In another contrasting example, whereas Tarasow et al. (1997) suggested thatthe inclusion of a modified nucleotide was the only reason they were able to select aDiels-Alder synthetase, Seelig and Jaschke (1999) later selected ribozymes from an un-modified pool. The modified selection yielded a catalyst with a kcat/KM of �4 M−1sec−1,while the unmodified selection yielded a catalyst with a kcat/KM of 167 M−1sec−1. How-ever, these selections were performed by different research groups with different pools,and thus, again, are not directly comparable. For a final example, see Critical Parameters.

To the extent that a modified nucleotide will be included to enhance catalytic functionality,the choice of modification should complement the desired function. For example, animidizole ring (with a pKa near neutrality; Battersby et al., 1999) may be beneficial ina pool to be screened for catalysis of an acid/base reaction, while the incorporation of athiolated residue (Jhaveri et al., 1998) could allow nucleic acids to participate in disulfidebond formation or rearrangement, reactions normally denied to them. Additionally, 5-bromo and 5-iodo modifications of pyrimidine bases can serve as moieties for UVcrosslinking to covalently link a modified aptamer to its targeted protein. In those caseswhere the target protein is not already known, this can lead to biomarker identification(Mallikaratchy et al., 2007). Crosslinking can also be used to increase the stringency ofwashing and thereby decrease background (Brody et al., 1999).

While most of the modifications mentioned thus far retain native Watson-Crick pair-ing, this may not always be the case. Introducing modifications with novel base pairingmay potentially provide additional chemical and functional properties, unrestricted byunmodified nucleotide base pairing. For example, Piccirilli et al. (1990) have shownthat diaminopyrimidine and xanthosine can form a stable base pair, although the fi-delity of the diaminopyrimidine during the polymerase chain reaction (PCR) was poor(Southworth et al., 1996). The Benner group developed an unnatural base pair (6-amino-5-nitro-3-(1′-β-d-2′-deoxyribofuranosyl)-2(1H)-pyridone and 2-amino-8-(1′-β-D-2′-deoxyribofuranosyl)-imidazo[1,2-α]-1,3,5-triazin-4(8H)-one) that is retained duringPCR (Yang et al., 2011), and thus is suitable for in vitro selection. Furthermore, Kimotoet al. (2013) have introduced the hydrophobic base pair 7-(2-thienyl)imidazole[4,5-b]pyridine and diol-modified 2-nitro-4-propynylpyrrole into selection experiments. No-tably, these modified aptamers against human vascular endothelial cell growth factor-165(VEGF-165) and interferon-γ (IFN-γ) have >100-fold improved dissociation constantsover aptamers with unmodified nucleotides.

Modified nucleotide pools can also potentially increase the overall binding affinities ofselected aptamers. For example, Vaught et al. (2010) selected a carboxamide-modifieddUTP aptamer against tumor necrosis factor superfamily number 9 (TNFRSF9); pre-viously unmodified DNA selections had failed. The increased affinity of this aptamerrelative to similar RNA aptamers suggested that the modifications offered some bindingadvantage. Similar enhancements of modified aptamer affinities relative to unmodifiedaptamers have also been observed for the human α-thrombin aptamer (Kato et al., 2005),vascular endothelial growth factor (VEGF) aptamer, and interferon γ aptamer (Kimotoet al., 2013).

In Vitro SelectionUsing Modified

or UnnaturalNucleotides

9.6.8


Finally, selections that targeted thirteen “difficult” human protein targets, which havepreviously failed to yield aptamers with native DNA pools, yielded high-affinity ap-tamers when the selections were carried out with pools that contained a modifica-tion at the 5-position of uridine. Across nearly 1500 human proteins, pools con-taining 5-tryptaminocarbonyl-dU or 5-benzylaminocarbonyl-dU enhanced the successrate of selection to >84% compared to selections containing the parental dU or dTnucleotide. The extent to which a given nucleotide modification was beneficial fora selection was highly dependent on the protein target (Gold et al., 2010). How-ever, there were still some proteins in the repertoire that failed to produce any ap-tamers, suggesting that improvements in methods and novel modifications could still bebeneficial.

Preparation and Use of Modified Nucleotides

This unit presents procedures in two main sections: determination of the suitability ofmodified nucleotides for in vitro selection, followed by the determination of the mod-ified pool most suitable for in vitro aptamer selection. Once the decision has beenmade that a modified nucleotide is necessary (see Fig. 9.6.1), the main considerationis how to incorporate the given modification (Fig. 9.6.2) and whether the given nucle-oside triphosphate analog can be reproducibly and specifically introduced (Fig. 9.6.3).For instance, can the modification be recognized and incorporated by a polymerase(i.e., written into the transcript)? If incorporated, can the modified nucleoside analoguebe specifically recognized by a polymerase and interpreted correctly without causingmutations (i.e., read from the transcript)? When potential enzymes have been identi-fied, Figure 9.6.3 describes a validation schema that can be utilized to ensure bothreproducibility and specificity of modified analogue incorporation. It will also be crit-ical to optimize reaction conditions to ensure sufficient yields for each round of theselection.

The protocols contained herein primarily focus on the enzymatic incorporation of mod-ified nucleotides and the use of RNA pools, which require reverse transcriptases, DNApolymerases, and transcriptases to regenerate the pools. For selections requiring onlyDNA polymerases, the protocols can be used as guidelines for designing similar work-flows with methods (e.g., PCR) described elsewhere (UNIT 9.2, Support Protocol 3 andBasic Protocol 2). Methods for chemical synthesis and screening without an amplificationstep are not covered in this unit.

The first set of procedures describes transcription reactions used to ensure that modifiednucleotides can act as substrates to produce full-length transcripts (see Basic Protocol1). Next, it is necessary to check the quality and prevalence of the modified nucleotidein the transcript (see Basic Protocol 2). Since selected transcripts must be amplified,it is necessary to verify that the modified transcript is a suitable template for reversetranscriptase (see Basic Protocol 3). If Basic Protocols 1, 2, and 3 produce successfulresults, then it is likely that the random sequence pool can make it through the preparativesteps leading to the selection. However, it is important to ensure the modification does notintroduce unwanted mutations or base conversions into the transcript (see Basic Protocol4). Finally, a mock selection (omitting the selective step) using a pool including themodifications should be performed with all the enzymatic steps in succession, and thenthe products should be sequenced to determine the fidelity of incorporation (see BasicProtocol 4 for a discussion). An overall outline of the process detailing key decisions ispresented in Figure 9.6.3.

Once a given modified nucleotide(s) has been validated for use in selection, the startingpool or pool candidates can be constructed, and trial selections can be performed. Inthe second set of protocols, double-stranded DNA that codes for the modified pool (as


9.6.9


target bindingpartitioning





transcription

transcriptionwith intermediate

PCRwith intermediates

synthesis withunmodified

phosphoramidites

Ch

emic

al s

ynth

esis

mod

ified

pho

spho

ram

idite

or c

onju

gatio

n po

st-s

ynth

esis

Co

nju

gat

ion

coup

ling

to m

odifi

edin

term

edia

te s

uch

as 2

� am

ine

En

zym

atic

dire

ct in

corp

orat

ion

with

mod

ified

nuc

leos

ide

trip

hosp

hate

assay to compareaffinity, functionality,stability, specificity

with unmodifiednucleoside triphosphates

in vitroselection

synthesis withmodified

phosphoramidites

modified nucleotide/substrate incorporation

modified oligonucleotideused as template

primerextension

substrateconjugation

substrateconjugation

PCRreverse

transcription

PCRreversetranscription

PCR

PCR

Figure 9.6.2 Mechanisms to incorporate modified nucleotides into in vitro selections. Three mechanisms are routinelyused to incorporate modified nucleotides into in vitro selections: enzymatic, conjugation, and chemical synthesis. Thesemechanisms and examples are discussed in detail in the Critical Parameters section. In this figure, the light grey bubbleindicates that a modified nucleotide will serve as a template. The dark grey bubble indicates that the modified nucleotidewill be the substrate for incorporation. If the modified nucleotide can serve both as a template and also be incorporatedby the same enzyme, the bubble is shaded from light to dark grey.

generated in UNIT 9.2) is first transcribed, and the products are gel purified using theoptimized set of conditions developed from Basic Protocols 1 and 3. Pool candidatespossessing different modifications can be tested side-by-side to determine which is bestfor a given target (see Basic Protocol 5). The sequential step in the overall selectionscheme, in vitro selections, is described in detail in UNITS 9.3, 9.4, & 9.5.



9.6.10


nono

yes

*Can the

conditions beoptimized?

[BP1]

*Are other

enzymes available?[BP1]

At theexpected ratio?

[BP2]

Can thetranscripts be read

by polymerase?[BP3]

Is the yieldsufficient?

[BP1 in UNIT 9.3]

STARTCan full lengthtranscripts be

obtained from adefined sequence?

[BP1]

Is the yield highenough for selection?

[BP1 in UNIT 9.3]

*Are other

enzymes available?[BP3]

*Can the conditions

be optimized?[BP3]

Is the clonemutated basedon the parent?

[BP4]

Does thesequence contain thedesired modification?

[BP2]

*Can the

modification beconjugatedinstead?

ENDThe modification must be

incorporated after selection

Is replicationfidelity maintained inthe pool compared to

native parent?[BP4]

Candifferent modificationsaccomplish the same

function?

yes

yes

yes

yes

yes

• incubation time• incubation temperature• reagent concentrations• additional reagents• increased reaction volume• promoter sequence

Conditions *

Legend

Start/Stop decision

* denotes guidelines inCritical Parameters

[BP#] basic protocol

yes

yes

yes

yes

yes

yes

yesyes

no nono

no no

no

no

no

no

no no

no

ENDCarry out the selections as

described in UNITS 9.3, 9.4, and/or9.5, synthesize binding variantswith the modifications and assay

for modifications and assayfor enhanced functionality.

Figure 9.6.3 Evaluation and optimization schema to validate the successful incorporation of modified nucleotidesinto a pool for in vitro RNA selection.

DETERMINATION OF SUITABILITY OF MODIFIED NUCLEOTIDES FORIN VITRO SELECTION

BASICPROTOCOL 1

Evaluation and Optimization of Transcription with a Modified Nucleotide

In the case of RNA selections, the ability of a modified nucleotide to be incorporated intoa transcript should be evaluated. It is best to perform a transcription reaction in which themodified nucleotide completely substitutes for its unmodified counterpart (e.g., 2′-deoxy-2′-fluorocytidine in place of cytidine) using a template of known sequence composition(initially). Once incorporation of the nucleotide has been validated, a small aliquot ofpool template should be used as template to verify fidelity of incorporation, thereby re-serving most of the pool for the optimized, large-scale reaction and selections. A positivecontrol reaction should always be performed in parallel on the same template with allunmodified nucleotides. Similarly, a negative control reaction, void of the modified nu-cleotide and unmodified counterpart, should always be performed on the same template.The comparative success of these reactions can be determined by including α-32P-labelednucleotides in the reaction mix, separating transcripts by denaturing polyacrylamide gelelectrophoresis, and visualizing the full-length transcripts with a phosphorimager or viaautoradiography.

This protocol also includes optimizing the transcription reaction using different modifiedand unmodified nucleotide concentrations, buffer components, starting double-strandedtemplate concentration, and different combinations of modified nucleotides (see CriticalParameters for examples and references). If, after optimization, yields (UNIT 9.3, BasicProtocol 1) are not high enough for selection (see UNIT 9.3, Basic Protocol 2), then consider


9.6.11


such options such as scaling up the reaction size, other modifications, or perhaps a singleround of selection (or screening), followed by high-throughput sequencing and sequenceanalysis for aptamer candidates (see Fig. 9.6.3 for a full range of options and validationschema).

Materials

AmpliScribe T7 high-yield transcription kit (Epicentre) containing:10× transcription (TNX) bufferT7 enzyme solution100 mM ATP, CTP, GTP, and UTP solutions10× reaction mix100 mM dithiothreitol (DTT)T7 RNA polymerase

DuraScribe T7 transcription kit (Epicentre) for modified NTP incorporation, kitincludes:

DuraScribe T7 enzyme solution50 mM ATP and GTP solutions50 mM 2′-deoxy-2′-fluoro-CTP and 2′-deoxy-2′-fluoro-UTP solutions10× reaction mix100 mM dithiothreitol (DTT)

100 mM modified nucleotide500 ng DNA clone of known sequence composition (dsDNA containing a T7

promoter)3000 Ci/mmol α-32P-labeled ATP (e.g., Perkin Elmer)Nuclease-free water2× denaturing stop dye (see recipe)10% (w/v) denaturing polyacrylamide gel (see recipe in APPENDIX 3B), 0.8-mm thick500 ng purified DNA pool or randomized clone per reaction (dsDNA containing a

T7 promoter) (UNIT 9.2)

Thermal cycler or 37°C to 42°C and 70°C water baths or heating blocksGel blotting paper (Bio-Rad)Plastic wrapGel dryer with vacuum (e.g., Bio-Rad)Phosphorimager screenPhosphorimager and image analysis software (e.g., GE Healthcare Life Sciences,

ImageQuant)

Additional reagents and equipment for denaturing polyacrylamide gelelectrophoresis (PAGE; APPENDIX 3B)

Perform transcription with modified and unmodified nucleotides

1. Prepare three transcription reactions: (1) In the experimental reaction, completelysubstitute the modified nucleotide for its unmodified counterpart (e.g., 2′-deoxy-2′-fluorocytidine in place of cytidine). (2) In the positive control reaction, use allunmodified nucleotides. (3) In the negative control reaction, use the modified nu-cleotide and omit unmodified counterpart. Assemble the transcription reactionsaccording to the kit manufacturer’s instructions (see also UNIT 9.3, Basic Protocol 1).Combine the following in three separate tubes (20 μL total volume):

2 μL 10 × transcription (TNX) buffer (AmpliScribe kit)2 μL 100 mM DTT (AmpliScribe or DuraScribe kit)1.5 μL each NTP (AmpliScribe kit) or modified nucleotideIn Vitro Selection

Using Modifiedor Unnatural

Nucleotides

9.6.12


4 μL (�500 ng) DNA clone of known sequence composition4 μL water1.5 μL T7 RNA polymerase (AmpliScribe or DuraScribe kit)0.5 μL 3000 Ci/mmol α-32P-labeled ATP (volume varies based on age and

subsequent activity of radiolabeled ATP)

The AmpliScribe kit (Epicentre), containing T7 RNA polymerase, is a high-yield, unmod-ified transcription kit. More affordable, conventional transcription reagents may be usedin place of the AmpliScribe kit; however, lower yields may result.

The DuraScribe kit (Epicentre), containing a proprietary preparation of the Y639Fvariant of T7 RNA polymerase, is a transcription kit optimized for the incorporation of2′-deoxy-2´-fluoro-CTP, 2′-deoxy-2´-fluoro-UTP, ATP, and GTP. Alternatively, other oradditional polymerases may be tested for modified nucleotide incorporation (see CriticalParameters and multiple examples provided in Table 9.6.1).

Many mutant T7 polymerases are not as active as their wild-type counterparts, andtherefore may require additional time, temperature, volume, and buffer considerationsthat should be optimized as described in Critical Parameters.

2. Incubate at 37°C to 42°C.

If α-32Phosphate substrates are not suitable or available (e.g., if all unmodified nu-cleotides are replaced by a modification), the polyacrylamide gel can be stained witha dye such as SYBR gold, or the template can be end-labeled with γ -32P. However, anadditional DNase step (such as with the use of Epicentre Baseline Zero DNase) will berequired to digest template DNA, thus permitting just the detection of the RNA transcripts.

If the template is the product of an extension or PCR reaction, care should be taken toremove the unincorporated dNTPs or assay extension efficiency in the absence of onenucleotide (preferably the modified nucleotide).

3. At several time points (e.g., 0.25, 0.5, 1, and 2 hr), gently mix and briefly centrifugethe sample to bring down the liquid. Transfer 1 μL of the reaction to separate tubescontaining 4 μL nuclease-free water and 5 μL of 2× denaturing stop dye.

Mixing and centrifuging the reaction ensures that each sample will contain a consistentportion of the reaction. The accumulation of condensation and the precipitation ofpyrophosphate during transcription can sometimes lead to variations in reaction volumesor concentrations.

4. Denature the time point samples by heating 3 min to 70°C and separate transcriptsfrom aborted transcripts and unincorporated nucleotides on a 0.8-mm thick, 10%denaturing polyacrylamide gel (APPENDIX 3B).

The concentration of acrylamide can be varied to efficiently separate different lengthproducts, depending on their sizes (APPENDIX 3B).

Pre-running the gel will heat it up and help to denature molecules that contain significantsecondary structure.

Analyze transcription reactions

5. Remove one of the two glass plates and place a sheet of gel blotting paper againstthe gel. Peel the gel off of the other glass plate (gel will stick to paper) and cover itwith plastic wrap. Dry gel under heat and vacuum using a commercial gel dryer.

Drying the gel can be omitted, and a wet gel can be directly used for exposure. If this isdone, leave the gel against one glass plate and carefully wrap with plastic wrap to avoidcontaminating the phosphorimager screen and the exposure cassette with radiation.

6. Expose the dried gel to a phosphorimager screen for 1 hr.CombinatorialMethods inNucleic AcidChemistry

9.6.13


The exposure time may need to be increased or decreased, depending on the specificactivity of the labeled nucleotide and the amount of transcript that is produced.

Gels can also be exposed to X-ray film, although this makes quantitation somewhat morecumbersome. A standard laboratory densitometer can be used for quantitation.

7. Develop the phosphorimager screen and calculate the relative amounts of RNAtranscripts produced using image analysis software such as ImageQuant.

8. Compare the intensities of product bands for the transcription reaction that includesthe modified nucleotide to those for transcription with the normal nucleotide todetermine the relative efficiency of incorporation for the modified nucleotide underthe conditions tested.

9. Plot the amount of product that accumulates versus time to determine the optimaltime for further transcriptions.

Optimize transcription reaction

10. Following the validation schema outlined in Figure 9.6.3, repeat steps 1 to 7 usingDNA pool or randomized clone (�500 ng) as template and different modified andunmodified nucleotide concentrations, buffer components, template concentrations,and modified nucleotide combinations (see Critical Parameters for examples andreferences). Take samples only at the optimal time determined in step 9.

If pool template is used for optimization, it is advisable to amplify only a small portionof the pool that will actually be used in the selection for these test transcriptions. As thisis a test reaction, do not commit the entire pool.

Pre-running the gel will heat it up and help to denature molecules that contain significantsecondary structure. Nonetheless, pools can sometimes appear “fuzzy” on a gel due tothe presence of molecules that differ slightly in length or that have different secondarystructures.

Since it is likely that the modified nucleotide will not work as well as the normal nucleotidebecause it binds to the polymerase more poorly, increasing the modified nucleotide con-centrations may be necessary to enhance transcription. When increasing the modifiednucleotide concentration, care should be taken that the amount of modified nucleotidedoes not unnecessarily restrict the magnesium concentration in solution. If the totalnucleotide concentration approaches that of magnesium, or the concentration of mod-ified nucleotide exceeds the normal concentration used for unmodified (i.e., 2.5 mM),then additional magnesium should be added to the reaction. Magnesium can, however,decrease the fidelity of the polymerase and should be assessed (see Basic Protocol 4).Alternatively, increasing the polymerase concentration may increase the efficiency of thePCR reaction incorporating modified nucleotides (see Kojima et al., 2013).

The initial sequence of the transcription product can also influence the incorporationof a modified nucleotide as can the promoter sequence when different polymerases areutilized (see Critical Parameters for examples and references).

5′ End-labeling of the transcription products with γ -32P-ATP in a kinase reaction (seeUNIT 9.3, Support Protocol 1) can also be utilized if internal labeling is undesirable.

11. Plot the amount of product (band intensity) versus modified nucleoside triphosphateconcentration or alternate reaction condition.

If the modified nucleotide is costly or difficult to produce, this step is important todetermine the minimum quantity of modified nucleotide to use to maximize the yield ofRNA. To conserve modified nucleotide, the lowest concentration of modified nucleotideshould be chosen at which near-maximal levels of RNA are produced.



9.6.14


BASICPROTOCOL 2

Confirmation of the Presence of Modified Nucleotides

The appearance of a full-length product in the transcription reaction (see Basic Protocol1) is encouraging, but does not indicate the level at which a modified nucleotide has beenincorporated into a pool. For example, full-length RNAs could be members of the originalpopulation that incorporated the modified nucleotide at only a few positions, could havebeen synthesized via contaminating unmodified nucleotides, could have been synthesizedvia the misincorporation of a unmodified nucleotide (e.g., via G:U mismatches), or couldbe the result of some combination of these possibilities. To determine with certaintythe level of modified nucleotide incorporation, transcription products should be isolated,fully digested, and separated by high-performance liquid chromatography (HPLC) toidentify the component nucleosides.

This protocol assumes that one has already gel-purified and calculated the yields ofthe full-length transcription products without radioisotopes (UNIT 9.3, Basic Protocol 1).The RNA will be digested to mononucleotides using nuclease P1 and then treated withalkaline phosphatase to generate nucleosides. Separation of individual nucleotides will becarried out by reversed-phase HPLC. Digestions will be compared to standards containingboth modified and unmodified nucleotides. Sample HPLC data are given in Figure 9.6.4(Robertson et al., 2001).

Materials

Transcribed RNA pools, modified and unmodified (see Basic Protocol 1 withoutradiolabeled ATP)

Nuclease P1 digestion mix (see recipe)Alkaline phosphatase reaction mix (see recipe)HPLC mobile phase solution: 5% (v/v) methanol in 0.1 M sodium phosphate, pH

6.0 (APPENDIX 2A)

Thermal cycler or 37°C and 50°C water bathsHPLC with reversed-phase C18 column (5 μm, 250 × 4.5–mm; Waters.) and UV

detector

1. Combine 200 pmol of each RNA pool (modified and unmodified) in separate reactionswith nuclease P1 digestion mix for a total reaction volume of 5 μL and incubate1.5 hr at 37°C.

This step will digest the nucleic acid into its component nucleotides. Nuclease P1 willdigest both RNA and DNA. It is possible that nuclease P1 will not digest after a specificmodified nucleotide. If incomplete digestion is seen in the HPLC results compared to anunmodified sample digestion, this in itself is convincing evidence of modified nucleotideincorporation.

2. Combine each entire nuclease P1 digestion reaction with alkaline phosphatase reac-tion mix in a final volume of 25 μL. Incubate 1.5 hr at 37°C, followed by 1 hr at50°C.

3. Inject 2.5 μL of one reaction onto a reversed-phase C18 HPLC column and separateusing a mobile phase flow rate of 1 mL/min and detection at 260 nm. Repeat with theother reaction.

If the pool initially had an equimolar mix of nucleotides, then the digested RNA shouldcontain �25% of the modified nucleoside and no peak corresponding to the replacednucleoside. The actual composition of pools can vary greatly, depending on the synthesisof the pool (e.g., see UNIT 9.2). Therefore, the composition of the original pool shouldfirst be determined either by sequencing individual clones (e.g., see Slatko et al., 1999)or by first digesting and separating a transcription product made solely from unmodifiednucleotides.


9.6.15


0

0

0.02

0.04

0.06

5 10

B

15 20Time (min)

A26

0 (a

u)

25

C

C

U

G

G

Ahm−U

A

30 35

A = 22.4% C = 25.1% G = 29.4% U = <0.5%hm-U = 23.1%

0

0.02

0.04

0.06

0.08A

A26

0 (a

u)

A = 22.0%C = 24.8%G = 29.0%U = 24.2%

0

0

0.02

0.04

0.06

5 10

D

15 20Time (min)

A26

0 (a

u)

25

Cim-U

G

A

30 35

A = 23.0% C = 26.1% G = 30.2% U = <0.5%pm-U = 20.7%

0

0.02

0.04

0.06

0.08

0.08 0.08

C

A26

0 (a

u)

A = 22.5% C = 25.5% G = 31.4% U = <0.5%im-U = 22.5%

G

A

pm-U

C

Figure 9.6.4 HPLC elution profile of pool RNA that has been treated with nuclease P1 andalkaline phosphatase as described in Basic Protocol 2. (A) Unmodified pool, (B) pool containing5-hydroxymethyluridine, (C) pool containing 5-imidazolemethyluridine, and (D) pool containing5-phenolmethyluridine (Robertson et al., 2001).

Nucleotides are detected at 260 nm as this is close to the absorbance maxima for theiraromatic bases. Nucleotides with heavily modified bases that disrupt their aromaticity mayabsorb at different maxima; however, A260 should still be used because the RNA will likelybe mostly unmodified.

The unmodified transcription reaction should be run separately through the same columnusing the same conditions to provide a series of standard retention times for comparison.This control can also be used to obtain the extinction coefficient of the modified nucleoside,if it is not otherwise known.

BASICPROTOCOL 3

Evaluation of Modified RNA as a Template for Reverse Transcriptase

During each round of in vitro selection, the selected RNAs must be amplified. Therefore,it is necessary to determine if an RNA transcript containing modified nucleotides willserve as a suitable template for reverse transcription (RT), the first step of the amplifi-cation process. A control reaction can be performed on the same template, except usingunmodified nucleotides. The relative success of the reverse transcription reactions canagain be visualized by incorporation of α-32P-labeled nucleotide, gel electrophoresis,and analysis on a phosphorimager.

The success of the reverse transcription reaction will depend primarily on the reversetranscriptase that is used. While reverse transcriptases, in general, tend to be somewhatforgiving with respect to template chemistry (e.g., they can recognize both RNA andDNA as templates), different reverse transcriptases may have distinct capabilities withmodified bases as substrates (for instance, higher reaction temperatures for structurallystable RNAs). If one reverse transcriptase does not prove efficient, then another one or acombination should be used. See Figure 9.6.3 for guidelines in optimizing this reaction.

Materials

Transcribed RNA pools, modified and unmodified (prepared as in Basic Protocol 1but without radiolabeled ATP)



9.6.16


100 μM reverse (3′-end) primer3000 Ci/mmol α-32P-labeled dATP (e.g., Perkin Elmer)10 mM dNTP mix (containing 10 mM each of dATP, dCTP, dGTP, and dTTP)ThermoScript reverse transcriptase kit (Invitrogen, Life Technologies) containing:

5× buffer100 mM DTTRNase OUT (ribonuclease inhibitor)Reverse transcriptase (RT) enzyme

2× denaturing stop dye (see recipe)10% (w/v) denaturing polyacrylamide gel (see recipe in APPENDIX 3B), 0.8-mm thick

Thermal cyclerGel blotting paper (Bio-Rad)Plastic wrapGel dryer with heat and vacuum (e.g., Bio-Rad)Phosphorimager screenPhosphorimagerImage analysis software (e.g., GE Healthcare Life Sciences ImageQuant)

Additional reagents and equipment for denaturing polyacrylamide gelelectrophoresis (PAGE; APPENDIX 3B)

Optimize RT reaction

1. Combine the following RT reactions in two separate tubes (13 μL total volume):

10.5 μL (10 pmol or �0.5 μg) transcribed RNA pool (modified in one tube,unmodified in the other)

1.0 μL 100 μM reverse (3′-end) primer0.5 μL α-32P-labeled dATP (volume varies based on the age and subsequent

activity of radiolabeled dATP)1.0 μL 10 mM dNTP mix

Do not use radiolabeled RNA as the input. It may be difficult to differentiate productfrom input if both are labeled. See UNIT 9.3, Basic Protocol 3, step 1 for suggested controlreactions.

2. Heat denature the RT reactions 5 min at 65°C in a thermal cycler and cool to roomtemperature over �10 min.

3. Add the following components to each reaction (20 μL total volume) and mix well:

4 μL 5× buffer (ThermoScript reverse transcription kit)1 μL 100 mM DTT (ThermoScript reverse transcription kit)1 μL RNaseOUT (40 U/μL, optional) (ThermoScript reverse transcription kit)1 μL ThermoScript reverse transcriptase (15 U/μL)

If optimizing difficult reverse transcription reactions, multiple reverse transcriptases (oreven combinations thereof) may be tested in parallel. In addition to ThermoScript (anAMV reverse transcriptase variant), AMV reverse transcriptase and SuperScript II reversetranscriptase (Invitrogen) have been used in modified nucleotide reverse transcriptionreactions.

RNA secondary structures may hinder the reverse transcription of some templates. Becauseof this, thermostable reverse transcriptases (such as ThermoScript) with reaction temper-atures at the primer annealing temperature are recommended to minimize secondarystructure formation. In extreme cases, a thermophilic DNA polymerase, Tth polymerase,has been shown to have significant reverse transcriptase activity and can potentially beused to copy recalcitrant RNA molecules.


9.6.17


4. Incubate reaction for 30 to 60 min at the primer annealing temperature (such as 50°C;up to 70°C). Heat inactivate the enzyme 5 min at 85°C.

5. Remove 2 μL and add it to 5 μL of 2× denaturing stop dye.

6. Denature by heating 3 min to 70°C and run on a 0.8-mm thick 10% denaturingpolyacrylamide gel (APPENDIX 3B).

A sample of radiolabeled RNA, as produced in the test transcriptions (see Basic Protocol1), can be used as a convenient size standard. RNA runs slightly slower on a gel than DNAof the same size.

The concentration of acrylamide can be varied to efficiently separate different products(APPENDIX 3B).

Analyze RT reaction

7. Remove one of the two glass plates. Place a sheet of gel blotting paper against thegel. Peel the gel off of the other glass plate (gel will stick to the paper) and cover withplastic wrap. Dry the gel in a gel dryer under heat and vacuum.

This latter step is not strictly necessary as a wet gel can be directly exposed to thephosphorimager plate. If this is to be done, carefully wrap the gel with plastic wrap toavoid contaminating the screen and exposure cassette with radiation.

8. Expose the dried gel to a phosphorimager screen for 1 hr.

The exposure time may need to be increased or decreased, depending on the specificactivity of the labeled nucleotide and the amount of transcript produced.

Alternatively, gels can be exposed to X-ray film, and the film developed to visualize bands.

9. Develop the phosphorimager screen and view the output with image analysis softwaresuch as ImageQuant.

Full-length cDNA products are desired from both modified and unmodified RNA templates.However, it is possible that smaller, discrete bands (i.e., incomplete cDNA products) willbe observed. These may be due to the failure of reverse transcriptase to read past RNAsecondary structures, which could artificially decrease the pool diversity in early roundsof selection. In this case, multiple reverse transcriptases should be examined (see step 3)to optimize the reaction for the modification.

BASICPROTOCOL 4

Determination of the Fidelity of Replication

If Basic Protocols 1, 2, and 3 give successful results, then it is likely that the randomsequence pool can make it through the preparative steps leading to the selection. How-ever, it is possible that the incorporation of modified nucleotides may alter the dynamicsof selection experiments in one of two ways: first, modified nucleotides may in and ofthemselves be extremely mutagenic; second, it may be difficult to either make full-lengthtranscripts that contain modified nucleotides or to fully copy transcripts containing modi-fied nucleotides. In the latter case, there will be an unseen selection against transcripts thatcontain modified nucleotides over the course of several cycles. This protocol determinesthe mutagenic or replicative potential of modified nucleotides in an in vitro selection.

To test the mutagenic potential, a single sequence or clone from the pool should besubjected to a “mock” round of selection (i.e., omitting the selective step). Refer toUNITS 9.3, 9.4, and/or 9.5 for in vitro selection protocols. Following the selection, �30clones should be sequenced (refer to Slatko et al., 1999 for sequencing protocol) and thenumber of mutations counted.

To test the replicative potential, compare the nucleotide ratios of the unselected pool toa pool participating in a “mock” round of selection (i.e., omitting the selective step).



9.6.18


To accomplish this, the pool is initially subjected to a round of “mock” selection (seeUNITS 9.3, 9.4, and/or 9.5 for selection protocols). Before and after the selection, �30 clonesshould be sequenced (see Slatko et al., 1999) and the base composition (i.e., relativeratio of nucleotides) of the pool should be analyzed compared to the initial ssDNA pooland RNA pool in which unmodified nucleotides were incorporated. In recent years, ithas become cost effective to perform high-throughput sequencing on the samples andcompare nucleotide ratios for many more sequences.

Through these protocols, it should be possible to isolate any bias in addition to theprocess responsible (also see UNIT 9.2, Support Protocol 2). If the nucleotide frequencyof the original pool is skewed, the entire pool should be resynthesized as describedin UNIT 9.2. If the nucleotide frequency of the pool incorporating modified nucleotidesis highly skewed relative to the original unmodified pool, then one of the enzymaticsteps should be investigated further. If the nucleotide composition is highly skewedafter the “mock” selection relative to before selection, then selection against a givenresidue may prove to be a problem. Even small skewing can prove to be significant overmany rounds of selection. For example, if a modified cytidine is present at 25% of therandom sequence positions in a starting pool, but is present at only 22.5% of the positionsfollowing one round of selection, after ten rounds of selection, its frequency may fall to0.25[(0.225/0.25)10] = 0.087, or <9% of the random sequence positions in a pool.

As mentioned previously, when examining pool bias, it may only be possible to detect sta-tistically significant differences in nucleotide frequencies by comparing high-throughputsequence data from sequences (perhaps hundreds of thousands). If such problems persistand inhibit the enrichment process, it may prove useful to perform a single round ofselection in the presence and absence of target, sequence clones, and use computationalanalyses to identify enriched motifs specific to incubation with the target that may besuggestive of function and that could either identify potential aptamer candidates forassay or assist with the design of a new, better pool.

DETERMINATION OF MODIFIED POOL MOST SUITABLE FOR IN VITROAPTAMER SELECTION

BASICPROTOCOL 5

Comparing Modified Nucleotide Pools and Selection Conditions

One of the key decisions to make when designing a selection experiment is to choose themodification (or set of modifications) to incorporate in the nucleic acid pool. Gold et al.(2010) showed that certain modifications work better than others with a given protein.Choosing a modified pool with an initially higher affinity for the target can potentiallyincrease the success of the selection. If choosing between equally beneficial modifiednucleotides that have passed each of the previous criteria (see Basic Protocols 1 through4), parallel “mini” selections of two to three rounds of selection (see UNITS 9.3, 9.4, and/or9.5) should be performed with the different modifications. Affinity assays can determinethe pool with the lowest dissociation constant to take forward in the full selection.Certain recalcitrant proteins may require more rounds of selection before a modified poolreveals preferential binding affinity. Alternatively, high-throughput sequencing, followedby sequence analysis and binding assays has previously been used to compare selectionconditions in early rounds when affinity assay detection limits prevent determination ofdifferences.

After performing two to three rounds of selection (see UNITS 9.3, 9.4, and/or 9.5) against atarget using different pool candidates, their dissociation constants should be estimatedusing the filter-based protocol provided below (also see UNIT 9.3, Support Protocol 2, forfilter-based assaying and UNIT 9.5, Support Protocol 1, for bead-based assaying). Combinatorial

Methods inNucleic AcidChemistry

9.6.19


Materials32P-end-labeled nucleic acid pools with specific modifications (modified pool in

one tube, unmodified in another; refer to UNIT 9.5, Support Protocol 1, for theradiolabeling protocol)

Binding buffer: this buffer should promote binding and be specific to the aptamerapplication; common examples are phosphate-buffered saline (PBS) and PCRbuffer, e.g., 10 mM Tris·Cl, pH 8.4, 50 mM KCl, and 1.5 mM MgCl2 (for morediscussion, see UNIT 9.5)

Target molecule (range of concentrations from 1 × 10−7 M to 1 × 10−12 M)Methanol0.5 M potassium hydroxide (KOH)

Thermal cycler or 37°C and 65°C water baths or heating blocksMinifold I dot blot system, 96-well plate vacuum manifold (e.g., Whatman)Nylon membranes (e.g., Hybond N+ nylon sheets)0.45-μm nitrocellulose transfer and immobilization membrane (e.g., Protran

BA-83, Whatman)Vacuum pump or water aspirator80°C ovenPhosphorimager screenPhosphorimager and image analysis software (e.g., GE Healthcare Life Sciences

ImageQuant)

Incubate pool with target

1. Heat denature the 32P-end-labeled nucleic acid pool 3 min at 65°C and then cool toroom temperature for at least 10 min.

This allows the RNA to assume a more stable structure.

Prepare about ten RNA aliquots for each pool candidate to be tested. A low RNA concen-tration of �1 × 10−11 M should be used for each reaction/aliquot.

2. Add the protein target in binding buffer to the RNA pool to generate protein-poolbinding reactions.

A range of about ten different concentrations (1 × 10−7 M to 1 × 10−12 M) of proteinshould be used. Prepare a “no protein control” to determine the pool’s background bindingto filters, beads, or other matrices. If background binding is sufficiently high (>10%), thentRNA or other nonspecific additives should be included.

3. Incubate the protein-pool binding reactions for 30 to 45 min at 37°C.

Alternatively, different incubation temperatures and times may be used depending on thedownstream applications for the selected aptamers.

Perform an affinity assay

4. Assemble the dot blot system, with the nitrocellulose sheet (top) and nylon mem-brane (below the nitrocellulose membrane) sandwiched between the dot blot plates(Fig. 9.6.5). Moisten the nylon membrane and the nitrocellulose sheet with bindingbuffer. Be sure the sheets are layered such that no air becomes trapped betweenthe sheets before final assembly of the system. Connect the assembled system to avacuum pump or water aspirator.

It may be necessary to soak the filters prior to assembly to enhance protein cap-ture (nitrocellulose) and unbound oligonucleotide (nylon). Typically, the nitrocellulosesheet is soaked in 0.5 M KOH for 5 min, whereas the nylon membrane is soaked in100% methanol for 5 min. Both filters should then be rinsed in water for 3 min andsoaked in binding buffer for at least 5 min. The buffer should not contain bovine serum



9.6.20


tovacuum

nylon

nitrocellulose

Figure 9.6.5 Dot-blot assembly for filter-based binding assays. See UNIT 9.3.

albumin (BSA) or other nonspecific blocking proteins that may saturate the nitrocellulosefilter.

To check for system leaks and verify proper assembly, prewash a few wells with bindingbuffer. The buffer should flow freely through the well at a rate of roughly 100 μL every3 sec. Bubbles can clog the well preventing filtration and may be unclogged by simplypipetting the applied solution up and down.

5. Apply each protein-pool binding reaction to a well and wash with 3 to 10 vol ofbinding buffer. When all of the reactions have passed through the filters, dry thefilters in an 80°C oven on a glass plate for 5 min.

Upon dispensing the protein-pool binding reaction into the well, wait for the solution tocompletely filter through the membranes before administering the washes.

When dispensing solution, pipet slowly to minimize the formation of bubbles, which mayclog the filter. Additionally, keep the pipet tip close to the membrane without touching themembrane, as contact may damage the surface.

6. Wrap the filters in plastic wrap and place in a developing cassette along with a blankedphosphorimager screen for 15 to 30 min.

Depending on the age of the radioactive nucleotide, a longer exposure time might benecessary.

7. Measure the radioactivity using a phosphorimager and calculate the dissociationconstants of the different pools (see UNIT 9.3 for calculations).

Calculate and compare the dissociation constants for each of the modified and unmodifiedpools tested. In general, the pool (modified or unmodified) with the lowest dissociationconstant may be the best candidate for continuing a selection.

Since the candidate pools are the result of several rounds of selection, additional roundsmay continue from this point using the best candidate pool. Alternatively, the selection canbe reinitiated from a naıve pool using either more or less stringent selection conditions,which may include negative selections, which may further refine aptamer selection. Referto UNITS 9.3, 9.4, and/or 9.5 for in vitro selection protocols.


9.6.21


REAGENTS AND SOLUTIONSUse deionized, distilled water in all recipes and protocol steps. For common stock solutions, seeAPPENDIX 2A; for suppliers, see SUPPLIERS APPENDIX.

Alkaline phosphatase reaction mix

50 mM Tris·Cl, pH 8.5 (APPENDIX 2A)0.1 mM EDTA2 U alkaline phosphatase per 25-μL reactionStore buffer without enzyme indefinitely at −20°CAdd enzyme fresh

Denaturing polyacrylamide gel, 10% (w/v)

1×TBE electrophoresis buffer (APPENDIX 2A) containing:10% (w/v) acrylamide0.5% (w/v) bisacrylamide7 M ureaPrepare fresh

See APPENDIX 3B for full details on pouring and running gels.

Denaturing stop dye, 2×Deionized formamide containing:0.1% (w/v) bromophenol blue50 mM EDTAStore indefinitely at –20°C

Daily-use aliquot may be stored up to 1 month at room temperature

Nondenaturing dye, 6×60% (v/v) glycerol0.6% (w/v) bromophenol blue10× ethidium bromideStore indefinitely at −20°C

Daily-use aliquot may be stored up to 1 month at room temperature

Nuclease P1 digestion mix

15 mM sodium acetate, pH 5.2 (APPENDIX 2A)0.1 mM zinc sulfate1 U P1 nuclease per 5-μL reactionStore buffer indefinitely at −20°CAdd enzyme fresh

COMMENTARY

Background InformationNucleic acid selection experiments have

generated a wide variety of binding species(aptamers) and catalysts (ribozymes and de-oxyribozymes). However, nucleic acids are,by and large, not as functional as proteins:most aptamers bind their ligands less well thancomparable antibodies, whereas ribozymes aregenerally slower catalysts than protein en-zymes. It is possible that these limitationsare a function of the relative infancy of ap-

tamer and catalyst selections relative to abetter-established, field-like protein engineer-ing; however, it is also possible that func-tional nucleic acids are inherently limited bythe diversity of the genetic alphabet. Thebinding interactions and chemical reactionsperformed by nucleic acid biopolymers maybe constrained by the functional groups theycontain. The ability to expand the functionalgroups available to a DNA or RNA poly-mer through the incorporation of modified



9.6.22


nucleotides could potentially open up bindinginteractions and reaction chemistries that werepreviously inaccessible.

Due to advances in chemical synthesis tech-nologies, decreases in the overall cost of start-ing materials, and the availability of mutantpolymerases, modified nucleotides can now beeasily incorporated within in vitro selections.Whereas once limited to the five canonical nu-cleotides, numerous aptamers, ribozymes, anddeoxyribozymes with a variety of functionsand applications have been selected that uti-lize a plethora of modified nucleotides (seeTable 9.6.1). The modifications regularly usedin in vitro selections can be coarsely dividedinto three main groups: ribose modifications,phosphate backbone modifications, and basemodifications (see Fig. 9.6.1). In particular,2′-ribose modifications (such as 2′-amino or2′-fluoro) and phosphate backbone modifica-tions offer increased stabilities, in large mea-sure because most ribonucleases polarize the2′-hydroxyl group to attack the phosphodi-ester linkage. Base modifications (in particularat the 5-position of pyrimidines nucleosides,such as uridine) can provide additional chem-ical functionalities that are accessible in themajor groove of helices.

Critical Parameters

How to choose a modified nucleotide

Modified versus unmodified nucleotideInitially, the most important decision is

what modification(s) to introduce. While mul-tiple pools containing different modified nu-cleotides can be used, depending on availabil-ity and incorporation efficiency, the potentialutility of not using modifications at all shouldbe kept in mind. It is entirely possible thatwhile modified nucleotides will increase theprobability of finding functional species, it isalso possible that simpler selections with un-modified pools will also find such species. Forexample, it has been shown that a pool contain-ing �1015 unique sequences may contain ex-tremely rare binding or catalytic species (e.g.,Bartel and Szostak, 1993).

Expanding the variety of modified nucleotidesWhereas there are numerous modified

nucleotides to choose from that have beensuccessfully used in in vitro selections (seeTable 9.6.1), additional modified nucleotideshave been proposed, but no prominentcases of their use can be found. Someexamples of these modified nucleotidesinclude 5-aminoallyl-2′-fluoro-dUTP, 5-

aminoallyl-UTP, and N(6) (6-aminohexyl)carbamoylmethyl-ATP (Schoetzau et al.,2003), as well as 5-iodo-dUTP (Brody et al.,1999), 2′-azido-dUTP (Padilla and Sousa,2002), 2′-Se-methyl nucleotides (Chellis-errykattil and Ellington, 2004; Siegmundet al., 2012), 6-amino-5-nitro-3-(1′-β-D-2′-deoxyribofuranosyl)-2(1H)-pyridone, and 2-amino-8-(1′-β-D-2′-deoxyribofuranosyl)-imi-dazo[1,2-a]-1,3,5-triazin-4(8H)-one) (Yanget al., 2011). Additionally, 7-deaza-dATP,which can be readily incorporated using TaqDNA polymerase, can be conjugated at thecarbon 7-position with aminopropyl, amino-propynyl, and Z-aminopropenyl side chains(Gourlain et al., 2001). Numerous conjugatesof 5-(3-aminopropenyl)-2′deoxyuridine suita-ble for selection have been identified(Sakthivel and Barbas, 1998) along withdeoxynucleoside triphosphate analogues ofall four unmodified nucleotides bearing basic,acidic, or lipophilic groups (Jager et al.,2005).

Understanding modified nucleotide chemistryThe incorporation of additional functional

groups into the context of an RNA backbone isexpected to increase the diversity of its avail-able interactions, both with itself and with itsdesired substrates. Thiol groups, for example,can participate directly in catalysis as nucle-ophiles. Additionally, disulfide bonds could beformed intramolecularly between thiols. Thismay add to the structural diversity of nucleicacids, normally limited to hydrogen bonding,salt bridges with metals, and stacking inter-actions. Charged groups, such as a lysine-likeside chain, could potentially add to the struc-tural repertoire of nucleic acids by allowingthe formation of electrostatic interactions andsalt bridges. The inclusion of chemical moi-eties with a pKa closer to neutrality, such asan imidazole group, is also expected to bene-fit nucleic acid chemistry. For instance, a pri-mary amine on 2′-amino nucleotides, with apKa of 6.2 and increased nucleophilicity com-pared to unmodified nucleotides, may reactwith an aldehyde near physiological condi-tions to generate 2′-imino-conjugated oligonu-cleotides (Bugaut et al., 2004, 2005). How-ever, unmodified nucleotide bases contain nofunctional groups with unperturbed pKa val-ues between 3.5 and 9.2 inherently limitingthe proton “push-pull” chemistry found in somany protein enzymes. The introduction ofthese and other functional groups can alsopotentially increase the abilities of nucleicacids to bind metals.


9.6.23


While attempts to associate nucleotidefunctionality with binding or catalytic prop-erties are at best still guesswork, it is nonethe-less true that failing to appreciate the chemicalproperties of modified nucleotides can poten-tially have an adverse effect on how a selectionexperiment proceeds. For example, the inclu-sion of thiolated nucleotides can potentiallylead to the unwanted formation of disulfide-bonded products if care is not taken to includea reducing agent in enzyme and/or selection re-actions. For example, deoxyribozyme ligasesthat form an unnatural internucleotide linkagebetween a 5′-iodinated pool and an oligonu-cleotide substrate with a 3′-phosphorothioatehave been selected from random sequencepools (Levy and Ellington, 2001, 2002). Un-less reducing agents are kept in the selectionreaction, the oligonucleotide substrates willdimerize, reducing their effective concentra-tion. Similarly, the inclusion of modified nu-cleotides that have altered pKas could veryeasily change the pH of a concentrated stocksolution, and care should be taken to makesure that all such stocks are at the desiredpH and appropriately buffered. The acciden-tal alteration of pH in enzyme or selectionreactions can lead to decreases in productyield. Finally, the inclusion of modified nu-cleotides that have new metal binding orchelating properties may alter the availablemetal concentrations in an enzyme or selec-tion reaction. In particular, the chelation ofmagnesium can lead to large changes in theefficiency of product formation by many dif-ferent polymerases.

Modified nucleotides and mutationsAlthough a selection starts with a large

number of sequences, this number is usually asmall fraction of the total number of sequencespossible for the length of the random region.Additionally, with each round, the number ofsequences within the pool is diminished. Assuch, it may be useful to explore the sequencespace around the selected winners to discoverfunctional variants. To some extent, this occursin any selection due to the inherent mutationsthat arise in the amplification process; how-ever, this background mutation rate is small,and one may wish to increase the frequencyof mutations. For these purposes, the muta-genic potential of a nucleotide analogue thatserves as a nonspecific template can be usedto increase the diversity of a pool (see BasicProtocol 4 for assessment of mutagenic po-tential). Similar to mutagenic PCR (UNIT 9.4),these techniques can be used at the outset of

a selection to mutate an existing ribozyme oraptamer for either optimization or reselectionfor altered specificity. For example, Kore et al.(2000) have used modified nucleotides to cre-ate a degenerate pool based on the hammer-head ribozyme, from which they selected avariant that cleaves at an alternate sequence.Additionally, a mutagenic step can be incor-porated between rounds in a selection to moreefficiently explore sequence space. However,it should be noted that in the latter case, onedoes not want the mutation rate to be so highthat a few of each winning sequence from oneround do not survive unchanged into the nextround.

One benefit of using modified nucleotidesrather than standard mutagenic PCR is that itis easy to control the amount of mutation in-troduced into the sample simply by adjustingthe relative rate of modified to unmodified nu-cleotide in the amplification process. As such,a higher rate of mutation can be achieved thanwith mutagenic PCR, which would be partic-ularly useful for diversifying an initial pool.Zaccolo et al. (1996) describe such a sys-tem using both a purine and pyrimidine ana-logue in a PCR and have calculated the fre-quencies of each base transition. When usingmutagenic-modified nucleotides, one wouldnot want them to be included in the active pool,as their decreased fidelity would make it lesslikely that they would be in the same positionduring the next round. For example, if a DNAselection were being performed, a second am-plification of the pool would be required usingonly unmodified nucleotides.

When to incorporate modified nucleotidesfor in vitro selection

Pre- versus post-in vitro selectionmodifications

Overall, the most important considerationsin deciding whether to use a modified nu-cleotide in a selection experiment are purelytechnical ones, i.e., can the modification be in-corporated and will it inhibit amplification ofthe nucleic acid pool? A number of modifiednucleotides have been incorporated with vary-ing degrees of success into selection experi-ments (Table 9.6.1), while other modified nu-cleotides cannot yet be enzymatically incorpo-rated. As previously described, there are threemain ways to incorporate modified nucleotidesinto the selection (summarized in Fig. 9.6.2)including enzymatic processing, conjugation,and chemical synthesis, each with their ownbenefits and disadvantages.



9.6.24


When modifications are enzymatically in-corporated into the naıve pool (i.e., pre-selection), nucleoside triphosphates (NTPs)are the pool building blocks, conferring un-modified or modified moieties. An advan-tage of this pre-selection incorporation is thatthe modified library yields oligonucleotidesthat exhibit better nuclease resistance fromthat start. Specifically, 2′-modifications (i.e.,nuclease-resistance modifications) are com-monplace in selections, as it is often moredifficult to modify an aptamer after the se-lection with nuclease stable nucleotides thanto perform the selection with a pool contain-ing these nucleotides. However, once aptamersare selected, downstream applications typi-cally require large-scale syntheses, thus re-quiring phosphoramidites (chemically equiva-lent to the original NTPs) for oligonucleotidesolid-phase synthesis, (see UNITS 9.3 & 9.4).Currently, the commercially available modi-fied NTPs are not equally available as modifiedphosphoramidites.

In addition to solid-phase synthesis, mod-ifications may be incorporated post-selectionusing conjugation techniques. While Bugautet al. (2006) used the conjugation method-ology to generate a naıve pool, the methodcould similarly be applied post-selection. Aswith all post-selection modifications and site-specific modified nucleotide incorporations,however, a detailed understanding, at the nu-cleotide position, of the chemical variant’simpact on binding, functionality, catalysis,stability, specificity, and other properties, isnecessary.

Furthermore, when enzymatic in-corporation is not possible, chemicallysynthesized libraries containing modifiednucleotides can instead be subjected tosuccessive, often less stringent, screens forfunction without amplification. Successfulmotifs can either be used directly, or a newlibrary can be synthesized and screenedbased on the perceived change in informationcontent. One advantage of this approach is thatchemically synthesized pools permit additionof multiple chemical modifications in parallel.As an example, He et al. (2012) screened forso-called X-aptamers that contained mod-ified nucleotides and a phosphorodithioateoligonucleotide backbone. However, ratherthan being replicated, X-aptamers weresynthesized by a split-bead method that ledto one aptamer variant clonally representedin many copies per bead. The beads werescreened for ligand-binding properties, andthe sequences from successful beads were

cleaved and sequenced. Using this method, adrug conjugate that bound CD44 was found ina single round of screening. The eliminationof PCR amplification removes the biases thatmay be related to amplification, such as thedisfavored amplification of structurally stablespecies.

Alternatively, Spiegelmers, or aptamerscomprised of L-ribose instead of the enzymat-ically incorporated D-ribose, are generated us-ing a novel approach to modified nucleotide in-corporation (Wlotzka et al., 2002; reviewed inEulberg and Klussman, 2003 and Wilson andKeefe, 2006). Specifically, Spiegelmer precur-sors are identified through in vitro selectionsagainst a biological protein’s enantiomer (syn-thesized using D-amino acids), thus gener-ating traditional aptamers (although againstnon-biological targets). Remarkably, themirror-image aptamer (i.e., Spiegelmer)recognizes the native biological protein

Efficiency of enzymatic incorporationIn general, modified nucleotides may in-

corporate less efficiently than unmodifiednucleotides, depending on the polymerase,modified nucleotide, and other reaction con-ditions. For example, in the authors’ expe-rience, the incorporation of 5-Br-dUTP withTaq DNA polymerase decreased the replicabil-ity of sequences and created an inherent andunintended selection in favor of those mem-bers of a population that had fewer modi-fied nucleotides. Lutz et al. (1998), using xan-thosine in the template, showed that diaminopyrimidine nucleoside triphosphate was in-corporated by the 9°N-7 thermostable DNApolymerase (Southworth et al., 1996) in asequence-specific fashion but with only 45%incorporation of the correct nucleotide ana-logue. Additionally, Vaught et al. (2010) com-pared primer extension yields of dUTP ana-logues modified at the 5-position and foundthat Vent DNA polymerase provided yieldssimilar to or better than TTP for four of thesix modified nucleotides tested, while KODXL DNA polymerase provided yields similarto TTP for only one of the six modified nu-cleotides tested.

Incorporation after amplification(conjugation or extension)

However, amplification complications us-ing modified nucleotides in aptamer selectionshave not stunted recent aptamer advances. In-stead, Vaught et al. (2010) developed a newDNA aptamer selection scheme to accom-modate the poor enzymatic incorporation of


9.6.25


modified nucleotides. The novel scheme in-volves an intermediate PCR step that relied onunmodified nucleotides to exponentially am-plify a selected pool. A subsequent primer ex-tension reaction reincorporates the modifiednucleotides to generate an enriched, modifiedpool (see Fig. 9.6.2). Somalogic has used thisstrategy (Ochsner et al., 2013) to successfullygenerate aptamers against >1000 human pro-teins (Mehan et al., 2013).

Alternatively, Bugaut et al. (2006) offereda “conjugation” approach to incorporate mod-ified nucleotides post-amplification into anin vitro selection (see Fig. 9.6.2). At eachround, following selection, amplification, andthe regeneration of the 2′-amino-2′-deoxyUTPpool, a reductive amination reaction conju-gated the aldehyde-functionalized ligands tothe 2′-amino groups. This unique approachto modified nucleotide incorporation was suc-cessfully used to generate novel modifiedaptamers against HIV-1 trans-activation re-sponse (TAR).

Making the most of modified nucleotideincorporation: Further optimizations

As has been apparent throughout, there arefew hard and fast rules regarding what will andwill not work when selecting functional nu-cleic acids. Because of this, researchers needto be somewhat versatile in their approach toselection experiments involving modified nu-cleotides. If a given approach does not work,this does not mean that the selection exper-iment inherently has no chance of working,but instead indicates that it may be necessaryto alter one or more parameters. Optimiza-tion of just three variables is all that may beneeded to improve the incorporation of mod-ified nucleotides: pool, polymerase, and reac-tion buffer.

Pool considerationsA good portion of this protocol has been de-

voted to testing reactions and optimizing con-ditions. Although this process will likely takeseveral weeks, it is time wisely spent. The firstround of selection is by far the most important,as this is the round when selection conditionswill query the greatest number of possibleanswers. Concomitantly, creating the largestpossible number of unique sequences (1013 to1015) for the first round of selection is a criti-cal task (UNIT 9.2). However, because creatinga large library requires a large amount of effortand large volumes of relatively costly reagents,initial optimization and practice should alwaysbe performed on a small scale.

Similarly, it is expected that only a smallnumber of the input molecules will survive anygiven round; therefore, it is essential that thesesuccessful sequences be efficiently carried tothe next round. Inefficient reverse transcrip-tion or amplification reactions may lead to aselection for molecules that can be efficientlyreplicated (amplification artifacts) rather thanto molecules that are highly functional, there-fore the optimization of these reactions andthe entire process as a whole is necessary (seeBasic Protocols 1 and 3).

Additionally, the length and compositionof a random sequence pool can vary accord-ing to the type of selection that is carried out(also see UNIT 9.2) and some attention shouldbe paid to the choice of constant sequence re-gions. While the largest possible pool is pre-ferred to begin the selection, shorter pools areeasier and less costly to generate (see UNIT

9.2). Similarly, aptamers and deoxy/ribozymestend to be synthetically produced for testingand downstream applications; thus, consid-ering solid-phase phosphoramidite synthesislimitations is important. For instance, whereasDNA pool lengths can be well >100 nt, RNApools should be kept <80 nt when full-lengthmodified variants are being considered for syn-thesis. In addition, current phosphoramiditereagent costs should be taken into account.For instance, 2′-deoxy-2′-fluoro purine nucle-oside phosphoramidites are considerably morecostly than their respective pyrimidines or 2′-O-methyl-containing phosphoramidites.

In consideration of composition, RNApolymerases tend to have difficulty incorpo-rating modified nucleotides prior to positions8 to 10 following the transcription start site(Milligan and Uhlenbeck, 1989; Kujau et al.,1997), most likely because the attempted in-corporation of modified nucleotides leads toincreased abortive initiation. Thus, it is wise todesign a pool that lacks or limits the incorpo-ration of the modified base within this region.Some modified T7 polymerases can overcomethis limitation through incorporating a purineleader sequence of 10 to 15 nucleotides afterthe T7 promoter. Inasmuch as the first 18 ormore nucleotides of a transcribed pool typ-ically remain constant to allow for second-strand DNA synthesis, this limitation shouldnot affect the overall diversity of the pool.

Finally, the sequence of the pool itself canhave a surprisingly large effect on selectionexperiments. Robertson and Ellington (1999)originally selected a relatively fast ribozymeligase from an RNA pool that contained 90random sequence positions. A variant of this



9.6.26


pool was generated in which only four po-sitions in the constant region were altered,substituting a GAAA tetraloop in place of aUUCG tetraloop. Selection experiments forribozyme ligases were initiated from ampli-fied aliquots of this pool; one aliquot wasused for the selection of ribozymes containingcanonical nucleotides, other aliquots of iden-tical complexity were used for the selectionof ribozymes containing one of three differentmodified nucleotides (Robertson, 2001). Aftersix rounds of selection and amplification, allof the pools had collapsed to contain a rela-tively few winning sequences. However, therates of all of these winning sequences wereremarkably slow (500-fold slower) comparedto the ribozyme ligase originally selected fromthe slightly different pool. There was no dis-cernible difference between the ribozymes thatcontained canonical or modified nucleotides;they were all very slow. Robertson (2001) hy-pothesized that the pools had somehow be-come artificially narrowed during the courseof the selection, and therefore repeated the se-lection experiments with new aliquots of theoriginal, amplified pool. Again, only very slowribozymes were obtained.

Choosing the right enzyme for modifiednucleotide incorporation

Different polymerases clearly have differ-ent potentialities for the incorporation of mod-ified nucleotides (see Table 9.6.1). For exam-ple, the Benner group has tested several ther-mostable DNA polymerases for their ability toincorporate a variety of modified nucleotides(Lutz et al., 1998, 1999). Likewise, Vaughtet al. (2010), using modified DNA templates,examined the extension efficiency of D. VentDNA polymerase and KOD XL DNA poly-merase to incorporate modified nucleotides(5-position of uridine). They found that VentDNA polymerase provided yields similar toor better than TTP for four of the six modi-fied nucleotides tested, while KOD XL DNApolymerase provided yields similar to TTPfor only one of the six modified nucleotidestested.

Similar variations in RNA polymerases toincorporate modified nucleotides, especially2′-modifications, have been observed. While2′-modified ribonucleotides can be incorpo-rated into RNA by the wild-type T7 RNA poly-merase, the kcat/KM values for incorporation of2′-modified ribonucleotides are substantiallyhigher than for unmodified nucleotides, witha preference order of 2′-OH > 2′-NH2 > 2′-F > 2′-H (Huang et al., 1997). To improve

incorporation, some groups have engineeredor evolved polymerases that can incorporatemodified nucleotides. For example, a singlemutation in T7 RNA polymerase (Y639F) re-duces discrimination at the 2′-position and al-lows more efficient incorporation of deoxyri-bonucleotides into RNA (Kostyuk et al., 1995;Sousa and Padilla, 1995). Additionally, theY639F T7 RNA polymerase, which is com-mercially available in DuraScribe kits (Epi-centre) allows the incorporation of NTPs withfluoro, amino, and thiol groups at their 2′-positions (see Basic Protocol 1). The Y639Fmutant of T7 RNA polymerase has been shownto have up to 24-fold greater specificity for in-corporation of 2′-modified NTPs, with a pref-erence order of 2′-OH > 2′-F > 2′-H > 2′-NH2

(Huang et al., 1997). It is, however, inefficientat incorporating bulkier 2′-groups such as 2′-OMe, that when incorporated, greatly enhancenuclease stability of transcripts.

Histidine 784 (H784) plays an importantrole in ribose discrimination by T7 RNApolymerase, as it is thought to cause prema-ture termination following the incorporationof a 2′-modified ribonucleotides (Padilla andSousa, 2002). An H784A mutation relaxes thisdiscrimination and, when coupled with theY639F mutation, allows for bulkier groupsto be incorporated such as 2′-O-methyl (2′-OMe) and 2′-azido (Padilla and Sousa, 2002).Further, the Y639F H784A double mutanthas reduced preference for NTPs over dNTPs(Brieba and Sousa, 2000) and 2′-F NTPs (un-pub. observ.).

A directed evolution approach was taken,in which R425, G542, Y639, and H784 wererandomized and the T7 RNA polymerase wasselected for normal in vivo activity (Chellis-errykattil and Ellington, 2004). Screening theresulting population yielded a particular mu-tant known as RGVG (R425, G542, Y639V,H784G plus additional E593G V685A muta-tions that arose during the selection), whichdemonstrated the ability to incorporate 2′-OMe-modified ribonucleotides. RGVG waslater shown to incorporate 2′-SeMe modi-fied ribonucleotides (Siegmund et al., 2011).RGVG was further improved by error-pronePCR and high-throughput screening. A par-ticular mutant “2P16” (which contained sevenadditional mutations beyond the RGVG back-ground) shows comparable 2′-OMe- and en-hanced 2′-SeMe-modified ribonucleotide in-corporation as compared to RGVG (Siegmundet al., 2012). A similar approach was usedto evolve the R425C mutant T7 RNA poly-merase, which is capable of synthesizing RNA


9.6.27


entirely from 2′-OMe-modified NTPs (Ibachet al., 2013).

Buffer optimizationThe buffer conditions for incorporation can

be adjusted and optimized for modified nu-cleotide incorporation. For example, Padillaand Sousa (1999) have systematically inves-tigated several buffer conditions that aid theincorporation of nucleotides modified specifi-cally at the 2′-position. These authors foundbetter incorporation upon supplementing atranscription reaction with 0.5 mM MnCl2,1 U/μL pyrophosphatase, and either 8 mMspermidine for plasmid templates or 8 mMspermine for short DNA templates. Simi-larly, Burmeister et al. (2006) optimized their2′-deoxy purine and 2′-O-methyl pyrimidine(dRmY) transcription reactions with 2 mMspermine, 9.6 mM MgCl2, and 2.9 mM MnCl2.However, Minakawa et al. (2008), in an at-tempt to optimize the transcription efficiencyof their fully modified 4′-thio RNA, failed toimprove the efficiency through MnCl2 addi-tion, instead of MgCl2, or with an increasedspermidine concentration. However, increas-ing the ratio of modified purines to unmodifiedpurines resulted in more modified nucleotidesbeing incorporated. While this did cause aslight decrease in transcription efficiency it re-sulted in the fully modified transcripts neededfor selection.

Archiving reactionsIn vitro selection experiments are particu-

larly grueling because the ultimate outcome,the successful isolation of binding species orcatalysts, may not be known for days, weeks,or even months. This is especially true whenworking with modified nucleotides, becauseof the strong possibility that one or more am-plification reactions or selection steps will notwork at some point during the course of theselection. Therefore, it is desirable to keep anarchived copy of each round. After the ini-tial round, there should be multiple copies ofeach winning sequence present in the selectedpopulation. Thus, it is not necessary to use allof the RNA, cDNA, or double-stranded DNAtemplate in a given selection or amplificationstep. Rather, some portion of the pool shouldbe saved at any convenient point (e.g., as adouble-stranded PCR product). In the event anunsuccessful round of selection should occur,it will not be necessary to repeat the entire ex-periment. Rather, selection can continue fromthe last successful round, with a minimal lossof time and effort.

A second added benefit of archiving eachround is that “branched” selections can be con-sidered. Such “branched” selections may be-gin with the same starting material (i.e., se-lected pool), but then parallel selections areperformed under different stringencies, selec-tion conditions, or using mutagenic condi-tions. Sampling a wider range of selectionconditions potentially increases the chances ofteasing out a high-affinity binding variant, es-pecially once the pool exhibits some bindingpotential.

Necessity of modified nucleotide: Did themodified nucleotide improve the functionaloligonucleotide?

To understand the contribution of a mod-ified nucleotide in an in vitro selection, par-allel modified nucleotide and unmodified nu-cleotide selections must be performed. An ex-ample of this is in seen in the Somalogicwork comparing modified (i.e., 5-position ofuridine) and unmodified nucleotide selectionsagainst thirteen difficult targets (Gold et al.,2010). The results indicated that only the se-lections with modified nucleotide generatedaptamers; however, the extent of benefit ofthe modified nucleotide was highly dependenton the modified nucleotide and target. Alter-native to selections in parallel with differentmodifications, the selections can be performedin parallel with one modification and differ-ent targets, then the targets that prove diffi-cult can be reselected with a different mod-ification. Similar modification contributionsmay be estimated upon substituting in un-modified nucleotides in place of modified nu-cleotides. Multiple examples of this are pro-vided above in Strategic Planning. However,nucleotide substitutions may interfere with theoligonucleotide fold and presentation to thetarget, thus obfuscating a direct comparisonof the modified with the unmodified species.Similarly, others have compared unmodifiedselections with modified selections performedusing different pools, conditions, etc. (see ex-amples in Strategic Planning). However, alter-ing multiple variables obscures the cause ofthe results, whether it is the superiority of thenucleotide, pool, and/or selection conditions.

Anticipated ResultsAs with any selection experiments (UNITS

9.3, 9.4, & 9.5), it is virtually impossible to an-ticipate the outcome of any given experiment.This is especially true when considering theincorporation of modified nucleotides, con-sidering the variety of modified nucleotides



9.6.28


available, and resulting combinatorial permu-tations. However, if the protocols for the pro-duction of RNA and DNA pools that have beenoutlined are followed, it can be anticipated thatit should be possible to generate and purifyupwards of at least 1014 different nucleic acidsequences (�10 μg) that contain a particularmodified nucleotide. For successful selectionexperiments, the population should be sievedby a factor of 100 to 1000 each round. Thatis, <1% of the total population will be re-covered following a round of selection. Thus,most successful selections will show signifi-cant functional improvement or be completedwithin five to ten rounds. However, in some se-lections, it may take very successful functionalsequences a longer period of time to over-come a great sea of mediocre functional se-quences. For example, it took 18 rounds of se-lection to isolate cGMP-dependent aptazymesfrom a random sequence population (Koizumiet al., 1999). Following selection, the popu-lation should contain a relatively small num-ber (one to ten) of sequence classes. A se-quence class is a set of related sequencesthat contain a common motif. To date, thesequences of aptamers or ribozymes selectedfrom pools containing modified nucleotides donot correspond to the sequences of aptamersor ribozymes selected from pools that containcanonical nucleotides. Thus, any selection thatincludes modified nucleotides is likely to pro-duce new sequence motifs.

Time ConsiderationsThe preparative portions of each round of

selection can be expected to take several days.A transcription could take several hours toovernight, followed by an �2-hr gel purifi-cation and an overnight elution from the ex-cised gel pieces. The amount of time spenton the selective steps will vary greatly witheach individual selection. However, in initialrounds, the reaction time alone is typicallyseveral hours to an entire day. The selectivesegregation of winning sequences from los-ing sequences will vary from a few to severalhours, depending on the method used. Reversetranscription followed by PCR, isolation, andquantitation of new template DNA will takea few more hours, depending mainly on thenumber of cycles required for amplification.The largest variable, the number of rounds ofselection required, will vary greatly depend-ing on the activity sought, from only a fewto �12 rounds. Cloning and assaying of indi-vidual sequences will likely also take severaldays. Thus, even if no problems are encoun-

tered during a selection, it is likely to take �1month to complete.

AcknowledgmentsGwendolyn Stovall was primarily sup-

ported by the Freshman Research Initiativeat the University of Texas at Austin and par-tially supported by both the National ScienceFoundation (Cooperative Agreement No. DBI-0939454) and the National Institute of Health(5R03HD068691). These methods were re-fined by undergraduate students (Robert Be-denbaugh and Shruti Singh) from the Fresh-man Research Initiative based on generousfunding from the National Science Foundation(CHE 0629136). Additionally, Robert Beden-baugh was partially funded by the National In-stitute of Health (5R03HD068691) and ShrutiSingh was partially funded by the NationalScience Foundation (Cooperative AgreementNo. DBI-0939454). Adam Meyer’s work wassupported by the National Security Scienceand Engineering Faculty Fellowship (FA9550-10-1-0169) and the Welch Foundation (F-1654). Any opinions, findings, and conclu-sions or recommendations expressed in thismaterial are those of the author(s) and do notnecessarily reflect the views of the NationalScience Foundation, The University of Texasat Austin, or Altermune Technologies, LLC.This work was a revision of the original proto-col written by Scott M. Knudsen and MichaelP. Robertson.

Literature CitedBartel, D.P. and Szostak, J.W. 1993. Isolation of

new ribozymes from a large pool of random se-quences. Science 261:1411-1418.

Battersby, T.R., Ang, D.N., Burgstaller, P., Jurczyk,S.C., Bowser, M.T., Buchanan, D.D., Kennedy,R.T., and Benner, S.A. 1999. Quantitative anal-ysis of receptors for adenosine nucleotides ob-tained via in vitro selection from a library in-corporating a cationic nucleotide analog. J. Am.Chem. Soc. 121:9781-9789.

Beaudry, A., DeFoe, J., Zinnen, S., Burgin, A.,and Beigelman, L. 2000. In vitro selectionof a novel nuclease-resistant RNA phosphodi-esterase. Chem. Biol. 7:323-334.

Bell, N.M. and Micklefield, J. 2009. Chemical mod-ification of oligonucleotides for therapeutic, bio-analytical and other applications. Chembiochem10:2691-2703.

Brieba, L.G. and Sousa, R. 2000. Roles of histidine784 and tyrosine 639 in ribose discrimination byT7 RNA polymerase. Biochemistry 39:919-923.

Brody, E.N., Willis, N.C., Smith, J.D., Jayasena,S., Zichi, D., and Gold, L. 1999. The use ofaptamers in large arrays for molecular diagnos-tics. J. Mol. Diagn. 4:381-388.


9.6.29


Bugaut, A., Toulme, J.J., and Rayner, B. 2004.Use of dynamic combinatorial chemistryfor the identification of covalently appendedresidues that stabilize oligonucleotide com-plexes. Angew. Chem. Int. Ed. Engl. 43:3144-3147.

Bugaut, A., Bathany, K., Schmitter, J.M., andRayner, B. 2005. Target-induced selection ofligands from a dynamic combinatorial libraryof mono- and bi-conjugated oligonucleotides.Tetrahedron Lett. 46:687-690.

Bugaut, A., Toulme, J.J., and Rayner, B. 2006. SE-LEX and dynamic combinatorial chemistry in-terplay for the selection of conjugated RNA ap-tamers. Org. Biomol. Chem. 4:4082-4088.

Burmeister, P.E., Lewis, S.D., Silva, R.F., Preiss,J.R., Horwitz, L.R., Pendergrast, P.S., Mc-Cauley, T.G., Kurz, J.C., Epstein, D.M., Wilson,C., and Keefe, A.D. 2005. Direct in vitro selec-tion of a 2′-O-methyl aptamer to VEGF. Chem.Biol. 12:25-33.

Burmeister, P.E. Wang, C., Killough, J.R., Lewis,S.D., Horwitz, L.R., Ferguson, A., Thompson,K.M., Pendergrast, P.S., McCauley, T.G., Kurz,M., Diener, J., Cload, S.T., Wilson, C., andKeefe, A.D. 2006. 2′-Deoxy purine, 2′-O-methylpyrimidine (dRmY) aptamers as candidate ther-apeutics. Oligonucleotides 16:337-351.

Chakravarthy, U., Adamis, A.P., Cunningham, E.T.,Goldbaum, M., Guyer, D.R., Katz, B., and Patel,M. 2006. Year 2 efficacy results of 2 random-ized controlled clinical trials of pegaptanib forneovascular age-related macular degeneration.Ophthalmology 113:1-25.

Chelliserrykattil, J. and Ellington, A.D. 2004. Evo-lution of a T7 RNA polymerase variant thattranscribes 2′-O-methyl RNA. Nat. Biotechnol.22:1155-1160.

Eulberg, D. and Klussman, S. 2003. Spiegelmers:Biostable aptamers. Chembiochem 4:979-983.

Faulhammer, D. and Famulok, M. 1997. Charac-terization and divalent metal-ion dependence ofin vitro selected deoxyribozymes which cleaveDNA/RNA chimeric oligonucleotides. J. Mol.Biol. 269:188-202.

Geyer, C. R. and Sen, D. 1997. Evidence forthe metal-cofactor independence of an RNAphosphodiester-cleaving DNA enzyme. Chem.Biol. 4:579-593.

Gold, L., Ayer, D., Bertino, J., Bock, C., Bock,A., Brody, E.N., Carter, J., Dalby, A.B., Eaton,B.E., Fitzwater, T., Flather, D., Forbes, A., Fore-man, T., Fowler, C., Gawande, B., Goss, M.,Gunn, M., Gupta, S., Halladay, D., Heil, J.,Heilig, J., Hicke, B., Husar, G., Janjic, N.,Jarvis, T., Jennings, S., Katilius, E., Keeney,T.R., Kim, N., Koch, T.H., Kraemer, S., Kroiss,L., Le, N., Levine, D., Lindsey, W., Lollo, B.,Mayfield, W., Mehan, M., Mehler, R., Nelson,S.K., Nelson, M., Nieuwlandt, D., Nikrad, M.,Ochsner, U., Ostroff, R.M., Otis, M., Parker,T., Pietrasiewicz, S., Resnicow, D.I., Rohloff, J.,Sanders, G., Sattin, S., Schneider, D., Singer, B.,Stanton, M., Sterkel, A., Stewart, A., Stratford,S., Vaught, J.D., Vrkljan, M., Walker, J.J., Wa-

trobka, M., Waugh, S., Weiss, A., Wilcox, S.K.,Wolfson, A., Wolk, S.K., Zhang, C., and Zichi,D. 2010. Aptamer-based multiplexed proteomictechnology for biomarker discovery. PLoS ONE.5:e15004.

Golden, M.C., Collins, B.D., Willis, M.C., andKoch, T.H. 2000. Diagnostic potential of photo-SELEX-evolved ssDNA aptamers. J. Biotech-nol. 81:167-178.

Gourlain, T., Sidorov, A., Mignet, N., Thorpe, S.J.,Lee, S.E., Grasby, J.A., and Williams, D.M.2001. Enhancing the catalytic repertoire of nu-cleic acids. II. Simultaneous incorporation ofamino and imidazolyl functionalities by twomodified triphosphates during PCR. NucleicAcids Res. 29:1898-1905.

Green, L.S., Jellinek, D., Bell, C., Beebe, L.A.,Feistner, B.D., Gill, S.C., Jucker, F.M., and Jan-jic, N. 1995. Nuclease-resistant nucleic acid lig-ands to vascular permeability factor/vascular en-dothelial growth factor. Chem. Biol. 2:683-695.

Griffiths, A.D., Potter, B.V., and Eperon, I.C.1987. Stereospecificity of nucleases towardsphosphorothioate-substituted RNA: Stereo-chemistry of transcription by T7 RNA poly-merase. Nucleic Acids Res. 15:4145-4162.

He, W., Elizondo-Riojas, M.A., Li, X., Lokesh,G.L.R., Somasunderam, A., Thiviyanathan, V.,Volk, D.E., Durland, R.H., Englehardt, J., Cava-sotto, C.N., and Gorenstein, D.G. 2012. X-Aptamers: A bead based selection method forrandom incorporation of druglike moieties ontonext-generation aptamers for enhanced binding.Biochemistry 51:8321-8323.

Hollenstein, M., Hipolito, C.J., Lam, C.H., and Per-rin, D.M. 2013. Toward the combinatorial selec-tion of chemically modified DNAzyme RNase Amimics active against all-RNA substrates. ACSComb. Sci. 15:174-182.

Huang, Y., Eckstein, F., Padilla, R., and Sousa, R.1997. Mechanism of ribose 2′-group discrim-ination by an RNA polymerase. Biochemistry36:8231-8242.

Ibach, J., Dietrich, L., Koopmans, K.R., Nobel, N.,Skoupi, M., and Brakmann, S. 2013. Identifi-cation of a T7 RNA polymerase variant thatpermits the enzymatic synthesis of fully 2′-O-methyl-modified RNA. J. Biotechnol. 167:287-295.

Jager, S., Rasched, G., Kornreich-Leshem, H., En-geser, M., Thum, O., and Famulok, M. 2005.A versatile toolbox for variable DNA function-alization at high density. J. Am. Chem. Soc.127:15071-15082.

Jellinek, D., Green, L.S., Bell, C., Lynott, C.K., Gill,N., Vargeese, C., Kirschenheuter, G., McGee,D.P., Abesinghe, P., Pieken, W.A., Shapiro, R.,Rifkin, D.B., Moscatelli, D., and Janjic, N. 1995.Potent 2′-amino-2′-deoxypyrimidine RNA in-hibitors of basic fibroblast growth factor. Bio-chemistry 34:11363-11372.

Jhaveri, S., Olwin, B., and Ellington, A.D.1998. In vitro selection of phosphorothio-lated aptamers. Bioorg. Med. Chem. Lett. 8:2285-2290.



9.6.30


Kasahara, Y., Irisawa, Y., Ozaki, H., Obika, H., andKuwahara, M. 2013. 2′,4′-BNA/LNA aptamers:CS-SELEX using a DNA-based library of full-length 2′-O,4′-C-methylene-bridged/linked bi-cyclic ribonucleotides. Bioorg. Med. Chem. Lett.23:1288-1292.

Kato, Y., Minakawa, N., Komatsu, Y., Kamiya, H.,Ogawa, M., Harashima, H., and Matsuda, A.,2005. New NTP analogs: The synthesis of 4′-thioUTP and 4′-thioCTP and their utility for SE-LEX. Nucleic Acids Res. 33:2942-2951.

Keefe, A.D. and Cload, S.T. 2008. SELEX withmodified nucleotides. Curr. Opin. Chem. Biol.12:448-456.

Kimoto, M., Yamashige, R., Matsunaga, K.,Yokoyama, S., and Hirao, I. 2013. Generationof high-affinity DNA aptamers using an ex-panded genetic alphabet. Nat. Biotechnol. 31:453-457.

King, D.J., Ventura, D.A., Brasier, A.R., and Goren-stein, D.G. 1998. Novel combinatorial selectionof phosphorothioate oligonucleotide aptamers.Biochemistry 37:16489-16493.

King, D.J., Bassett, S.E., Li, X., Fennewald, S.A.,Herzog, N.A., Luxong, B.A., Shope, R., andGorenstein, D.G. 2002. Combinatorial selectionand binding phosphorothioate aptamers target-ing human NF-kappa B RelA (p65) and p50.Biochemistry. 41:9696-9706.

Koizumi, M., Soukup, G.A., Kerr, J.N., andBreaker, R.R. 1999. Allosteric selection of ri-bozymes that respond to the second messengerscGMP and cAMP. Nat. Struct. Biol. 6:1062-1071.

Kojima, T., Furukawa, K., Maruyama, H., Inoue,N., Tarashima, N., Matsuda, A., and Minakawa,N. 2013. PCR amplification of 4′-thioDNA us-ing 2′-deoxy-4′-thionucleoside 5′-triphosphates.ACS Synth. Biol. 2:529-536.

Kore, A.R., Vaish, N.K., Morris, J.A., and Eckstein,F. 2000. In vitro evolution of the hammerheadribozyme to a purine-specific ribozyme usingmutagenic PCR with two nucleotide analogues.J. Mol. Biol. 301:1113-1121.

Kostyuk, D.A., Dragan, S.M., Lyakhov, D.L.,Rechinsky, V.O., Tunitskaya, V.L., Chernov,B.K., and Kochetkov, S.N. 1995. Mutants of T7RNA polymerase that are able to synthesize bothRNA and DNA. FEBS Lett. 369:165-168.

Kubik, M.F., Bell, C., Fitzwater, T., Watson, S.R.,and Tasset, D.M. 1997. Isolation and character-ization of 2′-fluoro-, 2′-amino-, and 2′-fluoro-/amino-modified RNA ligands to human IFN-gamma that inhibit receptor binding. J. Im-munol. 159:259-267.

Kujau, M.J., Siebert, A., and Wolfl, S. 1997.Design of leader sequences that improve theefficiency of the enzymatic synthesis of 2′-amino-pyrimidine RNA for in vitro selection.J. Biochem. Biophys. Methods 35:141-151.

Kuwahara, M., Takano, Y., Kasahara, Y., Nara, H.,Ozaki, H., Sawai, H., Sugiyama, A., and Obika,S. 2010. Study on suitability of KOD DNA poly-merase for enzymatic production of artificial nu-

cleic acids using base/sugar modified nucleosidetriphosphates. Molecules 15:8229-8240.

Latham, J.A., Johnson, R., and Toole, J.J. 1994.The application of a modified nucleotide in ap-tamer selection: Novel thrombin aptamers con-taining 5-(1-pentynyl)-2′-deoxyuridine. NucleicAcids Res. 22:2817-2822.

Lato, S.M., Ozerova, N.D.S, He, K., Sergueeva,Z., Shaw, B.R., and Burke, D.H. 2002. Boron-containing aptamers to ATP. Nucleic Acids Res.30:1401-1407.

Levy, M. and Ellington, A.D. 2001. Section of de-oxyribozyme ligases that catalyze the formationof an unnatural internucleotide linkage. Bioorg.Med. Chem. 9:2581-2587.

Levy, M. and Ellington, A.D. 2002. In vitro selec-tion of a deoxyribozyme that can utilize multiplesubstrates. J. Mol. Evol. 54:180-190.

Li, M., Lin, N., Huang, Z., Du, L., Altier, C.,Fang, H., and Wang, B. 2008. Selecting ap-tamers for a glycoprotein through the incorpo-ration of boronic acid moiety. J. Am. Chem. Soc.130:12636-12638.

Lin, Y., Qiu, Q., Gill, S.C., and Jayasena, S.D. 1994.Modified RNA sequence pools for in vitro se-lection. Nucleic Acids Res. 22:5229-5234.

Lutz, M.J., Horlacher, J., and Benner, S.A. 1998.Recognition of a non-standard base pair bythermostable DNA polymerases. Bioorg. Med.Chem. Lett. 8:1149-1152.

Lutz, S., Burgstaller, P., and Benner, S.A. 1999.An in vitro screening technique for DNApolymerases that can incorporate modified nu-cleotides. Pseudo-thymidine as a substrate forthermostable polymerases. Nucleic Acids Res.27:2792-2798.

Maden, B.E.H. 1990. The numerous modified nu-cleotides in eukaryotic ribosomal RNA. Prog.Nucl. Acid Res. Mol. Biol. 39:241-303.

Mallikaratchy, P., Tang, Z., Kwame, S., Meng, L.,Shangguan, D., and Tan, W. 2007. Aptamer di-rectly evolved from live cells recognizes mem-brane bound immunoglobin heavy mu chainin Burkitt’s lymphoma cells. Mol. Cell. Pro-teomics. 6:2230-2238.

Marshall, K.A. and Ellington, A.D. 1999. Molecu-lar parasites that evolve longer genomes. J. Mol.Evol. 49:656-663.

Mayer, G. 2009. The chemical biology of aptamers.Angew. Chem. Int. Ed. Engl. 48:2672-2689.

Mehan, M.R., Ostroff, R., Wilcox, S.K., Steele, F.,Schneider, D., Jarvis, T.C., Baird, G.S., Gold,L., and Janjic, N. 2013. Highly multiplexed pro-teomic platform for biomarker discovery, diag-nostics, and therapeutics. Adv. Exp. Med. Biol.734:283-300.

Milligan, J.F. and Uhlenbeck, O.C. 1989. Synthe-sis of small RNAs using T7 RNA polymerase.Meth. Enzymol. 180:51-62.

Minakawa, N., Sanji, M., Kato, Y., and Matsuda,A. 2008. Investigations toward the selection offully-modified 4’-thioRNA aptamers: Optimiza-tion of in vitro transcription steps in the presence


9.6.31


of 4’-thioNTPs. Bioorg. Med. Chem. 16:9450-9456.

Nakamaye, K.L., Gish, G., Eckstein, F., and Vos-berg, H.P. 1988. Direct sequencing of poly-merase chain reaction amplified DNA frag-ments through the incorporation of deoxynu-cleoside alpha-thiotriphosphates. Nucleic AcidsRes. 16:9947-9959.

Nieuwlandt, D., West, M., Cheng, X., Kirshen-heuter, G., and Eaton, B.E. 2003. The first ex-ample of an RNA urea synthetase: Selectionthrough the enzyme active site of human neu-trophile elastase. Chem. Biol. Chem. 4:649-662.

Ochsner, U.A., Katilius, E., and Janjic, N. 2013.Detection of Clostridium difficile toxins A, Band binary toxin with slow off-rate modifiedaptamer. Diagn. Microbiol. Infect. Dis. 76:278-285.

Padilla, R. and Sousa, R. 1999. Efficient synthe-sis of nucleic acids heavily modified with non-canonical ribose 2′-groups using a mutantT7RNA polymerase (RNAP). Nucleic Acids Res.27:1561-1563.

Padilla, R. and Sousa, R. 2002. A Y639F/H784A T7RNA polymerase double mutant displays supe-rior properties for synthesizing RNAs with non-canonical NTPs. Nucleic Acids Res. 30:e138.

Pagratis, N.C., Bell, C., Chang, Y.F., Jennings,S., Fitzwater, T., Jellinek, D., and Dang,C. 1997. Potent 2’-amino- and 2’-fluoro-2’-deoxyribonucleotide RNA inhibitors of ker-atinocyte growth factor. Nat. Biotechnol. 15:68-73.

Perrin, D.M., Garestier, T., and Helene, C. 2001.Bridging the gap between proteins and nucleicacids: A metal-independent RNAseA mimicwith two protein-like functionalities. J. Am.Chem. Soc. 123:1556-1563.

Piccirilli, J.A., Krauch, T., Moroney, S.E., and Ben-ner, S.A. 1990. Enzymatic incorporation of anew base pair into DNA and RNA extends thegenetic alphabet. Nature 343:33-37.

Robertson, M.P. 2001. Ph.D. dissertation. Engi-neered regulation of an RNA ligase ribozyme.University of Texas, Austin, Texas.

Robertson, M.P. and Ellington, A.D. 1999. In vitroselection of an allosteric ribozyme that trans-duces analytes to amplicons. Nat. Biotechnol.17:62-66.

Robertson, M.P., Hesselberth, J.R., and Ellington,A.D. 2001. Optimization and optimality of ashort ribozyme ligase that joins non-Watson-Crick base pairings. RNA 7:513-523.

Ruckman, J., Green, L.S., Beeson, J., Waugh,S., Gillette, W.L., Henninger, D.D.,Claesson-Welsh, L., and Janjic, N. 1998.2′-Fluoropyrimidine RNA-based aptamers tothe 165-amino acid form of vascular endothe-lial growth factor (VEGF165). Inhibition ofreceptor binding and VEGF-induced vascularpermeability through interactions requiringthe exon 7-encoded domain. J. Biol. Chem.273:20556-20567.

Sakthivel, K. and Barbas, C.F. 1998. Expandingthe potential of DNA for binding and catalysis:Highly functionalized dUTP derivatives that aresubstrates for thermostable DNA polymerases.Angew. Chem. Intl. Ed. 37:2872-2875.

Santoro, S.W. and Joyce, G.F. 1997. A general pur-pose RNA-cleaving DNA enzyme. Proc. Natl.Acad. Sci. U.S.A. 94:4262-4266.

Santoro, S.W., Joyce, G.F., Sakthivel, K., Gra-matikova, S., and Barbas, C.F. III. 2000.RNA cleavage by a DNA enzyme with ex-tended chemical functionality. J. Am. Chem.Soc. 122:2433-2439.

Schoetzau, T., Langner, J., Moyround, E., Roehl, I.,Vonhoff, S., and Klussmann, S. 2003. Aminomodified nucleobases: Functionalized nucleo-side triphosphates applicable for SELEX. Bio-conjug. Chem. 14:919-926.

Seelig, B. and Jaschke, A. 1999. A small catalyticRNA motif with Diels-Alderase activity. Chem.Biol. 6:167-176.

Siegmund, V., Santner, T., Micura, R., and Marx, A.2011. Enzymatic synthesis of 2′-methylseleno-modified RNA. Chem. Sci. 2:2224-2231.

Siegmund, V., Santner, T., Micura, R., and Marx,A. 2012. Screening mutant libraries of T7 RNApolymerase for candidates with increased accep-tance of 2′-modified nucleotides. Chem. Com-mun. 48:9870-9872.

Shoji, A., Kuwahara, M., Ozaki, H., and Sawai,H. 2007. Modified DNA aptamer that binds the(R)-isomer of thalidomide derivative with highenantioselectivity. J. Am. Chem. Soc. 129:1456-1464.

Slatko, B.E., Albright, L.M., Tabor, S., and Ju, J.1999. DNA sequencing by the dideoxy method.Curr. Protoc. Mol. Biol. 47:7.4A.1-7.4A.39.

Smith, D., Collins, B.D., Heil, J., and Koch, T.H.2003. Sensitivity and specificity of photoap-tamer probes. Mol. Cell. Proteomics 2:11-18.

Sousa, R. and Padilla, R. 1995. A mutant T7 RNApolymerase as a DNA polymerase. EMBO J.14:4609-4621.

Southworth, M.W., Kong, H., Kucera, R.B., Ware,J., Jannasch, H.W., and Perler, F.B. 1996.Cloning of thermostable DNA polymerasesfrom hyperthermophilic marine Archaea withemphasis on Thermococcus sp. 9°N-7 andmutations affecting 3′-5′ exonuclease activ-ity. Proc. Natl. Acad. Sci. U.S.A. 93:5281-5285.

Tarasow, T.M., Tarasow, S.L., and Eaton, B.E. 1997.RNA-catalysed carbon-carbon bond formation.Nature 389:54-57.

Teramoto, N., Imanishi, Y., and Ito, Y. 2000. Invitro selection of a ligase ribozyme carryingalkylamino groups in the side chains. Biocon-jug. Chem. 11:744-748.

Vaught, J.D., Bock, C., Carter, J., Fitzwater, T.,Otis, M., Schneider, D., Rolando, J., Waugh, S.,Wilcox, S.K., and Eaton, B.E. 2010. Expandingthe chemistry of DNA for in vitro selection. J.Am. Chem. Soc. 132:4141-4151.



9.6.32


Veedu, R.N. and Wengel, J. 2010. Locked nu-cleic acids: Promising nucleic acid analogs fortherapeutic applications. Chem. Biodivers. 7:536-542.

Waters, E.K., Genga, R.M., Schwartz, M.C., Nel-son, J.A., Schaub, R.G., Olson, K.A., Kurz,J.C., and McGinness, K.E. 2011. AptamerARC19499 mediates a procoagulant hemostaticeffect by inhibiting tissue factor pathway in-hibitor. Blood 117:5277-5278.

Wilson, C. and Keefe, A.D. 2006. Buildingoligonucleotide therapeutics using non-naturalchemistries. Curr. Opin. Chem. Biol. 10:607-614.

Wiegand, T.W., Janssen, R.C., and Eaton, B.E.1997. Selection of RNA amide synthases. Chem.Biol. 4:675-683.

Wlotzka, B., Leva, S., Eschgfaller, B., Burmeister,J., Kleinjung, F., Kaduk, C., Muhn, P., Hess-Stumpp, H., and Klussmann, S. 2002. In vivoproperties of an anti-GnRH Spiegelmer: An ex-ample of an oligonucleotide-based therapeuticsubstance class. Proc. Natl. Acad. Sci. U.S.A.99:8898-8902.

Yang, Z., Chen, F., Alvarado, J.B., Benner, S.A.2011. Amplification, mutation, and sequencingof a six-letter synthetic genetic system. J. Am.Chem. Soc. 133:15105-15112.

Zaccolo, M., Williams, D.M., Brown, D.M., andGherardi, E. 1996. An approach to random mu-tagenesis of DNA using mixtures of triphos-phate derivatives of nucleoside analogues. J.Mol. Biol. 255:589-603.

Zon, G. and Geiser, T.G. 1991. Phosphorothioateoligonucleotides: Chemistry, purification, anal-ysis, scale-up and future directions. AnticancerDrug Des. 6:539-568.


9.6.33


Documents

Current Protocols in Nucleic Acid Chemistry || In Vitro Selection Using Modified or Unnatural Nucleotides