Directed evolution of artificial enzymes (XNAzymes) from diverse repertoires of synthetic genetic polymers

Abstract

This protocol describes the directed evolution of artificial endonuclease and ligase enzymes composed of synthetic genetic polymers (XNAzymes), using 'cross-chemistry selective enrichment by exponential amplification' (X-SELEX). The protocol is analogous to (deoxy)ribozyme selections, but it enables the development of fully substituted catalysts. X-SELEX is initiated by the synthesis of diverse repertoires (here 10¹⁴ different sequences), using xeno nucleic acid (XNA) polymerases, on DNA templates primed with DNA, RNA or XNA oligonucleotides that double as substrates, allowing selection for XNA-catalyzed cleavage or ligation. XNAzymes are reverse-transcribed into cDNA using XNA-dependent DNA polymerases, and then PCR-amplified to generate templates for subsequent rounds or deep sequencing. We describe methods developed for four XNA chemistries, arabino nucleic acids (ANAs), 2′-fluoroarabino nucleic acids (FANAs), hexitol nucleic acids (HNAs) and cyclohexene nucleic acids (CeNAs), which require ∼1 week per round, and typically 10–20 rounds; in principle, these methods are scalable and applicable to a wide range of novel XNAzyme chemistries, substrates and reactions.

Introduction

XNAs are a novel class of unnatural nucleic acid–like polymers comprising base or backbone chemistries not found in nature¹. We have previously established a series of 'synthetic genetic systems' based on XNAs in which the canonical ribofuranose sugar of DNA and RNA is replaced by a range of five- or six-membered ring congeners, using engineered XNA polymerases and XNA reverse transcriptases (RTs) obtained through the reassignment of substrate specificity of the archaeal B-family polymerase Tgo^2,3. These systems allow sequence information to be transferred from DNA into a variety of XNAs (by the action of XNA polymerases) and 'retrieved' as DNA sequences on demand (by the action of XNA RTs). The capacity for both DNA-templated XNA synthesis and XNA-templated reverse transcription (RT) back into DNA implements an XNA replication cycle (via a DNA intermediate), which is reminiscent of retroviral replication, and which enables repertoire selection and Darwinian evolution of XNAs, as we have previously shown by the evolution of fully substituted nuclease-resistant XNA ligands (XNA aptamers)².

The protocols we describe here were developed and successfully used to demonstrate that several XNA systems (ANAs, FANAs, HNAs and CeNAs) have the capacity for directed evolution of artificial enzymes (XNAzymes) that catalyze RNA cleavage. Furthermore, we describe adaptation of the protocol for the isolation of ligase specific for RNA or XNA substrates and validated in the FANA system⁴. The methods can be used to obtain XNAzymes through an X-SELEX approach (Fig. 1), which is broadly similar to schemes used for ribozyme and DNAzyme selections by many groups⁵. We adapted these standard schemes to incorporate XNA synthesis and postselection XNA RT, leveraging methodologies described by several groups for 'mod-SELEX'⁶ of (partially) modified DNA⁷ and RNA aptamers⁸, as well as DNAzymes⁹; like these methods, XNAzyme sequences are directly recovered in X-SELEX, in contrast to display-based methods (in which library members carry DNA tags that must be maintained during selection and are used for indirect recovery of catalysts' sequences), which have also been suggested for the selection of modified DNAzymes¹⁰. Unlike previous methods, X-SELEX may therefore be used for the directed evolution of enzymes composed of fully unnatural backbones, as we have recently demonstrated⁴.

Figure 1: Overview of X-SELEX.

(a) Diverse repertoires of XNA molecules may be synthesized (using engineered DNA-dependent XNA polymerases. (b) Catalytic XNAs (XNAzymes) may be selected by reacting libraries of XNAs tagged with substrates of interest (e.g., a disease-related RNA) and isolating on the basis of change in substrate (e.g., urea-PAGE gel shift upon RNA hydrolysis). (c,d) XNAzymes may be subsequently reverse-transcribed using engineered XNA-dependent DNA polymerases (c), yielding cDNA that may be amplified to enable either deep sequencing or generation of templates for XNA synthesis (d), and further rounds of X-SELEX.

Applications

XNA selection technologies, in contrast to those involving solely natural nucleic acid backbones, expand selection technologies into novel structure space (i.e., removing structural constraints imposed by natural polymers during selection) and allow investigators to develop novel entities that possess physicochemical properties associated with XNAs, such as greatly enhanced RNA binding and resistance to nucleases. XNAs therefore have the potential to offer solutions to some of the long-standing problems associated with the application of nucleic acid–based therapeutics and sensors^11,12, not least instability in vivo or in the presence of biological fluids and materials. The protocols we present have been validated for the selection of RNA endonucleases and ligases (as well as XNA ligases)⁴, and they could in principle be applied to select highly stable XNAzymes capable of cleaving or ligating any RNA sequence of interest, including viral¹³, oncogenic¹⁴ or defective¹⁵ mRNA, using any backbone chemistry for which a polymerase and RT has been described (e.g., ANA, CeNA, 2′F-DNA, FANA, HNA, locked nucleic acid (LNA), 2′N₃-DNA, or threose nucleic acid (TNA))^2,3,4. Beyond RNA manipulation, it may be possible for these methods to be adapted for the selection of XNA analogs of many existing ribozymes and DNAzymes, such as those catalyzing phosphorylation, Diels-Alder cycloaddition, tRNA aminoacylation and peptide bond formation¹⁶, as well as offering the prospect of expanding the repertoire of nucleic acid–based catalysis to include reactions that cannot be performed by DNA or RNA, such as the assembly or modification of orthogonal synthetic genetic polymers for 'xenobiology'^17,18, as suggested by our demonstration of an XNA-XNA ligase XNAzyme in the FANA chemistry⁴.

Limitations

The first obstacle to the implementation of X-SELEX is the availability of XNA polymerases and xNTP building blocks. Several XNA polymerases have been described (a useful table is provided by Pinheiro and Holliger¹), including mutants of the archaeal B family polymerase Tgo developed in our group—these may be routinely expressed in Escherichia coli and purified using heparin Sepharose and ion-exchange chromatography^2,4. In our hands, we find that these may be stored at −20 °C for up to 10 weeks in polymerase storage buffer without substantial loss of activity. Some xNTPs are commercially available (e.g., FANAs, Metkinen Chemistry); for the other xNTPs (e.g., HNAs: 1, 5-anhydrohexitol nucleic acid triphosphates) for which synthesis procedures are published¹⁹ and feedstock molecules are available, in-house or custom synthesis can be performed within a reasonable budget.

It is important to note that XNA synthesis and replication was realized only recently, and as a technology it is still at an early stage of development—consequently, there are multiple technical challenges that should be considered during the design and implementation of X-SELEX experiments. The two issues that we believe contribute most to XNAzyme selection efficiency are undersampling and fidelity.

Although DNA can be amplified with potentially single-molecule sensitivity, for most XNA chemistries the RT521L-engineered XNA reverse transcriptase² has a much lower detection limit, which precludes the use of highly stringent selection conditions (such as very short reaction times), as a sufficient number of XNA molecules must accumulate for successful postselection amplification of the isolated XNA pool for subsequent selection rounds. The matter is further complicated by substantial (XNA-specific) sequence biases both in synthesis and RT, exacerbated in the case of HNA and CeNA by their considerably higher duplex stability than DNA or RNA and the formation of highly stable secondary structure motifs hampering efficient RT. Furthermore, as we have previously reported^2,4, the aggregate fidelity of the XNA systems is variable, but it is generally 10–100-fold lower than that of DNA systems used for directed evolution. Although the current error rates are quite possibly helpful at early stages of the selection process, they become more problematic in the later stages, as excessive drift makes it more challenging to home in and identify the XNA sequences encoding the best catalysts.

Experimental design Protocol overview.

Synthesis and preparation of single-stranded XNA libraries as described in Steps 1–20 yields XNAs with substrates attached in cis (Fig. 2). XNA libraries are challenged to react (with additional, biotinylated, substrates in trans in the case of ligase XNAzyme selections), as described in Steps 26–28, and the products are isolated (Step 29) by a combination of bead capture and urea-PAGE (Fig. 3). RT reactions must be performed in order to generate cDNA complementary to the selected XNAzymes (Steps 30–33), which can then be isolated and PCR-amplified (Steps 34–45), thus allowing the generation of templates for the synthesis of XNA for subsequent rounds of selection (Step 46A; Fig. 4), or sequencing Step 46B. These stages in the recovery of XNAzymes are discussed further below.

Figure 2: Synthesis and preparation of XNAzyme libraries (Steps 1–20).

(a–c) Diverse repertoires of XNA molecules tagged with substrates in cis may be prepared by DNA-templated XNA synthesis primed using substrate oligonucleotide sequences of interest, allowing, for example, selection for XNAzymes that catalyze RNA (or DNA) endonuclease (a), RNA-RNA ligase (b) or XNA-XNA ligase (c) reactions.

Figure 3: Reaction, selection and reverse transcription of XNAzyme libraries (Steps 26–34).

(a–c) Libraries of XNAs tagged with substrates of interest (e.g., a disease-related RNA) may be challenged to catalyze, for example, RNA hydrolysis (a), RNA-RNA ligation (b) or XNA-XNA ligation (c) reactions, isolated on the basis of change in substrate (e.g., urea-PAGE gel shift) and reverse-transcribed using engineered XNA-dependent DNA polymerases.

Figure 4: Amplification and generation of XNAzyme cDNA templates (Steps 35–46).

(a,b) Reverse-transcribed XNAzyme sequences (cDNA) from, for example, endonuclease (a) or ligase (b) selections, may be amplified in order to prepare single-stranded DNA templates for further rounds of XNA synthesis and selection. For selections involving HNA, an XNA chemistry that does not form stable heteroduplexes with DNA, RT primer binding and extension (Steps 30–33) may be improved by the incorporation of modifications that increase annealing temperature (e.g., 2′-O-methyl); amplification in this case requires a 'second' reverse transcription reaction using a polymerase able to tolerate the presence of the modification in the cDNA product of the XNA RT reaction.

Steps 1–20: library and primer design and preparation.

The XNA polymerases described previously^2,3,4 are able to initiate XNA synthesis from primers composed of, among others, DNA, RNA or FANA, thereby allowing the synthesis of a range of polyclonal chimeric libraries (i.e., pools of degenerate XNAs) tagged with different potential substrates in cis (Fig. 1a). The preparation of primers by solid-phase synthesis allows a range of 5′ chemical modifications (e.g., fluorophores, affinity tags, reactive groups) to be incorporated, provided that these are compatible with XNA synthesis conditions. For RNA endonuclease selections, primer substrates carry fluorophores and biotin tags so that libraries may be routinely purified and visualized. A stretch of 'irrelevant' DNA is also included, upstream of substrate ribonucleotide positions, in order to maximize the difference in molecular weight between cleaved and uncleaved sequences, which is subsequently exploited for isolation by gel shift. For RNA or XNA ligase selections, these features are instead present on the substrate to be ligated in trans, rather than on the primer used for XNA synthesis, which likewise yields a molecular weight difference upon ligation, although careful consideration must then be given as to which strand (biotinylated or nonbiotinylated) is to be recovered; in endonuclease selections, successful XNAzymes remove their biotin tags, whereas in ligase selections XNAzyme sequences become biotinylated (Fig. 2).

Libraries are typically composed of degenerate regions flanked by stretches of complementarity with substrates (although these may also contain randomized positions in order to allow for potential bias in cut or ligation site specificity); see Taylor et al.⁴ for XNAzyme library designs, as well as sequences of substrates and primers that were successful in our hands. We have found that XNAzymes may be selected from both highly diverse XNAzyme libraries (analogous to typical designs used in DNAzyme and ribozyme nuclease selections⁵), as well as 'patterned' libraries containing variants and sequence elements of previously characterized nucleic acid catalysts⁴. In the former approach, XNAzymes may be selected in an unbiased manner, whereas the latter strategy is to adapt an existing catalytic motif to the structural constraints imposed by the chosen XNA chemistry, provided that catalytic phenotypes reside in nearby sequence space.

Steps 21–25: activation of ligase substrates.

To select XNA-XNA ligase XNAzymes, activated XNA substrates must be prepared. In principle, this may be achieved using a variety of leaving groups; however, one must consider whether the activated substrate is to be used as a primer for library synthesis (i.e., attaching substrate to library in cis), in which case the leaving group must be stable during the synthesis reaction. For XNAzyme selections using pyrophosphate leaving groups⁴, which are sufficiently stable during downstream XNA synthesis, triphosphates may be added to primers and substrates as part of the oligonucleotide chemical synthesis procedure, according to the protocol described by Zlatev et al.²⁰. Alternatively, the activating group can be carried by the substrate to be ligated in trans. Our proof-of-principle selections for XNA-XNA ligase XNAzymes⁴ were performed using FANA oligonucleotides, which can be prepared in reasonable yields by chemical synthesis^4,21 using commercially available phosphoramidites (Glen Research). We describe a protocol for the activation of XNA (FANA) oligonucleotides with imidazolide leaving groups²² at the 3′, for use as XNA-XNA ligase substrates.

Steps 26–29: recovery of XNAzyme sequences.

XNA libraries are challenged to react (with additional, biotinylated, substrates in trans in the case of ligase XNAzyme selections) and the products are isolated by a combination of bead capture and urea-PAGE (Fig. 1b). The success or failure of a selection depends on the ability to prepare, recover and enrich active sequences. It is therefore crucial that (i) XNA libraries or pools that contain a sufficient number of active XNAzyme sequences be prepared; (ii) reactions be performed under conditions that allow the number of reacted XNAzymes to lie within the level of detectability of the system; and that (iii) any 'background' phenomena that facilitate the recovery of inactive XNA sequences be kept as low as possible, such that active sequences may be enriched over inactive sequences. Consideration must be given to the reaction duration and conditions. Therefore, if the library design (or limitations in preparation) is such that a very low level of activity is expected (e.g., a highly diverse XNA repertoire is synthesized, representing a sparse sampling of sequence space), or an XNA system with a higher error rate is used, then longer incubation times (days) or more favorable reaction conditions for the chosen chemistry (e.g., higher temperature, pH, divalent cation concentrations) could be used for initial selection rounds (e.g., several days in rounds 1–5), and then stringency could be increased later on (e.g., several hours in rounds 5–15, several minutes in round 15–20). This approach aims to increase the number of catalytic molecules recovered early on, thus allowing enrichment when active sequences are rare, and then later on selecting for (or evolving through point mutation, in an analogous manner to antibody affinity maturation) the most active. However, care must be taken when conditions that enhance substrate reactivity trade off with denaturation of catalysts (e.g., at high temperature, unstable secondary structure in 'immature' catalysts will be disfavored) or uncatalyzed reactions, which enhance 'background'.

In selections involving RNA substrates, contamination of reactions or reaction products with environmental nucleases must be avoided. Nuclease-free reagents and equipment must be used, RNase inhibitor must be added to buffers as appropriate and gloves must be worn throughout. PAGE plates, tanks and so on can be decontaminated with, e.g., RNaseZap (Ambion).

The selections for RNA-modifying XNAzymes described in Taylor et al.⁴ make use of long RNA substrates. The RT-PCR steps were designed in such a way that portions of these substrates must survive the selection conditions in order to provide PCR priming sites in the cDNA generated: for example, in the case of the RNA ligase XNAzyme selections, the successful ligation of the two RNA substrates generates the new priming site, which must be present to generate a successful amplicon; in the case of the RNA endonuclease XNAzyme selections, a sequence in the RNA substrate downstream of the cleavage site must remain intact.

This setup provides a useful safeguard against inadvertent recovery of inactive XNA sequences whose RNA substrates have been degraded by contaminating environmental nucleases (although the possibility of partial degradation yielding amplifiable molecules cannot be ruled out). Although background could, in principle, be reduced by the use of a single ribonucleotide embedded in a more robust substrate scaffold (e.g., DNA or 2′-O-methyl RNA), we do not recommend this if the goal is to select RNA-specific XNAzymes, as this setup has previously been shown to yield DNAzymes that were unable to use all-RNA substrates^23,24.

Steps 30–33: RT of XNAzymes (XNA or RNA-XNA cDNA).

Direct amplification (and/or sequencing) of XNA is not yet possible, and thus XNA RT reactions must be performed in order to generate complementary cDNAs (Fig. 1c), which can then be used as templates for PCR. In the procedure we describe, XNAs are reverse-transcribed using engineered XNA-dependent DNA polymerase RT521L (ref. 2). RT521L also has RNA-dependent DNA polymerase activity, and thus only a single polymerase is required for RT of chimeric RNA-XNA molecules during the selection of XNAzymes that manipulate RNA. The procedure describes the RT reaction for an HNA library as an example. The RT primers for endonuclease selections (Fig. 3a), but not ligase selections (Fig. 3b,c), are modified with a 5′ biotin-triethyleneglycol (TEG) moiety to allow capture and isolation of cDNA; for XNA chemistries that display highly stable XNA:DNA duplexes, this step allows removal of the XNA strand, which may otherwise inhibit or reduce the efficiency of cDNA amplification. In ligase selections, the ('S1') substrate carries a 5′ biotin-TEG moiety that becomes covalently linked to successful XNAzymes and can be likewise used to isolate (unbiotinylated) cDNA. In HNA, which displays weak affinity for cDNA, the RT primer for RT can be modified with a stretch of 2′-O-methyl RNA residues to improve primer binding. These 2′-O-methyl residues necessitate the inclusion of a 'second RT' step performed with a polymerase capable of 2′-O-methyl RNA-dependent DNA polymerase activity.

Steps 34–45: isolation and amplification of first-strand cDNA.

After RT and removal of (RNA-)XNA strands, isolated cDNA is amplified by three consecutive PCR steps (Fig. 4). In RNA endonuclease selections (Fig. 4a), the first ('out-nest') PCR is dependent on a priming site derived from the RNA substrate (recognized by 'primer 1') and a priming site (recognized by 'primer 2') inherited from the RT primer, and which is not present in the XNA library (i.e., a 'single out-nest' PCR). In ligase selections (Fig. 3b,c), the equivalent first PCR also requires the RT primer–derived out-nest site (recognized by 'primer 2'), but in this instance the forward priming site (recognized by 'primer 0') is derived from the ligated S1 RNA substrates (Fig. 3b,c), and thus these are 'double out-nest' PCRs, which are dependent on both ligation of the RNA substrate to the XNAzymes and subsequent RT.

The second PCR is designed to regenerate templates for XNA synthesis for additional rounds of selection. These reactions therefore use 'in-nested' priming sites to yield amplicons with the priming sites for downstream XNA synthesis (recognized by 'primer 1') and XNA RT (recognized by 'primer 3') at the 5′ and 3′ ends of the 'sense' strand, respectively.

Step 46A(i–vii): preparation of single-stranded cDNA templates for further rounds of XNA synthesis and selection.

Recovered DNA can be used as template for a third, large-scale preparative PCR (essentially a continuation of the second, 'in-nest' amplification in Step 43) in order to capture and isolate a single-stranded cDNA template for further rounds of XNA synthesis and selection using standard recombinant DNA methodologies (Fig. 1d).

Step 46B(i–vi): deep-sequencing XNAzyme selection pools.

Rather than the preparative PCR, an alternative reaction, described in Step 46B(i–vi), can be used to amplify cDNA from Step 44 and to append sequences necessary for bridge amplification and 'sequencing by synthesis' using the Illumina MiSeq next-generation sequencing platform²⁵. A small 3-nt degenerate region is added immediately downstream of the 'P5' sequence to increase the diversity of the first few sequencing cycles and facilitate clustering. Six-nucleotide 'barcode' sequences (e.g., ATCACG) can follow these degenerate positions to allow multiplexed sequencing of selection pools—a different P5-containing primer must therefore be used for each selection pool to be sequenced.

Steps 47–53: preparation of candidate XNAzyme sequences for analysis.

Once candidate sequences have been identified, XNAzymes may be synthesized (using the same procedure described for the synthesis and isolation of single-stranded XNA libraries, Steps 1–9) using DNA templates obtained by standard solid-phase synthesis. Consideration must be given to whether XNAzymes are active in cis (as selected) or in trans (i.e., without reaction substrates covalently linked to the XNAzyme). In order to prepare XNAzymes for the characterization of activity in trans in our initial selections⁴, we exploited the resistance of their backbone chemistries to alkaline hydrolysis, as described in Steps 47–53: XNA is synthesized using RNA primers, which are then completely hydrolyzed in high pH, yielding all-XNA sequences.

Controls.

As selections represent a substantial investment of time and effort, it is highly recommended to first perform a mock selection using pre-prepared XNA sequences designed to simulate reacted and unreacted molecules, if possible (e.g., full-length and mock-cleaved RNA-XNA sequences for an RNA endonuclease XNAzyme selection, prepared using either full-length primer-substrate, or a truncated version, in the XNA synthesis reaction). These control experiments allow an evaluation of the efficiency of recovery at each step for known ratios of mock-reacted to unreacted molecules, as well as the level of background. These mock sequences can also be used as positive and/or negative controls in RT-PCR proper, and as molecular-weight markers during the selection workup. Similarly, after the preparation of XNA libraries, we suggest that control samples be directly reverse transcribed and deep sequenced (i.e., without selection) in order to evaluate sequence diversity and library quality.

Once candidate XNAzymes have been identified, control reactions should include substitution with scrambled XNA and/or substitution of cognate substrate(s) with scrambled substrate(s). See Taylor et al.⁴ for further details of XNAzyme characterization experiments.

Materials

REAGENTS

• Acrylamide bis-acrylamide 19:1 (Severn Biotech, cat. no. 20-2400-10)

• Agarose (Life Technologies, cat. no. R0492)

• Ammonium chloride (Fisher Scientific, cat. no. A/3880/60)

• Ammonium sulfate (Fisher Scientific, cat. no. A/6480/53)

• Boric acid (Sigma-Aldrich, cat. no. 31146)

• dNTPs (GE Life Sciences, cat. no. 28-4065-52)

• DTT (Sigma-Aldrich, cat. no. D9779)

• N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC; Sigma-Aldrich, cat. no. E1769)

• EDTA (VWR International, cat. no. 20302.260)

• Ethanol (Sigma-Aldrich, cat. no. 32221)

• Extreme thermostable single-stranded DNA-binding protein (ET-SSB; New England BioLabs, cat. no. M2401S)

• ExoSAP-IT (Affymetrix, cat. no. 78250)

• Formamide (Sigma-Aldrich, cat. no. 47670)

• Gel extraction kit (Qiagen, cat. no. 28704)

• GelStar nucleic acid stain (Lonza, cat. no. 50535)

• Glycerol (VWR International, cat. no. 24388.320)

• Glycogen (Roche, cat. no. 10901393001)

• HEPES (Melford Laboratories, cat. no. B2001(S))

• IGEPAL CA-630 (Sigma-Aldrich, cat. no. 18896)

• Imidazole (Sigma-Aldrich, cat. no. 56748)

• Magnesium sulfate (Sigma-Aldrich, cat. no. 63138)

• Magnesium chloride (Fisher Scientific, cat. no. BP214-500)

• MiSeq v2 50-cycle reagent kit (Illumina, cat. no. 15033623)

• Nuclease-free stainless steel beads (Qiagen, cat. no. 69989)

• Nuclease-free water (Qiagen, cat. no. 129114)

• PCR purification kit (Qiagen, cat. no. 28104)

• PhiX genome library (Illumina, cat. no. 15017397)

• Potassium chloride (Fisher Scientific, cat. no. P/4280/53)

• RNasin (Promega, cat. no. N2511)

• RNaseZap (Ambion, cat. no. AM9780)

• Sodium acetate (AppliChem, cat. no. A3947,0100)

• Sodium chloride (Fisher Scientific, cat. no. S/3120/63)

• Stabilizer (H T Biotechnology/Cambio, cat. no. 020503)

• Streptavidin C1 MyOne Dynabeads (Invitrogen/Life Technologies, cat. no. 65001)

• SYBR Gold nucleic acid stain (Life Technologies, cat. no. S-11494)

• Tris-HCl (Sigma-Aldrich, cat. no. T3253)

• Triton X-100 (Sigma-Aldrich, cat. no. 93443)

• Tween 20 (Sigma-Aldrich, cat. no. P9416)

• Urea (VWR International, cat. no. 28877.260)

• xNTPs (2′-fluoroarabino nucleotide triphosphates: Metkinen Chemistry, cat. nos. 105-07-A, 105-07-C, 105-07-G, 105-07-U.)

• Yeast tRNA (Invitrogen/Life Technologies, cat. no. 54016)

Oligonucleotides

• See Taylor et al.⁴ for a complete list of DNA, RNA and XNA oligonucleotides used in our initial XNAzyme selection experiments. The DNA sequences (obtained from Integrated DNA Technologies) used in the procedure as examples are provided in Table 1

  Table 1 Oligonucleotide sequences.

Polymerases

• DNA-dependent XNA polymerase(s); they must be prepared in-house, e.g., Pol6G12, as described by Pinheiro et al.², Cozens et al.³ and Taylor et al.⁴

• OneTaq (New England BioLabs, cat. no. M0480S)

• SuperTaq (HT Biotechnology/Cambio, cat. no. 020503)

• ThermoScript reverse transcriptase (Invitrogen/Life Technologies, cat. no. 12236-014)

• XNA-dependent DNA polymerase(s), i.e., XNA reverse transcriptase(s); they must be prepared in-house, e.g., RT521, as described by Pinheiro et al.²

EQUIPMENT

• Agarose gel electrophoresis equipment

• Gel documentation system (Gel Doc, Bio-Rad)

• Transilluminator (Dark Reader, Clare Chemical Research)

• Large-format PAGE equipment ('sequencing' range, C.B.S. Scientific)

• High-voltage power supply

• Laser gel scanner (GE Healthcare Typhoon)

• Thermal cycler

• Next-generation sequencer (Illumina MiSeq)

• Magnetic tube stand and/or magnetic particle processor (Thermo KingFisher)

• (Optional) Bead mill homogenizer (Qiagen TissueLyser)

• Vacuum concentrator

• Centrifugal concentrator units, 3,000-Da MWCO (Millipore, cat. no. UFC500396)

• DNA LoBind 1.5-ml tubes (Eppendorf, cat. no. 0030 108.051)

• Spin filter units, 0.45 μM (Corning, cat. no. 8163)

• Protein LoBind 1.5-ml tubes (Eppendorf, cat. no. 0030 108.116)

REAGENT SETUP

XNA synthesis buffer (HNAs), 1×

• XNA synthesis buffer is 20 mM Tris-HCl (pH 8.8), 10 mM ammonium sulfate, 10 mM potassium chloride, 2 mM magnesium sulfate, 0.1% (vol/vol) Triton X-100 and 125 μM (each) xNTPs. A 10× stock can be made, filtered and stored at room temperature (21 °C) for up to 10 weeks. The 1× buffer should be freshly made.

XNAzyme selection buffer, 1×

• XNAzyme selection buffer is 30 mM HEPES (pH 8.5), 150 mM potassium chloride, 25 mM magnesium chloride and 0.5 U/μl RNasein. A 10× stock can be made (without RNasein), filtered and stored at room temperature for up to 10 weeks. The 1× buffer should be freshly made.

XNA RT buffer

• XNA RT buffer is 20 mM Tris-HCl (pH 8.8), 10 mM ammonium sulfate, 10 mM potassium chloride, 6 mM magnesium sulfate, 2% (vol/vol) Triton X-100, 200 μM (each) dNTPs and 1.5 mg/ml yeast tRNA. This buffer should be freshly made.

PCR buffer 1 (second-strand synthesis and 'out-nest' PCR)

• PCR buffer I is available with OneTaq: New England BioLabs, cat. no. M0480S, supplemented with magnesium chloride, or it can be prepared as follows and stored at −20 °C for up to 10 weeks. The buffer contains 20 mM Tris-HCl (pH 8.9), 22 mM ammonium chloride, 22 mM potassium chloride, 4 mM magnesium chloride, 5% (vol/vol) glycerol, 0.05% (vol/vol) Tween 20, 0.06% (vol/vol) IGEPAL CA-630 and 200 μM (each) dNTPs.

PCR buffer 2 ('in-nest' PCR)

• PCR buffer 2 is available with OneTaq: New England BioLabs, cat. no. M0480S, or it can be prepared as follows and stored at −20 °C for up to 10 weeks. The buffer contains 20 mM Tris-HCl (pH 8.9), 22 mM ammonium chloride, 22 mM potassium chloride, 1.8 mM magnesium chloride, 5% (vol/vol) glycerol, 0.05% (vol/vol) Tween 20, 0.06% (vol/vol) IGEPAL CA-630 and 200 μM (each) dNTPs.

PCR buffer 3 (preparative-scale PCR)

• PCR buffer 3 is available with SuperTaq: H T Biotechnology/Cambio, cat. no. 020503, or it can be prepared as follows and stored at −20 °C for up to 10 weeks. The buffer contains 10 mM Tris-HCl (pH 9.0), 50 mM potassium chloride, 1.5 mM magnesium chloride, 0.1% Triton X-100, 0.01% (vol/vol) stabilizer and 200 μM (each) dNTPs.

Streptavidin bead bind and wash buffer (BwB-Tw), 1×

• Streptavidin bead bind and wash buffer is 5 mM Tris-HCl (pH 7.5), 0.5 mM EDTA, 1 M sodium chloride, 0.1% (vol/vol) Tween 20 and 0.5 U/μl RNasein (added during binding steps involving RNA). A 5× stock can be made (without Tween), filtered and stored at room temperature for up to 10 weeks. The 1× buffer is freshly made.

PAGE loading buffer, 1–2×

• PAGE loading buffer is 10 mM Tris-HCl (pH 7.4), 95% formamide and 50 mM EDTA. It can be stored at room temperature for up to 10 weeks.

Polymerase storage buffer, 1×

• Polymerase storage buffer (1×) is 10 mM Tris-HCl (pH 7.4), 50 mM potassium chloride, 0.1 mM EDTA, 0.1 mM DTT, 0.1% (vol/vol) Triton X-100 and 50% (vol/vol) glycerol. A 2× stock can be made (without glycerol), filtered and stored at 4 °C for up to 10 weeks.

Tris-borate-EDTA, 1×

• Tris-borate-EDTA (TBE) is 89 mM Tris-HCl (pH 8.3), 89 mM boric acid and 2 mM EDTA. A 10× stock can be made, filtered and stored at room temperature for up to 10 weeks. The 1× buffer is freshly made.

Procedure

Synthesis of XNA libraries

Timing 3–6 h

Critical Step

The PROCEDURE describes a typical synthesis reaction for the preparation of an HNA library tagged with a substrate for the selection of RNA endonuclease activity⁴; for other XNAs, alternative triphosphates and polymerases are substituted—see Pinheiro and Holliger¹ for a table listing the synthetic genetic systems that have been currently implemented.

• 1 Mix 2 μM DNA template (e.g., 'N40temp', Table 1) and 1 μM primer (e.g., 'NucPrim', Table 1) in XNA synthesis buffer (without polymerase), supplemented with ET-SSB (4% vol/vol), in standard PCR tubes. Yields vary between XNA systems; 10–30 × 100 μl reactions may be necessary in order to obtain ∼1 nmol PAGE-purified single-stranded XNAzyme library for selection reactions (Step 26).

• 2 Heat the reaction to 95 °C for 30 s, and then immediately transfer it to ice for 1 min.

• 3 Add 1 μM DNA-dependent XNA polymerase (e.g., for HNA, pol6G12[I521L]) on ice.

• 4 Incubate the reaction at 50 °C for 0.75 min/nt, followed by incubation at 65 °C for 1.5 min/nt.

Preparation of single-stranded XNA libraries

Timing 16 h

• 5 Wash an appropriate quantity of streptavidin magnetic beads (use 5 μg beads per pmol biotinylated oligo to be captured) with 1× BwB-Tw (3 × 1 ml for 2 min at room temperature) using a magnetic stand or a magnetic particle processor to capture the beads, and allow the supernatant to be discarded after each wash. Vortex the beads thoroughly when resuspending them.

• 6 Dilute the solution of biotinylated molecules to be captured (e.g., XNA synthesis reaction from Step 4) with 5× BwB-Tw to a 1× final concentration, add 0.5 U/μl RNasin, and then add this mixture to the washed streptavidin magnetic beads from Step 5.

• 7 Incubate overnight at 4 °C on a tube rotator.

• 8 Wash the streptavidin beads from Step 7 with 1× BMB-Tw (3 × 1 ml, 1–2 min per wash) at room temperature.

• 9 Denature and recover single-stranded XNA libraries; for the preparation of nuclease or ligase libraries, use option A if primers are biotinylated or option B if templates are biotinylated, respectively (Fig. 2).

  • A Preparation of 5′ biotinylated strand (e.g., for endonuclease selections)

    • i Wash the beads (1 × 1 ml) with 0.1 N NaOH at room temperature using a magnetic stand or a magnetic particle processor to capture the beads, and allow the supernatant to be discarded after each wash.

      Critical Step

      If single-stranded chimeric RNA-XNA is being prepared, this step must be performed as quickly as possible, without incubation, to avoid RNA hydrolysis.

    • ii Wash the beads (1 × 1 ml) with BwB-Tw supplemented with 50 mM Tris (pH 7.4).

    • iii Wash the beads (1 × 1 ml) with nuclease-free water.

    • iv Elute single-stranded XNA from beads by incubation at 80 °C in nuclease-free water (use at least 0.5× original bead volume) for 2 min, and then use a magnetic stand to remove the beads; collect the supernatant and transfer it to a new tube. Alternatively, water may be substituted with PAGE loading buffer and incubated at 95 °C for 2 min if XNA is to be immediately purified by urea-PAGE (see below).

      Critical Step

      Eluted XNA must be rapidly separated from the beads to avoid rebinding. For complete removal of beads, the supernatant can be filtered through a 0.2-μm membrane, using, e.g., Spin-X centrifugal filters according to the manufacturer's instructions. A second elution can be performed to improve yields, by resuspending the beads in nuclease-free water and repeating Step 9A(iv).

  • B Preparation of the unbiotinylated strand (e.g., for ligase selections)

    • i Denature and elute single-stranded XNA from beads by resuspending them in 100 μl of 0.1 N NaOH at room temperature. Immediately pellet the beads using a magnetic stand or a magnetic particle processor, and then transfer the supernatant to a new tube.

      Critical Step

      If single-stranded chimeric RNA-XNA is being prepared, this step must be performed as quickly as possible, without incubation, to avoid RNA hydrolysis.

    • ii Add 60 μl of 1 M Tris (pH 7.4) to the supernatant to neutralize it.

      Pause point

      Samples can be stored at −20 °C for several days.

Purification of single-stranded XNA libraries

Timing 5 h

• 10 Incubate the recovered single-stranded XNA libraries at 95 °C in 1–2× PAGE loading buffer (the volume will depend on the PAGE gel size; 50–300 μl may be accommodated on large-format sequencing-type gels) for 5–10 min, and then quickly transfer to ice.

  Critical Step

  XNAs that form highly stable duplexes may not fully denature in formamide PAGE loading buffer unless they are desalted. In the case of HNA and CeNA XNAzyme libraries⁴, for example, this should be performed by ethanol precipitation (Box 1) rather than using desalting resins, as these XNAs readily bind to cross-linked dextran, thus reducing yields.

• 11 Run libraries in PAGE loading buffer on preparative-scale denaturing polyacrylamide gels (urea-PAGE) until good separation between primer, incomplete and full-length products can be observed. We typically use freshly cast 20 cm × 20 cm × 1.5 mm 10% (wt/vol) acrylamide/bis-acrylamide gels containing 8 M urea and 1× TBE, loaded with 100–200 μl of sample per well (containing 1–2× PAGE loading buffer), 4 or 8 wells per gel. Gels are prewarmed by running them for 30 min before loading the samples; samples are run for 1.5–2 h at 24 W. If necessary, gels can be removed from cassettes and stained by incubating them for 10 min at room temperature in 1× TBE with SYBR Gold stain (1:10,000 (vol/vol)).

  Troubleshooting

• 12 Excise relevant bands using a sterile blade and place them into clean 2-ml Eppendorf tubes. Note that the size of bands corresponding to full-length library will depend on individual library design, but it may be judged using a molecular-weight marker or by comparison with a DNA-equivalent library (prepared using the same templates and primers, but using dNTPs instead of xNTPs in the synthesis reaction), although note that, while this may be informative, DNA and XNA may not have identical mobility characteristics. Successful synthesis of full-length XNA product can be verified by, e.g., mass spectrometry or RT-primer binding.

• 13 Place a nuclease-free stainless steel bead into each tube and shake at frequency 28/s for 10–20 s using a bead mill homogenizer, until the gel pieces become a fine paste.

• 14 Add sufficient nuclease-free water to form a slurry (∼1.5 μl of water per mg gel).

• 15 Rapidly freeze-thaw the gel slurry by incubating the tubes on dry ice for 10 min, and then quickly transfer them to a heat block at 95 °C until defrosted.

• 16 Incubate the gel slurry at room temperature for 1 h, with gentle rotation.

• 17 Transfer the gel slurry (e.g., using a 1,000-μl pipette tip with the end removed) to one or more cellulose acetate spin-filter units.

  Critical Step

  Be careful not to overload the spin-filter unit, or the filter will clog; this will prevent the efficient recovery of oligos. With Spin-X units, use no more than 250 μl of slurry per unit.

• 18 Centrifuge at 16,000g for 10–20 min at room temperature, until the pulverized gel is visibly dry.

• 19 Pipette one volume of nuclease-free water onto the dry gel and mix it carefully (use a sterile pipette tip to gently stir the gel slurry, but avoid damaging the filter).

• 20 Centrifuge at 16,000g for 10–20 min at room temperature. Extracted oligos may be recovered from the filtrate by ethanol precipitation (Box 1).

Box 1: Precipitation and desalting of XNA oligonucleotides • TIMING 2 h

Activation of substrate XNA oligonucleotides (if necessary)

Timing 3 h

• 21 Prepare 3′ phosphorylated FANA oligonucleotides (deprotected and desalted) by solid-phase synthesis^4,21 using a 3′ phosphate CPG support (e.g., 'S1^X', Table 1).

• 22 Resuspend 3′ phosphorylated FANA oligonucleotides to 100 μM in 0.5 M imidazole (pH 6.0).

  Critical Step

  Primary amine-containing buffers (e.g., Tris) must be avoided, as these will react with both the carbodiimide and the phosphorylimidazolide, thus reducing yields of activated XNA.

• 23 Weigh out solid EDC (0.13 μmol per μl of oligo solution to be reacted) and add the oligonucleotide solution.

• 24 Incubate the reaction for 2 h at room temperature.

  Critical Step

  Check that the pH of the reaction remains 5–6, and adjust it with 0.1 M 2-(N-morpholino)ethanesulfonic acid (MES; pH 5) if necessary. Base modification (visible as smears and high-molecular-weight species by PAGE) may occur at a higher pH or at longer incubation times.

• 25 Desalt and remove unreacted EDC by concentrating and diluting the reaction mixture four or five times using a centrifugal concentrator unit with an appropriate molecular-weight cutoff. Imidazolide-activated oligos can be purified by urea-PAGE, although a non-Tris buffer system must be used, such as sodium boric acid²⁶.

  Critical Step

  Care must be taken to prevent subsequent hydrolysis of this activating group: avoid multiple freeze-thaw cycles, heating, precipitation or storage for prolonged periods. Activated FANA should be prepared freshly as required.

  Troubleshooting

Reacting XNA libraries

Timing up to 5 d

• 26 Resuspend ∼1 nmol of PAGE-purified single-stranded XNAzyme library from Step 20 in nuclease-free water to ∼20 μM in DNA LoBind Eppendorf tubes (or protein LoBind for reactions, which will involve imidazolides). For ligase selections, add additional S1 substrate (Fig. 3) at a 1:1 molar ratio (except imidazolide-activated substrates, which should not be added until after Step 27).

  Critical Step

  For the first round of selection, ∼1 nmol of library is a reasonable start; in subsequent rounds, XNA synthesis at this scale (using templates generated by PCR) is impractical—in our hands, 10–100 pmol of purified XNA (resuspended to ∼1 μM) is a more reasonable goal.

• 27 Anneal substrate-XNA libraries by incubation at 80 °C for 1 min, and then allow them to cool to room temperature over a period of 5 min.

  Critical Step

  (XNA-XNA ligase selection) Add imidazolide-activated substrates after XNAzyme annealing, as heating will cause significant hydrolysis of the phosphorylimidazolide.

• 28 Add XNAzyme selection buffer (to 1× final) to annealed XNAs and substrates (∼15 μM in round 1, ∼0.75 μM in subsequent rounds) and incubate the mixture at appropriate reaction temperature. Our initial RNA endonuclease and ligase XNAzymes⁴ were selected at 17 °C in order to minimize background RNA degradation and to allow the selection of XNAzymes with initially only weakly stable secondary structures. Incubation times were initially long (5 d), but they were incrementally reduced with each round, informed by the performance of the polyclonal pool.

  Critical Step

  To minimize buffer-dependent folding lag, when performing subsequent characterization experiments with candidate XNAzymes in trans, XNA and substrates should be first incubated separately in reaction buffer at reaction temperature and then mixed to start reactions.

Isolation of reacted XNAzymes

Timing 20 h

• 29 Stop the reactions and separate the reacted library members from the unreacted ones; the protocol will depend on whether you are selecting endonuclease XNAzymes (option A), which must be separated from unreacted library members bearing 5′ biotinylated primer-substrates, or ligase XNAzymes (option B), which are expected to form covalent bonds with 5′ biotinylated S1 substrates (Table 1) in trans (Fig. 3), and must be separated from unbiotinylated unreacted library members.

  • A Isolation of endonuclease XNAzymes

    • i Stop the reactions by adding four volumes of PAGE loading buffer.

    • ii Denature by incubation at 95 °C for 1 min, and then immediately transfer to ice.

    • iii Size-separate the reactions (and appropriate controls) by urea-PAGE (as described in Step 11).

      Troubleshooting

    • iv Compare the reaction samples before and after incubation with a molecular-weight marker (e.g., a mock reacted XNA control), and extract the area of the gel where cleaved library may be expected to run using a sterile blade. Recover the library from the gel, as described in Steps 13–20.

    • v Perform a streptavidin bead capture of any full-length, uncleaved XNA sequences, as described in Steps 5–7, using excess beads in BwB without Tween. The quantity of beads used should scale with the amount of library reacted and separated; use 5 μg of beads per pmol and aim to capture unreacted full-length library equivalent to 20% of the amount used in Step 26.

      Critical Step

      Beads should be prewashed in BwB-Tw before capture, but Tween is excluded from the capture solution in order to avoid interference in downstream steps.

    • vi Remove the bead supernatant and retain it.

    • vii To the remaining beads, add 100 μl of 0.1 N NaOH in order to elute any cleaved XNAzymes that are still hybridized to the product strand or nonspecifically bound to the beads. Quickly remove the beads (discard) and neutralize the supernatant with 60 μl of 1 M Tris (pH 7.4).

      Critical Step

      In principle, this step could be omitted if the application of a selection pressure for the catalysts to dissociate from products is desired. However, we note that streptavidin-coated beads (both paramagnetic and their agarose/Sepharose counterparts) exhibit substantial nonspecific affinity for nucleic acids, which would have to be blocked or minimized in order to prevent retention of successfully cleaved XNAzymes.

    • viii Combine the supernatants from Step 29A(vi,vii) and spin-filter the mixture to remove any remaining beads. Ethanol-precipitate the filtrate, as described in Box 1, and then proceed to the RT steps below (Steps 30–33).

  • B Isolation of ligase XNAzymes

    • i Stop the reactions by the addition of ten volumes of BwB-Tw supplemented with EDTA (50 mM final).

    • ii Perform a streptavidin bead capture of total S1 substrate (Table 1 and Fig. 3), which will include any ligated XNAzymes, as described in Steps 5–8.

    • iii Wash the beads (1 × 1 ml) with 0.1 N NaOH at room temperature.

      Critical Step

      A harsh denaturing wash is necessary to prevent carry-over of unligated XNA sequences either hybridized to substrate S1 or nonspecifically bound to beads. In the case of RNA ligase XNAzyme selections, subsequent amplification steps require the RNA portion of the recovered RNA-XNA molecules to remain intact; therefore, these steps must be performed as quickly as possible, at room temperature.

    • iv Wash the beads (1 × 1 ml) with BwB-Tw supplemented with 50 mM Tris (pH 7.4).

    • v Wash the beads (1 × 1 ml) with nuclease-free water.

    • vi Resuspend the beads in PAGE loading buffer (100–200 μl) and incubate them at 95 °C for 2 min to elute substrates and XNAzymes from beads.

    • vii Remove the beads (discard) and immediately transfer them to ice.

    • viii Size-separate the reactions (and appropriate controls) by urea-PAGE (as described in Step 11).

      Troubleshooting

    • ix Compare the reaction samples before and after incubation with a molecular-weight marker (e.g., a mock-reacted XNA control), and extract the area of the gel where ligated library may be expected to run using a sterile blade. Recover the library from the gel, as described in Steps 13–20, precipitate it as described in Box 1 and proceed to RT steps below (Steps 30–33).

RT of XNAzymes (XNA or RNA-XNA → DNA)

Timing 5 h

• 30 Resuspend selected XNA templates from Step 29 in 10 μl of PCR-grade (nuclease-free) water and add 0.2 μM RT primer (e.g., 'Tag4test7_2Me', Table 1) in XNA RT buffer (without polymerase).

• 31 Incubate the reaction at 95 °C for 30 s, and then immediately transfer it to ice for 1 min.

• 32 Add 0.2 μM XNA-dependent DNA polymerase RTI521L.

• 33 Incubate the reaction at 65 °C for 2.5 min/nt.

Isolation and amplification of first-strand cDNA

Timing 20 h

• 34 Perform a streptavidin bead capture of XNA RT reaction products using the procedure described in Steps 5–8, and then denature and recover the first-strand cDNA; if biotinylated RT primer was used (endonuclease selections), follow option A; if the XNAzyme strand itself is biotinylated (ligase selections) and unbiotinylated RT primer was used, follow option B.

  • A Isolation of biotinylated cDNA (endonuclease selections)

    • i Wash the beads (1 × 1 ml) with 0.1 N NaOH at room temperature.

    • ii Wash the beads (1 × 1 ml) with BwB-Tw supplemented with 50 mM Tris (pH 7.4).

    • iii Wash the beads (1 × 1 ml) with nuclease-free water.

    • iv Elute biotinylated cDNA from beads by incubation at 80 °C in nuclease-free water (use at least 0.5× the original bead volume) for 2 min. The eluate may be used directly in amplification reactions.

  • B Isolation of unbiotinylated cDNA (ligase selections)

    • i Denature and elute unbiotinylated cDNA from beads using 100 μl of 0.1 N NaOH at room temperature.

    • ii Add the eluate to 60 μl of 1 M Tris (pH 7.4) to neutralize.

    • iii Precipitate and desalt cDNA, as described in Box 1.

• 35 Resuspend the first-strand cDNA in 50 μl of PCR buffer 1, and add 0.5 μM (each) 'out-nest' amplification primer set (Fig. 4): primer 0 (ligase selections, Fig. 4b,c) or primer 1 (endonuclease selections, Fig. 4a); e.g., for the RNA endonuclease HNAzyme library example described above, primer 1 = 'dP2'; Table 1 and primer 2 (e.g., after RT with the primer 'Tag4test7_2Me' described in Steps 30–33, primer 2 = 'Tag4'; Table 1), in standard PCR tubes.

• 36 Incubate the mixture at 95 °C for 30 s, and then immediately transfer it to ice for 1 min.

• 37 Add the polymerase(s): a blend of 0.025 U/μl OneTaq hot start polymerase (a blend of Taq and Deep Vent_R) and 0.15 U/μl ThermoScript.

  Critical Step

  ThermoScript is an avian myeloblastis virus (AMV)-RT mutant with 2′-O-Me RNA-dependent DNA polymerase activity, used in this initial PCR when primer 'Tag4test7_2Me' described in Table 1 was used for XNA RT, thus enabling second-strand synthesis to proceed through the patch of 2′-O-Me residues in the cDNA template derived from this primer. It is reasonably tolerant to elevated reaction temperatures (70–80 °C for short periods), but it is denatured by incubation at 95 °C, and thus it must be added after the previous (primer annealing) step.

• 38 Perform the second-strand synthesis by thermal cycling: 80 °C for 30 s, 52 °C for 30 s and 72 °C for 15 min.

• 39 Without removing the tube(s) from the thermal cycler, incubate the mixture at the following temperatures: 94 °C for 1 min, 20–35 cycles (94 °C for 30 s, 54 °C for 30 s and 72 °C for 30 s), and 72 °C for 1 min.

  Critical Step

  In subsequent rounds, if active XNAzymes are becoming enriched in the selection pool, the number of cycles should be decreased. It is recommended to remove aliquots for analysis beginning at cycle 10 and to perform additional cycles only when an amplicon is not yet observed. Too many cycles at any of the PCR amplification steps described may risk preferential amplification of parasites (i.e., artifact sequences containing one or both primer sequences, which originate from mis-priming or primer dimerization) and/or by-products, which inhibits the recovery of XNAzyme sequences and potentially leads to failure of selections. Selection pools are especially prone to 'ladder-type' by-product formation²⁷ in later rounds, as sequences are designed with complementarity that may be expected to be even more extensive in active XNAzymes under enrichment, as secondary structures and XNAzyme:substrate hybridization regions evolve. Emulsion PCR has been suggested as a potential solution to amplification problems in aptamer SELEX²⁸ and phage display²⁹, but it has yet to be explored for XNAzyme selection.

• 40 Add 2 μl of ExoSAP-IT (a blend of exonuclease I and shrimp alkaline phosphatase) per 5 μl of PCR and incubate at 37 °C for 15 min to degrade the primers.

• 41 Incubate at 80 °C for 15 min to inactivate ExoSAP-IT.

• 42 Remove 1 μl of PCR and add it to 49 μl of PCR buffer 2 with polymerase: 0.025 U/μl OneTaq hot start polymerase.

• 43 Add the 'in-nest' amplification primer set (Table 1 and Fig. 3; 0.5 μM final, each): primer 1 and primer 3 (e.g., for the RNA endonuclease HNAzyme library example described above, primer 1 = 'dP2' and primer 3 = 'R3Test 7'; Table 1).

• 44 Perform 'in-nest' PCR by thermal cycling: 94 °C for 1 min, 10–20 cycles X (94 °C for 30 s, 54 °C for 30 s and 72 °C for 30 s) and 72 °C for 2 min.

  Critical Step

  Again, it is recommended to remove aliquots for analysis beginning at cycle 10 and to perform additional cycles only when an amplicon is not yet observed.

• 45 Size-separate 20–30 μl of the reaction by preparative-scale agarose gel electrophoresis, together with a DNA molecular-weight marker (we typically run 4% (wt/vol) agarose gels, containing 1× TBE and GelStar stain (1:10,000 (vol/vol)), in 1× TBE for 30–60 min at 150 V/cm) and excise bands (visualized using a blue light transilluminator) corresponding to the expected molecular weight of the amplicon. Recover DNA from agarose using a gel extraction kit (Qiagen) as per the manufacturer's instructions.

  Troubleshooting

  Pause point

  These purified dsDNA versions of XNAzyme selection pools can be stored at −20 °C indefinitely.

Use of first-strand cDNA

• 46 Amplified cDNA complementary to the selected XNAzymes can then be used for the generation of templates for synthesis of XNA for subsequent rounds of selection (option A; Fig. 4), or for sequencing (option B).

  • A Preparation of single-stranded cDNA templates for further rounds of XNA synthesis and selection

    • i Add an aliquot (∼100–200 ng) of the purified in-nest PCR products from Step 45 to 2 ml of PCR buffer 3 with polymerase: 0.05 U/μl SuperTaq.

    • ii Add 1 μM (each) biotinylated variants of the 'in-nest' amplification primer set (Table 1 and Fig. 4):

      For endonuclease selections, use 5′ biotin-primer 1 and (unbiotinylated) primer 3′ (e.g., for the RNA endonuclease HNAzyme library example described above, primer 1 = '(5′ biotin)dP2' and primer 3′ = 'Test 7-2'; Table 1). For ligase selections, use (unbiotinylated) primer 1 and 5′ biotin-primer 3′.

    • iii Perform PCR by thermal cycling: 94 °C for 1 min, 10–15 cycles (94 °C for 30 s, 54 °C for 30 s and 72 °C for 30 s), and 72 °C for 2 min.

      Critical Step

      Again, remove aliquots for analysis beginning at cycle 10—be cautious not to perform too many cycles; as a rough rule of thumb, do not cycle more than two or three times once an amplicon is observable by agarose gel electrophoresis.

    • iv Perform a streptavidin bead capture as described in Steps 5–8.

    • v Elute cDNA template strands as described in Step 9. For endonuclease selections, elute (unbiotinylated) ssDNA template in NaOH and neutralize it according to Step 9B; for ligase selections, remove nonbiotinylated template strand by a denaturing (NaOH) wash, and then elute the (biotinylated) ssDNA template in water according to Step 9A.

    • vi Precipitate and desalt eluted templates as described in Box 1.

    • vii Proceed to XNA synthesis for further rounds of selection (return to Step 1).

      Critical Step

      It is essential to precipitate and desalt even those ssDNA templates eluted in water, as polymerases used for XNA synthesis seem to be inhibited by material eluted directly from beads. We have found that XNA synthesis may be performed using biotinylated templates captured on streptavidin beads; however, in many cases this appears to be more problematic than reactions in solution.

    Timing 16 h

  • B Deep-sequencing XNAzyme selection pools

    • i Add an aliquot (∼10 ng) of the purified in-nest PCR product (from Step 44) to 100 μl of PCR buffer 2 with polymerase: 0.025 U/μl OneTaq hot start polymerase.

    • ii Add 0.1 μM (each) variant in-nest amplification primer set (Table 1): P5-primer 1 and P3-primer 3 (e.g., for the RNA endonuclease HNAzyme library example described above, P5-primer 1 = 'P5_P2' and P3-primer 3 = 'P3 Test 7-2'; Table 1).

    • iii Perform PCR by thermal cycling: 94 °C for 1 min, 10–15 cycles (94 °C for 30 s, 54 °C for 30 s and 72 °C for 30 s), and 72 °C for 2 min.

    • iv Remove the primers, polymerase and buffer using a PCR purification kit (Qiagen), as per the manufacturer's instructions.

    • v Prepare the libraries and sequence (12 pM final with 20% PhiX genome library) them using a MiSeq reagent kit and instrument (Illumina) as per the manufacturer's instructions.

    • vi Examine deep-sequencing data using a suitable bioinformatics tool, such as the Galaxy server^30,31,32. For example, to display MiSeq data (FASTQ format) by abundance (i.e., to rank the frequency with which sequences occur) using Galaxy tools, use the following workflow: ('Get Data' menu) Upload File → ('NGS: QC and manipulation' menu) FASTQ Groomer → (if multiplexed) Barcode Splitter → Upload File (data sets split by barcode) →; FASTQ to FASTA Converter → Collapse.

      Critical Step

      In our initial XNAzyme selections⁴, we were able to identify active XNAzyme sequences simply by screening some of the most abundant clones, although these are not necessarily the most active sequences. In future investigations, a more thorough examination of the enrichment dynamics of XNAzyme selections may allow candidates to be identified in fewer rounds^33,34.

    Timing 16 h

Preparation of candidate XNAzyme sequences for analysis

Timing 1–2 d

• 47 Obtain biotinylated DNA templates for candidate XNAzymes (for analysis of single sequences, use solid-phase–synthesized clonal oligos complementary to sequences identified in Step 46B(vi), with appropriate priming sites (see CRITICAL STEP below). Alternatively, for the analysis of bulk polyclonal activity of selection pools, use templates from Step 46A(vi) and synthesize XNA oligonucleotides using an appropriate primer, using the same procedure as Steps 1–4.

  Critical Step

  The primer in this reaction should be either the full primer-substrate ('NucPrim' in the example endonuclease selection; Table 1), as used in library synthesis (Steps 1–4) for subsequent preparation of candidate XNAzymes in cis format, i.e., with substrate attached as selected, or an RNA version of the primer 1 sequence ('rP2' in the example endonuclease selection; Table 1), for subsequent preparation in trans format, i.e., with substrates separate.

• 48 Prepare single-stranded RNA-XNA oligonucleotides by streptavidin bead capture and denaturing elution in NaOH, using the same procedure as Steps 5–9B; for XNAzymes to be analyzed in cis format, neutralize and purify as described in Steps 10–20; for XNAzymes to be analyzed in trans format, do not perform neutralization (Step 9B(ii)), but proceed directly to Step 49.

• 49 Add additional NaOH to a 0.7 N final concentration.

• 50 Incubate at 65 °C for 1 h to fully hydrolyze the RNA.

• 51 Add an equal volume of 1 M Tris (pH 7.4) to neutralize.

• 52 Precipitate and desalt as described in Box 1 and/or purify XNA using urea-PAGE (as described in Steps 10–20).

  Critical Step

  If XNA primers (or indeed full XNAzymes) can be prepared through solid-phase chemical synthesis (e.g., FANA^4,21), these may be used (with biotinylated DNA templates) in synthesis reactions (Steps 1–4) to generate the remainder of the XNAzyme sequence, which can then be isolated (Steps 5–9B) and purified (Steps 10–20).

• 53 React XNAzymes in cis or trans using the same procedure as Steps 26–28, annealing and adding substrates separately for reactions in trans. Analyze the reactions by urea-PAGE, as described in Steps 10 and 11. See Taylor et al.⁴ for further details of XNAzyme-characterization experiments.

  Troubleshooting

Troubleshooting

Troubleshooting advice can be found in Table 2.

Table 2 Troubleshooting table.

Timing

Steps 1–4, synthesis of XNA libraries: 3–6 h

Steps 5–9, preparation of single-stranded XNA libraries: 16 h

Steps 10–20, purification of single-stranded XNA libraries: 5 h

Steps 21–25, activation of substrate XNA oligonucleotides (if necessary): 3 h

Steps 26–28, reacting XNA libraries: up to 5 d

Step 29, isolation of reacted XNAzymes: 20 h

Steps 30–33, RT of XNAzymes (XNA or RNA-XNA → DNA): 5 h

Steps 34–45, isolation and amplification of first-strand cDNA: 20 h

Steps 46, use of first-strand cDNA: 16 h

Steps 47–53, preparation of candidate XNAzyme sequences for analysis: 1–2 d

Box 1, precipitation and desalting of XNA oligonucleotides: 2 h

A single XNAzyme selection cycle requires roughly 1 week, not including the XNAzyme reaction time. Our preferred workflow is to perform lengthy primer-extension reactions (XNA synthesis and RT) and bead-capture steps overnight. However, when RNA-XNA molecules (XNA libraries with RNA substrates in cis) are synthesized, storage and handling time should be kept to a minimum in order to reduce uncatalyzed hydrolysis of RNA. Longer breaks can be timed to coincide with the steps after RNA-XNA RT, as purified DNA versions of selection pools (i.e., after RT-PCR) can be stored at −20 °C.

Anticipated results

For the selections described in Taylor et al.⁴, starting from both 'patterned' and highly diverse libraries, efficiency of XNA synthesis and RT-PCR recovery improved substantially throughout initial rounds, presumably as 'difficult' sequences were quickly lost from the pools. Bulk polyclonal XNAzyme activities were compared at each round with that of initial libraries by PAGE. It was generally found that activities remained below detection levels until rounds 10–15. In pools displaying bulk polyclonal activity, examination of the most abundant sequences revealed by deep sequencing was sufficient to discover active XNAzymes; it may, however, be possible to identify sequences under enrichment through a more thorough examination of deep-sequencing data^33,34, thereby reducing the number of rounds of selection required.

Alternatively (e.g., if access to deep sequencing technology is limited), colony PCR and standard Sanger sequencing may be informative, but only in cases in which sequence convergence is very high, which may be possible if many rounds of selection are applied, at the risk of parasite sequence expansion. In our hands, the individual XNAzymes characterized in Taylor et al.⁴ typically represented <1% of the sequences in their selection pools, even after 10–15 rounds.