Systematic Analysis of 2′-O-Alkyl Modified Analogs for Enzymatic Synthesis and Their Oligonucleotide Properties

Enzymatic oligonucleotide synthesis is used for the development of functional oligonucleotides selected by in vitro selection. Expanding available sugar modifications for in vitro selection helps the functional oligonucleotides to be used as therapeutics reagents. We previously developed a KOD DNA polymerase mutant, KOD DGLNK, that enzymatically synthesized fully-LNA- or 2′-O-methyl-modified oligonucleotides. Here, we report a further expansion of the available 2′-O-alkyl-modified nucleotide for enzymatic synthesis by KOD DGLNK. We chemically synthesized five 2′-O-alkyl-5-methyluridine triphosphates and incorporated them into the oligonucleotides. We also enzymatically synthesized a 2′-O-alkyl-modified oligonucleotide with a random region (oligonucleotide libraries). The 2′-O-alkyl-modified oligonucleotide libraries showed high nuclease resistance and a wide range of hydrophobicity. Our synthesized 2′-O-alkyl-modified oligonucleotide libraries provide novel possibilities that can promote the development of functional molecules for therapeutic use.


Introduction
Enzymatic oligonucleotide synthesis is applied for the development of functional oligonucleotides so that their structure can contribute to their activities such as aptamers and nucleic acid enzymes.Aptamers are single-stranded oligonucleotides that specifically bind target molecules [1].Nucleic acid enzymes are single-stranded oligonucleotides that have catalytic activities such as the site-specific cleavage of RNA [2].These functional molecules are selected from 10 12-15 unique oligonucleotide sequences (oligonucleotide library) using in vitro selection [3,4].Since these functional oligonucleotides mainly consist of DNA or RNA, they are easily digested by nuclease [1,2], which diminishes the potential of using them as therapeutic agents.
The chemical modification of functional oligonucleotides enhances their properties such as their nuclease stability and binding affinity to target molecules.Therefore, expanding the available modified nucleotides for in vitro selection helps to develop functional oligonucleotides.Many nucleobase analogs have been developed and incorporated into aptamers to enhance their binding affinity.For example, incorporating hydrophobic nucleobase analogs like 5-modified uridine analogs into aptamers has been shown to improve their binding affinity [5][6][7][8].On the other hand, since nucleobase-modified analogs contribute less to improving nuclease stability [9], the expansion of available sugar modifications for in vitro selection is necessary to promote the development of the functional molecules for therapeutic use.However, wild-type polymerases have a low tolerance for sugar modifications.
Although several wild-type polymerases have the potential to polymerize sugarmodified nucleotides [18][19][20], sugar modification is still challenging to incorporate by wild-type DNA polymerase.A combination of multiple mutations in thermostable DNA polymerases can improve the polymerization of 2 -O-substituted nucleotides.Several polymerase mutants have been developed to synthesize 2 -OMe-modified oligonucleotides [21][22][23].Chen et al. developed Taq DNA polymerase mutants which could synthesize fully-2 -OMe-modified oligonucleotides [21].Moreover, Freund et al. recently developed Tgo polymerase mutants, which enzymatically synthesized not only fully-2 -OMe-modified oligonucleotides but also 2 -MOE-modified oligonucleotides [22].We also developed a KOD polymerase mutant, KOD DGLNK, for LNA-modified oligonucleotide synthesis, which can also efficiently synthesize fully-2 -OMe-modified oligonucleotides [23].KOD DGLNK had multiple mutations (N210D, Y409G, A485L, D614N, and E664K).The N210D substitution reduces exonuclease activity [24].The Y409 residue is predicted to sterically clash with the 2 -OH of ribonucleotide triphosphates [25].The Y409G substitution reduces this steric clash.The A485L substitution promotes the sending of triphosphates to the active site [26].The D614N and E664K substitutions enhance the binding affinity between the polymerase and primer-template duplex [23,26,27].These multiple mutations significantly promote the synthesis of fully-LNA-and 2 -OMe-modified oligonucleotides.Additionally, we developed a 2 -OMe-and-LNA-mix aptamer [23].However, there has been no systematic analysis of the tolerance range for the enzymatic synthesis of the 2 -O-alkyl modifications.And the properties of synthesized oligonucleotide will also be of interest.

Polymerase Incorporation of 2 -O-Alkyl 5-methyluridine Triphosphates into Oligonucleotide
First, we evaluated the incorporation efficiency of 2 -O-alkyl-5m UTPs using a primer extension assay against a DNA template (Template1) which had a 5nt consecutive adenosine.An extended FAM-labeled DNA primer (Primer1) was analyzed by denaturing polyacrylamide gel electrophoresis (PAGE) (Figure 2).Following an incubation period of 30 min at 74 • C with KOD DGLNK (50 ng/µL), we obtained a fully extended product with dTTP, and all of the 2 -O-alkyl-5m UTPs showed at least single nucleotide incorporation (Figure 2a).This result suggests that 2 -O-alkyl-5m UTPs potentially works as substrates for oligonucleotide polymerization.During the comparison of the relationship between the length of alkyl chains and incorporation efficiency, 2 -OBu-5m UTP, which is the longest alkyl chain, surprisingly showed almost the same incorporation efficiency as 2 -OMe-5m UTP.The efficiency of incorporation improved according to the length of the alkyl chain (incorporation efficiency: 2 -OBu-5m UTP > 2 -OPr-5m UTP > 2 -OEt-5m UTP) (Figure S1).On the other hand, the substitution of a carbon atom for an oxygen atom resulted in decreased incorporation efficiency at any carbon length (incorporation efficiency: 2 -OBu-5m UTP > 2 -MOE-5m UTP, 2 -OPr-5m UTP > 2 -HE-5m UTP) (Figure S1).We saw that 2 -MOE-5m UTP demonstrated higher incorporation efficacy into oligonucleotides than 2 -HE-5m UTP (Figure S1), and 2 -O i Pr-5m UTP showed the lowest incorporation efficiency (Figure S1).Although the steric hindrance of the 2 -O-alkyl modification is known to reduce incorporation efficiency [22], 2 -OBu-5m UTP was more effectively incorporated into DNA than 2 -OEt-5m UTP.Our results indicated that factors other than steric hindrance also impact the incorporation efficiency of 2 -O-alkyl modifications.Next, to enhance the extension efficiency of 2 -O-alkyl-5m UTPs, we attempted to optimize reaction conditions.The addition of Mn 2+ accelerates the incorporation of modified nucleotide triphosphates [21,27,31].Under stronger conditions such as a higher amount of KOD DGLNK (150 ng/µL), the addition of Mn 2+ (1.0 mM), and a longer incubation time (1 h), fully extended products were observed, except for 2 -O i Pr, in the incorporation of 2 -O-alkyl-5m UTPs (Figure 2b).On the other hand, during the incorporation of dTTP, we observed ladder-like bands.Moreover, we observed bands under the primer bands.These bands were probably caused by competition between the polymerization activity and exonuclease activity of KOD DGLNK.Extended products were also observed when Primer1 was replaced with Primer2, which was modified with 2 -OMe modifications (Figure S3).

Enzymatic Synthesis of 2 -O-alkyl-Modified Oligonucleotide Libraries
After the characterization of 2 -O-alkyl-5m UTPs for a short-length template, a fully modified 2 -O-alkyl-modified oligonucleotide synthesis for a longer template which had a 70 nt DNA sequence with a random region in the center (2 -O-alkyl-modified oligonucleotide libraries) was tested.For the incorporation of 50 nucleotide triphosphates, stronger reaction conditions than the reaction conditions for the incorporation of 5 nucleotide triphosphates are assumed to be required.However, such drastic conditions induce digestion on the part of DNA as we observed in Figure 2b.Therefore, primer extension was performed with 2 -OMe-modified primers with 2 -OMe-ATP, 2 -OMe-GTP, 2 -OMe-CTP, and synthesized 2 -O-alkyl-5m UTPs for giving exonuclease resistance to the extended oligonucleotide.In the random region, averagely, a 25% nucleotide would contain a target alkyl modification at the end.As we expected, extension efficiency was shown with the same tendency as the result of the short-length template (Figure 3).Except for the 2 -O i Pr modification, fully extended products were observed.Furthermore, by increasing the Mn 2+ concentration and reaction time, the 2 -O i Pr and 2 -HE modifications were also nicely incorporated and full-length products were observed under 3.0 mM of Mn 2+ following 4 h of incubation (Figure S4).We noticed that, when Mg 2+ and Mn 2+ coexist, 3.0 mM or more of Mn 2+ promotes the enzymatic synthesis of 2 -O-alkyl-modified oligonucleotides (Figure S4).The enzymatically synthesized 2 -O-alkyl-modified oligonucleotide library that include 2 -OEt-5-methyluridine was named ON_Et.Moreover, other 2 -O-alkyl-modified oligonucleotide libraries also named as the same manner (ON_Me, ON_Pr, ON_Bu, ON_ i Pr, ON_HE, and ON_MOE).

Enzymatic Synthesis of 2′-O-alkyl-Modified Oligonucleotide Libraries
After the characterization of 2′-O-alkyl-5m UTPs for a short-length template, a fully modified 2′-O-alkyl-modified oligonucleotide synthesis for a longer template which had a 70 nt DNA sequence with a random region in the center (2′-O-alkyl-modified oligonucleotide libraries) was tested.For the incorporation of 50 nucleotide triphosphates, stronger reaction conditions than the reaction conditions for the incorporation of 5 nucleotide triphosphates are assumed to be required.However, such drastic conditions induce digestion on the part of DNA as we observed in Figure 2b.Therefore, primer extension was performed with 2′-OMe-modified primers with 2′-OMe-ATP, 2′-OMe-GTP, 2′-OMe-CTP, and synthesized 2′-O-alkyl-5m UTPs for giving exonuclease resistance to the extended oligonucleotide.In the random region, averagely, a 25% nucleotide would contain a target alkyl modification at the end.As we expected, extension efficiency was shown with the same tendency as the result of the short-length template (Figure 3).Except for the 2′-O i Pr modification, fully extended products were observed.Furthermore, by increasing the Mn 2+ concentration and reaction time, the 2′-O i Pr and 2′-HE modifications were also nicely incorporated and full-length products were observed under 3.0 mM of Mn 2+ following 4 h of incubation (Figure S4).We noticed that, when Mg 2+ and Mn 2+ coexist, 3.0 mM or more of Mn 2+ promotes the enzymatic synthesis of 2′-O-alkyl-modified oligonucleotides (Figure S4).The enzymatically synthesized 2′-O-alkyl-modified oligonucleotide library that include 2′-OEt-5-methyluridine was named ON_Et.Moreover, other 2′-O-alkyl-modified oligonucleotide libraries also named as the same manner (ON_Me, ON_Pr, ON_Bu, ON_ i Pr, ON_HE, and ON_MOE).The hydrophobicity of the oligonucleotide libraries was evaluated by the retention time on the ion pair RP-HPLC (Figure 4).The longer alkyl chain modification increases the interaction between the oligonucleotide and C18 modification on the reverse phase column, resulting in slow elution.The substitution of carbon to oxygen in the alkyl chain induced fast elution in the column.The hydrophobic tendency of nucleotide triphosphates is reflected in the hydrophobic tendency of the oligonucleotide libraries (Figure S5).ON_Pr was shown to be hydrophobic as the fully-phosphorothioate-modified DNA library (PS_DNA).With only the substitution of 5-methyluridine, ON_Bu was shown to be more hydrophobic than PS_DNA.This property is highly beneficial in terms of expanding the hydrophobic range.

Hydrophobicity of the 2′-O-Alkyl-Modified Oligonucleotide Libraries
The hydrophobicity of the oligonucleotide libraries was evaluated by the retention time on the ion pair RP-HPLC (Figure 4).The longer alkyl chain modification increases the interaction between the oligonucleotide and C18 modification on the reverse phase column, resulting in slow elution.The substitution of carbon to oxygen in the alkyl chain induced fast elution in the column.The hydrophobic tendency of nucleotide triphosphates is reflected in the hydrophobic tendency of the oligonucleotide libraries (Figure S5).ON_Pr was shown to be hydrophobic as the fully-phosphorothioate-modified DNA library (PS_DNA).With only the substitution of 5-methyluridine, ON_Bu was shown to be more hydrophobic than PS_DNA.This property is highly beneficial in terms of expanding the hydrophobic range.

Nuclease Stability of the 2′-O-Alkyl-Modified Oligonucleotide Libraries
The nuclease stability of the 2′-O-alkyl-modified oligonucleotide libraries in 50% FBS was evaluated by denaturing PAGE (Figure 5).Following 4 h of incubation, the intact and digested products were observed.All the 2′-O-alkyl-modified oligonucleotides were more stable than PO_DNA.ON_Bu was as stable as PS_DNA.The substitution of the longer 2′-O-alkyl 5-methyluridine of the oligonucleotides tended to show a higher stability in 50% FBS (stability: ON_Bu > ON_Pr > ON_Et > ON_Me).The replacement of a carbon atom for an oxygen atom reduced stability in 50% FBS (stability: ON_Bu > ON_MOE, ON_Pr > Oligonucleotide BEH C18 column.The column temperature was 50 • C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A was a 100 mM triethylamine acetic acid solution (pH 6.9), and mobile phase B was acetonitrile.The column was initially maintained at 10% mobile phase B, and then at a gradient of 10-40% over 15 min.

Nuclease Stability of the 2 -O-Alkyl-Modified Oligonucleotide Libraries
The nuclease stability of the 2 -O-alkyl-modified oligonucleotide libraries in 50% FBS was evaluated by denaturing PAGE (Figure 5).Following 4 h of incubation, the intact and digested products were observed.All the 2 -O-alkyl-modified oligonucleotides were more stable than PO_DNA.ON_Bu was as stable as PS_DNA.The substitution of the longer 2 -O-alkyl 5-methyluridine of the oligonucleotides tended to show a higher stability in 50% FBS (stability: ON_Bu > ON_Pr > ON_Et > ON_Me).The replacement of a carbon atom for an oxygen atom reduced stability in 50% FBS (stability: ON_Bu > ON_MOE, ON_Pr > ON_HE).On the other hand, after 24 h of incubation, the ratio of the intact product was a little different.This was probably caused by the digestion of the 2 -OMemodified primer region.The column temperature was 50 °C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A was a 100 mM triethylamine acetic acid solution (pH 6.9), and mobile phase B was acetonitrile.The column was initially maintained at 10% mobile phase B, and then at a gradient of 10-40% over 15 min.

Nuclease Stability of the 2′-O-Alkyl-Modified Oligonucleotide Libraries
The nuclease stability of the 2′-O-alkyl-modified oligonucleotide libraries in 50% FBS was evaluated by denaturing PAGE (Figure 5).Following 4 h of incubation, the intact and digested products were observed.All the 2′-O-alkyl-modified oligonucleotides were more stable than PO_DNA.ON_Bu was as stable as PS_DNA.The substitution of the longer 2′-O-alkyl 5-methyluridine of the oligonucleotides tended to show a higher stability in 50% FBS (stability: ON_Bu > ON_Pr > ON_Et > ON_Me).The replacement of a carbon atom for an oxygen atom reduced stability in 50% FBS (stability: ON_Bu > ON_MOE, ON_Pr > ON_HE).On the other hand, after 24 h of incubation, the ratio of the intact product was a li le different.This was probably caused by the digestion of the 2′-OMe-modified primer region.Expanding the available chemical modifications to enzymatic oligonucleotide synthesis accelerates the development of aptamers and nucleic acid enzymes.By mixing 2 -OMe-adenosine, guanosine, and cytidine, oligonucleotide libraries containing 2 -OEt, 2 -OPr-, 2 -OBu or 2 -HE-5-methyluridine were successfully synthesized by KOD DGLNK (Figure 3).This indicated that KOD DGLNK is tolerant of 2 -O-linear alkyl modifications in polymerization.On the other hand, KOD DGLNK showed a lower tolerance for 2 -O-branched alkyl modification, such as 2 -O i Pr modification, than 2 -O-linear alkyl modifications.It would be interesting to explore the kinetics of the incorporation and identify the novel mutations associated with a tolerance for 2 -O-branched alkyl modification.Additionally, we investigated the impact of a partial substitution of 2 -O-alkyl modification on the nuclease stability and hydrophobicity of the oligonucleotide libraries.Despite only incorporating a 5-methyluridine base into the oligonucleotides, ON_Et, ON_Pr, ON_Bu, ON_HE, and ON_MOE demonstrated a different hydrophobicity (Figure 4).This variety of hydrophobicity difference can be a toolbox which offers the option for aptamer selection for individual targets.Enzymatic stability is also an important factor in the efficacy of aptamers and nucleic acid enzymes in vitro and in vivo.Indeed, approved antisense oligonucleotides and siRNA contain chemical modifications to improve nuclease stability [11].Moreover, the approved aptamers, Macugen (pegaptanib) [32] and IZERVAY (avacincaptad pegol) [33,34], have sugar modifications (2 -OMe and 2 -fluoro modifications) and 40 kDa PEG modifications by post-modification.However, following post-modification, even partial modification sometimes reduces their binding affinity to the target molecules.On the other hand, the 2 -O-alkyl-modified oligonucleotide libraries showed higher nuclease stability than DNA (Figure 5).Moreover, ON_Bu showed the equivalent stability to PS-modified DNA.These results indicate that our study can help develop high nuclease-stable functional oligonucleotides without post-modification.To develop such functional oligonucleotides through in vitro selection, it is important to explore the conversion of 2 -O-alkyl-modified oligonucleotide libraries to DNA.Furthermore, it is important to investigate the error rate of incorporation because the addition of Mn 2+ for the synthesis of 2 -O-alkyl-modified oligonucleotide libraries reduces fidelity [35].The 2 -O-alkyl-modified oligonucleotide libraries exhibited a wide range of hydrophobicity and improvement of nuclease stability by only the replacement of 5-methyluridines.The synthesis of 2 -O-alkyl-5m UTPs was systematic and easy; hence, our strategy can be applied for the synthesis of other 2 -O-alkyl 5m UTPs.Moreover, by applying the 2 -O-alkyl-modified oligonucleotides that are synthesized in this study to not only 5-methyluridine but also other nucleobases, 2 -O-alkyl-modified oligonucleotide libraries can show a further wide range of hydrophobicity and enhancement nuclease stability.

General Information
All reagents and solvents were purchased from commercial suppliers and used without purification unless otherwise specialized.All reactions were carried out under Ar atmosphere.PAGE images were recorded on Chemi-Doc (Bio-Rad, Hercules, CA, USA) and analyzed by Image Lab TM (Bio-Rad). 1H, 13 C and 31 P NMR spectra were recorded on an AVHD 400 NB (Bruker Daltonics, Billerica, MA, USA) using CDCl 3 , DMSO-d 6 , and D 2 O as the solvents.Mass spectra of all new compounds were measured on a JMS-700 instrument (JEOL, Tokyo, Japan) (for fast atom bombardment, FAB).For HPLC, Shimadzu SLC-20A3R, LC-20AD, CTO-20AC, SPD-20A, and FRC-10A were utilized.

Expression and Purification of KOD DGLNK
KOD DGLNK was expressed and purified according to a previous report [23].

Enzymatic Stability of Oligonucleotide Libraries
The reaction mixture [oligonucleotide libraries (0.1 µM) and 50% FBS] was prepared.Then, the reaction mixture was incubated at 37 • C.After incubation, the reaction mixture was heated at 94 • C for 5 min.Then, Proteinase K (36 mU/µL) was added and incubated at 37 • C for 30 min.The reaction was stopped by stop buffer (aqueous 3 mM EDTA containing 0.1% bromophenol blue and aqueous 7 M Urea).After adding Cy5-labeled DNA (83 nM) as an internal standard to the reaction mixtures, they were heated for 5 min at 95 • C. The reaction was analyzed by using a 10% denaturing urea-PAGE (30 min, 55 • C).

Ion Pair RP-HPLC Analysis of Oligonucleotide Libraries
Oligonucleotide libraries (5.0 pmol) were injected onto a XBridge Oligonucleotide BEH C18 column.The column temperature was 50 • C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A consisted of 100 mM triethylamine acetic acid solution, pH 6.9, and mobile phase B was acetonitrile.The column was initially maintained at 10% mobile phase B, and then at a gradient of 10% to 40% over 15 min.

Ion Pair RP-HPLC Analysis of Nucleotide Triphosphates
Nucleotide triphosphates (700 pmol) were injected onto a XBridge Oligonucleotide BEH C18 column.The column temperature was 50 • C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A consisted of 100 mM triethylamine acetic acid solution (pH 6.9), and mobile phase B was acetonitrile.The column was initially maintained at 5% mobile phase B, and then at a gradient of 5% to 20% over 15 min.

Conclusions
In conclusion, 2 -OEt, 2 -OPr, 2 -OBu, 2 -O i Pr, and 2 -HE-5m UTPs were newly synthesized and incorporated them into oligonucleotides by KOD DGLNK.By mixing the 2 -OMe modification, oligonucleotide libraries were enzymatically synthesized.These 2 -O-alkyl-modified oligonucleotides showed various hydrophobicities.Furthermore, these 2 -O-alkyl-modified oligonucleotide libraries were more stable than unmodified DNA.Our results can accelerate the development of aptamers or nucleic acid enzymes for therapeutic use.

Figure 4 .
Figure 4. Hydrophobicity analysis of 2′-O-alkyl-modified oligonucleotide libraries.Ion pair RP-HPLC was performed by the injection of 5.0 pmol of oligonucleotide libraries onto an XBridge Oligonucleotide BEH C18 column.The column temperature was 50 °C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A was a 100 mM triethylamine acetic acid solution (pH 6.9), and mobile phase B was acetonitrile.The column was initially maintained at 10% mobile phase B, and then at a gradient of 10-40% over 15 min.

Figure 4 .
Figure 4. Hydrophobicity analysis of 2 -O-alkyl-modified oligonucleotide libraries.Ion pair RP-HPLC was performed by the injection of 5.0 pmol of oligonucleotide libraries onto an XBridgeOligonucleotide BEH C18 column.The column temperature was 50 • C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A was a 100 mM triethylamine acetic acid solution (pH 6.9), and mobile phase B was acetonitrile.The column was initially maintained at 10% mobile phase B, and then at a gradient of 10-40% over 15 min.

Figure 4 .
Figure 4. Hydrophobicity analysis of 2′-O-alkyl-modified oligonucleotide libraries.Ion pair RP-HPLC was performed by the injection of 5.0 pmol of oligonucleotide libraries onto an XBridge Oligonucleotide BEH C18 column.The column temperature was 50 °C, with a flow rate of 1.0 mL/min, and detection at 260 nm.Mobile phase A was a 100 mM triethylamine acetic acid solution (pH 6.9), and mobile phase B was acetonitrile.The column was initially maintained at 10% mobile phase B, and then at a gradient of 10-40% over 15 min.