Download PDF
Synlett 2024; 35(06): 677-683
DOI: 10.1055/a-2212-7704
cluster
Special Issue to Celebrate the Centenary Year of Prof. Har Gobind Khorana

Facilitated Synthetic Access to Boronic Acid-Modified Nucleoside Triphosphates and Compatibility with Enzymatic DNA Synthesis

Germain Niogret‡
a   Institut Pasteur, Université Paris Cité, Department of Structural Biology and Chemistry, Laboratory for Bioorganic Chemistry of Nucleic Acids, CNRS UMR3523, 28, rue du Docteur Roux, 75724 Paris Cedex 15, France
c   Chimie ParisTech, PSL University, CNRS, Institute of Chemistry for Life and Health Sciences, Laboratory for Inorganic Chemical Biology, 75005 Paris, France
,
Pascal Röthlisberger  ‡
a   Institut Pasteur, Université Paris Cité, Department of Structural Biology and Chemistry, Laboratory for Bioorganic Chemistry of Nucleic Acids, CNRS UMR3523, 28, rue du Docteur Roux, 75724 Paris Cedex 15, France
,
Fabienne Levi-Acobas
a   Institut Pasteur, Université Paris Cité, Department of Structural Biology and Chemistry, Laboratory for Bioorganic Chemistry of Nucleic Acids, CNRS UMR3523, 28, rue du Docteur Roux, 75724 Paris Cedex 15, France
,
Frédéric Bonhomme 
b   Institut Pasteur, Université Paris Cité, Department of Structural Biology and Chemistry, Unité de Chimie Biologique Epigénétique, UMR CNRS 3523, 28 rue du Docteur Roux, CEDEX 15, 75724 Paris, France
,
Gilles Gasser 
c   Chimie ParisTech, PSL University, CNRS, Institute of Chemistry for Life and Health Sciences, Laboratory for Inorganic Chemical Biology, 75005 Paris, France
,
Marcel Hollenstein  ∗
a   Institut Pasteur, Université Paris Cité, Department of Structural Biology and Chemistry, Laboratory for Bioorganic Chemistry of Nucleic Acids, CNRS UMR3523, 28, rue du Docteur Roux, 75724 Paris Cedex 15, France
› Author Affiliations

G.N. gratefully acknowledges a fellowship from the doctoral school MTCI from Université Paris Cité. The authors gratefully acknowledge financial support from Institut Pasteur. M.H. acknowledges financially support from the ANR grant PEPR REV CNRS MOLECULARXIV.
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

Decorating nucleic acids with boronic acids can extend the usefulness of oligonucleotide-based tools to the development of medical imaging agents, the promotion of binding of aptamers to markedly more challenging targets, or the detection of (poly)saccharides. However, due to the hygroscopic nature and high intrinsic reactivity of boronic acids, protocols for their introduction into nucleic acids are scarce. Here, we have explored various synthetic routes for the crafting of nucleoside triphosphates equipped with phenylboronic acids. Strain-promoted azide–alkyne cycloaddition appears to be the method of choice for this purpose and it enabled us to prepare a modified nucleotide. Enzymatic DNA synthesis permitted the introduction of up to thirteen boronic acid residues in oligonucleotides, which bodes well for its extension to SELEX and related methods of in vitro selection of functional nucleic acids.


#

Key words

nucleoside triphosphates - click chemistry - DNA - boron chemistry - SELEX

Aptamers[1] [2] and DNAzymes[3] are single-stranded oligonucleotides capable of selective binding to specific targets with high affinity and catalysis of chemical transformations, respectively. These functional nucleic acids are all obtained by application of a Darwinian in vitro evolution method named SELEX (SElection of Ligands by EXponential enrichment).[4] Aptamers are now well-established tools in numerous applications,[2] [5] and their general usefulness has been undisputed since the approval of Macugen (pegaptanib sodium) by the US Food and Drug Administration.[6] [7] Similarly, although DNAzymes represent more recent additions to the range of functional nucleic acids, the first two clinical trials of DNAzymes against nodular basal-cell carcinoma[8] and asthma[9] have recently been reported. Nonetheless, functional nucleic acids suffer from shortcomings that restrict some in vivo applications. Consisting of DNA or RNA oligonucleotides, they are prone to rapid nuclease-mediated degradation, display poor cellular penetration capacities, and generally exhibit short in vivo residence times due to efficient renal filtration.[1] [10] These limitations can be at least partially remediated by introducing chemical modifications into the nucleoside scaffolds of aptamers, DNAzymes, and therapeutic oligonucleotides in general.[7] [11] [12] These additional groups can be introduced into functional nucleic acids either post-selection by automated solid-phase synthesis (the post-SELEX approach) or by supplementing selection experiments with modified nucleoside triphosphates (dN*TPs; the mod-SELEX approach).[12] [13] The latter approach avoids uncertain and tedious engineering of functional nucleic acids by allowing the direct selection of modified sequences.[14] [15] [16] The polymerization of dN*TPs broadens the chemical space that can be explored during SELEX, but also permits the introduction of functional groups alien to biology[17] that can be harnessed in applications ranging from electrochemical analysis[18] and surface immobilization[19] to artificial metal base-pair formation.[20] In this context, boronic acid moieties specifically interact with cis-1,2- or -1,3-diols under aqueous conditions to form the corresponding cyclic boronic esters in a covalent yet reversible manner.[21] Due to these intrinsic properties, this functional group has been extensively used for the development of biosensors,[22] catalysts,[23] [24] and imaging agents;[25] the construction of oligonucleotide scaffolds[26] and therapeutic agents;[27] and for surface immobilization.[28] In addition, the decoration of nucleotides with boronic acids has been suggested to be useful for SELEX to raise aptamers against more-demanding targets, such as glycosylated proteins.[23] [29] [30] In addition, appendage of boronic acids onto aptameric scaffolds could be harnessed for surface immobilization,[31] post-SELEX modification strategies,[32] or the development of potent tools for medical-imaging applications.[33] However, despite these intrinsic and alluring properties, synthetic protocols for facile and robust access to boronic acid-modified DNA nucleobases are scarce.[34] [35] [36]

Here, we have explored various synthetic strategies based on copper(I)-catalyzed alkyne–azide cycloaddition (CuAAC or click chemistry)[16] [35] [37] and strain-promoted copper-free click reactions [38] for the reliable introduction of boronic acid modifications onto the nucleobase of dN*TPs (Figure [1]). We demonstrate that fine tuning of the chemical nature of the substituents on the boronic residues is of critical importance in this approach. Also, we highlight the usefulness of copper-free click reactions for the preparation of such modified nucleotides. Lastly, we highlight the successful enzymatic incorporation of boronic-acid-modified dN*TPs into DNA oligonucleotides by polymerase-mediated synthesis.

Zoom Image
Figure 1 Chemical structures of dUBOHFTP (1), dUPhFTP (2), dUBOHTP (3), and dUCOBOHTP (4)

Our initial design was based on the (m-fluorophenyl)boronic acid motif, which is particularly adapted for efficient potential interactions between aptamer and sugar-containing targets, as it displays very high forward rate constants (k on of about 340 M–1 s–1) for the formation of boronates and favorable equilibrium constants (K eq in the 10–5 to 10–6 M–1 range) of the resulting boronates in aqueous media.[39] Particularly, whereas ortho-fluoro derivatives display reduced hydrolytic stabilities, (4-carboxy-3-fluorophenyl)boronic acid-based analogues are stable and have been used in various applications, including the immobilization of RGD-peptides on solid supports.[28] Despite the presence of copper salts, the CuAAC reaction was recently shown to be compatible with the introduction of boronic acids into DNA and RNA oligonucleotides.[23] [30] [34] [40] On the basis of these interesting properties, we attempted to examine whether triphosphates obtained by the modification of the highly useful synthon 5-ethynyl-2′-deoxyuridine 5′-triphosphate (EdUTP)[41,42] by the CuAAC reaction could serve as substrates for polymerases. Consequently, we set out to synthesize dUBOHFTP (1) by first converting the pinacol-protected (m-fluorophenyl)boronic acid derivative 5 into the corresponding azide 7 under standard amide-bond-forming conditions, following hydrolysis of the methyl ester group (Scheme [1]). Next, EdUTP was subjected to the CuAAC reaction with azide 7, which yielded the pinacol-protected triphosphate exclusively (data not shown). The protecting group could be removed by mild treatment of the nucleotide analogue with ammonium hydroxide. Surprisingly, besides the expected triphosphate 1 (18% yield), the deboronated derivative 2 was formed in a rather high yield (29%). The formation of undesired nucleotide 2 was attributed to the electron-withdrawing meta-fluorine of the arylboronic residue, which can induce efficient protodeboronation of arylboronic acids under the basic conditions required for deblocking of the pinacol protecting group.[43] This analysis of the resulting products was further complicated by the complex behavior of arylboronic acid derivatives under mass spectrometry conditions.[44]

Zoom Image
Scheme 1 Synthesis of modified triphosphates 1 and 2. Reagents and conditions: (i) THF, NaOH, RT, 4h; (ii) (a) (COCl)2, DMF (cat.), CH2Cl2, RT, 1 h; (b) 2-azidoethylamine, DMAP, NEt3, CH2Cl2, THF, RT, 12 h, 31% over 2 steps; (iii) (a) EdUTP, CuBr, tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine (TBTA), H2O, DMF, RT, 4 h, (b) NH4OH, RT, 2 h; 1 (18%) and 2 (29%).

The apparent intrinsic reduced stability of boronic acid 1 under the basic reaction conditions required for the removal of the protecting group prompted us to synthesize more-stable azide building blocks in which the meta-fluorine substituent was omitted (nucleotide 3 in Figure [1]). In addition, we compared synthetic protocols involving copper-dependent click reactions to those involving strain-promoted copper-free click reactions. The synthetic route to triphosphate 3 is similar to that of nucleotide 1 with the exception that the removal of the pinacol blocking group was carried out at the level of the azide derivative 9, prior to the CuAAC reaction (Scheme [2]). Characterization by NMR, HRMS, and MALDI-ToF of nucleotide 3 [45] revealed a predominant compound corresponding to an oxidatively deborylated species previously observed in other related nucleosides and nucleotides.[34] [35] [44]

Zoom Image
Scheme 2 Synthesis of modified triphosphate 3. Reagents and conditions: (i) 2-azidoethylamine, HATU, sym-collidine, DMF, RT, 19 h, 64%; (ii) NaIO4, NH4OAc, acetone, RT, 24 h, 98%; (iii) EdUTP, CuBr, TBTA, H2O, DMF, RT, 12 h, 29%.
Zoom Image
Scheme 3 Synthesis of modified triphosphate 4. Reagents and conditions: (i) DMF, RT, 12 h, 71%.

In order to rule out that protodeboronation occurred during the copper-catalyzed click reaction, we also carried out the addition of the arylboronic acid moiety to the nucleobase under milder, copper-free conditions. To do so, we first synthesized a nucleotide functionalized with an azadibenzocyclooctyne (ADIBO)[46] moiety [see the Supporting Information (SI)]. This was achieved by the application of standard amide-bond-formation conditions for the conjugation of an azadibenzocyclooctyne with the amino-modified dUTP nucleotide 11 in good yields (64%). The resulting ADIBO-modified nucleotide (dUCOTP) 12 was then subjected to copper-free click reaction conditions[47] to yield the nucleotide analogue 4 in 71% yield (Scheme [3]).[48]

With the modified nucleotides 1–4 in hand, we attempted to examine whether the presence of the modifications, particularly the boronic acids connected to the C5 position of the pyrimidine nucleobase through rigid triazole linkers, would be tolerated by DNA polymerases. To do so, we turned to primer/template systems that had previously been used to evaluate the substrate acceptance of dN*TPs under primer extension reaction (PEX) conditions.[15] [42] [49] First, we used a 31-mer DNA template T1 and a 15-nucleotide-long primer P1 carrying a fluorescein dye at the 5′-end (see SI)[49] and we tested family A and B DNA polymerases, along with the template-independent polymerase terminal deoxynucleotidyl transferase (TdT).[50]

Full extension of primer P1 was observed when triphosphate dUBOHFTP 1 was used in conjunction with Therminator, Vent (exo –), Deep Vent, or the T4 DNA polymerases, despite the propensity of dUBOHFTP 1 to deboronate (SI; Figure S1A). The remaining polymerases (i.e., Taq, Bst, and Kf exo –) were not efficient at accepting nucleotide 1 as a substrate. On the other hand, triphosphate dUPhFTP 2 acted as a moderate substrate for polymerases, because the formation of expected full-length products came at the expense of stalling of the polymerase leading to truncated species (SI; Figure S1B). As in the case of nucleotide 1, all the polymerases except the T4 DNA polymerase readily accepted triphosphate dUBOHTP 3 as a substrate and gave fully extended primers (SI; Figure S1C). Of all nucleotides, analogue dUCOBOHTP 4 displayed the best substrate capacity, because full-length products could be observed under all conditions, with little or no formation of truncated products (SI; Figure S2). As expected, the incorporation of all modified nucleotide analogues into DNA led to important retardation in the gel mobility and, consequently, shifted product bands. This can be explained by the presence of the rather bulky modifications combined with the differential interaction of the Lewis acid (boronic acid group) and the gel matrix.[35] [51] Interestingly, these differences in gel shifts were by far the largest when nucleotide 4 was supplemented to the reaction mixtures.

A similar trend in substrate acceptance was observed when primer extension reactions were performed with the TdT and 19-nucleotide (nt)-long, 5′-FAM-labeled primer P2. Indeed, longer reaction times, combined with Mn2+ or Co2+ as a cofactor and nucleotide 4, led to the formation of larger molecular-weight product distributions, which are typically observed with canonical nucleotides (SI; Figure S3). On the other hand, tailing reactions with nucleotides 1–3 produced only primers extended with one or two modified nucleotides (SI; Figure S4).

Encouraged by these initial results, we examined whether longer boronic acid-modified DNA sequences could be obtained, because this represents an important prerequisite for their potential use in selection experiments. Consequently, we used a 19-mer 5′-FAM-labeled primer P3 along with a 79-nucleotide-long template T2 to examine the substrate acceptance of nucleotide 4. In addition, we focused exclusively on the boronic acid containing nucleotide dUCOBOHTP 4 to synthesize longer DNA sequences, given the superior substrate acceptance of this analogue. Analysis of PEX reactions confirmed this, because the modified analogue was readily accepted as a substrate by all polymerases that were evaluated and the primers were fully converted into the expected products (Figure [2]). As noted above for PEX reactions with shorter templates and TdT-mediated tailing reactions, the introduction of thirteen boronic acid residues induced an important shift in electrophoretic mobility.

Zoom Image
Figure 2 Gel image (PAGE 20%) analysis of primer extension reactions with primer P3 and template T2 with dUCOBOHTP 4 (200 µM). Polymerases used: 1: Phusion; 2: HemoKlem Taq; 3: Q5 DNA polymerase; 4: Bst; 5: Taq; 6: Therminator; 7: Vent (exo –); 8: Dpo4; 9: Deep vent; 10: Klenow exo –. Control reactions: N+: all four natural dNTPs; N2 –: dATP, dCTP, dGTP; N1 ––: dATP, dGTP. Lane S represents a 5′-FAM-labeled, 71-mer synthetic oligonucleotide reference, lane 2 represents the PEX reaction product obtained with nucleotide 2, 71 corresponds to expected, 71-nt-long full-length product, and 18 corresponds to unreacted, 18-nt-long primer.

Given the excellent substrate capacity of nucleotide 4, we next evaluated the possibility of producing boronic-acid-modified DNA by using the polymerase chain reaction (PCR). To do so, we performed PCR with the 79-mer template T3 and primers P4 and P5 by using five different polymerases to see whether amplicons could be produced with nucleotide 4 (Figure [3]). This analysis revealed that Vent (exo –) and HemoKlem Taq accepted modified nucleotide 4 and produced full-length amplicons, whereas all the other polymerases failed to yield any significant amount of the expected product. Also, the boronic-acid-modified DNA showed an important shift in electrophoretic mobility, presumably due to a substantial interaction between the boronic acid moieties and the cis-diols of the galactose sugar of agarose.

Zoom Image
Figure 3 Agarose gel (4%) analysis of PCR products obtained with template T3, primers P4 and P5, and a mixture of dATP, dCTP, dGTP, and nucleotide 4 (all 200 µM). Polymerases used: 1: Vent (exo –); 2: Taq; 3: HemoKlem Taq; 4: Q5 DNA polymerase; 5: Phusion. Control reactions: +: PCR with all four natural nucleotides and –: PCR without any polymerase and with all four natural dNTPs. Full experimental details can be found in the SI.

To confirm the identity of the product stemming from the PEX reactions with nucleotide analogue 4, we carried out a digestion–LC/MS analysis protocol.[52] To do so, we first converted triphosphate 4 into the corresponding nucleoside with Shrimp Alkaline Phosphatase (rSAP). LC/MS analysis clearly revealed the presence of the expected nucleoside analogue equipped with a boronic acid moiety, further confirming the chemical nature of dUCOBOHTP 4. Next, we performed a PEX reaction with the 20-nucleotide-long template T4 and the 5′-FAM-labeled, 19-mer primer P6 with Vent (exo –) and dUCOBOHTP 4. The resulting modified DNA sequence containing a single dUCOBOHMP residue was then digested with nuclease and dephosphorylated, and the resulting deoxynucleosides were then analyzed by LC/MS (Figure [4]). This analysis clearly demonstrated the successful incorporation of the modified nucleotide without induction of the loss of the boronic acid moiety.

Zoom Image
Figure 4 Analysis of digested, modified dsDNA. (A) LC-MS chromatogram showing the nucleoside digestion profile of double-stranded DNA after PEX reaction with template T4 and primer P6. The chromatogram reveals five distinct peaks, four of which correspond to the canonical nucleosides of the template and primer: deoxycytidine (dC), deoxyadenine (dA), deoxyguanine (dG), and deoxythymidine (dT). The additional peak represents the modified deoxyuridine (dUCOBOH), indicating the successful insertion of dUCOBOHTP during PEX reactions and its subsequent digestion into dUCOBOH. (B) LC-MS chromatogram serves as the (dUCOBOH ) standard. (C) MS spectrum of dUCOBOH after digestion.

In conclusion, decorating nucleic acids with boronic acids is an alluring aim, as these functional groups can be harnessed for post-synthetic and orthogonal modification strategies, including conversion of oligonucleotides into medical-imaging devices. These functional groups might also permit aptamers to interact with highly glycosylated targets. However, due to their inherent reactivity, synthetic methods for the introduction of boronic acids, especially in longer oligonucleotides, are scarce. Here, we explored the effect of substituents of aryl boronic acid residues on the stability of these synthons in copper-catalyzed click reactions and we investigated the efficiency of the coupling strategy. Whereas the presence of ortho-fluoro substituents on phenylboronic acids might be useful for numerous applications, such synthons do not seem to be suitable for CuAAC reactions on nucleotides. In addition, due to the milder conditions and the absence of copper, strain-promoted azide–alkyne cycloaddition (SPAAC) appears as the method of choice for the crafting of nucleotides decorated with boronic acids. The resulting nucleotide dUCOBOHTP 4 proved to be an excellent substrate for a large number of DNA polymerases under primer extension reactions and could also be used under PCR conditions. The combination of this modified nucleotide with polymerase-mediated synthesis allowed the generation of DNA sequences containing up to thirteen boronic acid residues. Such an approach will be useful for the preparation of nucleotides modified with boronic acids in order to raise aptamers with enhanced properties and a broader target scope.


#

Conflict of Interest

The authors declare no conflict of interest.

Supporting Information

    Supporting information for this article is available online at https://doi-org.accesdistant.sorbonne-universite.fr/10.1055/a-2212-7704.
  • Supporting Information

Corresponding Author

Marcel Hollenstein
Institut Pasteur, Université Paris Cité, Department of Structural Biology and Chemistry, Laboratory for Bioorganic Chemistry of Nucleic Acids, CNRS UMR3523
28, rue du Docteur Roux, 75724 Paris Cedex 15
France   

Publication History

Received: 27 July 2023

Accepted after revision: 16 November 2023

Accepted Manuscript online:
16 November 2023

Article published online:
14 December 2023

© 2023. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


   Permissions and Reprints All articles of this category
Zoom Image
Figure 1 Chemical structures of dUBOHFTP (1), dUPhFTP (2), dUBOHTP (3), and dUCOBOHTP (4)
Zoom Image
Scheme 1 Synthesis of modified triphosphates 1 and 2. Reagents and conditions: (i) THF, NaOH, RT, 4h; (ii) (a) (COCl)2, DMF (cat.), CH2Cl2, RT, 1 h; (b) 2-azidoethylamine, DMAP, NEt3, CH2Cl2, THF, RT, 12 h, 31% over 2 steps; (iii) (a) EdUTP, CuBr, tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine (TBTA), H2O, DMF, RT, 4 h, (b) NH4OH, RT, 2 h; 1 (18%) and 2 (29%).
Zoom Image
Scheme 2 Synthesis of modified triphosphate 3. Reagents and conditions: (i) 2-azidoethylamine, HATU, sym-collidine, DMF, RT, 19 h, 64%; (ii) NaIO4, NH4OAc, acetone, RT, 24 h, 98%; (iii) EdUTP, CuBr, TBTA, H2O, DMF, RT, 12 h, 29%.
Zoom Image
Scheme 3 Synthesis of modified triphosphate 4. Reagents and conditions: (i) DMF, RT, 12 h, 71%.
Zoom Image
Figure 2 Gel image (PAGE 20%) analysis of primer extension reactions with primer P3 and template T2 with dUCOBOHTP 4 (200 µM). Polymerases used: 1: Phusion; 2: HemoKlem Taq; 3: Q5 DNA polymerase; 4: Bst; 5: Taq; 6: Therminator; 7: Vent (exo –); 8: Dpo4; 9: Deep vent; 10: Klenow exo –. Control reactions: N+: all four natural dNTPs; N2 –: dATP, dCTP, dGTP; N1 ––: dATP, dGTP. Lane S represents a 5′-FAM-labeled, 71-mer synthetic oligonucleotide reference, lane 2 represents the PEX reaction product obtained with nucleotide 2, 71 corresponds to expected, 71-nt-long full-length product, and 18 corresponds to unreacted, 18-nt-long primer.
Zoom Image
Figure 3 Agarose gel (4%) analysis of PCR products obtained with template T3, primers P4 and P5, and a mixture of dATP, dCTP, dGTP, and nucleotide 4 (all 200 µM). Polymerases used: 1: Vent (exo –); 2: Taq; 3: HemoKlem Taq; 4: Q5 DNA polymerase; 5: Phusion. Control reactions: +: PCR with all four natural nucleotides and –: PCR without any polymerase and with all four natural dNTPs. Full experimental details can be found in the SI.
Zoom Image
Figure 4 Analysis of digested, modified dsDNA. (A) LC-MS chromatogram showing the nucleoside digestion profile of double-stranded DNA after PEX reaction with template T4 and primer P6. The chromatogram reveals five distinct peaks, four of which correspond to the canonical nucleosides of the template and primer: deoxycytidine (dC), deoxyadenine (dA), deoxyguanine (dG), and deoxythymidine (dT). The additional peak represents the modified deoxyuridine (dUCOBOH), indicating the successful insertion of dUCOBOHTP during PEX reactions and its subsequent digestion into dUCOBOH. (B) LC-MS chromatogram serves as the (dUCOBOH ) standard. (C) MS spectrum of dUCOBOH after digestion.