Sequence-defined vinyl sulfonamide click nucleic acids (VS-CNAs) and their assembly into dynamically responsive materials

Abstract

Synthetic DNA analogues are of great interest for their application in information storage, therapeutics, and nanostructured materials, yet are often limited in scalability. Vinyl sulfonamide click nucleic acids (VS-CNAs) have been developed to overcome this limitation using the highly efficient thiol-Michael ‘click’ reaction. Utilizing all four nucleobases, sequence-defined click nucleic acids (CNAs) were synthesized using a simple and scalabale solution-phase approach. Employing a polyethylene glycol (PEG) support, synthesis of the CNA sequence, GATTACA, was achieved in high yields. CNA crosslinked hydrogels were assembled using multiarm PEG-CNAs resulting in materials that dynamically respond to temperature, strain, and competitive sequences.

DNA possesses a unique sequence-dependent assembly property that enables sophisticated material construction with unique function. The complementary binding sequences of DNA are the basis for several materials engineering platforms, such as aptamer responsive hydrogels,¹ colloidal crystals,² and DNA origami.³ The assembly properties of DNA are driven by hydrophobic interactions between nucleobases, which promote selective hydrogen bonding between base pairs.⁴ Specifically, thymine binds with adenine, and cytosine binds with guanine. Much like a computer stores information through binary code, DNA utilizes the selective hydrogen bonding to store genetic information along its double helical backbone. Despite the unique advantages of DNA as a material, it has limited engineering applications owing in part to its enzymatic degradability⁵ and high cost of production. To circumvent these challenges, nucleobase containing polymers^6,7 and synthetic DNA analogues⁸ have been developed in order to capture the unique assembly properties of natural DNA. Yet current designs also limit their applications. Although nucleobase containing polymers can be synthesized on the gram scale, they typically lack sequence control limiting their degree of assembly.⁹ Alternatively, DNA analogues that possess sequence control have been designed using solid supports¹⁰ or single monomer insertion strategies¹¹ but are limited in scale by the solid support loading capacity or the necessity for column chromatography for purification, respectively.

Recently, click nucleic acids (CNAs) have been developed with the aim of providing a robust strategy for sequence control of nucleobase containing polymers,¹² which are synthesized utilizing thiol-X click reactions. In general, click reactions possess a high thermodynamic driving force and reaction selectivity, enabling high conversion with minimal byproducts.¹³ In 2015, the Bowman group demonstrated a simple strategy to synthesize dimer thymine and adenine macromers, which were subsequentially homopolymerized via thiol–ene click chemistry to produce periodic nucleobase polymers.^12,14 Various analogues of CNAs were shown to mediate the assembly of colloidal nanoparticles,^14,15 drug delivery vehicles,^16,17 and organogels.¹⁸ Other derivatives of CNAs were even capable of binding to natural DNA by modulating the chemical spacing between the nucleobases.¹⁵

Despite the early success of CNA materials, they have been limited by inefficient coupling reactions between monomers. The conjugation reactions employed to create dimer and trimer nucleobase macromer are thiol-Michael¹² or thiol-halogen¹⁴ addition reactions, which tend to a have high thermodynamic driving force; however, the acrylamide or bromide thiol-acceptor in these cases exhibit slow reaction kinetics.^19,20 Additionally, only two of the four nucleobases (i.e., thymine and adenine) have been incorporated into the CNA synthetic scheme,¹² owing in part to synthetic challenges associated with nucleobase modifications.

While the promise of CNAs as DNA analogues has been established in seminal works,¹² both fundamental exploration of CNA assemblies and their applications in bulk materials or biosensing platforms have been limited by the lack of synthetic approaches for integrating all four nucleobases into sequence-defined materials with appropriate scalability. Herein, we have developed a facile strategy for synthesizing all four nucleobase CNA monomers, which additionally possess vinyl sulfonamide functionality for rapid, high yielding, and low byproduct thiol-Michael conjugations.

The general synthetic scheme presented in the literature for synthesizing thiol-Michael-type CNA monomers is to alkylate the nucleobase utilizing dibromoethane, resulting in a halogen handle for subsequent substitution reactions with a primary amine.¹² Pyrimidine nucleobases, thymine and cytosine, have limited reactivity towards these dihalogenethanes, resulting in low yields, and have been shown to undergo multiple side reactions and oligomerization states (Fig. S1, ESI†).²¹ Thus, only an adenine CNA monomer has been used with a thiol-Michael acceptor. To circumvent this synthetic limitation, aldehyde functional handles were incorporated onto the nucleobases.²² Aldehydes were generated by functionalizing the nucleobase with acetals, which were subsequently converted to aldehydes in the presence of acid.²³ The structures of the four aldehyde functional nucleobases are shown in Fig. 1A. Aldehydes can undergo reductive amination with the primary amine of a trityl-protected cysteamine to generate a secondary amine functional handle. The secondary amine can then be functionalized with a sulfonyl chloride to produce a Michael acceptor for thiol-Michael reactions (Fig. 1A). Notably, the synthetic procedures utilized for synthesizing all four aldehyde functional nucleobases require no chromatographic purification steps and are readily achieved on a multigram scale. Full synthetic procedures for the synthesis of these nucleobases are found in the ESI.†

Fig. 1.

Synthesis of vinyl sulfonamide functional click nucleic acids (VS-CNAs). (A) The synthesized aldehyde functional nucleobases were reacted with trityl-protected cysteamine via reductive amination forming a secondary amine. Reaction of this amine with 2-chloroethanesulfonyl chloride generated the vinyl sulfonamide functional group. (B) The vinyl sulfonamide functional CNA monomer was coupled with an allyl CNA monomer to form a dimer that, upon deprotection of the thiol, can either undergo additional CNA monomer additions via a thiol-Michael reaction or a radical initiated homopolymerization via a thiol–ene reaction.

To demonstrate the effectiveness of using vinyl sulfonamides as a Michael acceptor for conjugating CNA monomers, a simple reaction was performed between a thymine monomer possessing an allyl functional group¹² and one possessing the vinyl sulfonamide Michael acceptor (see Fig. 1B). In the presence of a catalytic amount of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), the reaction proceeded to full conversion within one hour at room temperature (Section 4 of ESI†). The resulting product was purified via flash chromatography allowing the excess unreacted thiol-Michael monomer to be recollected and reused in subsequent reactions, demonstrating the selectivity and sustainability of the process.

Thiol-Michael reactions present a rapid and highly efficient coupling strategy for building sequence-defined CNA oligomers. Typically, synthesis of sequence-controlled macromolecules is achieved using solid-supports for their ease of use and purification; however, this approach also has several drawbacks, including expensive supports, heterogenous reaction conditions, and the limited scale at which the reactions can be performed.²⁴ An alternative strategy is to use a soluble polymer support to preserve solution-phase kinetics.²⁵ Additionally, polymer supports can facilitate purification through common precipitation methods. Here, we used polyethylene glycol (PEG) as solution-phase support for its ability to be precipitated in diethyl ether and ethanol for purification of the sequence-defined CNA and subsequent removal of unreacted monomer and catalyst.

PEG was successfully applied as a polymer support for synthesizing sequence-defined CNAs by using sequential thiol-Michael coupling and thiol deprotection steps as depicted in Fig. 2A. The thiol functional PEG(5k) and excess thiol-Michael CNA monomer were coupled in the presence of DBU for 1 h, and the polymer product was precipitated in diethyl ether. The polymer and excess CNA monomer were collected through centrifugation and multiple diethyl ether washes. Excess CNA monomers were removed by dissolving the polymer in hot ethanol and precipitating the pure polymer via cooling in an ice bath. Excess monomer dissolved in the ethanol could be recollected and reused. The same precipitation purification procedure was adopted for the deprotection step that uses trifluoroacetic acid (TFA) to remove the trityl protected chain end and regenerate a thiol for subsequent thiol-Michael reactions.

Fig. 2.

Synthesis and characterization of sequence-defined VS-CNAs. (A) The iterative coupling and deprotection strategy used to form sequence-controlled CNAs tethered to a thiol PEG(5k) support. (B) MALDI-TOF mass spectroscopy revealed an increase in the number average molar mass (M_n) congruent to the mass of each CNA monomer addition.

The addition of the CNA monomers to the polymer support was confirmed via matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectroscopy and nuclear magnetic resonance (NMR) spectroscopy. The PEG-CNA synthesis evolution is shown via MALDI in Fig. 2B where the increases in number average molar mass (M_n) of the growing PEG-CNA polymer is in excellent agreement with the mass associated with the monomer addition (Table S2, ESI†). It is important to note that the dispersity observed is a result of the polymer support and not VS-CNAs. NMR further confirmed the successful addition of CNA monomers to the polymer support with the appearance of characteristic aromatic peaks associated with the newly coupled nucleobase monomers. Integration of the aromatic peaks with respect to the PEG peak suggested near quantitative conversions were achieved for both the coupling and deprotection steps. (Fig. S4-S9, ESI†) Post precipitations, the polymer was recovered in high yields (86% to 97%) because of the high coupling and deprotection efficiencies and the ease of the purification method (Fig. S4, ESI†). Post-purification yields were observed to be slightly lower when coupling thymine and guanine to the PEG as additional purification steps were sometimes necessary (Section 5 of ESI†). Although only a seven-mer VS-CNA was synthesized here, the robust strategy can be readily adapted for a variety of linear sequence compositions; however, owing to decreasing solubility with increasing CNA length, one would anticipate an eventual drop in the CNA monomer addition yield.

Using the iterative coupling and deprotection strategy described above, tetrafunctional PEG(20k) (tPEG) CNA macromers were synthesized with sequences of TTT, AAA, or ACA (Section 6 of ESI†). Sequences of three nucleobases were utilized with multiarm PEG as difficulties in solubilizing in hot ethanol for purification became evident with increasing CNA length. Two CNA-based hydrogels were formed with tPEG-TTT: one with tPEG-AAA and one with tPEG-ACA. Gels were formed at 10 wt% in a 1 : 1 DMSO: water solution at 40 °C for 20 h. The gels were then transferred to PBS and allowed to exchange overnight. The swollen hydrogels were subsequently used to examine the dynamic nature of the material under strain and temperature cycling (Fig. 3A) as well as degradation behavior via strand displacement (Fig. 3B). We speculate the driving force for hydrogel formation is a result of hydrophobic aggregation of the nucleobases, similar to that observed in DNA assembly.⁴

Fig. 3.

Thiolated 4-arm PEG(20k) functionalized with CNA sequences formed dynamic and responsive hydrogels. Illustrations of CNA hydrogels undergoing degelation (A) upon application of temperature (Δ) or strain (ε) or (B) by the introduction of a linear competitive CNA sequence via strand displacement (S.D.). (C) Temperature cycling (T = 25–75 °C) and (D) oscillatory strain (strain = 1– 50%) rheometry were conducted on swollen CNA hydrogels formed between tPEG-AAA and tPEG-TTT. (E) Swelling experiments (i.e., hydrogel volume divided by initial volume) were used to monitor the degradation of CNA hydrogels via S.D. in the presence of 1 mM linear PEG-TGT. Roman numerals represent statistical differences at 12 h where (I) is p < 0.0001 and represents statistical significance of the 1 mM TGT samples compared to both 0 mM controls and (II) is p < 0.005 where the statistical difference is between the two 1 mM TGT subjected samples. Error bars represent standard error (n = 3). For full statistical analysis see Table S1 (ESI†).

Thermal cycling of CNA hydrogels formed between tPEG-TTT and tPEG-AAA revealed recovery of the mechanical properties. The hydrogel temperature responsiveness was investigated by cycling the temperature between 25 °C and 75 °C. Upon reaching 40 °C, a crossover of the storage (G′) and loss (G″) moduli was observed, which is associated with a solid- to liquid-like transition via the disruption of CNA crosslinks. The thermoreversiblity of the network crosslinks was evidenced by an increase in G′ with decreasing temperature; however, full property recovery was not observed during the time allotted. A drop in the initial G′ was observed during the first temperature ramp and recovered slowly to 65% of its original value. Subsequent thermal cycling yielded similar moduli traces after the first ramp (Fig. 3C). As the tPEG-CNA hydrogels were initially formed in 1 : 1 DMSO to H₂O, followed by solvent exchange against aqueous PBS, we speculate that the decrease in the ultimate G′ attained following the initial temperature ramp is due to hydrophobic nature of the free TTT and AAA groups in a purely aqueous system. With the removal of DMSO, the CNA crosslinks become kinetically frozen, which rearrange into a lower energy state upon thermal cycling. This explanation is supported by O’Reilly and coworkers,⁷ in which nucleobase containing block copolymers were kinetically frozen via solvent exchange producing bicontinuous micelles that formed multilamellar vesicles after multiple annealing cycles.

CNA hydrogels with sequences tPEG-TTT and tPEG-AAA also displayed dynamic responsiveness of the mechanical properties under strain cycling (Fig. 3D). Upon application of strain, an immediate decrease in the G′ and an increase in the G″ occurred, where the crossover at 50% strain suggested hydrogel network disruption. Once the strain was lowered back to 1%, the hydrogel network showed excellent recoverability with minimal loss to network properties over several cycles. Upon strain release, the CNA crosslinks can reform into hydrophobic domains recovering material integrity.

Finally, we hypothesized that hydrogels with CNA crosslinks could be disrupted in the presence of a linear PEG-CNA sequence, analogous to strand displacement exhibited in natural DNA.²⁶ The degradation behaviour of hydrogels formed between tPEG-TTT and tPEG-AAA and hydrogels formed between tPEG-TTT and tPEG-ACA were compared in the absence and presence of 1 mM linear PEG-TGT stimuli (Fig. 3E). The hydrogel degradation was monitored using swollen gel volume, which is directly related to the network crosslink density through the Flory–Rehner theory.²⁷ As shown in Fig. 3E, the CNA hydrogels were stable in PBS in the absence of linear PEG-CNAs. In the presence of linear PEG-TGT, faster degradation of the tPEG-ACA hydrogels was observed as compared with the tPEG-AAA hydrogels (Fig. 3E). This result indicates that while complementary base pair interactions may accelerate degradation, the hydrophobic interactions are primarily responsible for the disruption of the physical crosslinks in these hydrogels. Nonetheless, the ability to tune degradation of the hydrogel based on crosslinker and displacement strand sequence composition is a unique property of DNA analogue based hydrogels.

A facile strategy was developed for synthesizing CNA monomers with vinyl sulfonamide functional handles and were incorporated into sequence-defined polymeric materials. Utilizing simple nonchromatographic purification strategies, all four nucleobases were functionalized with aldehyde groups for subsequent backbone addition. For the first time, a sequence specific oligomer was synthesized using all four CNA monomers, which used simple precipitation purification facilitated by a PEG support. The developed sequence-building strategy was applied to multiarm PEGs to form adaptive CNA crosslinked hydrogels. Utilizing the physical crosslinks between nucleobases, dynamic responsiveness was evaluated showing recovery of material properties under both temperature and strain. Finally, CNA hydrogels were subjected to degradation through strand displacement in the presence of linear PEG-CNA, revealing that sequence influenced degradation rate and thus providing a unique tunable handle for modulating material properties. With the development of these robust and facile CNA synthetic protocols, explorations into the assembly properties of CNAs with complementary strands and biological substrates can now be realized broadening the applications of CNA-based materials towards the development of scalable biosensors, programmable materials, and drug delivery vehicles. Additionally, the strategic use of polymer supports for building sequence-defined materials provides opportunities beyond the field of DNA analogues to create materials with unique function on a scale not previously attainable. Such a strategy enables the design of novel and inexpensive materials broadly impacting the fields of chemistry, biology, polymers, and biomaterials.

BPS would like to acknowledge support by the National Institutes of Health under Award Number P20GM104316 as well as P30GM110758-02 for additional core instrument support. AMK would like to acknowledge the support from the NIH Director’s New Innovator Award under Award Number DP2HL152424. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. PJL would like to thank the University of Delaware Dissertation Fellowship for financial support.