Toward highly effective loading of DNA in hydrogels for high-density and long-term information storage

Abstract

Digital information, when converted into a DNA sequence, provides dense, stable, energy-efficient, and sustainable data storage. The most stable method for encapsulating DNA has been in an inorganic matrix of silica, iron oxide, or both, but are limited by low DNA uptake and complex recovery techniques. This study investigated a rationally designed thermally responsive functionally graded (TRFG) hydrogel as a simple and cost-effective method for storing DNA. The TRFG hydrogel shows high DNA uptake, long-term protection, and reusability due to nondestructive DNA extraction. The high loading capacity was achieved by directly absorbing DNA from the solution, which is then retained because of its interaction with a hyperbranched cationic polymer loaded into a negatively charged hydrogel matrix used as a support and because of its thermoresponsive nature, which allows DNA concentration within the hydrogel through multiple swelling/deswelling cycles. We were able to achieve a high DNA data density of 7.0 × 10⁹ gigabytes per gram using a hydrogel-based system.

INTRODUCTION

In today’s digital information age, the global demand for data storage is growing at an exponential rate (1–3). However, further reduction of traditional storage architectures, such as hard drives and magnetic tapes, is a growing challenge (4, 5). Current devices have reached their physical limits and can hardly maintain pace with the ever-increasing digital storage requirements. These issues might be resolved by DNA data storage, which is an emerging technology that provides substantial DNA density (5–8) with a remarkable half-life (9, 10).

DNA storage is based on encoding data into DNA molecules expressed as a four-letter code, i.e., A, T, C, and G. On the basis of prior calculations, DNA can theoretically store up to 455 EB/g (4.55 × 10¹¹ GB/g), which are 10 billion times higher than the traditional storage media (100 GB/g) (11, 12). However, there are several hurdles that need to be overcome before approaching this theoretical storage density. At present, DNA powder is commonly used for DNA data storage, but this approach has the following limitations: First, the DNA powder needs to be associated with other DNA protective carriers, such as silica beads, to avoid contamination and loss. However, the carrier tends to have a much higher weight than the DNA itself, which reduces the effective DNA storage capacity (DNA weight/total weight). Second, this method is prone to DNA damage and contamination. Third, DNA may require coexistence with numerous DNA stabilizers, such as trehalose and polyvinyl alcohol, to enhance its storage life, which also reduces the effective DNA storage capacity (13–16). Furthermore, DNA powder may not mix well with the DNA stabilizer. Therefore, it is important to explore other methods that can improve the DNA loading efficiency during the preservation process.

DNA storage methods can be classified mainly into adsorption and encapsulation. Storing DNA by encapsulation produces relatively high DNA density. However, the challenge lies in the full utilization of encapsulated DNA owing to inefficient DNA recovery, which uses unfavorable conditions for DNA release, such as extremely high pH solvent use (17–20). Although adsorption of DNA to a surface does not generally encounter this challenge, the DNA loading density is often insufficient. Moreover, the support can easily adsorb impurities and negatively affect DNA adsorption. So far, the largest amount of data storage reported in a DNA carrier exceeded 10⁹ GB/g, and this was stored in the form of millions of oligonucleotides. Its high data density was achieved via layer-by-layer encapsulation of DNA in magnetic nanoparticles [multilayer: one layer of polyethyleneimine (PEI) and one layer of DNA]. However, the assembled multilayer structure was easily destroyed during the DNA release process (18). Therefore, it is not suitable for repeated storage and retrieval of information encoded in DNA. Hence, it is desired to develop an effective, simple, and convenient DNA storage method that achieves high data density while allowing its repeated use.

Hydrogels are three-dimensional cross-linked hydrophilic polymeric networks with the ability to hold large amounts of water, formed by physical and/or chemical cross-linking (21–24). Hydrogels used for controlled release of nucleic acids, entrapped either as conjugates or as polyplex (polymer–nucleic acid complex) particles, have been studied in several studies but are limited by low loading density because of their aggregation and instability during encapsulation by gelation (25–29). Although the use of hydrogel particles has been explored for storage of biomarkers from biological samples (30, 31), thus far, to the best of our knowledge, there are no existing studies on using hydrogel as an encapsulant for DNA data storage. DNA hydrogels seem to be a promising solution for data storage resulting from their precise molecular recognition capability, biocompatibility, and predictable stimuli response (32–34). However, compared to chemically cross-linked polymeric hydrogels, DNA hydrogels are mainly stabilized by noncovalent interactions such as hydrogen bonding, electrostatic interactions, metal-ligand coordination bonds, and host-guest complexation, and thus exhibit poor mechanical strength and stability (32). Hence, in this study, we will be focusing on chemically cross-linked hydrogels, more specifically on stimuli-responsive hydrogels.

Stimuli-responsive hydrogels are distinguished by their ability to change their physicochemical properties in response to changes in external stimuli such as temperature, pH, ionic strength, light, and electric field (35–37). Among them, poly(N-isopropylacrylamide) (PNIPAM)–based hydrogels have been extensively studied for their thermoresponsive behavior. PNIPAM hydrogels are well known for their lower critical solution temperature (LCST) of 32°C, below which they remain fully swollen. On heating above this temperature, the change in hydrophilicity of the polymeric networks results in the release of absorbed water. Because the size and porosity of PNIPAM hydrogels can be easily controlled by temperature, they are commonly used in controlled drug delivery applications. Moreover, it can be easily chemically modified through copolymerization as well as by forming a semi-interpenetrating network (semi-IPN) and an interpenetrating network (IPN) to improve functionality and to provide versatile properties (38, 39). Besides, they are considered good DNA carriers because of their excellent biocompatibility (40–45) and blend easily with DNA protectants (25, 46).

In this study, a thermally responsive functionally graded (TRFG) hydrogel has been explored for concentration and storage of DNA to obtain high DNA data density. The TRFG hydrogel used in this study was fabricated using a PNIPAM/poly sodium acrylate (PSA) semi-IPN hydrogel, formed by dispersing PSA within cross-linked PNIPAM matrix during gelation. A substantial advantage of using a semi-IPN architecture comes from its ability to rapidly absorb and release a large amount of water on heating above the LCST of PNIPAM, as compared to pure PNIPAM hydrogels, and can easily disperse negatively charged PSA in situ during synthesis (37). The TRFG hydrogel was then fabricated by swelling the negatively charged PNIPAM/PSA semi-IPN hydrogel in a protonated hyperbranched polyethyleneimine (HPEI) solution of different molecular weights and concentrations. PSA dispersed in the PNIPAM network further enables high loading of cationic HPEI through electrostatic interaction. In the past, nucleic acid was supplied in a carrier or after complexation with a cationic polymer to prevent its breakdown and enhance cellular uptake (47). The high charge density of HPEI further ensures high DNA loading capacity of the TRFG hydrogel and provides binding sites for negatively charged DNA through complexation with PEI. The excess amine functional groups of HPEI on the hydrogel surface were cross-linked by interfacial polymerization to form a thin polyamide (PA) layer. Hence, by subsequently swelling the TRFG hydrogel in the DNA solution at 20°C and deswelling at 60°C to release the absorbed solution, DNA is absorbed and retained in the TRFG hydrogel through interaction with PEI. Moreover, the cross-linked PA layer further enables retention of DNA within the hydrogel, hence providing high loading capacity and will be discussed further in this work. Figure 1 illustrates the overall process used for converting digital information into DNA sequences; synthesis, storage, and recovery of DNA sequences; and information retrieval at the end. DNA molecules are adsorbed and concentrated into and on the surface of the TRFG hydrogel using consecutive swelling/deswelling cycles until there is no change in the concentration of the DNA in the solution used. DNA from the TRFG hydrogel is then recovered using a buffer solution with high pH, and using sequencing of polymerase chain reaction (PCR)/digital polymerase chain reaction (dPCR), the recovered DNA molecules are then decoded. The thermoresponsivity of the formed TRFG hydrogels allows consecutive swelling/deswelling in the solution through multiple cycles until it gets saturated, and it was shown to reach a data density of 10⁹ GB/g, which is close to the highest data density reported in literature (18). Moreover, the stored DNA can be recovered using a nondestructive method simply by deprotonation of HPEI using high pH elution buffer. The effect of using different HPEI molecular weights and concentrations for TRFG hydrogel fabrication on DNA loading capacity will be discussed in detail in the following sections.

Fig. 1. A flowchart depicting the overall process of storing DNA in TRFG hydrogels.
The main steps include conversion of the digital information into DNA code, synthesis of DNA, adsorption and concentration of DNA in the TRFG hydrogel under several absorption cycles, recovery and release of DNA from the TRFG hydrogel, and PCR or dPCR and sequencing of the recovered DNA. The details of primes used in PCR are shown in table S1, and the details of data encoded by DNA are shown in table S2. The magnified image of the TRFG hydrogel at the bottom highlights the dual-responsive (i.e., both pH- and temperature-responsive) behavior of the TRFG hydrogel used in this study.

RESULTS

DNA encoding and decoding

As mentioned earlier, to fully use the large data storage ability of DNA storage technology, it is crucial to realize an efficient conversion between DNA code and digital information, as well as improve the storage density of the DNA container. To this end, we developed a Python program that efficiently converted digital information into DNA code and vice versa. The specific programming code and conversion process are provided in fig. S1A (encoder) and fig. S1B (decoder). The coding process takes 0.21 s, and the decoding process takes 0.26 s. In the coding step, the encoder coded the required information as a DNA sequence. In the decoding stage, the information from the sequencing samples was converted back into digital data via a decoder. The detailed encode and decode information is available in table S2.

Preparation and characterization of TRFG hydrogels

TRFG hydrogels were synthesized to enable high loading of DNA and controlled release using both temperature and pH. Figure 2 shows the schematic for fabrication of the TRFG hydrogel used in this study. First, the PNIPAM/PSA semi-IPN hydrogel is synthesized using free-radical polymerization of NIPAM in the presence of PSA at 10°C for 24 hours. The PNIPAM/PSA semi-IPN hydrogel can be obtained in different shapes and forms depending on the mold used. Synthetic DNA sequences contain phosphate groups and are typically negatively charged when placed in water. Therefore, to increase DNA loading density, hydrogels should be sufficiently positively charged. Hence, in this study, we used PNIPAM/PSA semi-IPN hydrogel as the support for two main reasons. First, the water recovery is higher using a semi-IPN architecture, and hence, the swollen hydrogel beads can be easily dewatered. Second, the negatively charged PSA can act as anchor points to interact with branched cationic polymer in the protonated state to ensure its high loading into the semi-IPN hydrogel and ensure its retention during deswelling.

Fig. 2. Fabrication and characterization of TRFG hydrogels.
(A) Schematic illustration of the TRFG hydrogel fabrication process and the swelling/deswelling principle of TRFG hydrogels. First, the PNIPAM/PSA semi-IPN hydrogel is synthesized, which is later successively immersed in aqueous PEI solution and hexane containing TMC for the formation of a cross-linked PA layer by interfacial polymerization. (B) 1 and 2: TRFG hydrogels (HPEI-2k, 2 wt %) before and after swelling, as seen by the naked eye. (C) Optical image of TRFG hydrogels obtained using wide-field fluorescence microscopy. (D) SEM image of the TRFG hydrogel. (E) FESEM images of the cross section of the HPEI-2k TRFG hydrogel. (F) DSC thermogram showing VPTT of the pure PNIPAM/PSA semi-IPN hydrogel and that of HPEI-2k and HPEI-25k TRFG hydrogels formed using different molecular weights of HPEI. (G) The equilibrium swelling ratio of TRFG hydrogels measured 2 days after immersion in excess of DI water at 20°C and 1 hour at 60°C. (H) Deswelling kinetics of TRFG hydrogels at 60°C.

In this study, HPEI is used as a cationic polyelectrolyte, with augmented positive charge density and abundant reactive functional groups (primary amine groups), because of its ability to condense DNA (27, 29, 38). Once the PNIPAM/PSA semi-IPN hydrogels are obtained, the fully swollen hydrogel beads are then immersed in an aqueous solution containing HPEI with different molecular weights, i.e., 2 and 25 kDa at 2 wt % for 30 min. Note that the pH of the solution was maintained at 5 to ensure that HPEI is in a protonated state. Subsequently, the hydrogel beads were then suspended in hexane containing 0.1 wt % trimesoyl chloride (TMC) for 30 min to allow the formation of cross-linked PA layer at the surface of the hydrogel via interfacial polymerization of amine groups of HPEI with TMC. This is done to ensure the retention of HPEI/DNA within the hydrogel during deswelling. The high charge density of HPEI provides enough reaction sites for the formation of a PA layer while interacting with the PSA dispersed within the hydrogel. The noncovalent interaction of branched PEI with PSA is known to enhance the PA layer stability on the hydrogel surface (38). Both the PA layer on the hydrogel surface and the HPEI loading in the hydrogel will critically affect the DNA loading density and hence will be discussed further. From here on, the TRFG hydrogel synthesized using HPEI with a molecular weight of 2 and 25 kDa will be referred to as HPEI-2k and HPEI-25k, respectively.

Figure 2B shows the dried and swollen HPEI-2k TRFG hydrogel bead synthesized in this study. As shown in Fig. 2C, TRFG hydrogels have a transparent white color. Upon 100× magnification and illumination with natural light, under a wide-field fluorescence microscope, the inner layer of the TRFG hydrogel appeared to be pure white, whereas the outer layers were transparent. The scanning electron microscopy (SEM) image of the lyophilized TRFG hydrogel can be seen in Fig. 2D. Figure 2E shows the field-emission scanning electron microscopy (FESEM) images of the cross section of the lyophilized HPEI-2k TRFG hydrogel. As can be seen, a dense PA layer is formed on the hydrogel surface conforming to the hydrogel pores. Results of the x-ray photoelectron spectroscopy (XPS) spectra of TRFG hydrogels are shown in fig. S2 and the elemental composition of the surface is shown in table S3. However, no notable difference was observed between TRFG hydrogels obtained using 2 or 25 kDa HPEI.

To see the effect of HPEI loading on the thermally responsive behavior of the TRFG hydrogel, the volume phase transition temperature (VPTT) of the hydrogel was measured using a differential scanning calorimeter (DSC) as shown in Fig. 2F. The VPTT of the pure PNIPAM/PSA semi-IPN hydrogel was found to be 35.37°C, which is higher than that of pure PNIPAM and can result from dispersion of PSA inside the PNIPAM matrix. However, the TRFG hydrogel shows no substantial change in the VPTT, compared to that of the PNIPAM/PSA hydrogel, indicating that the TRFG hydrogel is also thermally responsive even after loading of the protonated HPEI. This may result from the fact that the PNIPAM/PSA is a semi-IPN hydrogel network, where PSA and PEI are only dispersed in the PNIPAM network and do not form a part of the polymer backbone as in the case of the copolymerized hydrogel. The thermal response of the formed TRFG hydrogel can be further illustrated using the measured equilibrium swelling ratio (ESR) as shown in Fig. 2G. Both HPEI-2k and HPEI-25k showed a reduced swelling ratio at 60°C above its VPTT resulting from the change in hydrophilicity of the PNIPAM network. Figure 2H shows the rate of deswelling of the TRFG hydrogel at 60°C, indicating that most of the absorbed water is released in less than 30 min.

To study the effect of the surface cross-linked PA layer on DNA uptake, fluorescent-labeled DNA was used. The fabricated TRFG hydrogel beads were allowed to swell in a fluorescent-labeled DNA-containing solution at room temperature. Under a fluorescent chip scanner, the TRFG hydrogel did not produce any fluorescence in its initial state (Fig. 3A). However, after notable swelling, it produced a glowing green fluorescence, as shown in Fig. 3A. This was likely due to the complete absorption of the fluorescent-labeled DNA-containing solution within the hydrogel. This observation implied that most DNA molecules could enter the hydrogel from the solution. As the DNA forms a complex with the protonated amines of HPEI, it is adsorbed in the hydrogel, while the remaining solution can be released during deswelling by heating at 60°C. The presence of a protonated amine group on the TRFG hydrogel surface is further confirmed by the deconvolution of the N1 s peak from XPS spectra as shown in fig. S3. The swelling/deswelling process can be repeated a few number of times until the hydrogel gets saturated. Upon examination of the TRFG hydrogels with the fluorescent-labeled DNA via laser scanning confocal microscopy (LSCM), through nine swelling/deswelling cycles, it was clear that the HPEI-2k and HPEI-25k TRFG hydrogels with different HPEI molecular weights (2 kDa/25 kDa) demonstrated markedly different DNA uptake. A previous study (38) on the PNIPAM/PSA semi-IPN hydrogel with a PA layer has shown that both pH and molecular weight of PEI will affect the morphology of the surface cross-linked PA layer formed.

Fig. 3. Effect of HPEI molecular weight and concentration on fluorescent DNA absorption in TRFG hydrogels.
TRFG hydrogel (A) before and after three absorption cycles, as seen by a fluorescent chip scanner. Fluorescent DNA images of (B) HPEI-2k and (C) HPEI-25k TRFG hydrogels (through three cycles) using LSCM (Leica SP8). (D) Schematic illustration showing TRFG hydrogels with different HPEI molecular weights (2 kDa/25 kDa, 2 wt %) and its interaction with DNA. (E) FESEM images of HPEI-25k TRFG hydrogels with different wt % (0.5, 1, and 5) of 25-kDa HPEI used for PA layer formation. (F) Cross section of fluorescent DNA–loaded TRFG hydrogels using LSCM (Leica SP8) corresponding to different wt % (0.5, 1, and 5) of 25-kDa PEI (through six cycles) used in the fabrication of the TRFG hydrogel.

In the case of HPEI-2k, the DNA was absorbed mostly on the surface of the hydrogels, as shown in Fig. 3B. In the case of HPEI-25k, the DNA was absorbed mostly within the hydrogels, as shown in Fig. 3C. We believed that this may be related to the difference in the porosity of the formed PA layer. In the case of HPEI-2k, the formed PA layer was extremely dense, so the DNA was mostly adsorbed on the surface, as shown schematically in Fig. 3D. In the case of HPEI-25k, the formed PA layer was less dense and could result from diffusion limitation of the sterically hindered high–molecular weight HPEI. This results in a loose PA layer. Hence, the DNA molecules were able to permeate into the formed hydrogels, as shown in Fig. 3D. The above result is also supported by earlier work (38) where high rejection was observed for the PA layer formed by low–molecular weight PEI. Because high loading density of DNA was desired, we only focused on using the HPEI-25k TRFG hydrogel for further study.

Next, we investigated the effect of the HPEI loading in the TRFG hydrogel. For this, we used four different concentrations of the 25-kDa HPEI, i.e., 0.5, 1, 2, and 5 wt %, to form the TRFG hydrogel. To investigate the effect of the HPEI concentration on the DNA loading density, we again used fluorescent-labeled DNA. The surface morphology and cross section of the HPEI-25k TRFG hydrogel at different concentrations of HPEI were then observed via FESEM and LSCM, after nine swelling/deswelling cycles in DNA solution. Although the FESEM images (Fig. 3E) did not show much notable difference in the surface morphology of the formed hydrogels, the formed PA layer can be seen conforming to the pores of the hydrogel. As can be seen in Fig. 3F, the DNA loading density was found to increase with the increasing concentration of HPEI in aqueous solution used during the formation of the PA layer. The strong fluorescence seen within the hydrogel results from fluorescent DNA that binds to HPEI. The higher the diffusion of HPEI into the hydrogel during formation, the higher the loading density. We next investigated the maximum DNA loading capacity and DNA recovery for different TRFG hydrogels, which is discussed in the next section.

High DNA loading capacity and DNA recovery of TRFG hydrogels

Figure 4A shows the absorption/desorption mechanism of the formed TRFG hydrogel for DNA loading. DNA uptake occurs at pH 5 during swelling of the dried TRFG hydrogel bead in DNA solution at 20°C owing to interaction of protonated amines of HPEI with negatively charged DNA. The initial size of the HPEI-25k TRFG hydrogel (5 wt % HPEI) used was around 2 mm, and when observed with the naked eye, it appeared transparent white. The TRFG hydrogel swelled notably and turned into a pure white color. On heating above the VPTT of the TRFG hydrogel at 60°C, excess solution is released while retaining the absorbed DNA within the hydrogel. Through the subsequent swelling/deswelling cycle, DNA can be concentrated inside the TRFG hydrogel.

Fig. 4. Mechanism of DNA absorption and desorption cycles.
(A) A schematic representation of TRFG hydrogel DNA absorption via swelling/deswelling in DNA-containing solution until saturation. DNA is then released and recovered using an elution buffer with pH >9. (B) Graph showing one complete cycle of absorption, i.e., swelling at 20°C and deswelling at 60°C, at pH 5. (C) Changes in swelling ratio of the HPEI-25k TRFG hydrogel (5 wt % HPEI) for different DNA absorption cycles as a function of time. (D) Swelling time needed for complete absorption of 1.5 ml of initial solution containing DNA at pH 5 for different absorption cycles. (E and F) Graph showing change in the concentration of the DNA in the solution released from hydrogel after deswelling at 60°C after different cycles for DNA absorption and desorption, respectively. The inset shows the HPEI-25k TRFG hydrogel (5 wt % HPEI) cross section for different cycles, as observed by LSCM (Leica SP8). The inset schematic shows absorption of DNA at pH 5 due to interaction with protonated amine groups of HPEI and desorption due to deprotonation of HPEI at pH 10.

As shown in Fig. 4B, one single cycle of DNA absorption involves swelling of the hydrogel at 20°C in a solution containing DNA at pH 5 and deswelling the hydrogel by heating above VPTT of the hydrogel, i.e., 60°C. Figure 4C shows the changes in the swelling ratio of the HPEI-25k TRFG hydrogel (5 wt % HPEI) bead for different absorption cycles as a function of time. However, as shown in Fig. 4D, it can be observed that the time needed for swelling the hydrogel for a constant volume (1.5 ml) of the DNA solution used for the absorption study increases from 30 to 50 min, while the deswelling time is kept constant for 20 min at 60°C. This could be due to fewer protonated functional groups being available as a result of charge compensation following complexation of protonated amines of HPEI with DNA molecules.

To test the DNA binding ability of TRFG hydrogels, we determined the DNA concentration in the solution released from the TRFG hydrogel after deswelling at 60°C, as shown in Fig. 4E. The higher the DNA uptake by the hydrogel for a given cycle, the lower the amount of DNA in the released solution. Moreover, by labeling DNA molecules with fluorescent nucleic acid markers, we were able to observe the change in the distribution of DNA within the TRFG hydrogel during different cycles. The inset of Fig. 4E shows the cross section of the HPEI-25k TRFG hydrogel bead formed using 5 wt % HPEI after three, six, and nine DNA absorption cycles. As can be seen, the fluorescence intensity within the hydrogel bead is found to increase with each absorption cycle and gets saturated after nine cycles. However, the TRFG hydrogel still shows thermal responsiveness as evident from Fig. 4C where the change in the swelling ratio can still be observed. Moreover, once the TRFG hydrogel gets saturated, the absorbed DNA can be released using the elution buffer with pH >9, and the bead can be reused again as shown in Fig. 4F. At high pH, the deprotonation of amine results in the decomplexation of HPEI/DNA and, hence, the subsequent release of DNA from the TRFG hydrogel.

To determine the maximum DNA loading density of TRFG hydrogels, the standard DNA concentration curve obtained by the nucleic acid detection test was calibrated by the standard DNA template, as shown in Fig. 5A. The DNA absorption performance of TRFG hydrogels under different conditions was studied by evaluating the change in the concentration of DNA in the original solution and in the solution released from a single hydrogel bead after deswelling at 60°C. As shown in fig. S4, the zeta potential (millivolt) values of the TRFG hydrogel suggest that the hydrogels were positively charged when the solution pH was less than 9, and the entire reaction system remained relatively stable when the pH was less than 7 or greater than 10. Therefore, DNA absorption buffers containing 10 ng of DNA, with varying pH values around 7, were prepared to examine the DNA absorption ability of the TRFG hydrogels. Simultaneously, self-prepared elution buffers, with varying pH values around 10, were used to assess the DNA releasing ability of the TRFG hydrogels. Figure 5B illustrates that the hydrogel shows a maximum DNA loading capacity when the pH of the binding buffer was 5. Moreover, the absorbed DNA can be easily released using an elution buffer of pH 10, and when the pH of the elution buffer was below 8, the absorbed DNA could not be released from the TRFG hydrogels (Fig. 5D). The effect of temperature on the DNA binding ability of TRFG hydrogel was also examined. As depicted in Fig. 5C, the highest DNA uptake capacity of absorption was obtained at a temperature of 16°C, and it occurred in the least amount of time. This is due to the maximum swelling of the TRFG hydrogel bead (each binding cycle) at 16°C, such that it absorbed far more DNA in a single swelling cycle.

Fig. 5. Optimum conditions for DNA absorption and desorption.
(A) Standard DNA concentration curve, where red, black, and gray colors represent the DNA concentrations of the initial solution (absorption buffer), initial solution [binding buffer released from the HPEI-25k TRFG hydrogel (5 wt % HPEI)], and desorption buffer removed from the HPEI-25k TRFG hydrogel (5 wt % HPEI), respectively. a.u., absorbance unit (B) Effect of initial solution pH on the DNA absorption. (C) Effect of temperature on DNA absorption during swelling. (D) Effect of elution buffer pH on the DNA desorption. Note that the sum of concentration for the absorption and desorption cycles refers to the concentration of DNA in the released buffer solution from the HPEI-25k TRFG hydrogel on deswelling.

Because the volumes of the initial binding and constant-volume binding buffers were 1.5 ml (1500 μl) each, the DNA loading (in nanograms) of the TRFG hydrogels was calculated as follows: $\sum_{n = 1}^{n} C_{n - 1} - C_{n}$ × 1500 (ng), where C_n was the concentration of DNA after n absorption cycles. Generally, the buffer volumes are calculated by measuring the absorbance value A and by using the following formula: A = KCL = K′c, c = A/K′. If the measurement instrument directly provides C in the gram per milliliter unit, then 3000 μl becomes 3000 × C. At n = 0, the original concentration is C₀. When n = 1, C₁ represents the solution concentration after the first adsorption, and the amount of adsorption is C₀ − C₁. When n = n, the solution concentration after the nth adsorption is C_n, and the adsorption amount is (C₀ − C₁ + C₁ − C₂ + … + C_n−1− C_n), which is $\sum_{n = 1}^{n} C_{n - 1} - C_{n}$ . The specific calculation result for this is as follows.

The maximum DNA loading (in nanograms) of the HPEI-25k TRFG hydrogels (5 wt % HPEI) was 80 × 1500 = 1.20 × 10⁵ ng. As shown in fig. S5, it weighed 0.0094 g. Thus, the DNA loading density of the TRFG hydrogel was 1.30 × 10⁷ ng/g. The average weight of a base that make up DNA was approximately 10⁹ × 330/NA ≈ 5 × 10⁻¹³ (ng), wherein NA was the Avogadro constant. These data implied that 1 g of the TRFG hydrogels contains approximately 1.3 × 10⁷/5 × 10⁻¹³ = 2.60 × 10¹⁹ bases. Moreover, because in our previously unknown designed conversion program, 4 base pairs (bp) represented 1 byte, therefore, 1 g of the TRFG hydrogel developed in this study contained approximately 0.70 × 10¹⁹ bytes = 0.70 × 10¹⁹/1024 kB = 0.70 × 10¹⁹/1024/1024 MB = 0.70 × 10¹⁹/1024/1024/1024 GB ≍ 0.70 × 10¹⁰ GB. Therefore, the information density of the TRFG hydrogel developed in this study was nearly 7.0 × 10⁹ GB/g. Figures S6 and S7 present the data density of the TRFG hydrogel, i.e., 2.40 × 10⁹ GB/g (HPEI-2k, 2 wt % HPEI) and 3.98 × 10⁹ GB/g (HPEI-25k, 2 wt % HPEI), respectively. Compared to other methods of DNA storage, our approach was relatively simple, but it required extensive time for swelling/deswelling cycles (48). However, compared to other methods of DNA storage, TRFG hydrogels offered great benefit in terms of the relatively simple fabrication procedure, high data density, and ease of handling compared to other methods of DNA storage (18, 49–53).

Furthermore, to establish the minimum amount of DNA that can be recovered from the HPEI-25k TRFG hydrogel (5 wt % HPEI), we used a six-step concentration gradient (5, 10, 25, 50, 75, and 100 copies/ml), assessed DNA binding and release, and performed downstream PCR and dPCR. At this stage, it was necessary to include quantitative PCR (qPCR) technology for ultrasensitive DNA detection, because nucleic acid detection, with a detector, is generally limited. On the basis of the formula, single-stranded DNA: copies/ml = 10⁻³copies/μl = [(6.02 × 10²³) × (ng/μl × 10⁻¹²)]/(DNA length × 330), therefore, 1 ng/μl DNA was equal to nearly 10⁶ copies/ml. As depicted in Table 1, according to the Poisson distribution, because the DROPDx-2044HT dPCR system produces approximately 20,000 droplets in each reaction (N = 20,000), one can calculate the DNA copy number by counting the number of theoretical positive and negative droplets (N_p and N_n, respectively) for a given assay setup as follows: λ = −ln (N_n/N). When the target DNA concentrations of 5, 10, 25, 50, 75, and 100 copies were introduced to the binding reaction, the theoretical DNA results were 5, 10, 25, 50, 75, and 100 copies, respectively (if all the DNA copies were released from the TRFG hydrogel and added into the dPCR). Figure 6A 3 presents the percentage recovery for the actual results of the dPCR, which were 3, 8, 21, 47, 73, and 98, respectively. When the DNA sample concentration was below 5 copies/ml, the results were far from the theoretical value, and the DNA released from the TRFG hydrogel bead had a recovery rate of 60 to 84%, as depicted in Fig. 6A (1 and 2), which was not sufficient for accurate DNA quantification but was good enough for DNA sequencing. On the basis of these results, the DNA recovery limit using the TRFG hydrogel was approximately 5 copies. The results of all DNA templates are summarized in Table 1. When the DNA sample concentration was more than 50 copies/ml, the results were very close to the theoretical value, and the DNA released from the hydrogel show very good recovery (all more than 90%; see fig. S8). Although dPCR takes close to 50 min, if combined with our previous research results involving digital loop-mediated isothermal amplification (LAMP) simulation, the processing time can be reduced to about 20 to 30 min (54). The results of the HPEI-25k TRFG hydrogel (2 wt % HPEI) is shown in fig. S9, and the results are comparable to the HPEI-25k TRFG hydrogel (5 wt % HPEI). In summary, on the basis of our analysis, the minimum DNA amount that can be recovered from the TRFG hydrogels was between 5 and 10 copies.

DNA in mix (copies)	Theoretical result (N = 20,000)			Actual results using hydrogel bead
DNA in mix (copies)	λ	N_p	P (0)	N_p	N	λ	Copies	Recovery
5	0.00025	5	99.975%	3	19,076	0.00010	3	60%
10	0.00050	10	99.950%	8	19,234	0.00037	8	70%
25	0.00125	25	99.875%	21	19,522	0.00105	21	84%
50	0.00250	50	99.750%	47	19,087	0.00242	47	94%
75	0.00375	75	99.625%	73	19,104	0.00371	73	97%
100	0.00500	100	99.500%	98	18,993	0.00508	98	98%

Table 1. DNA recovery from the TRFG hydrogel.

DNA recovery from the TRFG hydrogel (25 kDa, 5 wt % HPEI) with varying initial DNA concentrations. N, total number of droplets; N_p, number of positive droplets; N_n, number of negative droplets; λ, copies per droplet = −ln (N_n/N) = copies in the reaction/total droplets = reaction concentration (copies/μl) × V (volume of each droplet); V = reaction volume/N = 0.0005 μl; P (0) = N_n/N = e^-λ, λ = −ln(N_n/N) = reaction concentration × V; reaction concentration = [−ln(N_n/N)]/V.

Fig. 6. DNA recovery, storage, and sequencing.
(A) dPCR results based on TRFG hydrogels (HPEI 25 kDa, 5 wt %). (1) Image of chemiluminescent droplets in dPCR; (2) counts and classification of chemiluminescent droplets. The DNA concentration of (1) and (2) is 5 copies/ml. (3) Percentage recovery of absorbed DNA from the TRFG hydrogels with different DNA concentrations. (B to D) Degradation kinetics of dry DNA storage. (B) Effect of temperature on DNA integrity (per qPCR) as a function of time using three different dry DNA storage technologies: DNA on FTA filter cards, DNA encapsulated in silica, and DNA in TRFG hydrogels (this study). (C) First-order decay rate constants derived thereof and activation energy of degradation processes assuming first-order kinetics. (D) Half-life of 158-bp DNA stored in TRFG hydrogels according to the Arrhenius equation with activation energies of about 155 kJ mol⁻¹ and compared to literature data on DNA in silica (50), desiccated DNA stability (55), and DNA stability in ancient moa bone (58). Data from the literature are scaled (58) by t_1/2158 nt = t_1/21 nt/158. t_1/2 means half-life of the DNA. (E) PCR results of the DNA recovered from TRFG hydrogels. Sanger sequencing results of fragments (lane 1) meant “Welcome to Nanyang Technological University.”

The TRFG hydrogel was further evaluated for long-term protection and stability of DNA for storage. For the HPEI-25k TRFG hydrogel (5 wt % HPEI) developed in this work, DNA plus 50% DNA shield (REIH) exhibited the best stability. This is likely due to the fact that TRFG hydrogels prevent further degradation of DNA by encapsulating DNA through a complex formation with PEI that enhances DNA stability. In addition, we also explored whether DNA stored in the solid state is more stable by comparing the activation energy for degradation kinetics of DNA stored in TRFG hydrogels with literature data on encapsulated DNA in silica (50), desiccated DNA (55), and DNA stability in ancient moa bone. As shown in Fig. 6B, compared to the storage of solid-state DNA without additional agents, all three DNA storage technologies decreased the DNA decay rates considerably. TRFG hydrogels show notable advantages on the time of storage compared with the filter card, while showing a similar effect on the time of storage compared with DNA in silica. From the temperature dependence of the decay rates, Arrhenius-type activation energies were calculated by assuming first-order kinetics, which were equivalent for all three storage formats (about 155 kJmol⁻¹ as shown in Fig. 6C; see table S4). Through calculation, the k₀ (s − 1) of the TRFG hydrogel is close to that of DNA in silica and much less than that of the filter card. It proves again that TRFG hydrogels show notable advantages on the time of storage compared with the filter card, while showing a similar effect on the time of storage compared with DNA in silica. As shown in Fig. 6D, under the same condition, the life expectancy of DNA storage in the TRFG hydrogel is comparable to existing technologies, enabling DNA storage for more than 1000 years (twice of half-life) at room temperature (20°C).

Verification of DNA recovery from the TRFG hydrogels by PCR and sequencing data

Last, the DNA released from the HPEI-25k TRFG hydrogels (5 wt % HPEI) was assessed via PCR (gel electrophoresis) before sequencing by GenScript Biotech Corp. (as a reference) or a dual mononucleotide sequencing system, as reported in our previous study (56). As depicted in fig. S10A, the PCR verification was successful for long sequence target fragments, around 300 bp (“Welcome to Nanyang Technological University,” the motto of the Nanyang Technological University and Harvard University), as well as short sequence target fragments, around 100 bp (the motto of Southeast University, Oxford University, and Columbia University). On the basis of the fluorescence qPCR (fig. S10B), the curves of two different cycles validated the effectiveness of the TRFG hydrogels used in this study. In terms of sequencing, although the accuracy of the Sanger sequencing (each run) was around 90%, following six times sequencing error correction, its accuracy reached 100%, as shown in Fig. 6E. While the accuracy of the dual mononucleotide sequencing system (each run) was around 93%, upon six times sequencing error correction, the accuracy reached 100%, as shown in fig. S10C. It depicts that the sequence “Welcome to Nanyang Technological University” is successfully decoded through our decoding program. It is worth noting that when two different target fragments are stored in a TRFG hydrogel at the same time with a different number of cycles (fragments that mean the motto of Southeast University with cycle 1 and fragments that mean “Welcome to Nanyang Technological University” with cycle 2), the different fragments could both be decoded correctly (fig. S11).

DISCUSSION

Conventionally, inorganic matrices have been used for long-term DNA storage because of easy processibility (18). However, they are limited by low loading density. Moreover, the encapsulated DNA cannot be fully recovered or needs complex recovery techniques (10, 18, 51). This work highlights the flexibility of creating a chemically cross-linked hydrogel network with desired functionality for high DNA uptake from the solution. By tuning properties of the hydrogels, we can enhance the DNA loading of the hydrogel from surface adsorption to absorption within the hydrogel with long-term stability. Moreover, the absorbed DNA can be easily recovered and the hydrogel can be reused for further DNA uptake. The concept presented in this work is different from previous studies whereby the DNA was condensed using cationic polymer to form polyplex nanoparticles and encapsulated into a degradable hydrogel matrix during gelation for gene therapy (25–29). The flexibility of modifying the hydrogel matrix for immobilizing multiple DNA molecules as well as its thermoresponsiveness that allows multiple swelling/deswelling cycle for concentrating DNA within hydrogel will hold great significance in the further development of hydrogels for DNA storage.

In this study, we comprehensively investigated the DNA storing process with an emphasis on enhancing the data density. First, a converter of digital information and genetic code was developed. Because four bases coded 1 byte, the DNA storage density was enhanced by improving the efficiency of the converter. TRFG hydrogels were rationally designed to have a high DNA loading capacity using a thermally responsive PNIPAM/PSA semi-IPN hydrogel as a support. The presence of PSA provided anchoring points for high loading of protonated HPEI within the hydrogel, which then enables DNA uptake via complexation. Both the concentration and molecular weight of HPEI played a critical role to enhance the DNA loading capacity of the TRFG hydrogel together with the thermoresponsive behavior of the TRFG hydrogel, which allows notable concentration of DNA within the hydrogel beads during multiple consecutive swelling/deswelling cycles in a DNA-containing solution. TRFG hydrogels release absorbed solution while retaining DNA through interaction with HPEI during deswelling at 60°C owing to its thermoresponsive characteristic. Moreover, the thin PA layer on the hydrogel surface further ensures retention of DNA during swelling/deswelling cycles. The DNA data density of the TRFG hydrogels (7.0 × 10⁹ GB/g) was one of the highest values reported to date using hydrogel. It is to be noted that because we cannot directly measure the concentration of DNA absorbed by the hydrogel, it might be possible that not all the absorbed DNA is released during deswelling. Moreover, the absorbed DNA can be released by the nondestructive method simply by changing the pH of the solution. In addition, as shown in table S5, TRFG hydrogels developed demonstrated advantages in terms of DNA stability, protection, and data recovery compared to other methods of data storage (10, 18, 20, 51, 57). Above all, the DNA storage system based on the TRFG hydrogels notably improved DNA data density, long-term accuracy, recovery, and ease of use. On the basis of these properties, our newly synthesized TRFG hydrogels have enormous application potential, particularly in the storage of DNA.

MATERIALS AND METHODS

Materials

N-isopropylacrylamide (NIPAM; Wako Pure Chemical Industries Ltd.) was recrystallized from n-hexane before use. For hydrogel synthesis, N,N′-methylene-bis-acrylamide (MBA), ammonium peroxydisulfate, and N,N,N′,N′-tetramethyl ethylenediamine (TEMED) were used as a cross-linker, initiator, and accelerator, respectively, and they were purchased from Sigma-Aldrich and used as received. PSA polymer with a molecular weight of 800 kDa was obtained from Nippon Shokubai Pte Ltd. and used as received. Branched PEI (Sigma-Aldrich) with a molecular weight of 2 and 25 kDa, TMC (Sigma-Aldrich), and hexane were used for the synthesis of the PA layer on the hydrogel surface. Droplet production oil and stabilizer were obtained from Suzhou Rainsure Scientific Co. Ltd. Adenosine 5′-phosphosulfate, adenosine 5′-triphosphate sulfurylase, Bacillus stearothermophilus (Bst) polymerase v2.0 Warm Start, and 10× Bst buffer were purchased from New England Biolabs (NEB), Beijing, China. d-Luciferin sodium was obtained from Promega (Shanghai, China). Avian myeloblastosis virus Reverse Transcriptase was purchased from Woosen Biotechnology (Shanghai, China).

DNA coding and DNA synthesis

First, a Python program was designed to realize the conversion between the DNA code and digital information. The DNA sequences implied the motto of five universities and “Welcome to Nanyang Technological University” (~300 bp).

Their amplified primers were synthesized and purified by GenScript Biotech Corp. The encoding and decoding processes are shown in fig. S1. The details of encoded information are shown in table S2. The synthesized DNA sequence (20 to 40 μg) was dissolved in 2 ml of DNA binding buffer (including some DNA shield solution) with a pH of 3.5 to 6.5 according to the process described in the “Verification of DNA recovery from TRFG hydrogels by PCR/dPCR for sequencing analysis” section. The details of amplified primers are shown in table S1.

Synthesis and characterization of the TRFG hydrogels

Synthesis of the PNIPAM/PSA semi-IPN hydrogel

The PNIPAM/PSA semi-IPN hydrogel used in this study was synthesized using the free-radical polymerization method. Typically, NIPAM (0.8 M), MBA (5 mol % of NIPAM), and PSA (0.2 M) were first dissolved in deionized (DI) water under constant stirring at room temperature. The mixture was then purged with nitrogen for 20 min and allowed to cool down to a polymerization temperature of 4°C for a few hours before polymerization. Keeping the mixture cold under a nitrogen atmosphere, 8 mM freshly prepared TEMED (5 vol %) and ammonium peroxydisulfate (APS) (10 wt %) solutions, which were also cooled to the desired polymerization temperature, were sequentially added with continuous stirring for 20 s. The mixture was then quickly transferred to a cylindrical mold (different molds could be used to synthesize hydrogels with various sizes and shapes) with a thickness of approximately 0.5 mm and placed in a desiccator purged with argon, before undergoing low-temperature polymerization at 10°C for 24 hours. The semi-IPN hydrogels were thoroughly washed by immersing them in DI water for 3 days with periodic water change, which was followed by drying at 60°C.

Fabrication of the PA layer

Gel-liquid interfacial polymerization was used to prepare the TRFG hydrogels with a PA layer. Fully swollen PNIPAM/PSA semi-IPN hydrogels were immersed in HPEI aqueous solutions (2 kDa/25 kDa) with varying concentrations (0.5, 1, 2, and 5 wt %) for 30 min. The pH value of the PEI solution was adjusted to 5 to ensure protonation of amine groups. After removing the excess solution from the hydrogel surface, the HPEI-containing hydrogel was suspended in hexane-containing TMC (0.1 wt %) for a 30-min polymerization time to fabricate the cross-linked PA layer on the hydrogel surface. After heating the TRFG hydrogel sample at 60°C for 7 min, it was thoroughly rinsed with hexane and water to eliminate any unreacted compounds. The gels were successively submerged in 20, 40, and 60% ethanol before being dried at 60°C for 24 hours.

Characterization of TRFG hydrogels

The microscopic morphology of the lyophilized fully swollen TRFG hydrogels was analyzed using FESEM (JEOL JSM-7600F, Germany) equipped with an EDS system, SEM (JEOL JSM-IT200, Japan), and BZ-X800E system (KEYENCE, Tokyo, Japan). Before the FESEM observation, the samples were sputter-coated with platinum to improve their conductivity. The distribution results of DNA molecules were characterized by LSCM (Leica SP8). The DROPDx-2044HT dPCR system was used from Suzhou Rainsure Scientific Co. Ltd., an ultraweak luminescence analyzer was purchased from from Jianxinlitou (Beijing, China), and an all-in-one BZ-X800E fluorescence microimaging system was used from Keyence (Shanghai, China).

The VPTT of the pure PNIPAM/PSA semi-IPN hydrogel and that of the TRFG hydrogel were measured by DSC (TA Q10, TA Instruments). The fully swollen hydrogel (5 to 10 mg) was sealed into an aluminum hermetic pan, equilibrated at 10°C, and heated to 80°C at 3°C/min. The binding energy of electrons at C1 s, N1 s, and O1 s core levels for the completely dried hydrogel surface was conducted using XPS (Kratos Axis Supra) with a monochromatic Al Kα x-ray source. The ESR of the fully swollen hydrogels in DI water was measured gravimetrically.

where w_t is the weight of the fully swollen hydrogel and w_d is the dry weight of the hydrogel. The ESR at 60°C was measured after 1-hour immersion in water at the respective temperature.

DNA capacity and DNA recovery of TRFG hydrogels

First, a standard DNA template (0 to 10 ng/μl) was prepared to calibrate the standard nucleic acid detection curve. Second, the synthesized TRFG hydrogels were used to absorb the entire initial solution (1.5 ml) under different temperatures between 16° and 26°C. Then, the temperature of the beads was raised to 60°C for nearly 20 min to release the entire water that was absorbed into the TRFG hydrogels with a thermal cycling instrument. The DNA concentrations of both initial solution and released water from the beads were measured by a nucleic acid detector. The entire absorption cycles were repeated until there was no notable change in the concentration of the original solution and the water released from the TRFG hydrogels. To facilitate DNA release, the TRFG hydrogels loaded with DNA after water release were placed in an elution buffer (2 ml) with different pH values (9 to 12). Subsequently, by changing the temperature from 18° to 60°C, the elution buffer was absorbed and released several times to achieve complete DNA release. The concentration of DNA in the elution buffer released from the TRFG hydrogels was tested by the nucleic acid detector. The TRFG hydrogels were loaded with a saturated amount of DNA and then dried for later use.

To determine the minimum amount of DNA that could be recovered from the TRFG hydrogels, the initial solution (1 ml) was used to prepare a six-step concentration gradient (5, 10, 25, 50, 75, and 100 copies/ml) for DNA binding and release, which was then used for downstream PCR and dPCR. The detection was conducted on a DROPDx-2044HT dPCR system (Suzhou Rainsure Scientific Co. Ltd.). This system produced nearly 20,000 droplets for one test, and the microfluidic chip matched the dPCR system. After PCR, the system was used to read the chemiluminescence signal.

Verification of DNA recovery from TRFG hydrogels by PCR/dPCR for sequencing analysis

The DNA recovered from TRFG hydrogels (25 kDa, 5 wt % HPEI) was used as the template for PCR. The 10-μl PCR amplification system contained 1 μl of DNA template with different concentrations, 1 μl of primer, and 8 μl of reaction mix [0.2 mM deoxynucleotide triphosphates, 1 μl of Golden Star T6 Super TaKaRa Taq DNA polymerase (2 U), 1 mM amplification buffer, and 1.5 mM MgCl₂]. The amplification conditions were as follows: predenaturation at 98°C for 30 s; 22 cycles of denaturation at 98°C for 30 s, annealing at 60°C for 30 s, extension at 72°C for 30 s, and another extension at 72°C for 1 min. Then, the PCR products were sent to GenScript Biotech Corp. for sequencing and further analysis and used for dual mononucleotide sequencing.

Acknowledgments

We would like to thank Nanyang Technological University, Singapore, for the research scholarship via the Interdisciplinary Graduate Programme.

Funding: This work was supported by the National Key Research and Development Program of China (no. 2020YFA0712104) and the National Natural Science Foundation of China (no. 61971123).

Author contributions: The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Conceptualization: P.X., X.H., Z.F., and N.G. Methodology and investigation: N.G. (hydrogel fabrication and characterization) and Z.F. and M.L. (encoding/decoding information in DNA, DNA storage, and recovery study in/from hydrogel). Funding acquisition and supervision: P.X. and X.H. Visualization: P.X., X.H., Z.F., and N.G. Writing—original draft: Z.F. and N.G. Writing—review and editing: P.X., X.H., Z.F., and N.G.

Competing interests: H.X. and N.G. are coinventors on a provisional patent application related to this work pertaining to the hydrogel filed by Nanyang Technological University (no. 10202251436A; filed 20 October 2022). The authors declare that they have no other competing interests.

Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.

Supplementary Materials

This PDF file includes:

Supplementary Text

Figs. S1 to S11

Tables S1 to S5

Download
1.61 MB

REFERENCES AND NOTES

N. Goldman, P. Bertone, S. Chen, C. Dessimoz, E. M. LeProust, B. Sipos, E. Birney, Towards practical, high-capacity, low-maintenance information storage in synthesized DNA. Nature 494, 77–80 (2013).

S. M. H. T. Yazdi, R. Gabrys, O. Milenkovic, Portable and error-free DNA-based data storage. Sci. Rep. 7, 5011 (2017).

L. Organick, S. D. Ang, Y. J. Chen, R. Lopez, S. Yekhanin, K. Makarychev, M. Z. Racz, G. Kamath, P. Gopalan, B. Nguyen, C. N. Takahashi, S. Newman, H. Y. Parker, C. Rashtchian, K. Stewart, G. Gupta, R. Carlson, J. Mulligan, D. Carmean, G. Seelig, L. Ceze, K. Strauss, Random access in large-scale DNA data storage. Nat. Biotechnol. 36, 242–248 (2018).

R. F. L. Evans, R. W. Chantrell, U. Nowak, A. Lyberatos, H. J. Richter, Thermally induced error: Density limit for magnetic data storage. Appl. Phys. Lett. 100, 102402 (2012).

S. H. Charap, P.-L. Lu, Y. He, Thermal stability of recorded information at high densities. IEEE Trans. Magn. 33, 978–983 (1997).

D. Carmean, L. Ceze, G. Seelig, K. Stewart, K. Strauss, M. Willsey, DNA data storage and hybrid molecular–electronic computing. Proc. IEEE 107, 63–72 (2019).

R. Heckel, G. Mikutis, R. N. Grass, A characterization of the DNA data storage channel. Sci. Rep. 9, 9663 (2019).

V. Zhirnov, R. M. Zadegan, G. S. Sandhu, G. M. Church, W. L. Hughes, Nucleic acid memory. Nat. Mater. 15, 366–370 (2016).

F. Farzadfard, DNA storage in everyday objects. Nat. Biotechnol. 38, 31–32 (2020).

S. L. Shipman, J. Nivala, J. D. Macklis, G. M. Church, CRISPR-Cas encoding of a digital movie into the genomes of a population of living bacteria. Nature 547, 345–349 (2017).

L. Ceze, J. Nivala, K. Strauss, Molecular digital data storage using DNA. Nat. Rev. Genet. 20, 456–466 (2019).

D. Clermont, S. Santoni, S. Saker, M. Gomard, E. Gardais, C. Bizet, Assessment of DNA encapsulation, a new room-temperature DNA storage method. Biopreserv. Biobank. 12, 176–183 (2014).

K. A. Alkhamis, Influence of solid-state acidity on the decomposition of sucrose in amorphous systems II (effect of buffer). Drug Dev. Ind. Pharm. 35, 408–416 (2009).

B. Zhu, T. Furuki, T. Okuda, M. Sakurai, Natural DNA mixed with trehalose persists in B-form double-stranding even in the dry state. J. Phys. Chem. B 111, 5542–5544 (2007).

M. Simbolo, M. Gottardi, V. Corbo, M. Fassan, A. Mafficini, G. Malpeli, R. T. Lawlor, A. Scarpa, DNA qualification workflow for next generation sequencing of histopathological samples. PLOS ONE 8, e62692 (2013).

S. E. Howlett, H. S. Castillo, L. J. Gioeni, J. M. Robertson, J. Donfack, Evaluation of DNAstable™ for DNA storage at ambient temperature. Forensic Sci. Int. Genet. 8, 170–178 (2014).

A. N. Zelikin, A. L. Becker, A. P. R. Johnston, K. L. Wark, F. Turatti, F. Caruso, A general approach for DNA encapsulation in degradable polymer microcapsules. ACS Nano 1, 63–69 (2007).

W. D. Chen, A. X. Kohll, B. H. Nguyen, J. Koch, R. Heckel, W. J. Stark, L. Ceze, K. Strauss, R. N. Grass, Combining data longevity with high storage capacity—layer-by-layer DNA encapsulated in magnetic nanoparticles. Adv. Funct. Mater. 29, 1901672 (2019).

D. Paunescu, M. Puddu, J. O. B. Soellner, P. R. Stoessel, R. N. Grass, Reversible DNA encapsulation in silica to produce ROS-resistant and heat-resistant synthetic DNA 'fossils'. Nat. Protoc. 8, 2440–2448 (2013).

A. X. Kohll, P. L. Antkowiak, W. D. Chen, B. H. Nguyen, W. J. Stark, L. Ceze, K. Strauss, R. N. Grass, Stabilizing synthetic DNA for long-term data storage with earth alkaline salts. Chem. Commun. 56, 3613–3616 (2020).

V. Calvino-Casilda, A. J. López-Peinado, E. Vaganova, S. Yitzchaik, I. E. Pacios, I. F. Piérola, Porosity inherent to chemically crosslinked polymers. Poly (N-vinylimidazole) hydrogels. J. Phys. Chem. B 112, 2809–2817 (2008).

J. Jang, Y. J. Seol, H. J. Kim, J. Kundu, S. W. Kim, D. W. Cho, Effects of alginate hydrogel cross-linking density on mechanical and biological behaviors for tissue engineering. J. Mech. Behav. Biomed. Mater. 37, 69–77 (2014).

L. Arens, J. B. Albrecht, J. Höpfner, K. Schlag, A. Habicht, S. Seiffert, M. Wilhelm, Energy consumption for the desalination of salt water using polyelectrolyte hydrogels as the separation agent. Macromol. Chem. Phys. 218, 1700237 (2017).

E. Kurt, T. Segura, Nucleic acid delivery from granular hydrogels. Adv. Healthc. Mater. 11, 2101867 (2022).

L. A. L. Fliervoet, J. F. J. Engbersen, R. M. Schiffelers, W. E. Hennink, T. Vermonden, Polymers and hydrogels for local nucleic acid delivery. J. Mater. Chem. B 6, 5651–5670 (2018).

A. Luchini, D. H. Geho, B. Bishop, D. Tran, C. Xia, R. L. Dufour, C. D. Jones, V. Espina, A. Patanarut, W. Zhou, M. M. Ross, A. Tessitore, E. F. Petricoin, L. A. Liotta, Smart hydrogel particles: Biomarker harvesting: One-step affinity purification, size exclusion, and protection against degradation. Nano Lett. 8, 350–361 (2008).

Y. Lei, S. Huang, P. Sharif-Kashani, Y. Chen, P. Kavehpour, T. Segura, Incorporation of active DNA/cationic polymer polyplexes into hydrogel scaffolds. Biomaterials 31, 9106–9116 (2010).

L. A. L. Fliervoet, H. Zhang, E. van Groesen, K. Fortuin, N. J. C. B. Duin, K. Remaut, R. M. Schiffelers, W. E. Hennink, T. Vermonden, Local release of siRNA using polyplex-loaded thermosensitive hydrogels. Nanoscale 12, 10347–10360 (2020).

S. Siegman, N. F. Truong, T. Segura, Encapsulation of PEGylated low-molecular-weight PEI polyplexes in hyaluronic acid hydrogels reduces aggregation. Acta Biomater. 28, 45–54 (2015).

W. Chen, T. Wang, Z. Dou, X. Xie, Self-driven “microfiltration” enabled by porous superabsorbent polymer (PSAP) beads for biofluid specimen processing and storage. ACS Mater. Lett. 2, 1545–1554 (2020).

F. Li, J. Tang, J. Geng, D. Luo, D. Yang, Polymeric DNA hydrogel: Design, synthesis and applications. Prog. Polym. Sci. 98, 101163 (2019).

V. Morya, S. Walia, B. B. Mandal, C. Ghoroi, D. Bhatia, Functional DNA based hydrogels: Development, properties and biological applications. ACS Biomater Sci. Eng. 6, 6021–6035 (2020).

S. Khajouei, H. Ravan, A. Ebrahimi, DNA hydrogel-empowered biosensing. Adv. Colloid Interface Sci. 275, 102060 (2020).

J. S. Kahn, Y. Hu, I. Willner, Stimuli-responsive DNA-based hydrogels: From basic principles to applications. Acc. Chem. Res. 50, 680–690 (2017).

M. C. Koetting, J. T. Peters, S. D. Steichen, N. A. Peppas, Stimulus-responsive hydrogels: Theory, modern advances, and applications. Mater. Sci. Eng. R Rep. 93, 1–49 (2015).

Q. Shi, H. Liu, D. Tang, Y. Li, X. Li, F. Xu, Bioactuators based on stimulus-responsive hydrogels and their emerging biomedical applications. NPG Asia Mater. 11, 64 (2019).

N. Gupta, Y. N. Liang, J. S. K. Lim, X. Hu, Design rationale for stimuli-responsive, semi-interpenetrating polymer network hydrogels—A quantitative approach. Macromol. Rapid Commun. 41, 2000199 (2020).

N. Gupta, Y. N. Liang, J. W. Chew, X. Hu, Highly robust interfacially polymerized PA layer on thermally responsive semi-IPN hydrogel: Toward on-demand tuning of porosity and surface charge. ACS Appl. Mater. Interfaces 13, 60590–60601 (2021).

A. J. F. Dalcin, I. Roggia, S. Felin, B. S. Vizzotto, M. Mitjans, M. P. Vinardell, A. P. Schuch, A. F. Ourique, P. Gomes, UVB photoprotective capacity of hydrogels containing dihydromyricetin nanocapsules to UV-induced DNA damage. Colloids Surf. B Biointerfaces 197, 111431 (2021).

A. Beyer, D. Cialla-May, K. Weber, J. Popp, Hydrogel decorated chips for convenient DNA test. Macromol. Chem. Phys. 217, 959–965 (2016).

E. P. Jonášová, A. Bjørkøy, B. T. Stokke, Toehold length of target ssDNA affects its reaction-diffusion behavior in DNA-responsive DNA-co-acrylamide hydrogels. Biomacromolecules 21, 1687–1699 (2020).

J. Gǎanin, C. V. Synatschke, T. Weil, Biomedical applications of DNA-based hydrogels. Adv. Funct. Mater. 30, 1906253 (2020).

A. T. Uzumcu, O. Guney, O. Okay, Nanocomposite DNA hydrogels with temperature sensitivity. Polymer 100, 169–178 (2016).

X. S. Hu, R. Liang, G. Sun, Super-adsorbent hydrogel for removal of methylene blue dye from aqueous solution. J. Mater. Chem. A 6, 17612–17624 (2018).

N. Gupta, Y. N. Liang, X. Hu, Thermally responsive ionic liquids and polymeric ionic liquids: Emerging trends and possibilities. Curr. Opin. Chem. Eng. 25, 43–50 (2019).

T. Tran, S. Pan, X. Chen, X. C. Lin, A. K. Blevins, Y. Ding, H. Lin, Zwitterionic hydrogel-impregnated membranes with polyamide skin achieving superior water/salt separation properties. ACS Appl. Mater. Interfaces 12, 49192–49199 (2020).

X. Tan, L. Ge, T. Zhang, Z. Lu, Preservation of DNA for data storage. Russ. Chem. Rev. 90, 280–291 (2021).

Y. Erlich, D. Zielinski, DNA Fountain enables a robust and efficient storage architecture. Science 355, 950–954 (2017).

K. Koch, S. Gantenbein, K. Masania, W. J. Stark, Y. Erlich, R. N. Grass, A DNA-of-things storage architecture to create materials with embedded memory. Nat. Biotechnol. 38, 39–43 (2020).

R. N. Grass, R. Heckel, M. Puddu, D. Paunescu, W. J. Stark, Robust chemical preservation of digital information on DNA in silica with error-correcting codes. Angew. Chem. Int. Ed. 54, 2552–2555 (2015).

L. C. Meiser, P. L. Antkowiak, J. Koch, W. D. Chen, A. X. Kohll, W. J. Stark, R. Heckel, R. N. Grass, Reading and writing digital data in DNA. Nat. Protoc. 15, 86–101 (2020).

S. Newman, A. P. Stephenson, M. Willsey, B. H. Nguyen, C. N. Takahashi, K. Strauss, L. Ceze, High density DNA data storage library via dehydration with digital microfluidic retrieval. Nat. Commun. 10, 1706 (2019).

Z. Fei, R. Wei, C. Cheng, P. Xiao, A novel approach to the bioluminescent detection of the SARS-CoV-2 ORF1ab gene by coupling isothermal RNA reverse transcription amplification with a digital PCR approach. Int. J. Mol. Sci. 22, 1017 (2021).

J. Bonnet, M. Colotte, D. Coudy, V. Couallier, J. Portier, B. Morin, S. Tuffet, Chain and conformation stability of solid-state DNA: Implications for room temperature storage. Nucleic Acids Res. 38, 1531–1546 (2010).

D. Pu, Y. Qi, L. Cui, P. Xiao, Z. Lu, A real-time decoding sequencing based on dual mononucleotide addition for cyclic synthesis. Anal. Chim. Acta 852, 274–283 (2014).

M. E. Allentoft, M. Collins, D. Harker, J. Haile, C. L. Oskam, M. L. Hale, P. F. Campos, J. A. Samaniego, M. T. P. Gilbert, E. Willerslev, G. Zhang, R. P. Scofield, R. N. Holdaway, M. Bunce, The half-life of DNA in bone: Measuring decay kinetics in 158 dated fossils. Proc. R. Soc. B Biol. Sci. 279, 4724–4733 (2012).