DNA storage in thermoresponsive microcapsules for repeated random multiplexed data access

The world's data generation far surpasses current storage capacity. Traditional media like hard disks and magnetic tapes suffer from limited durability and density. Therefore, there's increasing interest in alternative storage media, including DNA, due to its high information density and longevity. DNA's potential for archival digital data storage is significant, with coding schemes achieving densities exceeding magnetic and optical alternatives by orders of magnitude. Long-term DNA storage has been demonstrated in both natural fossils and synthetic protective shells. Large-scale DNA synthesis and next-generation sequencing (NGS) technologies have made DNA-based storage viable. PCR-based random access offers selective data retrieval from complex DNA pools, but suffers from drawbacks including consumption of the DNA pool during amplification and PCR bias, which skews sequence representation. Multiplexed retrieval is hampered by PCR bias and molecular crosstalk. While careful sequence design and redundancy can mitigate these issues, they increase synthesis and sequencing costs. Current strategies rely on physically separating reactions or using emulsion PCR, but these methods are not scalable or cost-effective for large-scale applications. This study introduces a new methodology, thermoconfined PCR, to address these challenges.

Existing DNA data storage methods face limitations in achieving scalable parallel random access. While advancements in coding schemes and synthesis/sequencing technologies have improved data density and throughput, challenges remain in efficiently retrieving specific data from a pool of DNA files. Emulsion PCR has been used to mitigate PCR bias and crosstalk, but it is complex, uses large quantities of organic solvents, and the resulting emulsion is not reusable. Other techniques, such as physical separation of reactions, also lack scalability.

This research utilizes thermoresponsive, semipermeable microcapsules (proteinosomes) for DNA storage and retrieval. Biotinylated DNA files are encapsulated within individual proteinosomes, which are composed of bovine serum albumin (BSA) and poly(N-isopropylacrylamide) (PNIPAm). The PNIPAm component's thermoresponsive properties control membrane permeability: at low temperatures, the membrane is permeable to enzymes, primers, and amplicons; at high temperatures (PCR temperatures), the membrane collapses, preventing molecular crosstalk. The heat-stable streptavidin analogue Tamavidin 2-HOT is used to ensure stable retention of biotinylated DNA within the proteinosomes even at high temperatures. The temperature-dependent permeability was characterized using confocal fluorescence microscopy. Multiplex PCR experiments were conducted to evaluate the reduction in chimera formation (a significant source of error in conventional multiplex PCR) using proteinosomes compared to bulk reactions. Quantitative PCR (qPCR) was used to assess the amplification efficiency and file representation. Next-generation sequencing (Illumina sequencing) was employed to determine the average coverage per file. The study further investigated repeated access of DNA-encoded files using multiple rounds of PCR, comparing proteinosomes, emulsion PCR, and bulk PCR approaches. Finally, fluorescent barcoding (using FITC, DyLight 405, Cy3, and Cy5) of the proteinosomes and fluorescence-activated cell sorting (FACS) were used to demonstrate selective retrieval of DNA files based on metadata. Lyophilization of DNA-containing proteinosomes was also tested to assess the long-term stability.

The study demonstrated the stable localization of biotinylated DNA within Tamavidin 2-HOT proteinosomes at high PCR temperatures. The temperature-dependent permeability of proteinosomes was confirmed, showing significantly reduced permeability at high temperatures. Thermoconfined PCR using proteinosomes significantly reduced chimera formation in multiplex PCR reactions compared to bulk PCR. The multiplexed amplification of 25 DNA files (1.5 million unique sequences) showed a more even sequence representation with proteinosome-localized reactions compared to bulk amplification. Repeated read operations were successfully demonstrated using magnetic bead-containing proteinosomes. Fluorescence-based barcoding of proteinosomes was successfully implemented using membrane labels and short, fluorescently labeled DNA strands; these barcodes enabled selective retrieval of files via FACS. Repeated access experiments showed that the proteinosome-based approach yielded the lowest sequence dropout compared to bulk and emulsion PCR, suggesting improved data retention. Lyophilization with trehalose showed no adverse effects on DNA integrity or proteinosome structure.

The results demonstrate the effectiveness of thermoconfined PCR within proteinosomes for high-fidelity multiplexed and repeated random access to DNA-encoded data. The significant reduction in PCR bias and chimera formation compared to bulk PCR and the successful implementation of repeated access highlight the advantages of this method. The use of fluorescent barcodes and FACS enables a further level of organization and retrieval. The method’s scalability and sequence-agnostic nature make it a promising alternative to emulsion PCR for large-scale DNA data storage. The ability to lyophilize the proteinosomes offers a path towards long-term archival storage.

This study presents a significant advancement in DNA data storage by introducing thermoconfined PCR using thermoresponsive microcapsules. This method addresses the limitations of existing approaches by reducing PCR bias, enabling repeated access to original data files, and facilitating selective retrieval via fluorescent barcoding. While the initial encapsulation process is time-consuming, the simplified and efficient repeated access more than compensates, making this a promising technique for future large-scale DNA data storage. Future research will focus on long-term stability studies and exploring alternative retrieval methods to improve efficiency and single-compartment retrieval.

While the study demonstrates significant improvements over existing methods, limitations exist. The FACS-based retrieval, although effective, results in some loss of encapsulated files and does not allow for single-compartment retrieval. Furthermore, the current implementation requires amplification from multiple proteinosomes for data retrieval, which affects the data density. Future work should explore methods to improve retrieval efficiency and increase the data density per proteinosome.