As the genetic material of life, DNA can be divided into linear and cyclic. Eukaryote genomic chromosomal DNA exists linearly, while organelles (mitochondria and chloroplasts), the vast majority of bacteria, and part viral genomic DNA exist in a circular form. Extrachromosomal circular DNA (eccDNA) refers to extrasomal, non-mitochondrial, non-chloroplast ring-like structure DNA found in eukaryotes. The discovery of eccDNA (1964)[1] predates the widely known mitochondrial DNA [2-4] and bacterial plasmid DNA [5,6], both of which were confirmed to be ring-shaped by microscopic observations in 1966 and 1967, respectively.
After more than half a century of research, there has been a consensus that eccDNA and genomic DNA sequences are highly homologous, but because of their huge differences in size: from hundreds of bps (base pairs) to hundreds of kbs (thousand base pairs), the researchers have proposed a variety of names, microDNA mainly refers to the ring DNA within 400 bp [7], ecDNA (extrachromosomal DNA) [8] is used to describe the discovery in cancer cells, Extrachromosomal DNA with a size of hundreds of kilobytes or more, its size is large enough to accommodate full-length genes and DNA replication initiation sites, so it can be copied and amplified autonomously, and is related to the amplification of oncogenes and the occurrence of cancer. EccDNA is being used to refer to all extrachromosomal cyclic DNA that is smaller than ecDNA, which is present in almost all eukaryotic cell lines and tissues and organs, and its abundance is extremely low and variable compared to chromosomal DNA. To date, there are still no answers to three basic questions about eccDNA: 1. Is the sequence homology of eccDNA and the genome specific, (from where)? 2, What is the production mechanism of eccDNA (how to come about)? 3, Does eccDNA have a biological function (what to do)?
On October 20, 2021, professor Yi Zhang's lab at Harvard Medical School and Boston Children's Hospital published a study titled eccDNAs are apoptotic products with high innate immunostimulatory activity in nature magazine in the form of a long article, answering the above three basic questions.
Professor Zhang Yi's laboratory has done a lot of outstanding work on the regulatory role of epigenetics in early embryonic development [9-15]. During early embryonic development, a large number of transposons and reversals in the genome will be activated. These activated transposons and reversal constituencies may be reinserted, destabilizing the genome. Therefore, how to avoid the active transposon and reverse the destruction of the genome by the possum is an important scientific question. One reasonable hypothesis is that post-activation transposons and reversal posspots end up in eccDNA mode, avoiding reinsertion to destroy the genome. To test this hypothesis, Professor Zhang Yi's lab began studying eccDNA five years ago.
Effective isolation and purification of research subjects is a prerequisite for research progress. eccDNA is extremely low in abundance in biological samples and its ring structure is extremely susceptible to damage, and breaks at any site on the double strand will make the ring linear, resulting in its loss. Effective isolation and purification has always been a major challenge in eccDNA research. The traditional eccDNA isolation and purification method is mainly divided into two steps: 1. alkaline lysis of complex coarse-extracted circular DNA; 2. Selective digestion of linear DNA by nucleic acid exonuclease. However, the authors found that traditional eccDNA purification methods are far from efficient in isolating cyclic DNA to satisfactory purity, and that 60-80% of the molecules in eccDNA samples remain linear, consistent with the results reported in nearly 10 years. The authors speculate that dna in natural biological samples may have a variety of modifications and complex structures, such as oxidation, cross-link, DNA replication fork, Cruciform, deoxyguanine tetracline (G-quadruplex) and so on. These complex structures most likely inhibit the action of nucleic acid exonucleases, making it impossible to rely solely on nucleic acid exonucleases to digest and obtain pure cyclic DNA. With this in mind, the researchers added a purification step to optimize the traditional alkaline crudely extracted cyclic DNA and linear DNA digested by the nucleic acid exonuclease: selectively recovering the cyclic DNA from the nucleic acid exonuclease digestion product. The new method enables more than 95% of DNA molecules to be ring-shaped when eccDNA samples are microscopically examined (Figure 1). At the same time, the new method also has the characteristics of short time and high stability: the traditional separation and purification process that takes one week is compressed to one day; the purification of eccDNA between parallel samples and batch samples shows good parallelism and high reproducibility.
To obtain accurate sequence information for each eccDNA molecular ring, the researchers also combined rolling cycle amplification with the third-generation Oxford Nanopore sequencing technology (Figure 1), enabling the full-length sequence of eccDNA to be obtained using only three generations of sequencing.
Figure 1: (a) New methods for purification and sequencing of cyclic DNA; (b) electrophoresis of eccDNA agarose gel; (c) atomic force microscopy of eccDNA molecules.
In order to avoid the influence of mutations, recombination, structural disorders and other factors that may exist in the genome of ordinary cell lines on the homologous comparison analysis of eccDNA and reference genome sequences, the authors chose to extract eccDNA from mouse embryonic stem cell lines with relatively stable genomes for research, and obtained the full-length sequence information and genome source location information of more than 1.6 million eccDNA molecules. The analysis found that eccDNA was randomly derived from genomic fragments, including single-fragment self-cyclization and multi-fragment junction cyclization, with a size concentration between 200 bp and 3 kb, and the distribution was similar to the size distribution law between oligo-nucleosomes. These results suggest that eccDNA may be a cyclization product of random fragments of genomic DNA.
Apoptosis is a programmed cell death that occurs widely in eukaryotes and in vitro cultured cells, and one of the important features is that during apoptosis, specific nucleases will cut off genomic DNA in the inter-nucleosome region to form a linear fragmented DNA (ladder-like) with a regular distribution of oligonucleosome size. Therefore, the authors compared the eccDNA content in normal cultured cells and cells that induced apoptosis, and found that the amount of eccDNA in apoptotic cells increased significantly, indicating that apoptosis can induce the occurrence of eccDNA.
In order to find out whether the production of eccDNA originated from fragmented DNA, the authors identified and knocked out the nuclease DNase1l3 responsible for the fragmentation of apoptotic DNA in mouse stem cells, and obtained a cell line with defects in apoptosis DNA fragmentation (DNase1l3 KO), the knockout of DNase1l3 did not affect the cell death caused by apoptosis, but the phenomenon of DNA fragmentation of its apoptosis oligonuclearoid body was completely lost, correspondingly, DNase1l3 KO cell lines also lost the ability of apoptosis-induced eccDNA to occur (Figure 2); this suggests that eccDNA occurrence stems from DNA fragments produced by apoptosis.
Linear DNA fragments are cyclized without DNA ligase (DNA ligase), and eukaryotes have three DNA ligase genes: Lig1, Lig3 and Lig4. Cell lines are knocked out by establishing dna ligases and multi-gene knockouts. The researchers found that the deletions of Lig1 and Lig4 did not hinder the production of eccDNA at all, while the yield of eccDNA in lig3-missing cell lines decreased dramatically, indicating that lig3 is the primary ligase responsible for DNA cyclization (Figure 2). This answers the question of how eccDNA happens.
Figure 2: (a-b) Apoptotic DNA fragmentation deletion cells (DNase1l3 KO) cannot produce eccDNA; (c) deletion of Lig3 inhibits eccDNA production.
Subsequently, the researchers challenged the third fundamental question of eccDNA: Does eccDNA function? Given that eccDNA does not exhibit distinct characteristics in sequence, the authors first ruled out the possibility of exploring the overall biological function of eccDNA based on sequence. By reviewing the literature, the authors found that important proteins in the innate immunity process, such as HMGB1, TLR9, etc., showed a higher affinity for the curvature (bending) structure of DNA, so it is speculated that the commonality of all eccDNA molecules - the cyclic structure may have a special role in the innate immune response.
Subsequent studies confirmed this hypothesis: the cyclic nature of eccDNA gives it a superior ability to stimulate the innate immune response. In typical immune cells, bone marrow derived dendritic cells (BMDC) and macrophages (BMDM), eccDNA induces higher cytokines (Figure 3) (type I interferon (IFN-α, β), interleukin 6 (TNF-α), compared to linear genomic DNA fragments of the same size, The expression of tumor necrosis factor (TNF-α, etc.) shows a super ability to stimulate the immune response. When the molecular ring of eccDNA is cut into the corresponding linear molecule by the unit point, the super ability to stimulate the immune response is lost (Figure 3), confirming that the strong immune stimulation ability of eccDNA depends on its cyclic structure. To further rule out eccDNA sequences and potential transcriptional effects on immunostimulation, the authors treated immune cells with synthetic cyclic DNA of similar size to eccDNA, but with randomly generated sequences, and unexpectedly, synthetic DNA loops perfectly reproduced the super stimulating ability of purified natural eccDNA to immune cells (Figure 3). In-depth studies have shown that the super innate immune stimulation ability of eccDNA relies on the intracellular Sting signaling pathway rather than the Myd88 signaling pathway.
Figure 3: (a) eccDNA has a super ability to induce the expression of interferon IFN-β (d) eccDNA loses the ability to induce interferon IFN-β expression after being linearized (Li-DNA); (c) synthetic random sequence DNA rings have a super ability to induce the expression of interferon IFN-β; (d) eccDNA stimulation of immunity depends on the Sting signaling pathway rather than the Myd88 signaling pathway.
In summary, the study gave clear answers in three most fundamental directions to the eccDNA puzzle that has lasted for more than half a century: 1, eccDNA is randomly derived from chromosomal genomic DNA and has no obvious location or sequence specificity (where it comes from), 2, eccDNA is a cyclization product of Lig3-mediated apoptosis DNA fragments (how to come), and 3, eccDNA has a super-strong ability to stimulate the innate immune response (what can be done). The study has a profound impact on the two major areas of apoptosis and innate immunity, especially for a variety of diseases that produce a large number of cell deaths in a short period of time, or disease treatment processes, such as bacteria, viruses, acute fatal infectious diseases, and immunotherapy for cancer. In addition, the study also provides new design ideas and paths for the design of nucleic acid vaccines and vaccine adjuvant.