Thank you for visiting Nature.com. The browser version you are using has limited CSS support. For the best experience, we recommend that you use an updated browser (or disable Compatibility Mode in Internet Explorer). In the meantime, to ensure continued support, we will render the site without styles and JavaScript.
Liquid biopsy (LB) is a concept that is rapidly gaining popularity in the biomedical field. The concept is mainly based on the detection of fragments of circulating extracellular DNA (ccfDNA), which are mainly released as small fragments after cell death in various tissues. A small proportion of these fragments originate from foreign (foreign) tissues or organisms. In current work, we have applied this concept to mussels, a sentinel species known for their high seawater filtration capacity. We use the ability of mussels to act as natural filters to capture environmental DNA fragments from a variety of sources to provide information about the biodiversity of marine coastal ecosystems. Our results show that mussel hemolymph contains DNA fragments that vary greatly in size, from 1 to 5 kb. Shotgun sequencing showed that a large number of DNA fragments are of foreign microbial origin. Among them, we found DNA fragments from bacteria, archaea, and viruses, including viruses known to infect a variety of hosts commonly found in coastal marine ecosystems. In conclusion, our study demonstrates that the concept of LB applied to mussels represents a rich but as yet unexplored source of knowledge about microbial diversity in marine coastal ecosystems.
The impact of climate change (CC) on the biodiversity of marine ecosystems is a rapidly growing area of research. Global warming not only causes important physiological stresses, but also pushes the evolutionary limits of the thermal stability of marine organisms, affecting the habitat of a number of species, prompting them to search for more favorable conditions [1, 2]. In addition to affecting the biodiversity of metazoans, CC disrupts the delicate balance of host-microbial interactions. This microbial dysbacteriosis poses a serious threat to marine ecosystems as it makes marine organisms more susceptible to infectious pathogens [3, 4]. It is believed that SS play an important role in mass deaths, which is a serious problem for the management of global marine ecosystems [5, 6]. This is an important issue given the economic, ecological and nutritional impacts of many marine species. This is especially true for bivalves living in the polar regions, where the effects of CK are more immediate and severe [6, 7]. In fact, bivalves such as Mytilus spp. are widely used to monitor the effects of CC on marine ecosystems. Not surprisingly, a relatively large number of biomarkers have been developed to monitor their health, often using a two-tier approach involving functional biomarkers based on enzymatic activity or cellular functions such as cell viability and phagocytic activity [8]. These methods also include the measurement of the concentration of specific pressure indicators that accumulate in soft tissues after the absorption of large amounts of sea water. However, the high filtration capacity and semi-open circulatory system of bivalves provide an opportunity to develop new hemolymph biomarkers using the concept of liquid biopsy (LB), a simple and minimally invasive approach to patient management. blood samples [9, 10]. Although several types of circulating molecules can be found in human LB, this concept is primarily based on DNA sequencing analysis of circulating extracellular DNA (ccfDNA) fragments in plasma. In fact, the presence of circulating DNA in human plasma has been known since the mid-20th century [11], but it is only in recent years that the advent of high-throughput sequencing methods has led to clinical diagnosis based on ccfDNA. The presence of these circulating DNA fragments is due in part to the passive release of genomic DNA (nuclear and mitochondrial) after cell death. In healthy individuals, the concentration of ccfDNA is normally low (<10 ng/mL) but can be increased by 5–10 times in patients suffering from various pathologies or subjected to stress, resulting in tissue damage. In healthy individuals, the concentration of ccfDNA is normally low (<10 ng/mL) but can be increased by 5–10 times in patients suffering from various pathologies or subjected to stress, resulting in tissue damage. У здоровых людей концентрация вккДНК в норме низкая (<10 нг/мл), но может повышаться в 5–10 раз у больных с различной патологией или подвергающихся стрессу, приводящему к повреждению тканей. In healthy people, the concentration of cccDNA is normally low (<10 ng/mL), but it can increase by 5–10 times in patients with various pathologies or under stress that leads to tissue damage.在健康个体中，ccfDNA 的浓度通常较低（<10 ng/mL），但在患有各种病理或承受压力的患者中可增加5-10 倍，从而导致组织损伤。在 健康 个体 中 ， ccfdna 的 浓度 较 低 （（<10 ng/ml） 但 在 各 种 病理 或 承受 压力 患者 中 可 增加 5-10 倍 ， 从而 组织。。。 损伤 损伤 损伤 损伤 损伤 损伤 损伤 损伤 损伤 损伤Концентрации ccfDNA обычно низкие (<10 нг/мл) у здоровых людей, но могут быть увеличены в 5-10 раз у пациентов с различными патологиями или стрессом, что приводит к повреждению тканей. ccfDNA concentrations are usually low (<10 ng/ml) in healthy individuals, but can be increased 5-10-fold in patients with various pathologies or stress, resulting in tissue damage. The size of ccfDNA fragments varies widely, but usually ranges from 150 to 200 bp. [12]. Analysis of self-derived ccfDNA, i.e., ccfDNA from normal or transformed host cells, can be used to detect genetic and epigenetic changes present in the nuclear and/or mitochondrial genome, thereby helping clinicians select specific molecular-targeted therapies [13] . However, ccfDNA can be obtained from foreign sources such as ccfDNA from fetal cells during pregnancy or from transplanted organs [14,15,16,17]. ccfDNA is also an important source of information for detecting the presence of nucleic acids of an infectious agent (foreign), which allows non-invasive detection of widespread infections not identified by blood cultures, avoiding invasive biopsy of infected tissue [18]. Recent studies have indeed shown that human blood contains a rich source of information that can be used to identify viral and bacterial pathogens, and that about 1% of the ccfDNA found in human plasma is of foreign origin [19]. These studies demonstrate that the biodiversity of an organism’s circulating microbiome can be assessed using ccfDNA analysis. However, until recently, this concept was used exclusively in humans and, to a lesser extent, in other vertebrates [20, 21].
In the present paper, we use the LB potential to analyze the ccfDNA of Aulacomya atra, a southern species commonly found in the subantarctic Kerguelen Islands, a group of islands atop a large plateau that formed 35 million years ago. volcanic eruption. Using an in vitro experimental system, we found that DNA fragments in seawater are quickly taken up by mussels and enter the hemolymph compartment. Shotgun sequencing has shown that mussel hemolymph ccfDNA contains DNA fragments of its own and non-self origin, including symbiotic bacteria and DNA fragments from biomes typical of cold volcanic marine coastal ecosystems. Hemolymph ccfDNA also contains viral sequences derived from viruses with different host ranges. We also found DNA fragments from multicellular animals such as bony fish, sea anemones, algae and insects. In conclusion, our study demonstrates that the LB concept can be successfully applied to marine invertebrates to generate a rich genomic repertoire in marine ecosystems.
Adults (55-70 mm long) Mytilus platensis (M. platensis) and Aulacomya atra (A. atra) were collected from the intertidal rocky shores of Port-au-France (049°21.235 S, 070°13.490 E .). Kerguelen Islands in December 2018. Other adult blue mussels (Mytilus spp.) were obtained from a commercial supplier (PEI Mussel King Inc., Prince Edward Island, Canada) and placed in a temperature controlled (4°C) aerated tank containing 10–20 L of 32‰ artificial brine. (artificial sea salt Reef Crystal, Instant Ocean, Virginia, USA). For each experiment, the length and weight of individual shells were measured.
A free open access protocol for this program is available online (https://doi.org/10.17504/protocols.io.81wgb6z9olpk/v1). Briefly, LB hemolymph was collected from abductor muscles as described [22]. The hemolymph was clarified by centrifugation at 1200×g for 3 minutes, the supernatant was frozen (-20°C) until use. For isolation and purification of cfDNA, samples (1.5-2.0 ml) were thawed and processed using the NucleoSnap cfDNA kit (Macherey-Nagel, Bethlehen, PA) according to the manufacturer’s instructions. ccfDNA was stored at -80°C until further analysis. In some experiments, ccfDNA was isolated and purified using the QIAamp DNA Investigator Kit (QIAGEN, Toronto, Ontario, Canada). Purified DNA was quantified using a standard PicoGreen assay. The fragment distribution of the isolated ccfDNA was analyzed by capillary electrophoresis using an Agilent 2100 bioanalyzer (Agilent Technologies Inc., Santa Clara, CA) using a High Sensitivity DNA Kit. The assay was performed using 1 µl of the ccfDNA sample according to the manufacturer’s instructions.
For sequencing of hemolymph ccfDNA fragments, Génome Québec (Montreal, Quebec, Canada) prepared shotgun libraries using the Illumina DNA Mix kit of the Illumina MiSeq PE75 kit. A standard adapter (BioO) was used. Raw data files are available from the NCBI Sequence Read Archive (SRR8924808 and SRR8924809). Basic reading quality was assessed using FastQC [23]. Trimmomatic [24] has been used for clipping adapters and poor quality reads. Shotgun reads with paired ends were FLASH merged into longer single reads with a minimum overlap of 20 bp to avoid mismatches [25]. Merged reads were annotated with BLASTN using a bivalve NCBI Taxonomy database (e value < 1e−3 and 90% homology), and masking of low-complexity sequences was performed using DUST [26]. Merged reads were annotated with BLASTN using a bivalve NCBI Taxonomy database (e value < 1e−3 and 90% homology), and masking of low-complexity sequences was performed using DUST [26]. Объединенные чтения были аннотированы с помощью BLASTN с использованием базы данных таксономии двустворчатых моллюсков NCBI (значение e < 1e-3 и 90% гомологии), а маскирование последовательностей низкой сложности было выполнено с использованием DUST [26]. Pooled reads were annotated with BLASTN using the NCBI bivalve taxonomy database (e value < 1e-3 and 90% homology), and low complexity sequence masking was performed using DUST [26].使用双壳类NCBI 分类数据库（e 值< 1e-3 和90% 同源性）用BLASTN 注释合并的读数，并使用DUST [26] 进行低复杂度序列的掩蔽。使用 双 壳类 ncbi 分类 （（（<1e-3 和 90% 同源） 用 用 用 注释 合并 读数 ， 并 使用 dust [26] 进行 复杂度 序列 的。。。。 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽 掩蔽Объединенные чтения были аннотированы с помощью BLASTN с использованием таксономической базы данных двустворчатых моллюсков NCBI (значение e <1e-3 и 90% гомологии), а маскирование последовательностей низкой сложности было выполнено с использованием DUST [26]. Pooled reads were annotated with BLASTN using the NCBI bivalve taxonomic database (e value <1e-3 and 90% homology), and low complexity sequence masking was performed using DUST [26]. Reads were divided into two groups: related to bivalve sequences (here called self-reads) and unrelated (non-self-reads). Two groups were separately assembled using MEGAHIT to generate contigs [27]. Meanwhile, the taxonomic distribution of alien microbiome reads was classified using Kraken2 [28] and graphically represented by a Krona pie chart on Galaxy [29, 30]. The optimal kmers was determined to be kmers-59 from our preliminary experiments. Self contigs were then identified by alignment with BLASTN (bivalve NCBI database, e value < 1e−10 and 60% homology) for a final annotation. Self contigs were then identified by alignment with BLASTN (bivalve NCBI database, e value < 1e−10 and 60% homology) for a final annotation. Затем собственные контиги были идентифицированы путем сопоставления с BLASTN (база данных двустворчатых моллюсков NCBI, значение e <1e-10 и гомология 60%) для окончательной аннотации. Self-contigs were then identified by matching against BLASTN (NCBI bivalve database, e value <1e-10 and 60% homology) for final annotation.然后通过与BLASTN（双壳贝类NCBI 数据库，e 值< 1e-10 和60% 同源性）对齐来识别自身重叠群以进行最终注释。然后通过与BLASTN（双壳贝类NCBI 数据库，e 值< 1e-10 和60% Затем были идентифицированы собственные контиги для окончательной аннотации путем сопоставления с BLASTN (база данных NCBI для двустворчатых моллюсков, значение e <1e-10 и гомология 60%). Self-contigs were then identified for final annotation by matching against BLASTN (NCBI bivalve database, e value <1e-10 and 60% homology). In parallel, nonself group contigs were annotated with BLASTN (nt NCBI database, e value < 1e−10 and 60% homology). In parallel, nonself group contigs were annotated with BLASTN (nt NCBI database, e value < 1e−10 and 60% homology). Параллельно чужеродные групповые контиги были аннотированы с помощью BLASTN (база данных nt NCBI, значение e <1e-10 и гомология 60%). In parallel, foreign group contigs were annotated with BLASTN (NT NCBI database, e value <1e-10 and 60% homology).平行地，用BLASTN（nt NCBI 数据库，e 值< 1e-10 和60% 同源性）注释非自身组重叠群。平行地，用BLASTN（nt NCBI 数据库，e 值< 1e-10 和60% 同源性）注释非自身组重叠群。 Параллельно контиги, не относящиеся к собственной группе, были аннотированы с помощью BLASTN (база данных nt NCBI, значение e <1e-10 и гомология 60%). In parallel, non-self group contigs were annotated with BLASTN (nt NCBI database, e value <1e-10 and 60% homology). BLASTX was also conducted on nonself contigs using the nr and RefSeq protein NCBI databases (e value < 1e−10 and 60% homology). BLASTX was also conducted on nonself contigs using the nr and RefSeq protein NCBI databases (e value < 1e−10 and 60% homology). BLASTX также был проведен на несамостоятельных контигах с использованием баз данных белка nr и RefSeq NCBI (значение e <1e-10 и гомология 60%). BLASTX was also performed on non-self contigs using the nr and RefSeq NCBI protein databases (e value < 1e-10 and 60% homology).还使用nr 和RefSeq 蛋白NCBI 数据库对非自身重叠群进行了BLASTX（e 值< 1e-10 和60% 同源性）。还使用nr 和RefSeq 蛋白NCBI 数据库对非自身重叠群进行了BLASTX（e 值< 1e-10 和60% 同源性）。 BLASTX также выполняли на несамостоятельных контигах с использованием баз данных белка nr и RefSeq NCBI (значение e <1e-10 и гомология 60%). BLASTX was also performed on non-self contigs using the nr and RefSeq NCBI protein databases (e value <1e-10 and 60% homology). The BLASTN and BLASTX pools of non-self-contigs represent the final contigs (see Supplementary file).
The primers used for PCR are listed in Table S1. Taq DNA polymerase (Bio Basic Canada, Markham, ON) was used to amplify the ccfDNA target genes. The following reaction conditions were used: denaturation at 95°C for 3 minutes, 95°C for 1 minute, set annealing temperature for 1 minute, elongation at 72°C for 1 minute, 35 cycles, and finally 72°C within 10 minutes. . PCR products were separated by electrophoresis in agarose gels (1.5%) containing SYBRTM Safe DNA Gel Stain (Invitrogen, Burlington, ON, Canada) at 95 V.
Mussels (Mytilus spp.) were acclimatized in 500 ml oxygenated seawater (32 PSU) for 24 hours at 4°C. Plasmid DNA containing an insert encoding the human galectin-7 cDNA sequence (NCBI accession number L07769) was added to the vial at a final concentration of 190 μg/μl. Mussels incubated under the same conditions without DNA addition were the control. The third control tank contained DNA without mussels. To monitor the quality of DNA in seawater, seawater samples (20 μl; three repetitions) were taken from each tank at the indicated time. For plasmid DNA traceability, LB mussels were harvested at the indicated times and analyzed by qPCR and ddPCR. Due to the high salt content of seawater, aliquots were diluted in PCR quality water (1:10) prior to all PCR assays.
Digital droplet PCR (ddPCR) was performed using the BioRad QX200 protocol (Mississauga, Ontario, Canada). Use the temperature profile to determine the optimum temperature (Table S1). Drops were generated using a QX200 drop generator (BioRad). ddPCR was carried out as follows: 95°C for 5 min, 50 cycles of 95°C for 30 s and a given annealing temperature for 1 min and 72°C for 30 s, 4°C for 5 min and 90°C within 5 minutes. The number of drops and positive reactions (number of copies/µl) were measured using a QX200 drop reader (BioRad). Samples with less than 10,000 droplets were rejected. Pattern control was not performed every time ddPCR was run.
qPCR was performed using Rotor-Gene® 3000 (Corbett Research, Sydney, Australia) and LGALS7 specific primers. All quantitative PCRs were performed in 20 µl using the QuantiFast SYBR Green PCR Kit (QIAGEN). qPCR was started with a 15 min incubation at 95°C followed by 40 cycles at 95°C for 10 seconds and at 60°C for 60 seconds with one data collection. Melting curves were generated using successive measurements at 95°C for 5 s, 65°C for 60 s, and 97°C at the end of the qPCR. Each qPCR was performed in triplicate, except for control samples.
Since mussels are known for their high filtration rate, we first investigated whether they could filter and retain DNA fragments present in seawater. We were also interested in whether these fragments accumulate in their semi-open lymphatic system. We resolved this issue experimentally by tracing the fate of soluble DNA fragments added to blue mussel tanks. To facilitate tracking of DNA fragments, we used foreign (not self) plasmid DNA containing the human galectin-7 gene. ddPCR traces plasmid DNA fragments in seawater and mussels. Our results show that if the amount of DNA fragments in sea water remained relatively constant over time (up to 7 days) in the absence of mussels, then in the presence of mussels this level almost completely disappeared within 8 hours (Fig. 1a,b). Fragments of exogenous DNA were easily detected within 15 min in intravalvular fluid and hemolymph (Fig. 1c). These fragments could still be detected up to 4 hours after exposure. This filtering activity with respect to DNA fragments is comparable to the filtering activity of bacteria and algae [31]. These results suggest that mussels can filter and accumulate foreign DNA in their fluid compartments.
Relative concentrations of plasmid DNA in seawater in the presence (A) or absence (B) of mussels, measured by ddPCR. In A, the results are expressed as percentages, with the borders of the boxes representing the 75th and 25th percentiles. The fitted logarithmic curve is shown in red, and the area shaded in gray represents the 95% confidence interval. In B, the red line represents the mean and the blue line represents the 95% confidence interval for the concentration. C Accumulation of plasmid DNA in the hemolymph and valvular fluid of mussels at different times after the addition of plasmid DNA. Results are presented as absolute copies detected/mL (±SE).
Next, we investigated the origin of ccfDNA in mussels collected from mussel beds on the Kerguelen Islands, a remote group of islands with limited anthropogenic influence. For this purpose, cccDNA from mussel hemolymphs was isolated and purified by methods commonly used to purify human cccDNA [32, 33]. We found that average hemolymph ccfDNA concentrations in mussels are in the low micrograms per ml hemolymph range (see Table S2, Supplementary Information). This range of concentrations is much larger than in healthy people (low nanograms per milliliter), but in rare cases, in cancer patients, the level of ccfDNA can reach several micrograms per milliliter [34, 35]. An analysis of the size distribution of hemolymph ccfDNA showed that these fragments vary greatly in size, ranging from 1000 bp to 1000 bp. up to 5000 b.p. (Fig. 2). Similar results were obtained using the silica-based QIAamp Investigator Kit, a method commonly used in forensic science to rapidly isolate and purify genomic DNA from low concentration DNA samples, including ccfDNA [36].
Representative ccfDNA electrophoregram of mussel hemolymph. Extracted with NucleoSnap Plasma Kit (top) and QIAamp DNA Investigator Kit. B Violin plot showing the distribution of hemolymph ccfDNA concentrations (±SE) in mussels. The black and red lines represent the median and the first and third quartiles, respectively.
Approximately 1% of ccfDNA in humans and primates has a foreign source [21, 37]. Given the semi-open circulatory system of bivalves, microbial-rich seawater, and the size distribution of mussel ccfDNA, we hypothesized that mussel hemolymph ccfDNA may contain a rich and diverse pool of microbial DNA. To test this hypothesis, we sequenced hemolymph ccfDNA from Aulacomya atra samples collected from the Kerguelen Islands, yielding over 10 million reads, 97.6% of which passed quality control. The readings were then classified according to self and non-self sources using the BLASTN and NCBI bivalve databases (Fig. S1, Supplementary Information).
In humans, both nuclear and mitochondrial DNA can be released into the bloodstream [38]. However, in the present study, it was not possible to describe in detail the nuclear genomic DNA of mussels, given that the A. atra genome has not been sequenced or described. However, we were able to identify a number of ccfDNA fragments of our own origin using the bivalve library (Fig. S2, Supplementary Information). We also confirmed the presence of DNA fragments of our own origin by directed PCR amplification of those A. atra genes that were sequenced (Fig. 3). Similarly, given that the mitochondrial genome of A. atra is available in public databases, one can find evidence for the presence of mitochondrial ccfDNA fragments in the hemolymph of A. atra. The presence of mitochondrial DNA fragments was confirmed by PCR amplification (Fig. 3).
Various mitochondrial genes were present in the hemolymph of A. atra (red dots – stock number: SRX5705969) and M. platensis (blue dots – stock number: SRX5705968) amplified by PCR. Figure adapted from Breton et al., 2011 B Amplification of hemolymph supernatant from A. atra Stored on FTA paper. Use a 3 mm punch to add directly to the PCR tube containing the PCR mix.
Given the abundant microbial content in seawater, we initially focused on the characterization of microbial DNA sequences in hemolymph. To do this, we use two different strategies. The first strategy used Kraken2, an algorithm-based sequence classification program that can identify microbial sequences with an accuracy comparable to BLAST and other tools [28]. More than 6719 reads were determined to be of bacterial origin, while 124 and 64 were from archaea and viruses, respectively (Fig. 4). The most abundant bacterial DNA fragments were Firmicutes (46%), Proteobacteria (27%), and Bacteroidetes (17%) (Fig. 4a). This distribution is consistent with previous studies of the marine blue mussel microbiome [39, 40]. Gammaproteobacteria were the main class of Proteobacteria (44%), including many Vibrionales (Fig. 4b). The ddPCR method confirmed the presence of Vibrio DNA fragments in the ccfDNA of A. atra hemolymph (Fig. 4c) [41]. To obtain more information about the bacterial origin of ccfDNA, an additional approach was taken (Fig. S2, Supplementary Information). In this case, reads that overlapped were assembled as paired-end reads and were classified as of self (bivalves) or nonself origin using BLASTN and an e value of 1e−3 and a cutoff with >90% homology. In this case, reads that overlapped were assembled as paired-end reads and were classified as of self (bivalves) or nonself origin using BLASTN and an e value of 1e−3 and a cutoff with >90% homology. В этом случае перекрывающиеся чтения были собраны как чтения с парными концами и были классифицированы как собственные (двустворчатые моллюски) или чужие по происхождению с использованием BLASTN и значения e 1e-3 и отсечения с гомологией> 90%. In this case, overlapping reads were collected as paired-ended reads and were classified as native (bivalve) or non-original using BLASTN and e value of 1e-3 and cutoff with >90% homology.在这种情况下，重叠的读数组装为配对末端读数，并使用BLASTN 和1e-3 的e 值和>90% 同源性的截止值分类为自身（双壳类）或非自身来源。在 这 种 情况 下 ， 重叠 读数 组装 为 配 末端 读数 ， 使用 使用 使用 blastn 和 1e-3 的 的 值 和> 90% 同源性 的 分类 自身 （双 壳类） 非 自身。。。。。。。。。 В этом случае перекрывающиеся чтения были собраны как чтения с парными концами и классифицированы как собственные (двустворчатые моллюски) или несобственные по происхождению с использованием значений e BLASTN и 1e-3 и порога гомологии> 90%. In this case, overlapping reads were collected as paired-ended reads and classified as own (bivalves) or non-original using e BLASTN and 1e-3 values and a homology threshold >90%. Since the A. atra genome has not yet been sequenced, we used the de novo assembly strategy of the MEGAHIT Next Generation Sequencing (NGS) assembler. A total of 147,188 contigs have been identified as dependent (bivalves) of origin. These contigs were then exploded with e-values of 1e-10 using BLASTN and BLASTX. This strategy allowed us to identify 482 non-bivalve fragments present in A. atra ccfDNA. More than half (57%) of these DNA fragments were obtained from bacteria, mainly from gill symbionts, including sulfotrophic symbionts, and from gill symbionts Solemya velum (Fig. 5).
Relative abundance at the type level. B Microbial diversity of two main phyla (Firmicutes and Proteobacteria). Representative amplification of ddPCR C Vibrio spp. A. Fragments of the 16S rRNA gene (blue) in three atra hemolymphs.
A total of 482 collected contigs were analyzed. General profile of the taxonomic distribution of metagenomic contig annotations (prokaryotes and eukaryotes). B Detailed distribution of bacterial DNA fragments identified by BLASTN and BLASTX.
Kraken2 analysis also showed that mussel ccfDNA contained archaeal DNA fragments, including DNA fragments of Euryarchaeota (65%), Crenarchaeota (24%), and Thaurmarcheota (11%) (Fig. 6a). The presence of DNA fragments derived from Euryarchaeota and Crenarchaeota, previously found in the microbial community of Californian mussels, should not come as a surprise [42]. Although Euryarchaeota is often associated with extreme conditions, it is now recognized that both Euryarchaeota and Crenarcheota are among the most common prokaryotes in the marine cryogenic environment [43, 44]. The presence of methanogenic microorganisms in mussels is not surprising, given recent reports of extensive methane leaks from bottom leaks on the Kerguelen Plateau [45] and possible microbial methane production observed off the coast of the Kerguelen Islands [46].
Our attention then shifted to readings from DNA viruses. To the best of our knowledge, this is the first off-target study of the virus content of mussels. As expected, we found DNA fragments of bacteriophages (Caudovirales) (Fig. 6b). However, the most common viral DNA comes from a phylum of nucleocytoviruses, also known as the nuclear cytoplasmic large DNA virus (NCLDV), which has the largest genome of any virus. Within this phylum, most DNA sequences belong to the families Mimimidoviridae (58%) and Poxviridae (21%), whose natural hosts include vertebrates and arthropods, while a small proportion of these DNA sequences belong to known virological algae. Infects marine eukaryotic algae. The sequences were also obtained from the Pandora virus, the giant virus with the largest genome size of any known viral genera. Interestingly, the range of hosts known to be infected with the virus, as determined by hemolymph ccfDNA sequencing, was relatively large (Figure S3, Supplementary Information). It includes viruses that infect insects such as Baculoviridae and Iridoviridae, as well as viruses that infect amoeba, algae and vertebrates. We also found sequences matching the Pithovirus sibericum genome. Pitoviruses (also known as “zombie viruses”) were first isolated from 30,000 year old permafrost in Siberia [47]. Thus, our results are consistent with previous reports showing that not all modern species of these viruses are extinct [48] and that these viruses may be present in remote subarctic marine ecosystems.
Finally, we tested to see if we could find DNA fragments from other multicellular animals. A total of 482 foreign contigs were identified by BLASTN and BLASTX with nt, nr and RefSeq libraries (genomic and protein). Our results show that among the foreign fragments of ccfDNA of multicellular animals DNA of bony bones predominates (Fig. 5). DNA fragments from insects and other species have also been found. A relatively large part of the DNA fragments has not been identified, possibly due to the underrepresentation of a large number of marine species in genomic databases compared to terrestrial species [49].
In the present paper, we apply the LB concept to mussels, arguing that hemolymph ccfDNA shot sequencing can provide insight into the composition of marine coastal ecosystems. In particular, we found that 1) mussel hemolymph contains relatively high concentrations (microgram levels) of relatively large (~1-5 kb) circulating DNA fragments; 2) these DNA fragments are both independent and non-independent 3) Among the foreign sources of these DNA fragments, we found bacterial, archaeal and viral DNA, as well as DNA of other multicellular animals; 4) The accumulation of these foreign ccfDNA fragments in the hemolymph occurs rapidly and contributes to the internal filtering activity of mussels. In conclusion, our study demonstrates that the concept of LB, which has so far been applied mainly in the field of biomedicine, encodes a rich but unexplored source of knowledge that can be used to better understand the interaction between sentinel species and their environment.
In addition to primates, ccfDNA isolation has been reported in mammals, including mice, dogs, cats, and horses [50, 51, 52]. However, to our knowledge, our study is the first to report the detection and sequencing of ccfDNA in marine species with an open circulation system. This anatomical feature and filtering ability of mussels may, at least in part, explain the different size characteristics of circulating DNA fragments compared to other species. In humans, most DNA fragments circulating in the blood are small fragments ranging in size from 150 to 200 bp. with a maximum peak of 167 b.p. [34, 53]. A small but significant portion of DNA fragments are between 300 and 500 bp in size, and about 5% are longer than 900 bp. [54]. The reason for this size distribution is that the main source of ccfDNA in plasma occurs as a result of cell death, either due to cell death or due to necrosis of circulating hematopoietic cells in healthy individuals or due to apoptosis of tumor cells in cancer patients ( known as circulating tumor DNA). , ctDNA). The size distribution of hemolymph ccfDNA that we found in mussels ranged from 1000 to 5000 bp, suggesting that mussel ccfDNA has a different origin. This is a logical hypothesis, since mussels have a semi-open vascular system and live in marine aquatic environments containing high concentrations of microbial genomic DNA. In fact, our laboratory experiments using exogenous DNA have shown that mussels accumulate DNA fragments in seawater, at least after a few hours they are degraded after cellular uptake and/or released and/or stored in various organizations. Given the rarity of cells (both prokaryotic and eukaryotic), the use of intravalvular compartments will reduce the amount of ccfDNA from self-sources as well as from foreign sources. Considering the importance of bivalve innate immunity and the large number of circulating phagocytes, we further hypothesized that even foreign ccfDNA is enriched in circulating phagocytes that accumulate foreign DNA upon ingestion of microorganisms and/or cellular debris. Taken together, our results show that bivalve hemolymph ccfDNA is a unique repository of molecular information and reinforces their status as a sentinel species.
Our data indicate that sequencing and analysis of bacterial-derived hemolymph ccfDNA fragments can provide key information about the host bacterial flora and the bacteria present in the surrounding marine ecosystem. Shot sequencing techniques have revealed sequences of the commensal bacteria A. atra gill that would have been missed if conventional 16S rRNA identification methods had been used, due in part to a reference library bias. In fact, our use of LB data collected from M. platensis in the same mussel layer at Kerguelen showed that the composition of gill-associated bacterial symbionts was the same for both mussel species (Fig. S4, Supplementary Information). This similarity of two genetically different mussels may reflect the composition of bacterial communities in the cold, sulfurous, and volcanic deposits of Kerguelen [55, 56, 57, 58]. Higher levels of sulfur-reducing microorganisms have been well described when harvesting mussels from bioturbated coastal areas [59], such as the coast of Port-au-France. Another possibility is that the commensal mussel flora may be affected by horizontal transmission [60, 61]. More research is needed to determine the correlation between the marine environment, seafloor surface, and the composition of symbiotic bacteria in mussels. These studies are currently ongoing.
The length and concentration of hemolymph ccfDNA, its ease of purification, and high quality to allow rapid shotgun sequencing are some of the many advantages of using mussel ccfDNA to assess biodiversity in marine coastal ecosystems. This approach is especially effective for characterizing viral communities (viromes) in a given ecosystem [62, 63]. Unlike bacteria, archaea, and eukaryotes, viral genomes do not contain phylogenetically conserved genes such as 16S sequences. Our results indicate that liquid biopsies from indicator species such as mussels can be used to identify relatively large numbers of ccfDNA virus fragments known to infect hosts that typically inhabit coastal marine ecosystems. This includes viruses known to infect protozoa, arthropods, insects, plants, and bacterial viruses (eg, bacteriophages). A similar distribution was found when we examined the hemolymph ccfDNA virome of blue mussels (M. platensis) collected in the same mussel layer at Kerguelen (Table S2, Supplementary Information). Shotgun sequencing of ccfDNA is indeed a new approach gaining momentum in the study of the virome of humans or other species [21, 37, 64]. This approach is particularly useful for studying double-stranded DNA viruses, since no single gene is conserved among all double-stranded DNA viruses, representing the most diverse and broad class of viruses in Baltimore [65]. Although most of these viruses remain unclassified and may include viruses from a completely unknown part of the viral world [66], we found that the viromes and host ranges of the mussels A. atra and M. platensis fall between the two species. similarly (see figure S3, additional information). This similarity is not surprising, as it may reflect a lack of selectivity in uptake of DNA present in the environment. Future studies using purified RNA are currently needed to characterize the RNA virome.
In our study, we used a very rigorous pipeline adapted from the work of Kowarski and colleagues [37], who used a two-step deletion of pooled reads and contigs before and after assembly of native ccfDNA, resulting in a high proportion of unmapped reads. Therefore, we cannot rule out that some of these unmapped reads may still have their own origin, primarily because we do not have a reference genome for this mussel species. We also used this pipeline because we were concerned about the chimeras between self and non-self reads and the read lengths generated by the Illumina MiSeq PE75. Another reason for the majority of uncharted readings is that much of the marine microbes, especially in remote areas such as Kerguelen, have not been annotated. We used Illumina MiSeq PE75, assuming ccfDNA fragment lengths similar to human ccfDNA. For future studies, given our results showing that hemolymph ccfDNA has longer reads than humans and/or mammals, we recommend using a sequencing platform more suitable for longer ccfDNA fragments. This practice will make it much easier to identify more indications for deeper analysis. Obtaining the currently unavailable complete A. atra nuclear genome sequence would also greatly facilitate the discrimination of ccfDNA from self and non-self sources. Given that our research has focused on the possibility of applying the concept of liquid biopsy to mussels, we hope that as this concept is used in future research, new tools and pipelines will be developed to increase the potential of this method to study the microbial diversity of mussels. marine ecosystem.
As a non-invasive clinical biomarker, elevated human plasma levels of ccfDNA are associated with various diseases, tissue damage, and stress conditions [67,68,69]. This increase is associated with the release of DNA fragments of its own origin after tissue damage. We addressed this issue using acute heat stress, in which mussels were briefly exposed to a temperature of 30 °C. We performed this analysis on three different types of mussels in three independent experiments. However, we did not find any change in ccfDNA levels after acute heat stress (see Figure S5, additional information). This discovery may explain, at least in part, the fact that mussels have a semi-open circulatory system and accumulate large amounts of foreign DNA due to their high filtering activity. On the other hand, mussels, like many invertebrates, may be more resistant to stress-induced tissue damage, thereby limiting the release of ccfDNA in their hemolymph [70, 71].
To date, DNA analysis of biodiversity in aquatic ecosystems has mainly focused on environmental DNA (eDNA) metabarcoding. However, this method is usually limited in biodiversity analysis when primers are used. The use of shotgun sequencing circumvents the limitations of PCR and the biased selection of primer sets. Thus, in a sense, our method is closer to the recently used high-throughput eDNA Shotgun sequencing method, which is able to directly sequence fragmented DNA and analyze almost all organisms [72, 73]. However, there are a number of fundamental issues that distinguish LB from standard eDNA methods. Of course, the main difference between eDNA and LB is the use of natural filter hosts. The use of marine species such as sponges and bivalves (Dresseina spp.) as a natural filter for studying eDNA has been reported [74, 75]. However, Dreissena’s study used tissue biopsies from which DNA was extracted. Analysis of ccfDNA from LB does not require tissue biopsy, specialized and sometimes expensive equipment and logistics associated with eDNA or tissue biopsy. In fact, we recently reported that ccfDNA from LB can be stored and analyzed with FTA support without maintaining a cold chain, which is a major challenge for research in remote areas [76]. The extraction of ccfDNA from liquid biopsies is also simple and provides high quality DNA for shotgun sequencing and PCR analysis. This is a great advantage given some of the technical limitations associated with eDNA analysis [77]. The simplicity and low cost of the sampling method is also particularly suitable for long-term monitoring programs. In addition to their high filtering ability, another well-known feature of bivalves is the chemical mucopolysaccharide composition of their mucus, which promotes the absorption of viruses [78, 79]. This makes bivalves an ideal natural filter for characterizing biodiversity and the impact of climate change in a given aquatic ecosystem. Although the presence of host-derived DNA fragments can be seen as a limitation of the method compared to eDNA, the cost associated with having such a native ccfDNA compared to eDNA is simultaneously understandable for the vast amount of information available for health studies. offset host. This includes the presence of viral sequences integrated into the genome of the host host. This is especially important for mussels, given the presence of horizontally transmitted leukemic retroviruses in bivalves [80, 81]. Another advantage of LB over eDNA is that it exploits the phagocytic activity of circulating blood cells in the hemolymph, which engulfs microorganisms (and their genomes). Phagocytosis is the main function of blood cells in bivalves [82]. Finally, the method takes advantage of the high filtering capacity of mussels (average 1.5 l/h of seawater) and two-day circulation, which increase the mixing of different layers of seawater, allowing the capture of heterologous eDNA. [83, 84]. Thus, mussel ccfDNA analysis is an interesting avenue given the nutritional, economic, and environmental impacts of mussels. Similar to the analysis of LB collected from humans, this method also opens up the possibility of measuring genetic and epigenetic changes in host DNA in response to exogenous substances. For example, third-generation sequencing technologies can be envisaged to perform genome-wide methylation analysis in native ccfDNA using nanopore sequencing. This process should be facilitated by the fact that the length of the mussel ccfDNA fragments is ideally compatible with long-read sequencing platforms that allow genome-wide DNA methylation analysis from a single sequencing run without the need for chemical transformations.85,86] This is an interesting possibility, as it has been shown that DNA methylation patterns reflect a response to environmental stress and persist over many generations. Therefore, it can provide valuable insight into the underlying mechanisms governing response after exposure to climate change or pollutants [87]. However, the use of LB is not without limitations. Needless to say, this requires the presence of indicator species in the ecosystem. As mentioned above, using LB to assess the biodiversity of a given ecosystem also requires a rigorous bioinformatics pipeline that takes into account the presence of DNA fragments from the source. Another major problem is the availability of reference genomes for marine species. It is hoped that initiatives such as the Marine Mammal Genomes Project and the recently established Fish10k project [88] will facilitate such analysis in the future. The application of the LB concept to marine filter-feeding organisms is also compatible with the latest advances in sequencing technology, making it well suited for the development of multi-ohm biomarkers to provide important information about the health of marine habitats in response to environmental stress.
Genome sequencing data has been deposited in the NCBI Sequence Read Archive https://www.ncbi.nlm.nih.gov/sra/SRR8924808 under Bioprojects SRR8924808.
Brierley A.S., Kingsford M.J. Impact of climate change on marine life and ecosystems. Cole Biology. 2009; 19: P602–P614.
Gissi E, Manea E, Mazaris AD, Fraschetti S, Almpanidou V, Bevilacqua S, et al. Consider the combined impacts of climate change and other local stressors on the marine environment. general scientific environment. 2021;755:142564.
Carella F, Antuofermo E, Farina S, Salati F, Mandas D, Prado P, et al. ). Science of the first of March. 2020;7:48.
Seront L, Nicastro CR, Zardi GI, Goberville E. Reduced heat tolerance under repetitive heat stress conditions explains the high summer mortality of blue mussels. Scientific report 2019; 9:17498.
Fey SB, Siepielski AM, Nussle S, Cervantes-Yoshida K, Hwan JL, Huber ER, et al. Recent changes in the frequency, causes and extent of animal deaths. Proc Natl Acad Sci USA. 2015;112:1083-8.
Scarpa F, Sanna D, Azzena I, Mughetti D, Cerruti F, Hosseini S, et al. Multiple non-species-specific pathogens may have caused mass mortality of Pinna nobilis. Life. 2020;10:238.
Bradley M, Coutts SJ, Jenkins E, O’Hara TM. Potential impact of climate change on Arctic zoonotic diseases. Int J Circumpolar health. 2005; 64:468–77.
Beyer J., Greene N.W., Brooks S., Allan I.J., Ruus A., Gomez T. et al. Blue mussels (Mytilus edulis spp.) as signal organisms in coastal pollution monitoring: a review. Mar Environ Res 2017; 130:338-65.
Siravegna G, Marsoni S, Siena S, Bardelli A. Integration of liquid biopsy in cancer treatment. Nat Rev Clean Oncol. 2017; 14:531–48.
Wan JCM, Massie C, Garcia-Corbacho J, Mouliere F, Brenton JD, Caldas C, et al. Liquid biopsy maturation: Allows tumor DNA to circulate. Nat Rev Cancer. 2017;17:223–38.
Mandel P., Metais P. Nucleic acids in human plasma. Meeting minutes of Soc Biol subsidiaries. 1948; 142:241-3.
Bronkhorst AJ, Ungerer W, Holdenrieder S. A new role for cell-free DNA as a molecular marker for cancer treatment. Quantification of biomolar analysis. 2019;17:100087.
Ignatiadis M., Sledge G.W., Jeffrey S.S. Liquid biopsy enters the clinic – implementation issues and future challenges. Nat Rev Clin Oncol. 2021; 18:297–312.
Lo Y.M., Corbetta N., Chamberlain P.F., Rai W., Sargent I.L., Redman C.W. and others. Fetal DNA is present in maternal plasma and serum. Lancet. 1997; 350:485-7.
Mufarray M.N., Wong R.J., Shaw G.M., Stevenson D.K., Quake S.R. Study of the course of pregnancy and its complications using circulating extracellular RNA in the blood of women during pregnancy. Dopediatrics. 2020;8:605219.
Ollerich M, Sherwood K, Keown P, Schütz E, Beck J, Stegbauer J, et al. Liquid biopsy: donor cell-free DNA is used to detect allogeneic lesions in a kidney graft. Nat Rev Nephrol. 2021; 17:591–603.
Juan F.C., Lo Y.M. Innovations in prenatal diagnostics: maternal plasma genome sequencing. Anna MD. 2016;67:419-32.
Gu W, Deng X, Lee M, Sucu YD, Arevalo S, Stryke D, et al. Rapid pathogen detection with next-generation metagenomic sequencing of infected bodily fluids. Nat Medicine. 2021;27:115-24.