Magnetotactic bacteria (MTB) represent a group of prokaryotes that can migrate along geomagnetic field lines (Blakemore, 1975; Bellini, 2009). MTB are diverse in their morphology, phylogeny, and physiology. There are single-cell forms (cocci, ovoid, rod, curved rod, and spirillum) and multicellular forms that are known as the multicellular magnetotactic prokaryotes (MMPs). The MMPs can be divided into two types based on their shape: spherical MMPs (sMMPs) and ellipsoidal MMPs (eMMPs) (Bazylinski et al., 2013; Amor et al., 2020). MTB exhibits great taxonomic diversity. Most are phylogenetically affiliated with phyla of Proteobacteria, Desulfobacterota (previously known as Deltaproteobacteria), Nitrospirae, Planctomycetes, the candidate phylum Omnitrophica, and the candidate phylum Latescibacteria (Bazylinski et al., 2013; Lin et al., 2018). Recent studies of reconstructed metagenome-assembled MTB genomes have expanded our knowledge of the taxonomy of MTB, several of which belonged to the phyla Nitrospinota, UBA10199, Bdellovibrionota, Bdellovibrionata_B, Fibrobacterota, Riflebacteria phyla, Elusimicrobia, and Candidatus Hydrogenedentes (Lin et al., 2020; Uzun et al., 2020).

The MTB can use intracellular biomineralization to synthesize special organelles called magnetosomes, which help them find and remain near suitably chemically stratified water columns or sediments. Magnetosomes are composed of magnetite or/and greigite (Zhang et al., 2014) and exhibit different shapes (e.g., elongated-prismatic, bullet/tooth-shaped, and cuboctahedron) (Bazylinski et al., 2013). There is a strong correlation between the morphology of biogenetic magnetosomes and MTB phylogeny (Li et al., 2020b). Alphaproteobacteria MTB always synthesizes cuboctahedral or elongated-prismatic magnetite magnetosomes, while MTB with bullet-shaped magnetite crystals is associated with the Desulfobacterota, Nitrospirae, and the candidate phylum Omnitrophica (Lin et al., 2014; Amor et al., 2020). Desulfobacterota MTB can also biomineralize greigite crystals of diverse morphologies (Lefèvre et al., 2011; Zhang et al., 2014). Inside the cell, magnetosomes are arranged as single chains, double chains, or multiple chains, and disorderly arrangements have also been reported (Amor et al., 2020). Magnetosome formation in MTB is a biomineralization process that is strictly controlled by conserved genes found in magnetosome gene clusters (MGCs) (Lin et al., 2020). Generally, there are four steps in the formation of magnetosomes, which arrange in chains: cytoplasmic membrane invagination forms vesicles, proteins are targeted to the vesicle (magnetosome) membrane, iron is transported to vesicles (membranous invagination of magnetosomes) and mineralized into magnetite crystals, and the crystals are assembled into chains of magnetosomes (Uebe and Schüler, 2016). Studies on pure cultures of the MTB, like Magnetospirillum gryphiswaldense MSR-1, revealed that the MGC genes are found in five operons: the mms6 operon, the mamGFDC operon, the mamAB operon, the mamXY operon, and the feoAB1 operon (Uebe and Schüler, 2016). Magnetosome genes that compose MGCs of different MTB species are not identical, as reflected in species-level differences in magnetosome type. However, almost all MGCs contain the mamAB operon (Lefèvre and Wu, 2013). This operon contains the core gene, mamABEKMOPQI, which is thought to play an important role in magnetosome formation (Lefèvre and Wu, 2013; Lin et al., 2018). After the death of a magnetotactic bacterium, the magnetosomes are released; over time, they accumulate in sediments, forming fossil magnetosomes (magnetofossils) (Lin et al., 2014). The magnetic mineral ultrastructure, morphology, composition, size, and other characteristics of magnetofossils can be observed and effectively distinguished using a transmission electron microscope (TEM) (Li et al., 2020a). The study of magnetofossils can provide paleoecological and paleoenvironmental information (Hesse, 1994).

The MTB is widespread in sediments and water at the oxic-anoxic interface (OAI, previously called the OATZ) of freshwater, brackish, marine, and hypersaline environments (Bazylinski et al., 2013). The abundance of marine MTB is usually higher in the intertidal zone, where most reports indicate that magnetotactic cocci are relatively common and the dominant species of MTB (Lin et al., 2012; Abreu et al., 2016). Most of these studied environments have near-neutral pH, moderate temperatures, and are concentrated in the Northern hemisphere. However, there are some reports of MTB in special habitats, including mangrove swamps, coral reefs, seamounts, deep seas, etc. (Torres de Araujo et al., 1986; Dong et al., 2016; Liu et al., 2017; Teng et al., 2018). Vibrio and rod-shaped MTB, but not cocci, have been found in hemipelagic sediments of the Santa Barbara Basin (Stolz et al., 1986). Other studies indicated that the MTB found in extreme environments often belonged to Desulfobacterota and Nitrospirae (Abreu et al., 2007; Nash, 2008; Lefèvre et al., 2010). In addition, biogenic magnetite, with elongated octahedral and prismatic morphologies in ferromanganese nodules, has also been reported (Hassan et al., 2020).

In deep-sea hydrothermal fields, which are rich in metal sulfides, microorganisms mainly obtain energy from reductive sulfur oxidation (McCollom and Shock, 1997; Amend et al., 2011; Meier et al., 2019), and chemolithoautotrophic microorganisms represent the only primary producers (Jannasch and Mottl, 1985). Simultaneously, most cultured MTB can grow chemolithoautotrophically using reductive sulfides. Desulfobacterota MTB are sulfate-reducing anaerobes that grow only chemoorganoheterotrophically (Bazylinski et al., 2013). MTB might dwell in and adapt to hydrothermal environments. To assess this possibility, we combined 16S rRNA gene amplicon and metagenomic data and TEM observations to study the presence, diversity, and characteristics of MTB in the Tangyin hydrothermal field of Okinawa Trough.

Since MTB was first discovered independently by Bellini and Blakemore (Blakemore, 1975; Bellini, 2009), the phylogenetic information reported for them has been based primarily on 16S rRNA gene sequence analysis. As it has proven to be difficult to cultivate MTB, the 16S rRNA gene analysis has been used to infer the diversity and phylogenetic affiliations of MTB in various environments (Lin et al., 2018; Teng et al., 2018; Amor et al., 2020). Tan et al. (2021) analyzed MTB communities in tropical marine environments using the 16S rRNA genes from the metagenomic analysis. MTB at seamounts was also identified by 16S rRNA gene analysis (Liu et al., 2017). Dong et al. (2016) used the 16S rRNA gene and magnetofossil analyses to infer the existence of MTB in deep-sea surface sediments. Most mature biogenic magnetosomes are single-domain and arranged in chains in most MTB cells. According to these two basic characteristics, combined with the size, crystallographic structure, and composition of the magnetosomes, TEM can be used to identify magnetofossils effectively (Kopp and Kirschvink, 2008; Li et al., 2020a). Previous studies found that biogenic magnetite was distributed widely in deep-sea sediments (Petersen et al., 1986; Yamazaki et al., 2019; He and Pan, 2020; Usui and Yamazaki, 2021), and the proportions of bullet-shaped magnetofossils increased in relatively reductive and less oxic environments, while isotropic magnetofossils dominated in relatively oxic environments (Hesse, 1994; Yamazaki and Kawahata, 1998; Yamazaki and Shimono, 2013; He and Pan, 2020; Lu et al., 2021). Recently, the known extent of MTB diversity has undergone a significant expansion due to the identification of magnetosome gene cluster (MGC)-containing genomes and studies screening for sequences homologous to known MGC genes (Lin et al., 2020; Uzun et al., 2020). Lin et al. (2020) performed a large-scale reconstruction of metagenome-assembled MTB genomes from diverse ecosystems, and 13 bacterial phyla were detected, six of which were not previously known, including MTB. Thus, the existing literature indicates that analyses of 16S rRNA gene sequences, magnetofossil diversity, and magnetosome gene-homologous sequences obtained from an environment can be used to predict the existence of MTB diversity.

In our study, 16S rRNA gene analysis showed that MTB affiliated with Desulfobacterota, Alphaproteobacteria, and Nitrospirae had been found. FORC diagrams have no central ridge and the coercivity obtained from FORC is smaller than 20 mT, which is different from common FORCs of magnetofossils (Jovane et al., 2012). Vali and Kirschvink (1989) found partial magnetofossil dissolved in deep-sea sediment. Particle corrosion and clumping probably decreased the coercivity and increased the level of interparticle interaction, which is shown in our FORC diagram (Figure 3A). We also observed that partial magnetofossils experienced moderate corrosion in our sample (Supplementary Figure S1). Moreover, the diversity of magnetofossil was also affected by early diagenesis (Rodelli et al., 2019; Yamazaki, 2020). Yamazaki (2020) concluded that bullet-shaped magnetofossils were dissolved easier than prismatic ones, which may explain that the proportions of bullet-shaped magnetofossils are only 24.8% in our sample (Figure 3D). According to the diversity of magnetofossils observed from TEM images, the ancient MTB living in the Tangyin hydrothermal field of the Okinawa Trough might belong to Alphaproteobacteria, Etaproteobacteria, Gammaproteobacteria, Desulfobacterota, Nitrospirae, and the candidate phylum Omnitrophica (Figure 4) (Amor et al., 2020; Liu et al., 2021a,b). Moreover, we had detected some homologous sequences of magnetosome genes, and these homologous magnetosome genes did not form a complete magnetosome gene cluster. However, these homologous sequences have the same domain as the known magnetosome genes. This indicates they may have the same functions, and these homologous sequences could be magnetosome genes. Then, the phylogenetic trees of MamE, O, and Q protein sequences showed that these homologous sequences are evolutionarily distinct from known MTB (Figure 6). The homologous sequences of MamE and MamO representative sequences were homologous to Desulfobacterota. These branches were far from known MTB in the phylogenetic tree. Whereas, compared with the known MTB, most of these homologous Mam protein sequences have conserved regions (Supplementary Figure S2). Therefore, the MamE, O, and Q in hydrothermal have different evolutionary statuses and show high diversity. Of course, in consideration of these homologous sequences having big differences evolutionarily from known MTB, the homologous sequences of MamE and mamO are not necessarily Desulfobacterota either; they might be originated from other MTB identified by 16S rRNA gene sequences shown in Figure 2 (Alphaproteobacteria or Nitrospirae). Meanwhile, the possibility that they are paralogous genes or pseudogenes from non-MTB cannot be ruled out. Nevertheless, combining the 16S rRNA gene results with the magnetofossils, we are more inclined that these are magnetosome genes from MTB. Here, our analysis of MTB-related 16S rRNA gene sequences, magnetofossils produced by MTB, and homologous sequences of magnetosome genes indicate that MTB could inhabit the deep-sea hydrothermal sediments.

Our analysis revealed that 65.3% of the 16S rRNA gene sequence reads associated with MTB belonged to OTU1477, which is affiliated with the MMP branch. All MMP-related 16S rRNA gene sequence reads accounted for 96.2% of all MTB-related reads. Moreover, the homologous sequences of magnetosome genes are most similar to known MTB belonging to Desulfobacterota, and MMPs accounted for 44.2% of all magnetosome gene homology sequences. The proportions of bullet-shaped magnetic particles are higher than those of MTB magnetosomes from the intertidal zone of Huiquan Bay in Qingdao (unpublished data) (24.8% vs. 4.0%). Bullet-shaped magnetosomes have only been found in the MTB belonging to Desulfobacterota, Nitrospirae, and the candidate phylum Omnitrophica (Kolinko et al., 2012; Chen et al., 2015; Li et al., 2015; Qian et al., 2019). Interestingly, both rock magnetic results and TEM analysis indicate that the abundance of BH components (bullet and elongated prisms) is ~1.7 times higher than that of BS components (cuboctahedron and short prisms) (Figures 3B,D). All the homologous sequences of mam genes identified herein correspond to the core genes of magnetosome synthesis. Given this, 16S rRNA gene sequences, magnetofossils, and homologous sequences of magnetosome genes consistently indicate that MMPs might be the dominant species among the MTB in this region. The hydrothermal field is an anoxic environment and is rich in reducing sulfides and sulfates (Orcutt et al., 2011). Previous research found that sulfate-reducing bacteria were the dominant species of the microbial community in the Tangyin hydrothermal field (Wang et al., 2018). Among the MTB, MMPs are anaerobic sulfate-reducing bacteria (Kolinko et al., 2014). Therefore, MMPs may represent the dominant group of MTBs at the sampled location, likely reflecting their adaptation to the hydrothermal environment.

As we all know, the hydrothermal zone is a candidate place for the origin of life (Trolard et al., 2022). Meanwhile, although the origins of MTB remain unclear, previous research has shown that the MTB of Nitrospirae and Proteobacteria differentiated near the Archean, and this suggests that MTB has existed at least in the Archean Eon (Lin et al., 2017). The geochemical conditions in the hydrothermal zone are thought to be similar to those on the early Earth (Baross and Hoffman, 1985; Trolard et al., 2022). The evolution of MTB here may be different from known MTB, which has been implied by the phylogenetic tree of 16S rRNA and homolog sequences of magnetosome genes. The dominant MTB in the hydrothermal field may represent a particular MTB species associated with the early Earth-like environment. There are microorganisms of ancient origin in hydrothermal vents (Takai and Nakamura, 2011), and MMPs are also an important model for studying evolution in prokaryotes (Keim et al., 2007). If MMPs are present in the hydrothermal field, the MMPs may be a potential model microorganism for understanding the early evolution of life on Earth.

In this study, a total of 709 MTB-related 16S rRNA gene sequence reads were found in the Tangyin hydrothermal field. The 20 related OTUs represent Desulfobacterota, Alphaproteobacteria, and Nitrospirae. MMPs represented the largest number total of 121 homologous magnetosome gene sequences were annotated. The results collectively suggest that MTB exists in the Tangyin hydrothermal field and MMPs might be the dominant MTB in this region.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/genbank/, OM108105-OM108124 and https://www.ncbi.nlm.nih.gov/genbank/, PRJNA514953.

TX, WYZ, and WCZ designed the research. MY and X-HZ collected the sample and performed the metagenomic sequencing. WYZ and MY carried out a metagenomic analysis. SC carried out rock magnetic measurement and TEM experiments. SC and KH carried out the magnetism and TEM data analysis. SC, WYZ, KH, TX, and WCZ carried out data and statistical analysis. SC, TX, WYZ, WCZ, KH, and L-FW prepared the manuscript. All authors participated in the discussion of the results. All authors contributed to the article and approved the submitted version.

This study was supported financially by the National Natural Science Foundation of China (U1706208 and 41976137).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.