Phylogenetic Revisit to a Review on Predatory Bacteria

Predatory bacteria, along with the biology of their predatory behavior, have attracted interest in terms of their ecological significance and industrial applications, a trend that has been even more pronounced since the comprehensive review in 2016. This mini-review does not cover research trends, such as the role of outer membrane vesicles in myxobacterial predation, but provides an overview of the classification and newly described taxa of predatory bacteria since 2016, particularly with regard to phylogenetic aspects. Among them, it is noteworthy that in 2020 there was a major phylogenetic reorganization that the taxa hosting Bdellovibrio and Myxococcus, formerly classified as Deltaproteobacteria, were proposed as the new phyla Bdellovibrionota and Myxococcota, respectively. Predatory bacteria have been reported from other phyla, especially from the candidate divisions. Predatory bacteria that prey on cyanobacteria and predatory cyanobacteria that prey on Chlorella have also been found. These are also covered in this mini-review, and trans-phylum phylogenetic trees are presented.


Introduction
As comprehensively reviewed by Pérez et al. (2016) [1], predatory bacteria are a group of prokaryotes that can actively hunt and consume other bacteria as their food source. By doing so, they can alter the abundance and diversity of the prey bacteria and thus influence the overall structure of the microbial community. In addition to predatory bacteria, protists and bacteriophages can also have significant impacts on the biomass, structure, and function of microbial communities, though their impacts differ in size, prey specificity, and hunting tactics [2]. Among their interwoven interactions, this mini-review, as an update of Pérez et al. (2016) [1], focuses on predatory bacteria with reference to phylogenetic aspects, particularly after the proposal in 2020 of the new phyla Bdellovibrionota and Myxococcota, which show distinct hunting strategies of predation [3]. Pérez et al. (2016) [1] reviewed the hunting strategies of predators of the order Bdellovibrionales, which physically attach to prey cells with flagella-based motility and penetrate into the periplasm of the prey cells, and the order Myxococcales, which are known for a "group attack" with gliding motility, the secretion of lytic enzymes, and the release of antibiotics. Pérez et al. (2016) [1] also reviewed the genomes, transcriptomes, and comparative genomics of predators, including the idea of the "predatome", i.e., the protein families in phenotypes of predatory bacteria [4]. Through detailed analysis of the predation-related proteins and the encoding genes, predatory properties are predicted for the clades whose predations are not yet known in the phyla Bdellovibrionota and Myxococcota [3]. Moreover, detailed analyses on the correlation between antibiotics biosynthesis and predation indicate that myxobacteria may be prioritized for the discovery of unexplored natural products [5][6][7].
After Pérez et al. (2016) [1], ecological significances and industrial applications of predatory bacteria have been increasingly studied. For example, a study on the potential use of predatory bacteria as alternatives to antibiotics showed that intrarectal inoculations Trans-phylum phylogenetic analysis of predatory bacteria was conducted in some of the PubMed-searched publications in and after 2016. In metagenomics of the intertidal soils along the Peruvian coastline, about 0.5% of 16S rRNA gene sequences are ascribed to predatory bacteria of multiple phyla, and 30 antibiotic-producing strains are kept in cultures with Escherichia coli XL1 Blue or Pheobacter inhibens DSM17395 as prey [13]. In the V4-V5 microbiomics of the Chesapeake Bay sediment, amplicons are affiliated with the predatory genus Haliangium and the class Polyangia (phylum Myxococcota), the genuslevel clade of the phylum Bdellovibrionota, and the Bradymonadales clade of the phylum Desulfobacterota, along with the abundant amplicons affiliated with Ca. Electrothrix of the phylum Desulfobacteriota [14]. A metatranscriptomic study combined with a set of curated genomes revealed increased and decreased expressions of proteases by the members of the phyla Bacteroidota-Actinobacteriota and Myxococcota, respectively [15].

Phylogenetics on the Members of BALOs, or the Phylum Bdellovibrionota, in and after 2016
Most of the PubMed-searched publications on predatory bacterial phylogenetics deal with selected taxa such as BALOs in the phylum Bdellovibrionota, for example, with a co-occurrence microbial network, particularly with Myxococcales of the phylum Myxococcota [16]. Two major BALO clades, i.e., the Bacteriovorax and Bdellovibrio clades, are targeted to decipher their roles and interplays in the microbiomes of wastewater treatment plants [17]. Comparative genomics of ten BALO genomes and one genome of a marine Bacteoidota predator (Saprospira grandis) allowed the genome-mining of 18 putative "predatomes" [18]. Comparative genomics of 152 BALO genomes from databases and 5 metagenome-assembled genomes (MAGs) from the Mariana Trench deep-sea water suggested that the chitinase-possessing members of Oligoflexia, the third class of the phylum Bdellovibrionota, may also act as predatory bacteria [19].
Phylogenetic aspects have also been viewed from the prey side. A wide prey range of a marine, prey-generalist BALO, Halobacteriovorax marinus, is discussed in evolutionary reference to horizontal gene transfer [20], although other Halobacteriovorax isolates (no phylogenetic information available) are rather prey-specific to Vibrio parahaemolyticus, a seafood-associated pathogen, and not to other seafood-associated pathogens, such as Vibrio vulnificus, Vibrio alginolyticus, Escherichia coli O157:H7, and Salmonella enterica serovar Typhimurium DT104 [21]. Bdellovibrio bacteriovorus is also known as prey-specific to Gram-negative bacteria by burrowing through the outer membrane and peptidoglycan cell wall and entering the periplasmic space of the prey cells [22].
Prey-specificity or prey-generality may be correlated to the alpha diversity of the BALO-associated microbiomes as suggested by a correlation study based on nine microbiomes from animals and environments [23], although this correlation is currently a correlation, not strong enough to be a prediction. Among the BALOs, alpha and beta diversities of the Halobacteriovorax clades in an estuary microbiome may be driven by inputs of pollutants, such as dissolved inorganic phosphorus and NH 4 + -N, as implicated by an analysis of the 676F-1193R region of the 16S rRNA gene sequence [24], though they are probably indirectly driven via bottom-up trophic control [9].
Detailed genotypic and phenotypic comparisons have also been performed for a soil isolate Bdellovibrio sp. NC01 and the type strain Bdellovibrio bacteriovorus HD100, indicating that reduced predatory activities of the isolate NC01 are probably due to the absence of ten genes in its genome [25].

Phylogenetics on the Members of the Predatory Phylum Myxococcota in and after 2016
A microbiomic study at a wastewater treatment plant using 13 C-labeled Escherichia coli ESS5 and Pseudomonas putida ESE1 as preys discovered the dominance of Haliangium and the mle1-27 clade of the phylum Myxococcota, contrary to the conventional Bdellovibrionotadominant view, which is further confirmed by analyzing the global datasets [26] from the Global Water Microbiome Consortium [27]. A large-scale study on agricultural soil microbiomes revealed that, out of all 6151 bacterial operational taxonomic units (OTUs), 242 are the myxobacterial OTUs affiliated with the phylum Myxococcota and that the alpha and beta diversities of the myxobacterial communities are more sensitive to geography (location and climate) than fertilization [28].

Candidate Divisons
Co-cultured with the methanogenic archaeon, Methanosaeta, which is probably the globally prevalent methane producer [29], an epibiotic bacterial strain OP3 LiM that preys on the archaeon Methanosaeta has been characterized by genomics and proteomics as well as fluorescence and electron microscopy. The strain OP3 represents the candidate genus and a species, "Ca. Velamenicoccus archaeovorus" gen. nov., sp. nov., affiliated with the phylum "Ca. Omnitrophica" (candidate division OP3) [30].
Another one is the ultra-small epibiont to the gammaproteobacterial photoautotroph Chromatium minus, inhabiting the karstic lakes in northeastern Spain, that was tentatively named Vampirococcus in 1986 [36]. In the same article, a Gram-negative, facultatively anaerobic epsilonproteobacterium, Daptobacter, was also mentioned but has been less studied and unvalidated. In 2013, Chromatium (Halochromatium)-like bacteria and associated ultrasmall (550 nm × 220 nm) epibionts, which morphologically matched to the Vampirococcus reported in 1986, were sampled in a hypersaline lake in northeastern Spain and subject to a two-member consortium mini-metagenomic analysis, yielding a genome similar to the genomes affiliated with the superphylum "Ca. Patescibacteria" (the candidate phyla radiation, CPR) or "Ca. Absconditabacteria" (formerly SR1). Based on the putative Vampirococcus genome, "Ca. Vampirococcus lugosii" has been proposed [37].

Phyla Bdelovibrionota and Myxococcota
The 16S rRNA gene sequences of a soil bacterium strain LBG001 and the type strain Bdellovibrio bacteriovorus HD100 share a 97% similarity, which does not discriminate based on a traditional 97% threshold [38] but does discriminate based on an updated view [39]. The genomic features of LBG001, i.e., its average nucleotide identity, average amino acid identity, and digital DNA-DNA hybridization values with other Bdellovibrio members, are as low as <79%, <72% and <17%, respectively, enough to be discriminated, and thus LBG001 is described as the new species Bdellovibrio reynosensis [40], whose specific epithet indicates its place of origin, Reynosa, a Mexican city on the southern bank of the Rio Grande.
One of the five new species of the genus Myxococcus described in 2020 is also named after its place of origin, i.e., a settlement on the island of Anglesey in North Wales, UK. The settlement's name appears in the specific epithet as Myxococcus llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogochensis, whose 16S rRNA gene sequence is not explicitly available from its shotgun-sequenced genome [41].
Two new predatory strains isolated from a freshwater pond, Bacteriovorax stolpii HI3 (phylum Bdelovibrionota) and The term "myxobacteria" (or "slime bacteria") has been ambiguously defined but may be regarded as the members of the orders Myxococcales and Polyangia in the phylum Myxococcota [43]. Omics studies, including "predatomics" of myxobacteria based on 163 publicly available genomes and 24 newly added genomes, were reviewed in 2021 [44]. More extensive review on myxobacteria with >400 references has recently been published [43].

Phylum Planctomycetota
The phylum Planctomycetota, along with Verrucomicrobiota and Chlamydiota, composes the superphylum PVC, whose member is hypothesized to be an ancestor of the protoeukaryotic cell via symbiogenesis with an archaeon [45,46]. Within the phylum Planctomycetota, a bacterium that exhibits eukaryote-like phagocytosis predation was discovered from the seawater of Palau. The bacterium is described as "Ca. Uab amorphum", whose genus is named after a giant of Palauan mythology, indicating its size as giant (4.5-7.8 µm × 2.8-5.5 µm), and it is capable of engulfment or phagocytosis [47].

Order Bradymonadales in the Phylum Desulfobacteriota
In addition to the two major taxa of predatory bacteria, i.e., the phyla Bdelovibrionota and Myxococcota, the order Bradymonadales may represent a third type of predation. The order Bradymonadales was formerly ascribed to the class Deltaproteobacteria of the phylum Proteobacteria but is currently affiliated with the phylum Desulfobacteriota [3]. The bacteria of this order, or Bradymonabacteria, show a unique predation type that is different from the so-called "obligate" and "facultative" as defined in Pérez et al. (2016), [1] which are characterized as "completely prey-dependent" and "prey-independent" [48]. The predation type of Bradymonabacteria is characterized as "facultatively prey-dependent" and termed lately as "facultative" [48], and the previous term "facultative" from Pérez et al. (2016) [1] was newly replaced with "opportunistic".
Within the order Bradymonadales, a new genus and new species has been proposed for the wide-ranging predatory strain YN101 from the sediment of a marine solar saltern, which is described as Persicimonas caeni gen. nov., sp. nov. in the family Bradymonadaceae [49].
Another new genus has also been proposed during the descriptions of two new species, Lujinxingia litoralis gen. nov. sp. nov. and Lujinxingia sediminis sp. nov., in the family Bradymonadaceae [50]. The genus name Lujinxingia is named after a Chinese microbiologist, Jin-Xing Lu, or Lu Jin-Xian in the Chinese name order.
A detailed comparison between Lujinxingia sediminis and a related strain, V1718, from a Chinese island, has led to the proposal of two new families, in addition to the family Bradymonadaceae, within the order Bradymonadales. During the description of the strain V1718 as Microvenator marinus gen. nov., sp. nov., two new families, Microvenatoraceae fam. nov. and Lujinxingiaceae fam. nov., were proposed. The genus Lujinxingia has been reclassified to the latter [51].

Order Herpetosiphonales in the Phylum Chloroflexota
Predatory species of the phylum Chloroflexota are known in the genus Herpetosiphon established in 1968 during the description of the filamentous gliding bacterium, Herpetosiphon aurantiacus gen. et sp. n. (=gen. nov., sp. nov.) [52]. One of the original strains used for the description was isolated from the slimy coating of Chara sp., a charophyte green alga, growing in Birch Lake, Minnesota, which implies predation on the slime-forming bacteria. The genus Herpetosiphon is placed in the family Herpetosiphonaceae within the order Herpetosiphonales, whose detailed descriptions were given by Gupta et al. (2013) [53].
An environmental strain, Hp g472 (DSM 52871), was isolated from a sandy soil in the beach of Poel island, Germany, and described as Herpetosiphon gulosus sp. nov., whose specific epithet derives from its gluttony, i.e., strong predatory activity [54].
Another environmental strain, CA052B, represents the fifth species of the genus Herpetosiphonaceae, described as Herpetosiphon llansteffanense sp. nov., whose specific epithet derives from the name of the origin, i.e., soil from the edge of a stream near the village of Llansteffan in Wales, UK [55]. This species demonstrates efficient predatory activity against a diverse array of prey microbial species using a "wolf pack" strategy that is possibly mediated by the secreted outer membrane vesicles containing a variety of hydrolytic enzymes [56].

Family Chitinophagaceae in the Phylum Bacteroidota
Biological soil crusts, or biocrusts, are regarded as promising mitigants for arid and semiarid lands. Filamentous cyanobacteria such as Microcoleus spp. play major roles in biocrust formation by stabilizing soil particles [57] and performing nitrogen fixation as well as carbon fixation via photosynthesis [58]. Biocrusts are applied to the restoration of damaged or disturbed lands [59]. The inoculation of "nursery grown" biocrusts have been attempted; however, inoculated biocrusts are often devastated by pathogen-like agents, leaving "plaques" [60]. From the diseased biocrusts, a cyanobacteria predator was enriched with Microcoleus vaginatus (PCC 9802), characterized by microscopy and physicalbiochemical tests, genome-sequenced, and described as a new species, Ca. Cyanoraptor togatus. sp. nova., in the family Chitinophagaceae within the phylum Bacteroidota [61]. Bethany et al. (2022) [61] estimate that predation by Cyanoraptor reduces photosynthetic production of cyanobacteria by >10%.

Class Melainabacteria in the Phylum Cyanobacteria
Cyanobacteria can be predators, although they are preyed by predatory bacteria, such as Ca. Cyanoraptor togatus. However, the preys are not bacteria but eukaryotic microalgae such as Chlorella. The known predatory cyanobacterium is Vampirovibrio chlorellavorus [62] in the class Melainabacteria [63]. The members of the class Melanibacteria are affiliated with the phylum Cyanobacteria and characterized by nonphotosynthetic metabolisms and dark habitats, such as the human gut and groundwater [64]. Affiliation of the predatory Vampirovibrio chlorellavorus with the class Melanibacteria agrees with the nonphotosynthetic nature of this taxon. The genome of Vampirovibrio chlorellavorus was reconstructed in 2009 from the old co-culture with Chlorella vulgaris deposited in 1978 and dealt with in an academic article in 2015 [65]. There have been established strategies to defend commercially produced Chlorella from predation by Vampirovibrio chlorellavorus [66].

Phylogenetic Tree of Predatory Bacteria
A total of 136 sequences of predatory bacterial 16S rRNA genes were collected from 12 phyla (including candidate phyla) of Actinobacteriota, Bacteroidota, Bdellovibrionota, Chloroflexota, Cyanobacteria, Desulfobacteriota, Myxococcota, Ca. Omnitrophica (OP3), Ca. Patescibac-teria (CPR) or Ca. Absconditabacteria (SR1), Planctomycetota, Pseudomonadota, and Ca. Saccharibacteria (TM7). The available 16S rRNA gene sequence of "Ca. Vampirococcus lugosii" (accession number MW286273, 1071 bp) [37] was the shortest among the collected sequences, and the phylogenetic trees with and without "Ca. V. lugosii" were constructed, along with the 35 reference sequences from current and former bacterial phyla. The sequences were aligned online with MEGA11 (https://www.megasoftware.net/; accessed on 20 May 2023) [67], and the phylogenetic trees based on the maximum likelihood method were drawn online with iTOL v6 (https://itol.embl.de/; accessed on 20 May 2023) [68] ( Figure 1 and Figure S1). Figure 1 shows the phylogenetic tree based on the sequences of about 1.6 kb after alignment excluding the shortest 1071 bp sequence of "Ca. V. lugosii". The 1.6 kb length, instead of the generally cited 1.5 kb, resulted from the alignment of 135 (136 minus 1, Ca. V. lugosii) full-length and near-full-length sequences that contain "gaps". Figure S1 displays the tree based on sequences of about 0.6 kb including "Ca.  Table S1 along with the hunting strategies of the corresponding predators. The reference 16S rRNA gene sequences from 35 representative, current, and former bacterial phyla  are listed in Table S2.
Another notable irregularity is "CP075895 Chitinophagaceae bacterium" at 9:10 in Figure 1 and 9:05 in Figure S1. This bacterium, Ca. Cyanoraptor togatus LGM1 [61], is the only known obligatory predator within the phylum Bacteroidota and is the only known endobiotic invader outside the Bdellovibrionota (orange)-Pseudomonadota (pale pink) cluster from a little before 3:00 to about 5:25. It is also the first obligatory, intracellular predator of cyanobacteria.
Only some of the predatory species of the phyla Bdelovibrionota and Myxococcota are shown in Figures 1 and S1, which would have exhibited more significant proportions of the phyla if all the predatory species were included. However, the importance of the phyla in the phylogeny of predatory bacteria is already explicit in the current Figures 1 and S1 with only selected species.   A myxobacterial species, Sorangium cellulosum, is not a typical epibiotic or endobiotic predator but a cellulolytic consumer as suggested by its genome [109], and its strong "lytic" activity may be regarded as predatory [155]. No such "lytic predators" are included in Figures 1 and S1.

Future Perspective
According to the mitochondrial endosymbiotic theory, the origin of mitochondria is regarded as the endosymbiosis of an ancestral alphaproteobacterial linage [156]. A popular idea about the mechanisms of mitochondria acquisition is the engulfment or phagocytosis of proto-mitochondria by proto-eukaryotes [157]. Asgard archaea are regarded as a paraphyletic lineage of proto-eukaryotes [158]. However, Asgard archaea are genomically predicted to be nonphagocytotic [159], although a cultured Asgard archaeon "Ca. Prometheoarchaeum syntrophicum" is hypothesized to entangle-engulf-endogenize aerobic bacteria as metabolic partners like proto-mitochondria [160]. In any case, the endogenized proto-mitochondria have to stay "undigested", and a new idea arose from the study of predatory bacteria. That is, the proto-mitochondrion was not undigested food but an "attenuated predator" [79,161,162]. More detailed studies on the known alphaproteobacterial predators, such as Ensifer adhaerens [111,112,163], as well as the finding of novel alphaproteobacterial predators, will shed more light on the relationship between the origins of mitochondria and predatory bacteria.  Table S1 [4,18,20,31,37,42,47,49,50,61,, except [37] and 35 sequences from representative, current, and former bacterial phyla listed in Table S2  . Predation properties are indicated by the symbols as follows: •, obligate, epibiotic; •, obligate, endobiotic or direct invasion; , opportunistic, epibiotic; , opportunistic, group attack; , facultative, epibiotic; and, , facultative, unknown.
A myxobacterial species, Sorangium cellulosum, is not a typical epibiotic or endobiotic predator but a cellulolytic consumer as suggested by its genome [109], and its strong "lytic" activity may be regarded as predatory [155]. No such "lytic predators" are included in Figures 1 and S1.

Future Perspective
According to the mitochondrial endosymbiotic theory, the origin of mitochondria is regarded as the endosymbiosis of an ancestral alphaproteobacterial linage [156]. A popular idea about the mechanisms of mitochondria acquisition is the engulfment or phagocytosis of proto-mitochondria by proto-eukaryotes [157]. Asgard archaea are regarded as a paraphyletic lineage of proto-eukaryotes [158]. However, Asgard archaea are genomically predicted to be nonphagocytotic [159], although a cultured Asgard archaeon "Ca. Prometheoarchaeum syntrophicum" is hypothesized to entangle-engulf-endogenize aerobic bacteria as metabolic partners like proto-mitochondria [160]. In any case, the endogenized proto-mitochondria have to stay "undigested", and a new idea arose from the study of predatory bacteria. That is, the proto-mitochondrion was not undigested food but an "attenuated predator" [79,161,162]. More detailed studies on the known alphaproteobacterial predators, such as Ensifer adhaerens [111,112,163], as well as the finding of novel alphapro-teobacterial predators, will shed more light on the relationship between the origins of mitochondria and predatory bacteria.
Cell sizes of predatory bacteria would be another perspective. Predatory bacteria of the candidate phylum radiation (CPR) represent ultra-small predators. Flat and stacked cells of 500-600 nm diameters and 200-250 nm heights are observed with the epibiotic predator or "sucker", Ca. Vampirococcus lugosii (phylum Ca. Absconditabacteria or candidate division SR1), of photosynthetic anoxic bacteria [37]. Coccoid cells of 200-300 nm diameters are seen for the facultative epibiotic predator Ca. Velamenicoccus archeaovorus (phylum Ca. Omnitrophica or OP3) of Bacteria and Archaea [30,110]. Coccoid cells of 200-300 nm diameters along with rod cells are found for the epibiont predator Ca. Nanosynbacter lyticus (phylum Ca. Saccharibacteria or TM7) of human oral Actinomyces odontolyticus [31,32]. A recent transcriptomic study suggests that the N. lyticus-A. odontolyticus relationship is symbiotic rather than predatory [164]. The occurrences of these small predators (symbionts, parasites, or pathogens) support the idea that the ultra-small cell sizes of the CPR bacteria are associated with the small genome sizes, leading to dependence on other prokaryotes with fully (or more) functional genomes and larger cells [165].
Cells of Bdelovibrio, the obligate invading/attenuating predators, are also as small as 0.2-0.5 µm × 0.5-2.5 µm (200-300 nm × 500-2500 nm), larger than the ultra-small CPR bacteria but significantly smaller than the prey cells [22]. In contrast to these small-sized predators, the flat, round, or oval cells of the engulfing bacterium Ca. Uab amorphum (phylum Planctomycetota) are as large as~4-5 µm in diameter, reaching 10 µm in diameter after the engulfment of bacterial preys [47]. Spherical to ovoid cells of Gemmata obscuriglobus, the not-bacteria-but-protein engulfing bacterium [166] within the phylum Planctomycetota, are in the moderate size range of 1.4-3.0 µm × 0.5-3.0 µm [167]. The type species of the phylum Planctomycetota is Planctomyces bekefii, which represents a rare example of as-yetuncultivated bacteria with validly published names but no known identity [168]. The isolation and cultivation of more Planctomycetota species may assist in elucidating the origin and function of engulfment (endocytosis, phagocytosis) for predation [169,170].
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms11071673/s1, Figure S1: Trans-phylum phylogenetic tree based on 136 (about 0.6 kb) sequences of predatory bacterial 16S rRNA genes listed in Table S1 and 35 sequences from representative, current, and former bacterial phyla listed in Table S2;  Table S1: List of 136 sources of 16S rRNA gene sequences and hunting strategies of predatory bacteria; Table S2  Data Availability Statement: All the source information of the used DNA sequences are available in Tables S1 and S2. The processed DNA sequences for the construction of the phylogenetic trees (Figures 1 and S1) are available in Supplementary Data S1 to S8.
Acknowledgments: Ryosuke Nakai at National Institute of Advanced Industrial Science and Technology, Japan, kindly reviewed and edited the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.