Advancements in Characterizing Tenacibaculum Infections in Canada

Tenacibaculum is a genus of gram negative, marine, filamentous bacteria, associated with the presence of disease (tenacibaculosis) at aquaculture sites worldwide; however, infections induced by this genus are poorly characterized. Documents regarding the genus Tenacibaculum and close relatives were compiled for a literature review, concentrating on ecology, identification, and impacts of potentially pathogenic species, with a focus on Atlantic salmon in Canada. Tenacibaculum species likely have a cosmopolitan distribution, but local distributions around aquaculture sites are unknown. Eight species of Tenacibaculum are currently believed to be related to numerous mortality events of fishes and few mortality events in bivalves. The clinical signs in fishes often include epidermal ulcers, atypical behaviors, and mortality. Clinical signs in bivalves often include gross ulcers and discoloration of tissues. The observed disease may differ based on the host, isolate, transmission route, and local environmental conditions. Species-specific identification techniques are limited; high sequence similarities using conventional genes (16S rDNA) indicate that new genes should be investigated. Annotating full genomes, next-generation sequencing, multilocus sequence analysis/typing (MLSA/MLST), matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF), and fatty acid methylesters (FAME) profiles could be further explored for identification purposes. However, each aforementioned technique has disadvantages. Since tenacibaculosis has been observed world-wide in fishes and other eukaryotes, and the disease has substantial economic impacts, continued research is needed.

The genus Tenacibaculum is critically understudied. Significant knowledge gaps exist regarding bacterial diversity, distribution, and pathogenicity of Tenacibaculum species. Given that the economic impact of mouthrot seems to be increasing, it is imperative that research focuses on the pathogenesis of disease and the role of the various bacterial species that induce mouthrot. Therefore, the principal objective of this publication is to review the ecology, identification, and impacts of potentially pathogenic

Genus Introduction
In the 21st century, Tenacibaculum species and strain identification are primarily based on 16S rDNA sequences and bacterial morphology. The genus Tenacibaculum was initially identified using 16S rDNA and gyrB sequences [41]. The phylogenetic analysis (neighbour-joining method) by Suzuki et al. (2001) using the partial gyrB gene sequence demonstrated that Flexibacter species were not closely related to Tenacibaculum species [41]. Flexibacter flexilis had gyrB sequence similarities between 69.4% and 76.4% compared to other Flexibacter species, including the reclassified Flexibacter maritimus (Tenacibaculum maritimum) and Flexibacter ovolyticus (Tenacibaculum ovolyticum) [41]. However, the sequence similarity comparing gyrB sequences of T. ovolyticum to T. maritimum was 95.2%. 16S rDNA sequences were similar among species tested and demonstrated a closely related phylogeny to the gyrB gene, where the phylogenetic sister to Tenacibaculum was proposed to be Polaribacter [41]. DNA-DNA hybridization also occurred; within a Tenacibaculum species, reassociation values were generally at or above 86 ± 2%, while outside the species and genus, reassociation values were less than 35 ± 6% and 18 ± 1%, respectively [41]. Overall, Suzuki et al. (2001) concluded that Tenacibaculum species were separate from Flexibacter and occupied a novel genus [41]. Other genes have also been applied for the identification of Tenacibaculum species, including atpA, dnaK, glyA, gyrB, ileS, infB, rlmN, tgt, trpB, tuf, and yqfO [35]. Phylogenies generated through multilocus sequence analysis (MLSA) and 16S rDNA were different [35]. If MLSA was used, there were potentially three monophyletic clades instead of several polyphyletic clades. Using MLSA, Clade 1 consisted of T. mesophilum, T. aestuari, T. lutimaris, T. litoreum, T. discolor, and T. gallaicum, Clade 2 consisted of T. aiptasiae, T. ovolyticum, T. dicentrarchi, and T. soleae, Clade 3 consisted of T. geojense, T. skagerrakense, T. amylolyticum, and T. jejuense, and there were independent lineages of T. adriaticum and T. maritimum together, and T. litopenaei and T. crassotreae separately [35]. MLSA is a more accurate technique compared to traditional phylogenies using only 16S rDNA because as more protein-encoding genes are compared, the resolution of phylogenetic analyses improves. However, the trade-off is the time allocated and the expenses applied. MLSA was also used with 7 of 11 housekeeping genes [35] to investigate the diversity of Norwegian Tenacibaculum isolates [42]. The MLSA phylogeny proposed by [42] differs from that proposed by [35]; differences include the relation and position of select species. These differences may be related to novel isolates used and different numbers (7 [42] vs. 11 [35]) of housekeeping genes used for comparisons. A recent review by Fernández-Álvarez et al. (2018) focused on Tenacibaculum species identification techniques using four methods: (1) culture-based, (2) serological studies and immunological, (3) genotyping and molecular, and (4) proteomic and chemotaxonomic [43].
A potential way to infer some aspects of Tenacibaculum spp. biology and one way to identify Tenacibaculum species are through complete genome sequence analyses. The full DNA sequence of T. maritimum NCIMB 2154 T has a 3,435,971 bp chromosome predicted to contain 3071 genes encoding 2866 proteins [44]. Complete genomes of other Tenacibaculum spp. have been shown to have similar characteristics (i.e., bp, GC%, number of genes and proteins) to T. maritimum NCIMB 2154 T (Table A1). The potential pathogenicity of T. maritimum NCIMB 2154 T has been predicted through complete genome sequencing and will be further discussed in Section 2.3.4 [44]. Full genome analyses would provide valuable information, and is slowly becoming cheaper; however, it is still too expensive to routinely sequence isolates collected from the field.
Bacteria have been traditionally identified by comparing physical morphology and biochemical characteristics; however, these comparisons alone do not always differentiate between species or genera. Most Tenacibaculum spp. are yellow; rod-shaped; of similar width (0.2-0.7 µm) and similar length (typically < 10 µm); lack flagella; and are capable of gliding motility (Table A2). They are also gram-negative, catalase-positive, oxidase-positive, strictly aerobic, and have similar pH tolerances (Table A2). In comparison, temperature and salinity tolerance and the ability to reduce nitrate are more variable characteristics among the identified species (Table A2). When describing bacterial morphology from culturing, one should note that variations could occur based on the media applied [43]. Medias applied to culture Tenacibaculum spp. include marine agar [9,16,43,[45][46][47], Flexibacter maritimus medium [43,[45][46][47], Anacker and Ordal agar with modifications [43,45,48,49], tryptone agar with modifications [43,47], thiosulphate-citrate-bile-sucrose agar [43,47], Marine Luria Broth medium prepared with seawater [41,43], Cytophaga agar with seawater [50], blood agar with modifications [16,25], and more; different chemical compositions are likely to be selective for particular isolates of Tenacibaculum. Selection for Tenacibaculum isolates can also occur by using media with aminoglycoside antibiotics such as kanamycin. The minimum inhibitory concentrations to kanamycin for Tenacibaculum are high (>30 µg per disc [51][52][53][54] up to 500 µg per disc [40] and 50 µg mL −1 [55]), relative to other bacterial groups around netpens [56]. Variations in bacterial characteristics can also change depending on when an isolate of Tenacibaculum is selected for sampling; prolonged periods in sub-cultured media can lead to the generation of spherical cells [53,57]. While there are multiple methods to identify and characterize Tenacibaculum spp. and strains, few molecular diagnostic techniques have been developed for this genus. Further research needs to focus on developing accurate and fast identification techniques for Tenacibaculum species.

Distribution and Diversity
From the limited research performed to date, the genus Tenacibaculum appears to be very diverse and many more species are likely to be described. Thirty-two named species are currently described, predominatly through 16S rDNA sequencing (Table A2). Since 2006, 22 new species have been described.
The genus Tenacibaculum has a cosmopolitan distribution within saltwater; however, local distributions of Tenacibaculum spp. are largely unknown. Currently, 20 out of 32 species have only been identified in Asia, and five species are unique to Europe (Table A2). The restricted distribution of these species may be a result of the lack of investigation to date. The remaining seven species have broader distributions; T. maritimum has been found in marine waters in Canada (East and West coast), Chile, Japan, Norway, Ireland, Spain, and Australia; T. dicentrarchi has been identified in Antarctica, Canada (East and West coast), Chile, Norway, Spain, and potentially Australia; T. finnmarkense has been identified in Chile and Norway; T. soleae has been identified in Canada, the USA, Europe, and Australia; T. xiamenense has been identified in China and Chile; T. mesophilum has been identified in Japan, the USA, and China; and T. ovolyticum has been identified in Japan, the USA, and Norway (Table A2). Further research will likely reveal that many species have more cosmopolitan distributions.

Host Relationships
Approximately three-quarters of known Tenacibaculum spp. have been found in only one or two hosts and have not been reported to be pathogenic (Table A2). Non-pathogenic Tenacibaculum spp. have been described from algae, tunicates, tidal sediments, seawater, mollusks, and crustaceans (Table A2). The remaining Tenacibaculum spp. have proposed pathogenic relationships with multiple fishes and few bivalves; these include T. maritimum, T. dicentrarchi, T. finnmarkense, T. gallaicum, T. discolor, T. ovolyticum, T. mesophilum, and T. soleae (Table A2). However, a pathogenic relationship fulfilling Koch's postulates with a host species has only been demonstrated for T. maritimum [6] and T. finnmarkense [11,40] in S. salar. T. maritimum has also been reported in other animals such as sea lice (Lepeophtheirus salmonis) and mauve stingers (Pelagia noctiluca); while T. dicentrarchi has been identified in epidermal tissue from wild killer whales (Orcinus orca) and these animals may act as vectors for the bacteria [58][59][60] (Table A3). Recently, mortality events have also occurred in the kelp industry, Kombu (Saccharina japonica) seedlings experiencing green rotten disease had an increase in Tenacibaculum spp. from 0.8% to 4.5%; however, several other bacterial genera also experienced increases in abundance [61]. In conclusion, most Tenacibaculum sp. are not considered to be pathogenic and have few identified specific hosts; however, the eight pathogenic or potentially pathogenic Tenacibaculum sp. are often associated with fishes but have also been identified from a vast array of species that may act as vectors.

Identification of Pathogenic Species
In this section, the identification of potentially pathogenic species in the order of T. ovolyticum, T. gallaicum, T. discolor, T. finnmarkense, T. mesophilum, T. soleae, T. dicentrarchi, and T. maritimum is described. In addition, BLAST comparisons from the National Center for Biotechnology Information (NCBI; https://blast.ncbi.nlm.nih.gov/Blast.cgi) were used and the top 100 top hits described, with the search parameter "Organism" left blank unless otherwise mentioned.

T. gallaicum and T. discolor
Identification techniques for T. gallaicum and T. discolor include 16S rDNA sequence comparisons, MLSA, MALDI-TOF, and FAME profiles. Representatives of these species (T. gallaicum A37.1T, T. discolor LL04 11.1.1T) were identified in the same study using morphology, GC content comparisons, DNA-DNA hybridization, and 16S rDNA sequences [12,13]. The 16S rDNA sequence of T. gallaicum BE263 (AN: LT601375.2) is most similar to three other T. gallaicum isolates (A37.1 T , BE228, BE045), which had percent identities above 99.12% (Supplementary Materials Figure S1). Another comparison using T. gallaicum A37.1 T (AN: NR_042631.1) indicated several isolates of T. litoreum, T. discolor, T. sediminilitoris, and T. ascidiaceicola have high percent identities above 97.5% (Supplementary Materials Figure S1). 16S rDNA sequences of T. discolor 9A5 (AN: JQ231117.1), and LL04 11.1.1 T (AN: NR_042576.1) are most similar to five T. discolor isolates, T. litoreum CL-TF13, and T. ascidiaceicola RSS1-6, with percent identities above 99.10% (Supplementary Materials Figure S1). MLSA places T. gallaicum in Clade one, and the sister species was T. litoreum, while T. discolor was also found in Clade one, and was the sister to T. mesophilum [35]. Complete genomes of T. gallaicum DSM 18841, T. discolor DSM 18842 and IMLK18 are available online at NCBI (Table A1). FAME profiles for both T. discolor and T. gallaicum did not record species-specific differences [62]. There were also no specific masses through MALDI-TOF that allowed for distinction between T. discolor and T. gallaicum; however, the overall mass-spec could be used to identify either species [63]. The same study also noted that phyloproteomics indicated that isolates of T. gallaicum and T. litoreum had been misidentified and were proposed to be T. discolor [63].
T. finnmarkense T. finnmarkense has been identified using 16S rDNA sequences, MLSA and complete genome sequence comparisons. Comparisons using the 16S rDNA sequence of T. finnmarkense TNO006 (AN: MN699389.1) and S2F6 (AN: MF192947.1) demonstrated high sequence similarities among several Tenacibaculum species, similar to aforementioned comparisons using T. gallaicum and T. discolor (Supplementary Materials Figure S1). Other species had highly similar sequences, including T. dicentrarchi, T. aestuarivivum, T. insulae, T. soleae, and a T. ovolyticum clone (Supplementary Materials Figure S1). Genome (Illumina) sequencing and average nucleotide identity (ANI) of various T. dicentrarchi and T. finnmarkense isolates revealed that they were highly similar and were proposed to be sister-species [64]. However, it was also reported that T. dicentrarchi AYD7486TD was proposed to be within the species T. finnmarkense based on ANI [64] and MLSA [71]. Recently, MLSA and genome (Illumina) sequencing also indicated that the T. finnmarkense clade consists of two species (T. finnmarkense and T. piscium) and the T. finnmarkense species consists of two genomovars (T. finnmarkense genomovar finnmarkense and T. finnmarkense genomovar ulcerans) [72]. Complete genomes of T. finnmarkense (HFJ T and TNO006), specific genomovars of T. finnmarkense, and T. piscium (TNO020 T , TNO070, TNO063, TNO066, TNO064) are available on NCBI (Table A1) and more diagnostic techniques are required. FAME profiles of T. piscium (TNO020 T ), T. finnmarkense genomovar ulcerans (TNO010 T ), T. finnmarkense genomovar finnmarkense (TN0OO6 T ), T. finnmarkense genomovar (HFJ T ), and T. dicentrarchi USC 35/09 T identified that numerous fatty acids were similar; however, the summed feature for T. piscium is greater than the other isolates, A35:0 is greater for T. finnmarkense genomovar ulcerans and T. finnmarkense genomovar finnmarkense, while T. finnmarkense genomovar finnmarkense had greater levels of A35:1 ω6c. More isolates need to be tested through FAME profiles to determine if any fatty acid could be used as a chemotaxonomic marker. MALDI-TOF [63] could be potentially useful for further describing clades of T. finnmarkense and closely related species.

T. mesophilum
For the identification of T. mesophilum, 16S rDNA comparisons and MLSA have been applied. The 16S rDNA sequence of T. mesophilum MBIA3140 (AN: NR_024736) was most similar to 13 T. mesophilum isolates with percent identities above 99.04%; however, T. lutimaris DI 83II and Actinobacterium YH73 had percent identities of 97.44% and 99.29% (Supplementary Materials Figure S1). As mentioned previously, MLSA described a close relationship between T. mesophilum and T. discolor within Clade 1 [35]. Three full genomes of T. mesophilum are available on NCBI (Table A1). As for T. finnmarkense, more diagnostic tests are required, as well as studies to demonstrate the pathogenic potential of the bacteria.
T. soleae T. soleae has been identified using 16S rDNA sequences, PCR, MLSA, and MALDI-TOF. The 16S rDNA sequence of T. soleae LL04 12.1.7 (AN: NR_042630) indicated that all T. soleae isolates had percent identities above 98.15% (Supplementary Materials Figure S1). However, isolates of T. dicentrarchi, T. aestuari, T. lutimaris, T. insulae, and T. discolor had percent identities above 97.04% (Supplementary Materials Figure S1). BLAST comparisons indicate that PCR primer sequences using 16S rDNA [73] and 16S-23S internal spacer region (ISR) [74] were most similar to T. soleae; however, other outgroups had a percent identity of 100% for one of the primers. Both PCR assays have been applied before [75,76] and appears to be a reliable tool for detecting T. soleae. However, in another study, a new PCR assay specific to T. soleae was developed for multiplex-PCR due to previous PCR amplicons for other species of Tenacibaculum being a similar size [76]. The phantom band described in some samples [67] may be related to the amplicon of a potential contaminate, false positive, or another section within the genome that may amplify. For both PCR assays, melting curve analyses would help determine if the generation of multiple products is occurring. MLSA phylogenies placed T. soleae closest to T. ovolyticum [35]. The complete genome of T. soleae UCD-KL19 is available online (Table A1). The only non-genetic test, MALDI-TOF, applied to identify T. soleae was effective; a characteristic peak mass was identified at 9048.66 m/z [63]; meaning that a single peak could distinguish this species from the rest of the genus [63].
T. dicentrarchi T. dicentrarchi has been identified using 16S rDNA sequences, PCR, qPCR, MLSA, and MALDI-TOF. The 16S rDNA sequence of T. dicentrarchi 35/09 T (AN: NR_108475.1) indicated that isolates labeled as T. dicentrarchi had a percent identity above 98.21% (Supplementary Materials Figure S1). The most similar Tenacibaculum sequence of another named species would be T. aestuariivivum JDTF-79 and T. ovolyticum with a percent identity of 98.13% and 96.82% (Supplementary Materials Figure S1). A 16S rDNA PCR assay (Tenadi) to identify T. dicentrarchi [77] has limited application because it also identified T. finnmarkense [78]. A BLAST comparison of the Tenadi primers indicated that several outgroups including T. finnmarkense were matches (Supplementary Materials Figure S1). A new T. dicentrarchi PCR-specific assay was developed for multiplex-PCR due to the Tenadi assay amplicon being of a similar size to other species-specific PCR assays [76]; however, validation in other studies is needed. A 16S rDNA qPCR assay has been developed to identify T. dicentrarchi and was reported to be specific but needs to be validated and investigated for potential false positives [79]. MLSA demonstrated that T. dicentrarchi is in Clade two and was the phylogenetic sister to T. ovolyticum and T. soleae [35]. Subsequently, MLSA determined that the sister to T. dicentrarchi was an unnamed Tenacibaculum species (T. finnmarkense), while T. soleae and T. ovolyticum were still closely related in a monophyletic clade [35,42,64]. Four complete genomes of T. dicentrarchi are available on NCBI for comparison and include AY7486TD, TNO021, TD3509 = 35/09 T , and TdChD05 (Table A1). Non-genetic identification techniques for T. dicentrarchi are limited to MALDI-TOF [63], which detected a peak mass unique to the species at 2579.41 m/z from T. dicentrarchi NCIMB14598 [63]. More species and isolates need to be tested to confirm if MALDI-TOF can distinguish between T. finnmarkense and T. dicentrarchi.

T. maritimum
Many different techniques have been used for the identification of T. maritimum including rDNA sequence comparisons, PCR, nested PCR, qPCR, viability-qPCR (v-qPCR), MLSA, serology, FAME profiles, and MALDI-TOF. The 16S rDNA sequence of T. maritimum TmarCan1 (AN: KY428892.1) was most similar to~30 T. maritimum isolates with a percent identity above 97.52% and the next closest comparison was Polaribacter sp. 7002-035 with a percent identity of 95.99%; other Tenacibaculum species have percent identities below 95.62% (Supplementary Materials Figure S1). Two 16S rDNA PCR assays [80,81] are widely applied to identify T. maritimum [82][83][84][85][86][87][88]. qPCR and v-qPCR for 16S rDNA [89][90][91], gyrB gene [92], and outer membrane protein A (ompA) gene sequences [7] have been developed. The 16S rDNA assay [89] was determined to be more sensitive than the ompA gene assay [7]. Using MLSA, T. maritimum was shown to be an independent lineage separated from monophyletic clades of other Tenacibaculum species, a classification that is also supported by the 16S rDNA alignment [35]. There are 25 complete genome sequences available on NCBI (Table A1). Most genetic identification techniques are specific to T. maritimum, with identification not being a primary concern for this species; novel studies may want to focus on the biology of T. maritimum.
Non-genetic identification techniques for T. maritimum include serology, FAME profiles, and MALDI-TOF. T. maritimum can be divided into at least four serogroups, which were identified among European or Asian isolates isolated from fishes [12,47,[93][94][95]. Antigenic heterogeneity of T. maritimum was proposed to be host-specific [93]; however, at least two distinct serological groups of T. maritimum from Atlantic salmon in BC have been described [55,96]. Additional serological studies of isolates of Tenacibaculum spp. identified from mouthrot cases are needed to allow better characterization. Future studies should also focus on comparing isolates from the Americas against those from Europe and Asia to determine if the known serological scheme can be expanded. MALDI-TOF on 22 isolates for T. maritimum revealed that there was a species-specific peak mass at 9408.33 m/z and that there was also a characteristic peak mass at 11,356.67 m/z for 17 out of 22 isolates [63]. More recent MALDI-TOF applications identified 18 monomorphic and nine polymorphic biomarkers within the species, having potential use for species and strain typing [97]. The same study used MLST-like approaches combining isoform numbering (1-5) corresponding to the MALDI profile to designate a MALDI-type (MT1-20), and used the clustering of MALDI-types to identify MALDI-groups (MG1-4) [97]. Using the aforementioned technique, trends were identified between the geographical origin of the strain and the designated MT, in agreement with previous MLST results [35,97]. FAME profile could also distinguish T. maritimum from other species [62]; however, there were two main clusters (Ia and Ib) for T. maritimum and the author proposed that these clusters may be based on host species or geographic origin [62]. Based on the fatty acids iso-A35:1 G and iso-A35:0 3-OH comprising a greater mean percent composition for the tested T. maritimum isolates compared to the other species, it was also interpreted that those two fatty acids may be used as chemotaxonomic markers [62].
T. ovolyticum T. ovolyticum has been found in deep waters off the coast of Japan [103], in a lobster culture associated with epizootic shell disease [104], as a component of sardine egg microflora [105] and has been reported to be an opportunistic pathogen in eggs and larvae of Atlantic halibut (Hippoglossus hippoglossus L.) [106,107]. In one study, when T. ovolyticum represented less than 30% of the epifloral community, halibut eggs hatched to larvae; however, when values rose above 30%, substantial increases in mortality at the hatching stage occurred [106]. T. ovolyticum was able to dissolve the chorion and damage the zona radiata through enzymatic activity [106]. Halibut but not turbot eggs immersed in baths of 10 5 -10 6 bacteria mL −1 of T. ovolyticum before hatching had significantly increased mortality compared to controls [107]. In a phylogenetic study, of 89 Tenacibaculum isolates collected from disease outbreaks in Norway, isolate TNO089 had the greatest genetic similarity to T. ovolyticum (95%) and was cultured from halibut fry; demonstrating the potential for this organism to induce tenacibaculosis in fish past the larval stage [42]. Overall, T. ovolyticum has been found in environmental samples, invertebrates displaying disease and as part of the microflora of marine fish eggs; however, few studies to date have demonstrated that T. ovolyticum is a fish pathogen.
T. gallaicum and T. discolor T. gallaicum and T. discolor were first identified together in Spain; T. gallaicum was isolated from seawater taken from a turbot (Psetta maxima) holding tank, and T. discolor was identified in the kidney of a deceased sole (Solea senegalensis) [12,13]. Both bacterial species are proposed pathogens because T. discolor L0LO4.11.1.1T and T. gallaicum A37.1 T experimentally induced tenacibaculosis in both turbot and sole [12,13]. For T. discolor and T. gallaicum, mortalities ranged 60-100% following intraperitoneal (IP) injections of 10 5 -10 7 colony forming units (CFU) fish −1 [12,13]. Diseased sole and turbot displayed an eroded mouth, necrotic fins, ulcerations on the flanks, and pale internal organs [12,13]. Based on these findings, T. gallaicum and T. discolor are likely fish pathogens; however, more research is needed to validate a relationship between the presence of bacteria and diseased fishes and if these species are primary pathogens or opportunistic pathogens.
T. mesophilum T. mesophilum was initially identified on a sponge (Halichondria okadai) [41], from sediment samples [109], and associated with the microbiome of Pacific white shrimp (Litopenaeus vannamei) [110]. There are few studies focused solely on T. mesophilum. Research has demonstrated that T. mesophilum induces a humoral immune response in gilthead seabream (Sparus aurata) [111] and produces a unique linear siderophore (bisucaberin (B) without macrocyclic counterparts [112,113]. Recently, the bacterium has been associated as the agent responsible for black-spot shell disease in Akoya pearl oysters (Pinctada fucata) [114]. A BLAST comparison (https://blast.ncbi.nlm.nih.gov/Blast.cgi) of 16S rDNA sequences of Tenacibaculum sp. Pbs-1 (NCBI Accession number: LC342074) cultured from diseased Akoya pearl oysters indicated that five separate sequences in the complete genome of T. mesophilum DSM 13764 (NCBI Accession number: CP045192) are identical (query cover of 100%, an E-value of 0, and a percent identity of 100%). Even though Tenacibaculum sp. strain Pbs-1 was thought to be the agent responsible for black-spot shell disease, reproduction of the disease including mortality requires additional factors, including a compromised shell [114]. More work is needed to demonstrate that T. mesophilum can be a pathogen to select groups of animals.
T. dicentrarchi T. dicentrarchi is a potential pathogen that has received attention due to mortality events in Chile [14,98], and in BC, where the bacterium has been identified as a common isolate from S. salar with lesions similar to mouthrot [119]. In addition to Atlantic salmon [14], T. dicentrarchi has been documented in red conger eel (Genypterus chilensis) [98], and sea bass (Dicentrarchus labrax) [57]. Some Tenacibaculum sp. isolates from cod (Gadus morhua), wrasse species, and lumpfish (Cyclopterus lumpus L.) were also re-identified as T. dicentrarchi using MLSA [42]. Clinical signs in fishes infected with T. dicentrarchi included external ulcers, frayed fins, hemorrhagic organs, and damaged gills [14,57,98]. A bath immersion for 1 h using 3.78 × 10 5 CFU mL −1 of T. dicentrarchi TdChD05 induced 65% mortality in S. salar, and 93% mortality in rainbow trout (Oncorhynchus mykiss); around 50% of the mortalities occurred in one day for O. mykiss, and four days for S. salar [14]. However, 50% cumulative mortality in O. mykiss observed within one day is unusually fast for the development of tenacibaculosis in fishes in comparison to other studies and may indicate that other factors were involved [6,7,14,15]. Coho salmon (Oncorhynchus kisutch) in the same experiment experienced no adverse effects [14], but outbreaks by Tenacibaculum spp. have been reported in coho salmon [71]. In a separate study, 30 Atlantic salmon smolts were IP injected with 10 7 CFU fish −1 and then co-housed with another 197 salmon [120]. The water was then decreased from 1400 L to 400 L and the fish underwent bath immersion with 10 6 CFU mL −1 for 30 min [120]. None of the fish in this trial died or displayed ulcers [120]. In a second trial, 26 of 36 fish were scarified using a scalpel blade to removes scales, and three drops of the bacterial culture were added to 20 of these scarified fish. All fish were then bath immersed in 10 7 CFU fish −1 for 2 h. Ultimately, 32 fish died and 2 were euthanized as moribund [120]. Of the four survivors, one was scarified without the addition of bacteria and the three others only experienced the bath immersion [120]. All the fish were reported to have extensive scale loss, small hemorrhagic lesions, ascites, and dark livers [120]. Recently, exposure trials of Atlantic salmon to Neoparamoeba perurans (causative agent of amoebic gill disorder) identified a weak positive correlation between the presence of N. perurans and T. dicentrarchi in diseased fish [79]. In some lesions of diseased fish, 70.7% of the mean bacterial abundance was T. dicentrarchi, and there was reduced species richness and diversity indexes in diseased fish compared to naïve fish, supporting that dysbiosis may have implications for tenacibaculosis [79]. In another recent study, two isolates (QCR29 and QCR41) at 3.1 and 3.7 × 10 4 CFU mL −1 were exposed to red conger eel (n = 12) for 2 h through bath exposure [121]. Eel mortalities began four days post-exposure and by the end of the experiment (30 d) 8 fish head died, and fish presented with epidermal ulcers, hemorrhagic fins, mouth and operculum, irritation around the head and yellow plaques around the jaws [121]. These experiments support the potential for T. dicentrarchi to be a pathogen of concern for aquaculture.
T. maritimum T. maritimum is the most documented Tenacibaculum species in the literature. The bacteria have been described worldwide, and in at least 30 host species (Table A3). In S. salar smolts, mortality rates in successful experimental infections often exceed 50% [6,7,9,26,122]. Variable success has been recorded using different infection protocols with the variability likely due to the isolate tested, host species used, infection methodology applied, and the concentration of infective dose [6,7,9,26,122].

Pathogenesis
There is a limited amount of information available on the pathogenesis and virulence factors associated with the genus Tenacibaculum except for T. maritimum. The similarity of the clinical signs caused by several of the proposed pathogens of this genus suggest that common themes will apply as for T. maritimum. The pathogenesis section is described in the order of in-situ infections including the clinical presentation of mouthrot in BC, subcutaneous (SC) injections, IP injections, and bath immersion infections.
In-situ infections of T. maritimum are believed to infect fishes through damaged epithelia initially [26], where in Atlantic salmon, mouthrot is often associated with the mouth of fish [6,7]. In an S. salar smolt that died two months after being transferred to saltwater and was proposed to have died from mouthrot, histopathology and scanning electron microscopy (SEM) revealed lesions around the mouth surrounded by filamentous bacteria and dislodged teeth, with bacteria occupying gingival pockets [7]. Based on similar lesions from experimental bath infections using T. maritimum isolates (TmarCan15-1, TmarCan16-1, TmarCan16-2, and TmarCan16-5), it was proposed that for mouthrot, bacteria become systemic via the vascularized tooth pulp [7]. Histology from a jaw section of an Atlantic salmon with mouthrot had ulceration of the mucosal epithelium with plaques of basophilic filamentous bacteria and inflammation of the dermis, while bacteria were also dispersed throughout the compromised epithelium and dermis surrounding teeth (Figure 1). The histological section of the infected jaw presented here is similar to those of Frisch et al. [7]. While these authors also identified plaques of filamentous bacteria and necrosis of the gill tissue of affected Atlantic salmon. T. maritimum NCIMB 2153 in-situ infected farmed Senegalese sole (S. senegalensis L.) have also been studied by histology and SEM [47]. The flanks of infected fish exhibited a loss of epidermis, dermis, and hypodermis, extensive necrosis of superficial muscles, severe hemorrhages, and the presence of macrophages at infection sites [47]. SEM demonstrated that some lesions first impacted the epithelium in the middle of a scale exposing the fibroid bone and then progressed outwards. SEM also revealed that areas without epithelium had copious amounts of rod-shaped bacteria present around scales and a reduction in the micro-ridges of the surface epithelia [47]. Further studies are required as there are substantial differences between the reports from [7,47] including fish species and bacterial isolates used, different geographic locations, and different transmission routes, resulting in different clinical signs of infection. From Atlantic salmon post-smolts sampled at BC netpen sites, microbial profiling using the 16S rDNA sequences indicated that T. maritimum was identified in healthy, diseased, and post-treated fish [123]. Diseased fish had reduced microbial diversities with two sequence variants of T. maritimum dominating the community [123]. Surviving fish with high proportions of T. maritimum, in association with the abundance Vibrio spp. and the presence of mouthrot led researchers to provide evidence that mouthrot is a complex multifactorial disease characterized by dysbiosis [123]. With several Tenacibaculum species reported to have associations with dysbiosis, more research should focus on what multifactorial processes initiate dysbiosis linked to the presence of mouthrot. In BC, the most common clinical presentation of tenacibaculosis is denoted as mouthrot. The few common clinical signs of mouthrot include yellow plaques and ulcers on the mandibles, gills, and infrequently on the flank (Figure 2). The distinguishing feature of mouthrot compared to other presentations of tenacibaculosis is that at netpen sites, small plaques often occur on the jaws and gills with few conspicuous ulcers ( Figure 2). Ulcers can be found on other epidermal surfaces, and laboratory trials identify more severe ulcerations on the fish [6][7][8]. In other presentations of tenacibaculosis, conspicuous ulcers are often located around the head, flanks, and fins of the fish [26,49,99,100].    SC and IP infection trials using T. maritimum has been documented in several fishes. In turbot, SC and IP injections of 10 8 -10 9 CFU fish −1 of T. maritimum LL01.8.3.8 induced anorexia and lethargy as early as 3 h post-injection (hpi) [99]. At this time, histopathology of tissue at the site of SC injection demonstrated clusters of bacteria distributed through the connective tissue of the hypodermis and degeneration of muscle without inflammatory responses [99]. The first gross lesion was recorded at 24 hpi and consisted of discoloration at the injection site [99]. The formation of an ulcer at the inoculation site occurred around 48 hpi, while at 72 hpi, ulcers obtained the characteristic circular appearance and developed peripheral hyperemia [99]. At all-time points, histopathology demonstrated degeneration and necrosis of muscles, detached or absent epidermis and dermis, and inflammation of the ulcerated area [99]. The expansion of the lesion was evident based on the spread of grossly discolored tissues [99]. At seven days post-infection, most fish displayed ulcerative dermatitis and hyperemia and some fish displayed diffuse hyperemia and hemorrhages on the fins [99]. Throughout the experiment, bacteria were primarily found from three organs (skin, kidney, and spleen) using PCR, culture, and immunohistochemistry [99]. A subsequent study used SC injections of 10 8 CFU fish −1 T. maritimum LL01.8.3.8 on turbot, and similar pathological signs were noted [10]. In another study, Dover sole (Solea solea L.) injected subdermally with 10 7 cells fish −1 T. maritimum developed epidermal lesions and experienced 100% mortality in four days [125]. Difficulties were also reported based on bacterial cultures flocculating leading to inconsistent quantification of bacteria and numerous mortalities in both pilot trials [125]. In the first pilot trial, SC-injected fish had 30% mortality in 48 h and the rest were euthanized; in the second pilot trial, the proportion of mortalities among controls were 30% and 50% [125]. The high number of mortalities within such short time frames for both pilot and experimental trials may indicate that other factors may be involved. In black sea bream (Acanthopagrus schlegei), SC injection of 10 2 -10 6 cells fish −1 using T. maritimum A4 induced ulcers at the injection site [126]. However, since the highest mortality using 10 6 cells fish −1 was 52.6% while the control group experienced 40.8%, it is possible that the infection method may have contributed to mortality [126]. Based on these results, the experimental reliability and effectiveness of SC injections are questionable and additional studies are needed to validate current work in these fish species.
IP injections of 10 8 -10 9 CFU fish −1 of T. maritimum strain LL01.8.3.8 in turbot at 168 hpi resulted in splenitis and capsular necrosis, necrosis and hemorrhage of the liver, head kidney, and intestine, reduced hematopoietic tissue in the head and kidney, and enteritis with transmural inflammation [99]. No grossly observable, macroscopic epidermal lesions were recorded, but T. maritimum was identified in the liver, heart, gastrointestinal tract, and gills from 6-48 hpi and the spleen and kidney from 6-168 hpi using immunohistochemistry [99]. In Atlantic salmon, isolated extracellular proteins (>250 µg/fish) from T. maritimum 89/4762 injected-IP induced 100% mortality, with the author concluding that extracellular products released from T. maritimum induce disease [122]. Overall, it appears that IP injections of T. maritimum can produce internal lesions, and extracellular products may play a significant role in the pathogenesis of disease. However, other studies were unsuccessful in reproducing tenacibaculosis through IP injections [9,127].
For both IP and SC infection methodologies, bacteria were detected in most organs as early as 3 h until the end of the experiment; bacteria were also detected in blood vessels, possibly demonstrating bacteremia [99]. However, IP and SC injections do not mimic natural horizontal transmission, and researchers have had variable success inducing infection [6,7,9,127]. Subcutaneously injected fish displayed similar clinical signs to affected fish from production settings [99].
Bath infections, mimicking horizontal transmission, using T. maritimum, also resulted in clinical signs similar to those described from outbreaks on fish farms [6]. Infection trials with turbot using T. maritimum isolate ACC6.1 suggested that an immersion of 18 h using 5 × 10 3 CFU mL −1 was needed [128]. However, in Atlantic salmon, successful bath infections have occurred at 1.5 and 5 h using various Canadian T. maritimum isolates at 10 7 CFU mL −1 [6]. Isolate virulence and bath concentration are likely important factors that impact the period required to allow infection [6]. Bath infections with S. salar smolts produced atypical behavior (erratic swimming and loss of equilibrium), oral ulcers, and yellow plaques on the external surfaces such as the mouth and gill [6,7]. In several fishes including S. salar, bath infections resulted in ulcers externally on the flanks and fins of fishes, pale organs, friable livers, congested kidneys, and eventually mortality [16,26,100,102,129]. Numerous Tenacibaculum sp. infections in laboratory practices use variably concentrated infective doses, often at high concentrations (above 10 3 CFU mL −1 ), which may indicate that the bacteria are not primary pathogens, but instead other variables, including dysbiosis [123], may influence infection.

Virulence Factors
Virulence factors are required to allow bacteria to invade, induce disease, and evade host defenses [130]. A complete genome analysis of T. maritimum NCIMB 2154 T identified categories of virulence genes related to motility, adhesion, quorum sensing/quenching, metabolism, iron acquisition, stress response, transport/secretion systems, and toxins [44].
T. maritimum genes predicted to express proteins for gliding machinery (14 gld genes [gldA to gldN] and 10 spr genes [sprA to sprE, and five sprF paralogs]) allowing mobility on multiple surfaces have been identified. Seventeen other genes in T. maritimum code for various adhesins, factors related to the biosynthesis of exopolysaccharides, and lectin or carbohydrate-binding motifs [44]. These three groups (adhesins, exopolysaccharides, and binding motifs) may allow adhesion to multiple biotic and abiotic surfaces [44]. Several isolates of T. dicentrarchi and T. maritimum from S. salar lesions formed biofilms on abiotic surfaces such as polystyrene [131,132]. During the same experiment, biofilm formation indexes were the greatest at 24 h for all strains tested, but there was significant variability between strains over time (120 h) [131,132]. More studies are needed to determine what adhesins are necessary to bind to specific surfaces, as one researcher has highlighted the difficulty in creating infection models for fishes based on a lack of understanding of the adhesive properties of Tenacibaculum species [6]. An understanding of the adhesive properties of T. maritimum may help identify the mechanisms that lead to flocculation in media, as this can lead to unreliable estimates in bacterial concentration using spectrophotometry and has downstream implications for experimental trials.
Genes in T. maritimum related to metabolism include a complete glycolysis pathway, a tricarboxylic acid cycle, sugar transporters, sugar enzymes, and several proteases, among others [44]. The identification of various proteases, which degrade proteinaceous compounds such as gelatin and casein, and the capability to use amino acids as a carbon and nitrogen source, support this organism's capability as a pathogen and its ability to survive off the host [44,50]. However, the identification of genes related to carbohydrate processing is a unique finding and contradicts previous studies that demonstrated that the bacterium is unable to process simple and complex carbohydrates [9,44,50]. Determining the function of these genes may explain how nutrients are obtained and utilized by Tenacibaculum.
Iron acquisition genes have been identified in T. maritimum and include the production of the bisucaberin siderophores and transporters, heme-related proteins, iron-regulation proteins, as well as a Fur regulator [44]. The identification of these genes is in agreement with the results obtained from iron-limitation experiments and assays [135]. T. maritimum isolates utilized several iron sources (hemin, hemoglobin, ferric ammonic citrate, and transferrin) when added to iron-deficient media, were able to bind to hemin, demonstrating the presence of heme-related proteins, and had siderophores identified in universal colorimetric chemical assays [135]. Hypothetical genes involved in iron acquisition [44] and the ability to remove iron from other sources [135] may play important roles in obtaining iron from the blood and tissue, as T. maritimum are reported to undergo bacteremia and can produce lesions in the liver and spleen [118,128,129]. Future studies should investigate how these genes aid iron regulation in Tenacibaculum species and should identify the expression of these genes in in-vivo/vitro models.
Bacterial stressors can include chemical (i.e., reactive oxygen species (ROS), heavy metals), physical (i.e., temperature), and biological interactions. T. maritimum encodes three superoxide dismutases (SodA, SodB, and SodC) and two catalase/peroxidase enzymes (KatA and KatG), indicating that the organism can cope with oxidative stress [44]. Applications of hydrogen peroxide, which generate ROS, did not dramatically reduce T. maritimum infections but speculated that hydrogen peroxide inadvertently promoted tenacibaculosis through the stress that the fish experienced during treatment [136]. Several genes related to heavy metal resistance have also been identified and were proposed to remove cationic heavy metals to limit ROS production [44]. Temperature is another stressor, but there is considerable variation in the range of temperatures tolerated by Tenacibaculum species (Table A2). Studies have reported fish mortalities caused by Tenacibaculum spp. following either a decrease or increase in water temperature [26,103]. Additional research should occur to identify genes related to stressor response.
Transport systems are useful for pathogens, as they allow proteins to be brought to the cell surface. Genes encoding an ATP binding cassette type transport system, a Sec-dependent transport system, a twin-arginine transport system, and a type IX secretion system (T9SS) were identified in T. maritimum [44]. Extracellular products of T. maritimum reported to induce mortality in S. salar by [117] are possibly transported using these systems. The role of each transport system in Tenacibaculum sp. needs further research; if toxins can be prevented from reaching the surface of the bacterial cell, Tenacibaculum infections may be attenuated.
T. maritimum genes have been found that code for many enzymes, including cholesterol-dependent cytolysin, collagenase, sphingomyelinase, ceramidase, chondroitin AC lyase, streptopain family protease, and proteins related to sialoglycan degradation/uptake [44,97]. Many of these enzymes are classified as toxins because they damage cells. For example, cholesterol-dependent cytolysins are cytolytic pore-forming toxins; however, these are also predicted to interact with the phagosome (as with Listeria monocytogenes) or cause translocation of enzymes (as with Streptococcus pyogenes) [137]. Sphingomyelinases are multi-functional and can aid phagosomal escape or avoidance, tissue colonization, infection establishment, and evasion from host immune responses [138]. Ceramidase in Pseudomonas aeruginosa has been reported to have functions linked to hemolysis in mammals [139]. Given that heme-related genes were identified in T. maritimum, ceramidase may be linked to iron acquisition if ceramidase can lyse cells rich in iron, such as erythrocytes in the bloodstream of fishes. Chondroitin AC lyase hydrolyzes chondroitin; a reduction in the rigidity of connective tissues caused by loss of chondroitin allows for easier dissemination of bacteria throughout the host [140]. Collagenase is an enzyme that breaks down collagen [141]. Since the skin, cartilage, and bones of finfish are rich in chondroitin sulphate and collagen [141][142][143], chondroitin AC lyase and collagenase may play a role in the development of external lesions and invasion into deeper tissues. Genes for sialidase were reported, and their products may allow foraging for host glycoproteins [44,144]. Further research is needed to determine how the genes identified by [44] are related to the pathogenesis of disease induced by T. maritimum.

Conclusions
Members of the genus Tenacibaculum are Gram-negative, filamentous, marine bacteria that are likely cosmopolitan and ubiquitous. Most bacterial species are non-pathogenic or have not been reported in mortality events, while eight other species (T. ovolyticum, T. gallaicum, T. discolor, T. finnmarkense, T. mesophilum, T. soleae, T. dicentrarchi, and T. maritimum) have been related to finfish or shellfish mortality events. Most potential pathogens are identified using 16S rDNA sequencing, and few diagnostic tests have been developed to identify each species, except T. maritimum. Similar clinical signs of infection in fishes induced by Tenacibaculum spp. include external ulcers, atypical behaviors, and mortality, and indicate that the term tenacibaculosis should be expanded to encompass Tenacibaculum species. Imitations of tenacibaculosis outbreaks from aquaculture sites are repeatable using experimental infections via bath immersions, where SC and IP injections have had less success. Variations in observed infections can be related to the bacterial isolate, host, geographic origin, and mode of transmission. More research is needed to define local distributions of bacteria, increase the number of diagnostic tests for pathogenic species, and clarify the pathogenesis of Tenacibaculum species.

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. Complete genome characteristics among members of the genus Tenacibaculum. The species, strain or isolate designation, sequence length, GC%, number of genes, number of proteins, and the submission info are described. The domain name and constant directory for all inquiries is 'https: //www.ncbi.nlm.nih.gov/assembly/GCF_', while the unique directory is below. No plasmids were described for any Tenacibaculum species. NA = Not available.    Table A2. Tenacibaculum species description: species, strain or isolate, location, source of collection, size (µm), gliding motility (+/−), temperature range and optimum ( • C), pH range and optimum, NaCl range and optimum (% w/v) and enzymatic nitrate reduction (+/−). All entries were Gram-negative, composed of rod-shaped bacteria, lacked a flagellum, and were either beige, yellow, or green. All entries tested for catalase were positive and all entries tested for oxidase were positive except [136,141]. + = Yes or positive, − = No or negative, NA = Not available.     Appendix C