bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Unpicking the mysterious symbiosis of Mycoplasma in salmonids 2 3 Cheaib, Ba,b *, Yang P c, Kazlauskaite Ra, Lindsay Ea, Heys Ca, De Noa Ma, Patrick 4 Schaala Dwyer Ta, Sloan W b, Ijaz UZb, Llewellyn, MSa 5 6 7 * Corresponding author: [email protected] 8 a Institute of Behaviour, Animal Health and Comparative Medicine, Graham Kerr Building, 9 University of Glasgow, Glasgow, G12 8QQ. 10 b School of Engineering, University of Glasgow, Glasgow, G12 8QQ 11 c Laboratory of Aquaculture, nutrition and feed, Fisheries College, Ocean University of China, 12 Hongdao Rd, Shinan District, Qingdao, Shandong, China 13 14 Importance (144/150 words) 15 16 Mycoplasma is the smallest self-replicating and cell wall deficient life form. Several strains of 17 this bacterial genus can parasitise a wide array of vertebrates, including the human body, 18 causing several diseases. Unfortunately, in aquaculture, the role of mycoplasmas in the 19 gastrointestinal tracts (GI) tract of Atlantic salmon (Salmo salar) remains unclear. However, 20 recent microbiome studies have demonstrated their dominance in the acidic compartments of 21 salmon GI. The continued increase in production of farmed Atlantic salmon, have accentuated 22 the need to unravel the potential adaptive function of the mycoplasmas, and to classify their 23 symbiose between commensalism and mutualism. From the pyloric caecum of Atlantic 24 salmon, we assembled a complete genome (~0.57 MB) via shotgun-metagenomics. We 25 discovered encoding genes of riboflavin pathway and sugars transporters. Their small 26 genome size, lack of pathogenicity factors and mobile genetic elements suggest a symbiotic 27 relationship between Mycoplasma and the Atlantic salmon. 28 29 30 Abstract (245/250 words) 31 32 Lacking a peptidoglycan cell wall, mycoplasmas are the smallest self-replicating life forms. 33 Members of this bacterial genus are known to parasitise a wide array of metazoans including 34 vertebrates. Whilst much research has been significant targeted at parasitic mammalian 35 mycoplasmas, very little is known about their role in other vertebrates. In the current study, we 36 aim to explore the biology and evolution of Mycoplasma in salmonids, including cellular niche, 37 genome size structure and gene content. Using Fluorescence in-situ hybridisation (FISH), 38 mycoplasmas were identified in epithelial tissues across the digestive tract (stomach, pyloric 39 caecum and midgut) during the developmental stages (eggs, parr, subadult) of farmed 40 Atlantic salmon (Salmo salar), showing a high abundance in acidic compartments. With high 41 throughput sequencing from subadults farmed Atlantic salmon, we assembled a nearly 42 complete genome (~0.57 MB) via shotgun-metagenomics. The phylogenetic inference from 43 the recovered genome revealed successful taxonomic proximity to Mycoplasma penetrans 44 (~1.36 Mb) from the recovered genome. Although, no significant correlation between genome 45 size and its phylogeny was observed, we recovered functional signatures, especially, 46 riboflavin encoding genes pathway and sugars transporters, suggesting a symbiotic 47 relationship between Mycoplasma and the host. Though 247 strains of Mycoplasma are 48 available in public databases, to the best of our knowledge, this is the first study to 49 demonstrate ecological and functional association between Mycoplasma and Salmo salar bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

50 which delineates symbiotic reductive evolution and genome erosion primarily and also serves 51 as a proxy for salmonid health in aquaculture processes (cell lines, in vitro gut models). 52 53 54 Introduction 55 56 Commensal associations between bacteria and metazoan hosts are already well-established 57 as being ubiquitous in nature. The extent to which microbes have adapted to persist in the 58 intra-host environment varies considerably between taxa: some are opportunistic 59 commensals, with others being obligate parasites or symbionts (1). Mycoplasmas are a 60 diverse group of bacteria that are known to parasitise a wide array of metazoans, plants, 61 invertebrates and vertebrates, including fish (1). In vertebrates, the mucosal surfaces of the 62 alimentary canal, respiratory and genital tract are the primary site of colonisation (2). 63 Mycoplasma sp. are a source of human and mammalian diseases. Of particular interest, but 64 not limited to, is their implication in immunocompromised human cohorts (3, 4). It is generally 65 thought that mycoplasmas have strict host associations, resulting in low zoonotic potential (5– 66 7). Whilst there has been significant research effort targeted at parasitic mammalian 67 Mycoplasma species, less is known about their importance and role in other vertebrates. 68 Mycoplasmas, as well as related taxa included in the class Mollicutes (Spiroplasmas, 69 Ureaplasma and Acholeplasmas), are recognized as the smallest and simplest free-living and 70 self-replicating forms of life (6, 8). Mycoplasmas lack a peptidoglycan cell wall and are 71 bounded by a simple cell membrane (7). In addition to being physically small, mycoplasmas 72 have the smallest genomes of any free-living organism(2). Mycoplasma genitalium, in 73 particular, has a genome size of 580 kilobases comprising of only 482 protein-coding 74 genes(9), whilst Mycoplasma mycoides, typically has 473 protein-coding genes, of which 149 75 still have no known function(9). The relative simplicity of Mycoplasma genomic contents and 76 structure has made this genus the target of scientific community’s efforts to design and 77 synthesize a minimal bacterial genome, de novo, to establish the minimum requirements for 78 biological life(10). 79 80 To further support this argument, the small size and simplicity of mycoplasmas, as well as 81 their close association with metazoan hosts and their ability to survive as free-living species, 82 irrespective of host, has led them to be considered as a target species to explore genome 83 erosion or reductive evolution (11, 12) which refers to genes loss from an organism’s genome. 84 Dependence on host organisms can theoretically lead to mutual interdependence of 85 metabolic processes. This results in its relaxed selection amongst the pool of bacterial 86 genomes, with the main process being the accumulation of loss-of-function mutations in 87 coding genes, and the eventual loss of genetic material from the bacterial genome (13). 88 Genetic drift can also play a significant role as host-associated microbes have relatively fewer 89 opportunities to exchange genetic material (14). Enhanced mutational pressure from impaired 90 DNA repair machinery could also be a factor (15). Isolation from microbial congeners and host 91 dependence may be further enhanced in mycoplasmas that exploit an intracellular niche, 92 which several species have been show to do within the literature (2, 16). Mycoplasma 93 penetrans, for example, is predominantly important because of its ability to penetrate the host 94 cells via an organelle specialised for host cell adherence (17). Mycoplasmas, likely owing to 95 their dependence on their hosts, have fastidious requirements for in vitro culture. Culture-free 96 approaches for microbial identification, especially, with the advent of DNA sequencing 97 approaches, have markedly increased in the recent years to identify new Mycoplasma-like 98 organisms(18–21). bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

99 100 Several studies have identified Mycoplasma from marine teleosts using culture-free 101 approaches. Mudsucker (Gillichthys mirabilis) and pinfish (Lagodon rhomboids), for example, 102 have been identified as having gut microbiomes rich in Mycoplasma (22). However, salmonids 103 in particular are frequently reported to be colonised by Mycoplasma (23, 24). This is 104 especially true in Atlantic salmon (Salmo salar), both in wild and in farmed settings and in 105 farmed settings (23, 25). In some cases, Mycoplasma phylotypes can comprise >70% of the 106 total microbial reads recovered from salmon intestines (24, 26).The distribution and biological 107 role of Mycoplasma in the intestines of salmonids is far from clear, and requires further 108 exploration. Nonetheless, demographic modelling of microbial communities suggest 109 colonisation of salmonid guts by these organism as non-neutral, i.e. the rate at which these 110 organisms colonise the gut, indicates a significant degree of specific adaptation to the host 111 environment (26, 27). 112 113 In the current study, we aimed to explore the characteristics of Mycoplasma in salmonids, 114 including cellular niche, taxonomic affiliations, genome structure and gene content. We 115 focused on the genetic features and metabolic functions which may help us to explain the role 116 of reductive evolution in the close association of the Mycoplasma with the host, through 117 different physiological and physio-chemical adaptations to survival within the digestive tract of 118 Salmo salar. We also explored the phylogenetic relatedness of the Mycoplasma in salmonids 119 as compared to all the known and sequenced mycoplasmas to date. 120 121 122 Materials and Methods 123 124 Sample collection 125 126 Farmed Atlantic salmon (Salmo salar) subadults (3 to 5 kg) were sampled from marine cages 127 at an aquaculture facility at Corran Ferry, near Fort William, Scotland, in Autumn 2017 in 128 collaborations with MOWI Ltd. Salmo salar freshwater parr and ova were sampled at the 129 Institute of Biodiversity, Animal Health and Comparative Medicine aquarium facility , 130 University of Glasgow. Animals were euthanised by blunt cranial trauma under a Schedule 1 131 procedure and gut compartments (stomach, pyloric caecum, and midgut) samples were flash 132 frozen in liquid nitrogen and stored in -80 C. 133 134 Fluorescence in-situ hybridisation (FISH) 135 136 Previous work has established the dominance of Mycoplasma in marine Salmo salar GI 137 (gastrointestinal tract) (26). To explore their physical distribution in different gut compartments 138 and life cycle stages, FISH was undertaken on salmon tissues. Samples were fixed in a 139 freshly made sterile-filtered solution of 4% paraformaldehyde in PBS (pH 7.4) for 16-24 hours 140 and maintained at room temperature for 16-48 hours. Fixed samples were then washed with 141 sterile-filtered PBS (pH 7.4) three times before being fixed the sample in 70% ethanol. 142 Samples were then gradually dehydrated in a series of ethanol-xylene-paraffin treatment 143 steps (28). Prior to sectioning, samples were embedded in paraffin and stored at 4℃. At least 144 four 3-4 µm sections were taken from each embedded tissue block, rehydrated in sterile 145 ddH20, and mounted on slides for pepsin treatment and straining. Pepsin treatment was 146 undertaken in a 0.05% pepsin solution and 0.01M HCL. Samples were DAPI stained to target 147 cell nuclei of host cells, and FISH probes were hybridised at 55℃ to the 16S rDNA small bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

148 subunit of bacterial cells. Multiple FISH probes labelled with Cy3 and Cy5 dyes were 149 deployed to distinguish Mycoplasma from other microbes present in samples (Table 1). To 150 improve the visualisation of non-Mycoplasma bacteria, multiple probes were deployed using 151 the same dye. A Mycoplasma (Myc1-1, Table 1) probe was designed based on Illumina 152 amplicon sequences based upon the most abundant operational taxonomic (OTU) sequence 153 identified in Adult Salmon that we identified in previous work (Heys, Cheaib et al 2020). 154 Samples were visualised at 20-30x magnification on a DeltaVision-Core microscope (Applied 155 Precision, GE), equipped with a CoolSNAP HQ camera (Photometrics) and operated with 156 SoftWoRx software (Applied Precision, GE). 157 158 DNA extraction, library annotation and sequencing 159 160 DNA was extracted from the pyloric caecum homogenate derived from a single individual on 161 which FISH analyses had identified the presence of Mycoplasma-like organisms, based on 162 their labelling with a targeted 16S probe. A sequence library for Illumina Next-Seq WGS 163 (whole genome shotgun) sequencing was prepared using a sonication protocol and a TruSeq 164 library protocol and sequencing adaptors. Sequencing was undertaken at the University of 165 Glasgow Polyomics facility. 166 167 Data preprocessing, assembly, binning and annotations 168 169 The short paired-end NextSeq Illumina reads (2 X 63 million reads) were preprocessed for 170 quality filtering using sickle V1.2 (https://github.com/najoshi/sickle). Decontamination of good 171 quality reads was performed by mapping reads against the Salmo salar genome (available at 172 NCBI sequence archive with the accession number GCF_000233375.1) using Deconseq V 173 0.4.3 (29) based on BWA mapper V 0.5.9 (30). The decontaminated paired-end reads (~18 174 million of bacterial reads) were assembled using the Megahit V1.1 software (31). The 175 assembled contigs (~93400 ) were processed for genomic binning using MetaBAT V2.12.1 176 (32). Quality assessment for completeness and contamination of sequence bins was 177 performed using CheckM V1.0.18 software (33). Annotation of gene content was performed 178 using the pipeline ATLAS-metagenome (34), which involves the prediction of open reading 179 frames (ORFs) using Prodigal (35). Translated gene products were clustered using linclust 180 (36) to generate non-redundant gene and protein catalogues, which were mapped to the 181 eggNOG catalogue (37) using DIAMOND (38). 182 183 184 Phylogenetic analyses 185 186 Two approaches were undertaken to construct phylogenetic trees: a) MLST-based (Multi 187 Locus Sequence Typing); and b) 16S gene markers (recovered from the genome) of the 188 Mycoplasma MAG (metagenome assembled genome) from this study as well as what is 189 previously available in the literature. Using CheckM software, the MLST-based strategy 190 focused on a concatenation of 21 conserved housekeeping genes annotated in the 191 Mycoplasma bin supplemented with the orthologues available for all the Mycoplasma genera, 192 to date. The MLST-based dataset included 55 orthologues of protein sequences of 193 concatenated 21 markers. The 16S rDNA sequence dataset included: one sequence of 16s 194 annotated in the Mycoplasma bin; five Operational Taxonomic Units (OTUs) sequences of 195 mycoplasmas characterised form the same farmed salmon system (26); 101 and 17 196 sequences of 16S rDNA from the Mycoplasma and Sprioplasma genomes respectively from bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

197 IMG database; and 11 sequences from environmental studies detected in marine species 198 including shrimp, fish and isopods. DNA and protein sequences were aligned using MAFFT 199 version 6.24 (39). Phylogenetic inference was performed using PhyML version 3.0 (40) and 200 MrBayes V.3.2.6 (41). The evolutionary model was chosen using MODELTEST(42), and 201 parameters were iteratively estimated in PhyML using the GTR+I+G model for nucleotide 202 sequence of 16s trees and the LG+I+G model for amino-acid sequence of concatenated 203 markers trees (43). Bootstrap values were calculated with 100 replicates (44). With MrBayes, 204 posterior probability values were calculated using an average standard deviation of partition 205 frequencies < 0.01 as a convergence diagnostic (45). MrBayes runs consisted of eight 206 simultaneous Markov chains, each with 1,000,000 generations, a subsampling frequency of 207 1000, and a burn-in fraction of 0.15. Trees were then visualized and adapted for presentation 208 in FigTree version1.4.3 as a graphical viewer of phylogenetic trees (http://tree.bio.ed.ac.uk). 209 210 211 Metabolic pathways comparison and genome reduction analysis 212 213 All pFam V.32 (comprehensive and accurate collection of protein domains and families) 214 annotations were predicted with Prodigal and analysed in terms of function categories and 215 metabolic content (focusing on Enzyme EC numbers). The 530 genes identified were 216 associated with 746 pFam functions. The pFam function led to the recovery of Gene Ontology 217 (GO) terms and were then mapped to the KEGG database. Simultaneously, the alternate 218 approach involving MetaCyc database was employed to elucidate metabolic pathways from 219 all domains of life (46) . The EC numbers of the coding sequence regions in Mycoplasma 220 penetrans was extracted from the KEGG database and was then compared with those 221 annotated within the MAG of Mycoplasma from Salmo salar in this study. The mapping of 222 metabolic pathways from both genomes were visualized using the iPath (47). From the IMG 223 genomic database, all available metadata on sequenced Mycoplasma strains were then 224 collected and compared to the MAG for the genome size, GC content, gene content and their 225 preference (e.g. intracellular, free-living etc). Annotations for the assembled Mycoplasma 226 genome were submitted to CG view (48) for radial visualisation of genomic structure of the 227 assembled Mycoplasma bin. The 570 predicted genes were compared at the DNA and protein 228 sequence levels against all the available genes of Mycoplasma penetrans using BLAST+ V 229 2.8.1 (49). Best hits for each query were represented in a radial plot using Circoletto software 230 version V.069-9 (50). Complimentary annotations were performed using RAST software 231 which, consisted of subsystem classification of microbial functions available in the curated 232 database, i.e. SEED subsystem (51). 233 234 235 Results 236 237 Fluorescence in situ hybridization (FISH) of mycoplasmas in the farmed salmon 238 239 The set of probes used in FISH for the identification of bacterial populations is summarized in 240 the Table 1. Only the Myc1-1 probe was shown to be specific to mycoplasmas (Supp. Figure 241 S1). FISH visualization in salmon ova demonstrated low abundance of bacteria and no signal 242 of mycoplasmas (Figure 1-a; Supp. Figure 2.1). In Salmo salar freshwater parr, Mycoplasma 243 aggregates were observed on the stomach lining (Figure 1-b; Supp. Figure 2.2), as well as 244 on the muscularis mucosae, and epithelium of the pyloric caecum (Figure 1-c; Supp. Figure 245 2.3). In the midgut of salmon parr (distal to the pyloric caecum) we found some evidence that bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

246 the Mycoplasma may be clustering intracellularly (Supp. Figure 2.4). In the stomach (Figure 247 1-d; Supp. Figure 2.5) and pyloric caecum (Figure 1-e-f; Supp. Figure 2.6) of adult salmon, 248 Mycoplasma signals were clearly clustered in small aggregates in the lumen around the 249 nuclei of epithelial cells. Figure 1 indicates this intracellular clustering most clearly. In the 250 midgut of adult salmon, Mycoplasma showed low abundance and the signals of Mycoplasma 251 showed aggregations near epithelium cell nuclei (Supp. Fig 2.7). The specificity of the 252 Mycoplasma probe against pure culture Escherichia coli and Mycoplasma Muris was 253 evaluated (Supp. Figure S1). The Mycoplasma probe Myc1-1 showed specific hybridization, 254 giving a positive signal solely with cultured Mycoplasma muris. We made multiple attempts in 255 both solid and liquid culture mycoplasmas from the salmon intestines, but without success. 256 257 Binned Mycoplasma genome features and orthologs 258 259 Using a total of 63,180,207 reads, and after decontamination, 93397 contigs were assembled 260 using megahit software (see materials and methods). The assembled contigs were binned, 261 annotated and assessed for completeness (see materials and methods). The best quality 262 assembled bins corresponded to a nearly complete genome assigned to Mycoplasma (see 263 bin sequences in Supp. File 1). The completeness of this metagenome-assembled genome 264 (MAG) was estimated at 98 % with 0.38 % of contamination and 0.0% of heterogeneity (Table 265 2). 266 267 The average size of the assembled genome was estimated to be 0.557 Mb and comprised a 268 set of 570 predicted genes accounting for a total of 694 CDS regions found on the 5’3’ and 269 3’5’ ORFs (supplementary file 2). The GC percentage was estimated to be 39.2% (Table 2). 270 Circular representation of the genomic structure of the MAG highlights CDS annotations (694) 271 on the negative (Figure S3-a) and positive (Figure S3-b) strands, respectively. To further 272 resolve CDS annotations, a supplementary annotation framework was applied using the 273 curated SEED database and the RAST server (52). The results showed 600 CDS across the 274 negative (275 CDS) and positive strands (325 CDS). Amongst these CDS regions (Supp. File 275 2 ), 390 had functional annotations, and within these, three annotated CDS regions (> 85% of 276 similarity threshold against SEED) were identified as: Riboflavin kinase (EC 2.7.1.26/EC 277 2.7.1.26 ;1278 bp) along with two Riboflavin/purine transporters of length 1383 bp and 1608 278 bp, respectively. In addition, other functions required for host-microbiota symbiosis, such as 279 ribonucleotide reductase, were annotated with SEED and are reported (Supp. Table 1). 280 281 282 Phylogenetic proximity to Mycoplasma penetrans 283 284 The recovered phylogenetic tree based on the MLST approach as well as 16S rDNA, 285 corroborated the same genetic relatedness of the assembled Mycoplasma to the closest 286 lineage represented by Mycoplasma penetrans. The 16S rDNA tree includes four OTUs of 287 Mycoplasma detected in the digestive tract of farmed salmon previously (26) .The 288 phylogenetic distances indicate that the Mycoplasma MAG is closer to the two OTUs from the 289 same farmed system as compared to M. penetrans (see sequence alignment in Supp. File 290 3). Furthermore, clusters containing these taxa are supported with medium to high posterior 291 probabilities (>0.5) according to the Bayesian approach (Figure 2). 292 293 To further ascertain the above clustering of 16S rDNA sequences of Mycoplasma, and the 294 phylogenetic relatedness, a second analysis based on MLST approach using 21 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

295 concatenated housekeeping genes (see PFAM IDs of markers and their functions in Supp. 296 File 4) increased our confidence in M. penetrans being close to the recovered Mycoplasma 297 MAG (Figure 3; see sequence alignment in Supp. File 5). These 21 markers are detected 298 in single copies and are conserved in the bacteria and the Mycoplasma lineage (). The MLST 299 tree shows high posterior probabilities in support of this argument (post prob > 0.9). Tip labels 300 of the selected Mycoplasma samples are further annotated with the genome size information 301 in Mbp. It should be noted that the genome sizes did not appear to correlate well with the 302 spatial distribution patterns of Mycoplasma species in the tree. The genome size of M. 303 penetrans (1.36 Mb) is approximately double to that of the binned Mycoplasma and, is the 304 highest amongst the Mycoplasma genomes. 305 306 Orthology, metabolic pathways and genome reduction analyses 307 308 A core genome analysis including the amino acid sequences of predicted CDS and all the 309 available CDS from the closely related M. penetrans, available on NCBI repository, were 310 blasted against the COG database. Circular track of the genome including the orthology 311 clearly showed the difference in genome size between the binned Mycoplasma and M. 312 penetrans. We observed heterogeneity across the different regions of the genomic structure 313 in terms of GC content and GC skew (Figure 4). An orthology analysis based on SEED 314 annotations indicated 14 functions (oxidative stress, periplasmic stress, protein biosynthesis, 315 detoxification, ribonuclease H, cation transporters, ABC transporters) specific to the binned 316 genome, 144 functions specific to M. penetrans, and 156 functions that are common to 317 Mycoplasma MAG and M. penetrans. The shared functions between these two genomes 318 belong to nine different general subsystems including those related to symbiosis and 319 intracellular lifestyles such as: riboflavin metabolism; intracellular resistance; and resistance 320 to antibiotics and toxic compounds (RATC) (Supp. Table 2). We only found two similarity hits 321 associated to RATC. Complimentary analysis pointed out a bifunctional riboflavin kinase/FMN 322 FMN adenylyltransferase among the best reciprocal similarity’s hits between the Mycoplasma 323 MAG in this study and Mycoplasma penetrans (Figure 5; Supp. Table 3). 324 325 To understand genome reduction in Mycoplasma lineage, the genome size and genes count 326 were compared across 247 strains (Figure 6-a; Supp. Table 4) of the available Mycoplasma 327 database (IMG) isolated from a wide variety of human and animal sources and comprising 328 both parasitic and commensal strains. In view of the collected data (Supp. Table 4), gene 329 content and genome size are strongly correlated. The average genome size of the 247 330 available mycoplasmas was 0.87 Mbp ± 0.15, and the average genes count was 790 ± 157 331 genes; however, this was not the case with all considered genomes. For instance, 8 genomes 332 are lower than 0.8Mb, accumulating somewhere between 829 and 1036 genes. Further 333 analysis revealed that pseudogenes count was not a significant factor, whilst both the 334 transmembrane proteins and GC content were correlated with Mycoplasma genome sizes 335 (Figure 6-b). Furthermore, the average count of pseudogenes was significantly higher in free- 336 living than within intracellular mycoplasmas (Supp. Figure S4), although available databases 337 contain incomplete information with regards to Mycoplasma lifestyles. Enzyme content was 338 analyzed in terms of metabolic pathways by comparing the annotated EC numbers of the 339 Mycoplasma MAG and M. penetrans. Common pathways of both genomes are highlighted in 340 red lines (Figure 7) and include the riboflavin biosynthesis pathway. 341 342 343 Discussion bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

344 345 Mycoplasma are hyper-abundant commensals of salmonid guts. Our study suggests, based 346 on FISH data, that in Salmo salar, these organisms grow intracellularly in the epithelial and 347 possibly muscular lining of the fish’s GI tract, both in freshwater and marine lifecycle stages. 348 Mycoplasma sequences recovered from Salmo salar, including the Mycoplasma MAG 349 reported here, had a strong phylogenetic closeness to M. penetrans. Comparative analysis of 350 genome size and content across Mycoplasma strains suggest that the genome we recovered 351 in this study is among the smallest ever observed, to the best of our knowledge. Comparative 352 genomics analyses between the Mycoplasma MAG and M. penetrans were undertaken and 353 provide insight into the potential host-microbe interaction. 354 355 Mycoplasmas have been widely reported within Salmo salar (23), and other teleosts (27). It is 356 not uncommon to find that communities of gut microorganisms are dominated by 357 mycoplasmas (53). The modelling approaches comparing environmental and intestinal 358 frequency distributions of these organisms undertaken in this study have previously 359 suggested that salmon mycoplasmas are well adapted to colonisation of their hosts (26). On 360 the flip side, culture-based approaches have had been less successful in isolating these 361 organisms (54) and despite numerous attempts, we failed to obtain pure cultures of 362 Mycoplasma from the adult salmon used in this study (data not shown). This may be 363 attributed to a potential source of bias arising from cell wall deficiency (55) in mycosplasmas 364 which decrease their growth in presence of inhibitors such as nucleoside and nucleobase 365 analogs as demonstrated in Mycoplasma pneumoniae (56) and others mycoplasmas (57). 366 FISH data from the current study, however, indicate that many mycoplasmas are sequestered 367 within the basal the epithelial cells, suggesting potential unknown parameters in symbiosis 368 with Salmo salar which were missed from the culture media trials and reduce their cultivability. 369 Although only a qualitative assessment is possible by employing FISH, our data suggest that 370 Mycoplasma comprised the majority of the resident microbes (Figure 2), and the binding 371 efficiency between Mycoplasma and the universal probes differed only slightly in the cultured 372 controls (Supp Figure S1). 373 374 The intracellular niche of in the epithelial cells of the GI tract of salmon may represent a 375 strategy to avoid hydrolysis in the intestinal environment. Sequestration in the muscularis 376 under the basal of epithelium, as we observed in the stomach, is thought to offer a protective 377 niche in the case of invading Helicobacter pylori in human stomach diseases (58). Also similar 378 to H. pylori, our functional annotation of the Mycoplasma MAG (Supp Table 1) identified 379 components of the urea cycle subsystem that may assist acid tolerance via the hydrolysis of 380 urea to ammonia (59). Further investigations of gene complexes involved in acid 381 neutralization in the Mycoplasma could potentially be achieved via targeted transcriptomics. 382 383 A high level of adaptation to, and dependence on the host organism, is a key feature of many 384 Mycoplasma species (60). The exploitation of an intracellular niche, dependence on the host, 385 and relative isolation from the other microorganisms and mobile genetic elements are thought 386 to have contributed to genome decay in mycoplasmas (61) . One result of this decay is a 387 reduction in genome size and the number of genes, and the Mycoplasma MAG in this study 388 appears to have been potentially affected by such processes in comparison to the other 389 mycoplasmas (Figure 5). According the phylogenetic tree, we did not observe any specific 390 relationship between the tree topology and the genome sizes of mycoplasmas (Figure 2). 391 Indeed, the closely related M. penetrans was over three times larger than the size of the 392 Mycoplasma MAG in this study. Despite sharing a recent ancestor with the human pathogen bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

393 M. penetrans, a long and close evolutionary association of this Mycoplasma strain and 394 salmonids is possible given the similarity of another Mycoplasma MAG sourced from the 395 Norwegian sea salmon and identified to M. penetrans (62). We were also able to identify 396 Mycoplasma in freshwater parr via the FISH method in this study. One potential route for 397 vertical transmission of the Mycoplasma among salmon could be observed during oviposition. 398 We were not able to identify microbes colonising eggs in this study, although our sample size 399 was limited. Further development on specific Mycoplasma strain markers could potentially 400 reveal their abundance as well as their epidemiology, and potential routes of inter and intra 401 generational transmission. 402 403 Many well-characterised mycoplasmas are pathogens (17, 63, 64), with several Mycoplasma 404 sp. being responsible for human, animal and plant diseases; however some species are 405 considered to be commensal organisms (2, 65). The role of Mycoplasma in the context of 406 Salmo salar is not well established. Koch’s postulates were not applied in this study (66). 407 Given the challenges encountered in culturing these microorganisms, it seems quite likely that 408 they may never be applied. Furthermore, the apparent abundance of Mycoplasma in the 409 healthy salmonids (23, 24, 26) , and lack of any clear associated pathology in gut tissues, 410 implies that there is not a significant impact on the host health or fitness. Commensal 411 exploitation of the host intracellular niche is potentially the most parsimonious description of 412 the host-microbe interaction. The ultimate metabolic adaptation to an intracellular lifestyle (i.e 413 Buchnera, Wigglesworthia and Blochmannia) appears to be solely regulated by the metabolic 414 activity of the host cells to which the bacteria may actively contribute to, by delivering 415 essential metabolites that are limited in their habitats and are not produced by the hosts (67). 416 Nonetheless, the presence of an apparently complete riboflavin pathway could potentially 417 indicate benefit from the host perspective of Mycoplasma colonisation. Riboflavin, known as 418 the precursor for the cofactors flavin mononucleotide (FMN) and flavin adenine dinucleotide, 419 is an essential metabolite in organisms (67, 68), although vertebrates cannot synthesise it on 420 their own (69). The Mycoplasma likely plays a role in riboflavin supplementation in salmon, as 421 observed in several deep-sea snailfish (70). Riboflavin supplementation is not limited to the 422 mycoplasmas; in the bedbug Cimex lectularius, the Gram-negative Wolbachia can synthesize 423 biotin and riboflavin which, are crucial for the host growth and reproduction (71, 72). 424 Riboflavin biosynthesis is common for symbiotic associations and therefore occurs even in 425 small and optimized genomes size like Wolbachia (~ 1.48 Mb) and Mycoplasma (0.51 ~1,38 426 Mb). 427 428 The small genomes sizes reduce the metabolic capabilities of Mycoplasma, although they can 429 utilise sugars as a source of carbon and energy via glycolysis (Halbedel 2007). The simple 430 sugar pentose phosphate pathway and genes encoding enzymes of the tricarboxylic acid 431 cycle are absent from Mycoplasma genomes. So, sugars are transported by the 432 phosphotransferase system which play a regulatory role via the regulatory function of HPr 433 kinase/phosphorylase, which, was also annotated in the assembled MAG Mycoplasma in this 434 study (Supp. Table 1). This carbon source is available in the epithelial tissues of hosts and its 435 metabolism leads to the formation of hydrogen peroxide, the major virulence factor of several 436 species of Mollicutes (Halbedel 2007). Up until this point, the functions detected in the 437 Mycoplasma MAG, such as riboflavin, and phosphotransferase, support the argument for the 438 strong symbiotic association of mycoplasmas with the host epithelial tissues, however, we 439 were unable to find virulence factors in this study. 440 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

441 Also, it is reported that many Mycoplasma species can modify their surface antigenic 442 molecules with high frequency (64, 73) which may likely play a key role in outmaneuvering the 443 host immune system. This ability may generate phenotypic heterogeneity in colonising 444 Mycoplasma populations and provide fitness benefits such as evasion of host immune 445 responses and to adaptation to the environmental changes (73, 74). The majority of the 446 variable surface antigenic molecules of mycoplasmas are lipoproteins (74–76), which, 447 depending on the species, are encoded by single or multiple genes (64, 77). The expression 448 of these lipoproteins, due to extensive antigenic variation, is thought to be a major factor for 449 immune evasion, for example the P35 lipoprotein and its paralogs, which are distributed 450 across the surface of M. penetrans cells, are immunodominant (78–80). Two lipoprotein 451 encoding genes were found only in Mycoplasma penetrans but not in the Mycoplasma’s MAG 452 in this study (Supp. Table 2). 453 454 455 Our study clearly demonstrates a potentially important ecological and functional association 456 between Mycoplasma and Salmo salar that merits further investigation. This association 457 probably reflects a mutualism rather than commensalism. Targeted meta-transcriptomics and 458 strain-specific screening for this organism could improve our understanding of its role, biology, 459 and function. Furthermore, targeted studies involving genome reduction and their association 460 with the host dynamics are also necessary to fully understand the evolution of Mycoplasma 461 symbiosis in Salmo salar. Further Omics investigations are also needed to assess the 462 population genomics of mycoplasmas associated with Atlantic salmon to identify genetic 463 variants in antibiotic resistant genes and the evolution of riboflavin pathway. New analytical 464 genomic tools that can reveal further insights as well as bespoke experiments driven by the 465 recommendations given in this study may possibly lead to the development of practices that 466 can improve aquaculture industry especially whether we can demonstrate in the future a 467 probiotic potential of mycoplasmas in salmonids. 468 469 470 References 471 472 1. Razin S. 1992. Peculiar properties of mycoplasmas: The smallest self-replicating prokaryotes.

473 FEMS Microbiol Lett 100:423–431.

474 2. Razin S, Yogev D, Naot Y. 1998. Molecular Biology and Pathogenicity of Mycoplasmas.

475 Microbiol Mol Biol Rev 62:1094–1156.

476 3. Heilmann C, Jensen L, Jensen JS, Lundstrom K, Windsor D, Windsor H, Webster D. 2001.

477 Treatment of resistant mycoplasma infection in immunocompromised patients with a new

478 pleuromutilin antibiotic. J Infect 43:234–238. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

479 4. Preiswerk B, Imkamp F, Vorburger D, Hömke RV, Keller PM, Wagner K. 2020. Mycoplasma

480 penetrans bacteremia in an immunocompromised patient detected by metagenomic sequencing: a

481 case report. BMC Infect Dis 20:7.

482 5. Baseman JB, Tully JG. Mycoplasmas: Sophisticated, Reemerging, and Burdened by Their

483 Notoriety - Volume 3, Number 1—March 1997 - Emerging Infectious Diseases journal - CDC.

484 6. Bové JM. 1993. Molecular Features of Mollicutes. Clin Infect Dis 17:S10–S31.

485 7. Miyata M, Ogaki H. 2006. Cytoskeleton of Mollicutes. J Mol Microbiol Biotechnol 11:256–264.

486 8. Trachtenberg S. 2005. Mollicutes. Curr Biol 15:R483–R484.

487 9. Citti C, Blanchard A. 2013. Mycoplasmas and their host: emerging and re-emerging minimal

488 pathogens. Trends Microbiol 21:196–203.

489 10. Hutchison CA, Chuang R-Y, Noskov VN, Assad-Garcia N, Deerinck TJ, Ellisman MH, Gill J,

490 Kannan K, Karas BJ, Ma L, Pelletier JF, Qi Z-Q, Richter RA, Strychalski EA, Sun L, Suzuki Y,

491 Tsvetanova B, Wise KS, Smith HO, Glass JI, Merryman C, Gibson DG, Venter JC. 2016. Design

492 and synthesis of a minimal bacterial genome. Science 351:aad6253.

493 11. Fadiel A, Eichenbaum KD, El Semary N, Epperson B. 2007. Mycoplasma genomics: tailoring the

494 genome for minimal life requirements through reductive evolution. Front Biosci J Virtual Libr

495 12:2020–2028.

496 12. Rocha EPC, Blanchard A. 2002. Genomic repeats, genome plasticity and the dynamics of

497 Mycoplasma evolution. Nucleic Acids Res 30:2031–2042. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

498 13. Boscaro V, Kolisko M, Felletti M, Vannini C, Lynn DH, Keeling PJ. 2017. Parallel genome

499 reduction in symbionts descended from closely related free-living bacteria. 8. Nat Ecol Evol

500 1:1160–1167.

501 14. Moran NA. 1996. Accelerated evolution and Muller’s rachet in endosymbiotic bacteria. Proc Natl

502 Acad Sci 93:2873–2878.

503 15. Itoh T, Martin W, Nei M. 2002. Acceleration of genomic evolution caused by enhanced mutation

504 rate in endocellular symbionts. Proc Natl Acad Sci 99:12944–12948.

505 16. Yavlovich A, Tarshis M, Rottem S. 2004. Internalization and intracellular survival of Mycoplasma

506 pneumoniae by non-phagocytic cells. FEMS Microbiol Lett 233:241–246.

507 17. Sasaki Y, Ishikawa J, Yamashita A, Oshima K, Kenri T, Furuya K, Yoshino C, Horino A, Shiba T,

508 Sasaki T, Hattori M. 2002. The complete genomic sequence of Mycoplasma penetrans, an

509 intracellular bacterial pathogen in humans. Nucleic Acids Res 30:5293–5300.

510 18. Aceves AK, Johnson P, Bullard SA, Lafrentz S, Arias CR. 2018. Description and characterization

511 of the digestive gland microbiome in the freshwater mussel Villosa nebulosa (Bivalvia:

512 Unionidae). J Molluscan Stud 84:240–246.

513 19. Bokma J, Pardon B, Deprez P, Haesebrouck F, Boyen F. 2020. Non-specific, agar medium-related

514 peaks can result in false positive Mycoplasma alkalescens and Mycoplasma arginini identification

515 by MALDI-TOF MS. Res Vet Sci 130:139–143.

516 20. Costello EK, Carlisle EM, Bik EM, Morowitz MJ, Relman DA. 2013. Microbiome Assembly

517 across Multiple Body Sites in Low-Birthweight Infants. mBio 4. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

518 21. Martin DH, Zozaya M, Lillis RA, Myers L, Nsuami MJ, Ferris MJ. 2013. Unique Vaginal

519 Microbiota That Includes an Unknown Mycoplasma-Like Organism Is Associated With

520 Trichomonas vaginalis Infection. J Infect Dis 207:1922–1931.

521 22. Egerton S, Culloty S, Whooley J, Stanton C, Ross RP. 2018. The Gut Microbiota of Marine Fish.

522 Front Microbiol 9.

523 23. Holben WE, Williams P, Saarinen M, Särkilahti LK, Apajalahti JHA. 2002. Phylogenetic Analysis

524 of Intestinal Microflora Indicates a Novel Mycoplasma Phylotype in Farmed and Wild Salmon.

525 Microb Ecol 44:175–185.

526 24. Llewellyn MS, McGinnity P, Dionne M, Letourneau J, Thonier F, Carvalho GR, Creer S, Derome

527 N. 2016. The biogeography of the atlantic salmon (Salmo salar) gut microbiome. ISME J

528 10:1280–1284.

529 25. Zarkasi KZ, Taylor RS, Abell GCJ, Tamplin ML, Glencross BD, Bowman JP. 2016. Atlantic

530 Salmon (Salmo salar L.) Gastrointestinal Microbial Community Dynamics in Relation to Digesta

531 Properties and Diet. Microb Ecol 71:589–603.

532 26. Heys C, Cheaib B, Busetti A, Kazlauskaite R, Maier L, Sloan WT, Ijaz UZ, Kaufmann J,

533 McGinnity P, Llewellyn MS. 2020. Neutral Processes Dominate Microbial Community Assembly

534 in Atlantic Salmon, Salmo salar. Appl Environ Microbiol 86.

535 27. Cheaib B, Seghouani H, Ijaz UZ, Derome N. 2020. Community recovery dynamics in yellow

536 perch microbiome after gradual and constant metallic perturbations. Microbiome 8:14.

537 28. Copper JE, Budgeon LR, Foutz CA, van Rossum DB, Vanselow DJ, Hubley MJ, Clark DP,

538 Mandrell DT, Cheng KC. 2018. Comparative analysis of fixation and embedding techniques for bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

539 optimized histological preparation of zebrafish. Comp Biochem Physiol Toxicol Pharmacol CBP

540 208:38–46.

541 29. Schmieder R, Edwards R. 2011. Fast Identification and Removal of Sequence Contamination

542 from Genomic and Metagenomic Datasets. PLoS ONE 6:e17288.

543 30. Li H. 2013. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM.

544 ArXiv13033997 Q-Bio.

545 31. Li D, Liu C-M, Luo R, Sadakane K, Lam T-W. 2015. MEGAHIT: an ultra-fast single-node

546 solution for large and complex metagenomics assembly via succinct de Bruijn graph.

547 Bioinformatics 31:1674–1676.

548 32. Kang DD, Froula J, Egan R, Wang Z. 2015. MetaBAT, an efficient tool for accurately

549 reconstructing single genomes from complex microbial communities. PeerJ 3:e1165.

550 33. Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. 2015. CheckM: assessing the

551 quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome

552 Res 25:1043–1055.

553 34. Kieser S. 2019. ATLAS: a Snakemake workflow for assembly, annotation, and genomic binning

554 of metagenome sequence data. Academic.

555 35. Hyatt D, Chen G-L, LoCascio PF, Land ML, Larimer FW, Hauser LJ. 2010. Prodigal: prokaryotic

556 gene recognition and translation initiation site identification. BMC Bioinformatics 11:119.

557 36. Steinegger M, Söding J. 2018. Clustering huge protein sequence sets in linear time. 1. Nat

558 Commun 9:2542. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

559 37. Huerta-Cepas J, Szklarczyk D, Heller D, Hernández-Plaza A, Forslund SK, Cook H, Mende DR,

560 Letunic I, Rattei T, Jensen LJ, von Mering C, Bork P. 2019. eggNOG 5.0: a hierarchical,

561 functionally and phylogenetically annotated orthology resource based on 5090 organisms and

562 2502 viruses. Nucleic Acids Res 47:D309–D314.

563 38. Buchfink B, Xie C, Huson DH. 2015. Fast and sensitive protein alignment using DIAMOND. Nat

564 Methods 12:59–60.

565 39. Katoh K, Standley DM. 2013. MAFFT Multiple Sequence Alignment Software Version 7:

566 Improvements in Performance and Usability. Mol Biol Evol 30:772–780.

567 40. Guindon S, Gascuel O. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies

568 by maximum likelihood. Syst Biol 52:696–704.

569 41. Huelsenbeck JP, Ronquist F. 2001. MRBAYES: Bayesian inference of phylogenetic trees.

570 Bioinformatics 17:754–755.

571 42. Posada D, Crandall KA. 1998. MODELTEST: testing the model of DNA substitution. Bioinforma

572 Oxf Engl 14:817–818.

573 43. Le SQ, Gascuel O. 2008. An improved general amino acid replacement matrix. Mol Biol Evol

574 25:1307–1320.

575 44. Felsenstein J. 1985. CONFIDENCE LIMITS ON PHYLOGENIES: AN APPROACH USING

576 THE BOOTSTRAP. Evol Int J Org Evol 39:783–791.

577 45. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, Larget B, Liu L, Suchard

578 MA, Huelsenbeck JP. 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model

579 choice across a large model space. Syst Biol 61:539–542. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

580 46. Caspi R, Billington R, Fulcher CA, Keseler IM, Kothari A, Krummenacker M, Latendresse M,

581 Midford PE, Ong Q, Ong WK, Paley S, Subhraveti P, Karp PD. 2018. The MetaCyc database of

582 metabolic pathways and enzymes. Nucleic Acids Res 46:D633–D639.

583 47. Yamada T, Letunic I, Okuda S, Kanehisa M, Bork P. 2011. iPath2.0: interactive pathway explorer.

584 Nucleic Acids Res 39:W412-415.

585 48. Grant JR, Stothard P. 2008. The CGView Server: a comparative genomics tool for circular

586 genomes. Nucleic Acids Res 36:W181–W184.

587 49. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. J

588 Mol Biol 215:403–410.

589 50. Darzentas N. 2010. Circoletto: visualizing sequence similarity with Circos. Bioinformatics

590 26:2620–2621.

591 51. Overbeek R, Olson R, Pusch GD, Olsen GJ, Davis JJ, Disz T, Edwards RA, Gerdes S, Parrello B,

592 Shukla M, Vonstein V, Wattam AR, Xia F, Stevens R. 2014. The SEED and the Rapid Annotation

593 of microbial genomes using Subsystems Technology (RAST). Nucleic Acids Res 42:D206–D214.

594 52. Glass EM, Wilkening J, Wilke A, Antonopoulos D, Meyer F. 2010. Using the Metagenomics

595 RAST Server (MG-RAST) for Analyzing Shotgun Metagenomes. Cold Spring Harb Protoc

596 2010:pdb.prot5368.

597 53. Dehler CE, Secombes CJ, Martin SAM. 2017. Seawater transfer alters the intestinal microbiota

598 profiles of Atlantic salmon ( Salmo salar L.). 1. Sci Rep 7:13877. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

599 54. Llewellyn MS, Boutin S, Hoseinifar SH, Derome N. 2014. Teleost microbiomes: the state of the

600 art in their characterization, manipulation and importance in aquaculture and fisheries. Front

601 Microbiol 5.

602 55. Razin S. 1995. MOLECULAR PROPERTIES OF MOLLICUTES: A SYNOPSIS, p. 1–25. In

603 Razin, S, Tully, JG (eds.), Molecular and Diagnostic Procedures in Mycoplasmology. Academic

604 Press, San Diego.

605 56. Sun R, Wang L. 2013. Inhibition of Mycoplasma pneumoniae growth by FDA-approved

606 anticancer and antiviral nucleoside and nucleobase analogs. BMC Microbiol 13:184.

607 57. Wehelie R, Eriksson S, Wang L. 2004. Effect of fluoropyrimidines on the growth of Ureaplasma

608 urealyticum. Nucleosides Nucleotides Nucleic Acids 23:1499–1502.

609 58. Haesebrouck F, Pasmans F, Flahou B, Chiers K, Baele M, Meyns T, Decostere A, Ducatelle R.

610 2009. Gastric Helicobacters in Domestic Animals and Nonhuman Primates and Their Significance

611 for Human Health. Clin Microbiol Rev 22:202–223.

612 59. Jones MD, Li Y, Zamble DB. 2018. Acid-responsive activity of the Helicobacter pylori

613 metalloregulator NikR. Proc Natl Acad Sci 115:8966–8971.

614 60. Faucher M, Nouvel L-X, Dordet-Frisoni E, Sagné E, Baranowski E, Hygonenq M-C, Marenda M-

615 S, Tardy F, Citti C. 2019. Mycoplasmas under experimental antimicrobial selection: The

616 unpredicted contribution of horizontal chromosomal transfer. PLoS Genet 15:e1007910.

617 61. Sirand-Pugnet P, Lartigue C, Marenda M, Jacob D, Barré A, Barbe V, Schenowitz C, Mangenot S,

618 Couloux A, Segurens B, de Daruvar A, Blanchard A, Citti C. 2007. Being Pathogenic, Plastic, and

619 Sexual while Living with a Nearly Minimal Bacterial Genome. PLoS Genet 3. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

620 62. Jin Y, Angell IL, Sandve SR, Snipen LG, Olsen Y, Rudi K. 2019. Atlantic salmon raised with diets

621 low in long-chain polyunsaturated n-3 fatty acids in freshwater have a Mycoplasma-dominated

622 gut microbiota at sea. Aquac Environ Interact 11:31–39.

623 63. Meseguer MA, Álvarez A, Rejas MT, Sánchez C, Pérez-Dı́az JC, Baquero F. 2003. Mycoplasma

624 pneumoniae: a reduced-genome intracellular bacterial pathogen. Infect Genet Evol 3:47–55.

625 64. Rosengarten R, Citti C, Glew M, Lischewski A, Droeße M, Much P, Winner F, Brank M, Spergser

626 J. 2000. Host-pathogen interactions in mycoplasma pathogenesis: Virulence and survival

627 strategies of minimalist prokaryotes. Int J Med Microbiol 290:15–25.

628 65. Siqueira FM, Thompson CE, Virginio VG, Gonchoroski T, Reolon L, Almeida LG, da Fonsêca

629 MM, de Souza R, Prosdocimi F, Schrank IS, Ferreira HB, de Vasconcelos ATR, Zaha A. 2013.

630 New insights on the biology of swine respiratory tract mycoplasmas from a comparative genome

631 analysis. BMC Genomics 14:175.

632 66. Falkow S. 2004. Molecular Koch’s postulates applied to bacterial pathogenicity — a personal

633 recollection 15 years later. 1. Nat Rev Microbiol 2:67–72.

634 67. Fuchs TM, Eisenreich W, Heesemann J, Goebel W. 2012. Metabolic adaptation of human

635 pathogenic and related nonpathogenic bacteria to extra- and intracellular habitats. FEMS

636 Microbiol Rev 36:435–462.

637 68. Gutiérrez-Preciado A, Torres AG, Merino E, Bonomi HR, Goldbaum FA, García-Angulo VA.

638 2015. Extensive Identification of Bacterial Riboflavin Transporters and Their Distribution across

639 Bacterial Species. PLOS ONE 10:e0126124. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

640 69. Vitreschak AG, Rodionov DA, Mironov AA, Gelfand MS. 2002. Regulation of riboflavin

641 biosynthesis and transport genes in bacteria by transcriptional and translational attenuation.

642 Nucleic Acids Res 30:3141–3151.

643 70. Lian C-A, Yan G-Y, Huang J-M, Danchin A, Wang Y, He L-S. 2020. Genomic Characterization of

644 a Novel Gut Symbiont From the Hadal Snailfish. Front Microbiol 10.

645 71. Kubiak K, Sielawa H, Chen W, Dzika E. 2018. Endosymbiosis and its significance in

646 dermatology. J Eur Acad Dermatol Venereol 32:347–354.

647 72. Moriyama M, Nikoh N, Hosokawa T, Fukatsu T. 2015. Riboflavin Provisioning Underlies

648 Wolbachia’s Fitness Contribution to Its Insect Host. mBio 6.

649 73. Horino A, Sasaki Y, Sasaki T, Kenri T. 2003. Multiple Promoter Inversions Generate Surface

650 Antigenic Variation in Mycoplasma penetrans. J Bacteriol 185:231–242.

651 74. Halbedel S, Hames C, Stülke J. 2007. Regulation of Carbon Metabolism in the Mollicutes and Its

652 Relation to Virulence. J Mol Microbiol Biotechnol 12:147–154.

653 75. Chambaud I, Wróblewski H, Blanchard A. 1999. Interactions between mycoplasma lipoproteins

654 and the host immune system. Trends Microbiol 7:493–499.

655 76. Wise KS. 1993. Adaptive surface variation in mycoplasmas. Trends Microbiol 1:59–63.

656 77. Rosengarten R, Wise KS. 1990. Phenotypic switching in mycoplasmas: phase variation of diverse

657 surface lipoproteins. Science 247:315–318.

658 78. Distelhorst SL, Jurkovic DA, Shi J, Jensen GJ, Balish MF. 2017. The Variable Internal Structure

659 of the Mycoplasma penetrans Attachment Organelle Revealed by Biochemical and Microscopic

660 Analyses: Implications for Attachment Organelle Mechanism and Evolution. J Bacteriol 199. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

661 79. Neyrolles O, Eliane J-P, Ferris S, Ayr Florio da Cunha R, Prevost M-C, Bahraoui E, Blanchard A.

662 1999. Antigenic characterization and cytolocalization of P35, the major Mycoplasma penetrans

663 antigen. Microbiol Read Engl 145 ( Pt 2):343–355.

664 80. Wang RY-H, Hayes MM, Wear DJ, Lo SC, Shih JW-K, Alter HJ, Grandinetti T, Pierce PF. 1992.

665 High frequency of antibodies to Mycoplasma penetrans in HIV-infected patients. The Lancet

666 340:1312–1316.

667 668 669 670 Acknowledgements 671 672 This research was supported in part by a research grant from Biotechnology and Biological 673 Sciences Research Council (BBSRC) grant number BB/P001203/1, Science Foundation 674 Ireland, the Marine Institute and the Dept. for the Economy, N. Ireland, under the Investigators 675 Programme Grant No. SFI/15/IA/3028, and the Scottish Aquaculture Innovation Centre. The 676 authors gratefully acknowledge the Glasgow Imaging and PolyOmics Facility for their support 677 & assistance in this work’. Umer Zeeshan Ijaz is supported by NERC Independent Research 678 Fellowship NERC NE/L011956/1 as well as Lord Kelvin Adam Smith Leadership Fellowship 679 (Glasgow). 680 681 682 Author Contributions 683 684 ML conceived and designed the study. PY, RK, MDN, TD, CH, PS, and EL collected the 685 samples, performed FISH experiments, and analysed image data. BC performed the 686 bioinformatics and interpreted the results. BC and ML wrote the original draft of the 687 manuscript. All authors contributed to revisions of the manuscript. 688 689 Legends of Figures and tables 690 691 Figure 1. FISH visualization of Mycoplasma in acidic gastrointestinal tracts of salmon 692 Parr and adults. The images were an overlay of DAPI signals (blue), hybridization signals of 693 Gam-1, FIR-1, EUB338, EUB338 II, EUB338 III probes (Cy5, red) and Mycoplasma Specific 694 Myc1-1 probe (Cy3, orange). (A) Low Mycoplasma distribution in Salmon egg, the image is an 695 overlay of Cy3, Cy5, and DAPI filter set showed bright signal scaled at 10 μm.(B) Distribution 696 of Mycoplasma in the stomach of Salmon Parr, scaled at 10 μm. Orange signals indicate 697 mycoplasmas were clustered in small groups. (C) Distribution of Mycoplasma in the 698 epithelium of Pyloric caecum of Salmon Parr at 10 μm. (D) Distribution of Mycoplasma in 699 Stomach of adult Salmon scaled at 50 μm. (E, F) Distribution of Mycoplasma in Pyloric 700 caecum of adult salmon scaled at 10 μm (E) and 5μm (F) respectively. Mycoplasmas signals 701 were aggregated on the muscularis mucosae, lamina propria (E) and clustered in high bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

702 abundance around epithelial cells nuclei (F). (*) indicatesa is frame of micrograph detail. 703 Horizontal bars indicate that these images are scaled at 10 μm for (A, B, C, D,E ), and 5 μm 704 for (F). 705 706 Figure 2. Phylogenetic tree based on 16S rRNA gene sequences of mycoplasmas. Sequence 707 name abbreviation of tree tips labels and clade A (including spiroplasmas) are reported in 708 Supplementary File 6. 709 710 Figure 3. Phylogenetic tree of mycoplasmas based on 21 MLST markers with the details 711 given in the supplementary data. Sequence name abbreviation of tree tips labels are 712 explained in Supplementary File 7. 713 714 Figure 4. Circular track of the core genomes. This figure highlights core genes shared 715 between the Mycoplasma MAG from this study and related Mycoplasma penetrans species. 716 717 Figure 5. Circular track of best BLAST similarities. This figure highlights orthologous 718 CDS regions between the Mycoplasma MAG from this study and Mycoplasma Penetrans. 719 The orthologous CDS (WP.011077747.1) labelled in red encodes for bifunctional riboflavin 720 kinase/FMN adenylyltransferase in Mycoplasma penetrans. See abbreviations of orthologs 721 from this figure in Supp.Table 3. 722 723 Figure 6. Genomic features of Mycoplasmas. (A) Plot of genome size and genes count in 724 the Myoplsamas lineage. (B) Plot of functions and GC contents against genome size in the 725 Mycoplasmas lineage 726 727 Figure 7. Pairwise Metabolic pathways comparison of Mycoplasma MAG from this 728 study and related Mycoplasma penetrans species. Red color represents shared and 729 conserved pathways between the two genomes, whereas blue color represents the metabolic 730 pathways of the Mycoplasma MAG from this study and the green color represent the 731 metabolic pathways of Mycoplasma penetrans. 732 733 Table 1. Summary of FISH probes used in this study 734 735 Table 2. Summary of mycoplasma metagenome assembled genome (MAG) features. 736 737 738 Legends of Supplementary Figures and Tables 739 740 Supp. Figure S1. Evaluation of the specificity of Mycoplasma probe against pure culture 741 Escherichia coli and Mycoplasma muris. Mycoplasma probe Myc1-1 showed specific 742 hybridization, giving a positive signal solely with Mycoplasma muris. 743 744 Supp. Figure S2.1. The distribution of bacteria in farmed Salmon egg. 745 Very low abundance of bacteria was observed. Sections of egg samples were hybridized with 746 Myc1-1 probe (Cy3, orange), Gam-1, FIR-1, EUB338, EUB338 II, EUB338 III probes (Cy5, 747 red) and stained with DAPI (blue). a) Overlay of Cy3, Cy5, and DAPI filter set showed bright 748 signal of Mycoplasma in Salmon egg, scaled at 10 μm. b) No signal of Mycoplasma was 749 detected in egg samples, scaled at 50 μm. 750 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

751 Supp. Figure S2.2. Distribution of Mycoplasma in the stomach of Salmon Parr. 752 (a) Image of DAPI signals (blue), (b) Gam-1, FIR-1, EUB338, EUB338 II, EUB338 III probes 753 with Cy5 (red), (c) Image of Mycoplasma Specific probe Cy-3(orange), (d) Overlay image of 754 all channels, (e) and (f), Mycoplasma signals were clustered in small groups. These are 755 scaled at 100 μm for (a,b,c,d,) 20 μm for (e), and 10 μm for (f). 756 757 758 Supp. Figure S2.3. Distribution of Mycoplasma in the Pyloric caecum of Salmon Parr. 759 The images were an overlay of DAPI signals (blue), hybridization signals of Gam-1, FIR-1, 760 EUB338, EUB338 II, EUB338 III probes (Cy5, red) and Mycoplasma Specific Myc1-1 probe 761 (Cy3, orange). (a) Signals of Mycoplasma aggregates on the muscularis mucosae, (b) 762 epithelium of pyloric caecum. These are scaled at 10 μm for (a) and 10 μm for (b). 763 764 Supp. Figure S2.4. Distribution of Mycoplasma in the midgut of Salmon Parr. Overlay of 765 all channels. (a), Mycoplasma showed low abundance with small amount bacteria detected in 766 midgut sections. (b), Signals of Mycoplasma aggregated near epithelium cell nuclei. These 767 are scaled at 20 μm for (a) and 10 μm for (b). 768 769 Supp. Figure S2.5. Distribution of Mycoplasma in Stomach of adult Salmon. (a), DAPI, 770 (b), MycI probe, (c) all other bacteria probes, and (d) all layers overlaid. More signals were 771 observed using universal probes in b than c using Mycoplasma specific probe. These are 772 scaled at 50 μm. 773 774 Supp. Figure S2.6. Distribution of Mycoplasma in Pyloric caecum of adult Salmon. 775 Signals of bacterial aggregates on the muscularis mucosae, lamina propria (a, arrows) and 776 epithelium (b) of the pyloric caecum. (c) and (d), overlay of all channels and differential 777 interference contrast (DIC) image showed Mycoplasma signals were clustered in high 778 abundance around epithelium cell nuclei. These are scaled at 50 μm for (a), 20 μm for (b), 5 779 μm for (c) and (d) respectively. 780 781 Supp. Figure S2.7. Distribution of Mycoplasma in midgut of adult Salmon. Overlay of all 782 channels scaled at 20µm. (a), Mycoplasma was distributed closely to epithelium cells. (b), 783 Overlay of all channels and DIC image confirmed Mycoplasma were in epithelium cells. 784 785 Supp. Figure S3. Circular track highlighting the function annotations of the Mycoplasma 786 MAG from this study. The functions are predicted on (a) the negative strand 3’-5’ and (b) 787 the positive strand 5’-3’. 788 789 Supp. Figure S4. Relationships of lifestyle and genome reduction. These include (a) plot of 790 genes count and genome size, (b) plot of genes count and pseudogenes, and (c) the 791 proportion of non-coding DNA across lifestyle. 792 793 Supp. Table 1. Summary of SEED subsystems annotations of the Mycoplasma MAG from 794 pyloric caecum of Atlantic salmon (Salmo salar) 795 796 Supp. Table 2. Summary of shared and unique SEED subsystems after pairwise genomic 797 comparison of the Mycoplasma MAG from this study and Mycoplasma Penetrans 798 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

799 Supp. Table 3. Summary of best CDS orthologs based on best reciprocal BLAST hits 800 between the Mycoplasma MAG in this study and Mycoplasma Penetrans 801 802 Supp. Table 4. Summary of genomic features from 247 strains of mycoplasmas. These data 803 were collected from IMG database 804 805 Supplementary files 806 807 Supp. File 1. The complete DNA sequence of the Mycoplasma MAG recovered from the 808 pyloric caecum of the Atlantic salmon (Salmo salar) 809 810 Supp. File 2. Predicted coding regions (CDS) of genes in Mycoplasma MAG recovered from 811 Salmo salar 812 813 Supp. File 3. Sequence alignment in (PHYLIP format) format of 137 16s genes from 814 environmental and host-associated Mycoplasmas and Spiroplasmas. 815 816 Supp. File 4. Accession numbers and functions of the 21 markers concatenated in the MLST- 817 based tree. 818 819 Supp. File 5. Sequence alignment (PHYLIP format) of 55 protein from the 21 concatenated 820 from free-living and parasitic mycoplasmas 821 822 Supp. File 6. Sequence name abbreviation of tips labels within the 16s gene tree 823 824 Supp. File 7. Sequence name abbreviation of tips labels within the concatenated markers 825 MLST-based tree 826 827 828 829 830 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which Table 1. Fishwas not probes certified by and peer review) sequences is the author/funder. deployed All rights in reserved. this study. No reuse allowed without permission.

Probes Target group Sequence (5'-3') Reference Myc1-1 Mycoplasma GCGGTAATACATAGGTYGCAAGCG This study Gam-1 Gammaproteobacteria GCCTTCCCACATCGTTT Manz et al,1992 FIR-1 Firmicutes GGAAGATTCCCTACTGCTG Hallberg et al,2006 EUB338 All bacteria GCTGCC TCCCGTAGGAGT Amann et al,1990 EUB338 II Planctomycetes GCAGCCACCCGTAGGTGT Daims et al,1999 EUB338 III Verrucomicrobia GCTGCCACCCGTAGGTGT Non EUB338 None CGACGGAGGGCATCCTCA Wallner et al,1993 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Table 2. Summary of mycoplasma metagenome assembled genome (MAG).

Mycoplasma bin (TMDNneg-S20-001) detected in Adult Salmon Genome Salmo Salar Completeness 92.18 Contamination 0.38 Strain heterogeneity 0 Unique markers (of 43) 39 Multi-copy 0 k__Bacteria; p__Tenericutes ;c__Mollicutes; Taxonomy (contained) o__Mycoplasmatales; f__Mycoplasmataceae ;g__Mycoplasma Taxonomy of sister lineage s__Mycoplasma_penetrans GC content 24.98204716 Genome size (mbp) 0.577903 Gene count 570 Coding density 0.938088226 Length 577903 N50 14796 Comment 1 Genome with >90% completeness and < 5% contamination Comment 2 Binned with Metabat and quality assessed with CheckM

bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1-a. Salmon eggs 1-b. Stomach of Salmon Parr 1-c. Pyloric caecum of Salmon Parr

* *

1-d. Stomach of adult Salmon 1-e. Pyloric caecum of adult Salmon 1-f. Pyloric caecum of adult Salmon * bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

T.M.cav1 1 T.M.ins T.M.fas T.M.pne T.M.pne 1 T.M.tes 0.99 T.M.amp M.M.tul T.M.imi 0.73 T.M.gal3 OTU_141 0.99 M.sal.gut 0.92 T.M.mic T.M.mur 0.57 0.62 T.M.iow 0.67 T.M.pen 0.51 OTU_357 0.79 MSalar.16s 0.87 OTU_6660 0.71 OTU_6827 M.iso.hep2 M.iso.gut M.sed.hyd M.iso.hep1 1 M.shr.gut M.cra.gut1 M.cra.gut2

1 M.fish.gut T.S.syr T.S.chr 1 T.S.mel 0.96 T.S.ins 0.81 Clade A bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Mpen_F1.36 0.9444 Msalar_0.57 Mimi_0.92 1 Mgal_F1.01 Mgal_S0.93 Malv_0.84 0.8889 Mtes_1.32 Mgen_0.58 Mgen_F0.58 Mpne_F0.82 Mpne_S0.82 Mput_F0.83 Mmyc_F1.21 1 Mcap_F1.01 Mlea_F1.02 Mora_0.71 Msal_F0.71 Mart_F0.82 Mans_0.74 1 Mclo_0.66 1 Mspu_0.84 Malk_0.77 Marg_0.62 Mhom_0.69 Mhom_F0.66 Mmob_F0.78 1 Mpul_F0.96 Mele_0.77 Mall_F0.97 0.8889 Mana_0.93 Mfelis0.81 Mleo_0.90 1 Mcan_0.86 0.7778 Mcan_S0.87 Mcolu_0.91 Mcri_0.85 Msyn_0.74 Msyn_F0.80 1 Mfer_F1.12 Msim_0.85 0.8889 Mgan_0.83 Mcolu_0.75 Mine_0.77 Mfelif0.76 Mpri_0.90 Mbov_F1.0 Maga_F1.0 Maga_S0.92 Mcoll_0.90 Mmol_0.84 MhyorF0.81 1 MhyorS0.84 Movi_F1.02 Mflo_0.76 MhyopF0.9 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Predicted CDS on negative strand Predicted CDS on positive strand Mycoplasma penetrans (GCA_000011225.1) GC content GC skew TMDNneg-S20-001.5603

TMDNneg-S20-001.5189

TMDNneg-S20-001.4650

TMDNneg-S20-001.3179

TMDNneg-S20-001.2649

TMDNneg-S20-001.2428

TMDNneg-S20-001.1722

WP.011077204.10331

WP.011076915.10114 TMDNneg-S20-001.1473 WP.011077643.10473

TMDNneg-S20-001.1092 WP.044891297.1

WP.011077003.1 WP.011077318.10853 TMDNneg-S20-001.679 WP.011077191.1

WP.011077020.10427 WP.011077514.1 TMDNneg-S20-001.511 WP.011077657.10760 TMDNneg-S20-001.14

TMDNneg-S20-001.19

TMDNneg-S20-001.61 TMDNneg-S20-001.500WP.129374373.10466

TMDNneg-S20-001.5 WP.011077781.10121 WP.011076857.10791 TMDNneg-S20-001.443 WP.011077056.10841

WP.011077376.10300

TMDNneg-S20-001.9 TMDNneg-S20-001.429 WP.011077200.1 WP.011077467.10637 WP.011077792.10743 TMDNneg-S20-001.397 TMDNneg-S20-001.41

WP.011077184.10113 WP.011077564.1 TMDNneg-S20-001.269 TMDNneg-S20-001.317 WP.044891387.10035

WP.011076873.1 WP.011077379.1 TMDNneg-S20-001.270 WP.011077366.10192

TMDNneg-S20-001.0 TMDNneg-S20-001.147WP.011077759.10509

TMDNneg-S20-001.146WP.011077106.10404

WP.011076830.1 TMDNneg-S20-001.117WP.052270422.1

TMDNneg-S20-001.1 TMDNneg-S20-001.105WP.011077407.1

WP.011076904.1 TMDNneg-S20-001.99 WP.011076905.1

WP.011077682.1 TMDNneg-S20-001.2 TMDNneg-S20-001.85

WP.044891317.10803 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. TMDNneg-S20-001.77 WP.011077730.10417 WP.011077116.1

TMDNneg-S20-001.71 TMDNneg-S20-001.3 WP.011077431.1

WP.011077709.1 TMDNneg-S20-001.67 WP.011077468.10136

TMDNneg-S20-001.4 TMDNneg-S20-001.62 WP.011076921.1 WP.011077747.1

TMDNneg-S20-001.59 WP.011077513.10975 TMDNneg-S20-001.6 WP.011077675.1 TMDNneg-S20-001.56 WP.011076990.10752

TMDNneg-S20-001.7 WP.083755096.1 TMDNneg-S20-001.54WP.011077717.10098

WP.011077774.1

TMDNneg-S20-001.52 WP.083755103.10404TMDNneg-S20-001.10

WP.011077160.1

TMDNneg-S20-001.51 WP.011077806.1 TMDNneg-S20-001.11

WP.011077214.10314 WP.011077038.1 TMDNneg-S20-001.49 WP.044891383.1 TMDNneg-S20-001.12 WP.011077025.1

WP.011076954.1 TMDNneg-S20-001.40 TMDNneg-S20-001.13 WP.011077736.10091

WP.011077509.1

WP.011077168.10131 WP.011077392.1 TMDNneg-S20-001.15

TMDNneg-S20-001.39 TMDNneg-S20-001.16

TMDNneg-S20-001.18

TMDNneg-S20-001.37 WP.011076852.1 WP.011077230.10560

WP.011077623.1

TMDNneg-S20-001.35 WP.011077850.1

TMDNneg-S20-001.34

WP.011076899.10029

TMDNneg-S20-001.32

TMDNneg-S20-001.30

TMDNneg-S20-001.23

TMDNneg-S20-001.21 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

A. 1600 B. ● ● ● ● ● ● 1500

● Gene count

● ●● ●● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● GC count ● ● ● ● ● ● 1200 ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● 1000 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● CDS count ● ● ● ● ● ●●●● ● ● ●●●●● ●● ● ● ● ● ● ● ● ● ● ● ●●●● ●● ● ● ● ●● ● ● ● ●●● ●● ● ● ● ●●● ● ●● ●●●● ● ●●●● ● ●● ●●●●●●●● ● ●● ● ● ●●● ● ●● ●●● ● ● ● ● ● ●●●●●● ●●●●●● ● ● ●● ● ●●●● ● ●● ●●●●●●● ● ● ● ●●● ● ●●● ● ● ● ●● ●● ●● ●●● ● ● ●● ● Values ●● ●● ● ●● ● ● ● ● ●● ●●●●● ● ●● ● ● ● ●● ●●●●●● ● ● ● ● Functions count ● ● ● ●●●●● ● ●● ● ● ● ● ●●●●● ●● ●●● ●● ● ●● ●● ● ● ● ● ● ● ●●● ●●●●●● ●● ●●● ●● ●● ● ● ●●●● ●●● ●●● ● ● ●●● ●●●● ●●● ●● ● ● ●● ● ● ●● ●● ● ●● ● ● ● ● ● ●●● ●●●● ● ● ● ●● ● ●● ● ● ● ● ●●● ●●● ●●● ● ●●● ● ● ● ● ●● ● ●● ● ●● ●● ●●● ●● ● ● ● ●● ●●● ● ● ● ●● ● ● ● ●●●● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ●●● ●● ● ●● ● ● ●● ● ● ● ● ● ● ●●●●●● ●●●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ●●●●●● ●● Transmembrane proteins ● ● ● ● ● ●●● ● ●● ●●● 500 ●●●● ● ●● ●●● ●● ● 800 ● ● ●●● ● ● ●●●●● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ●● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ●●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● Genes count ● ● ● ● ● ● ●● ● ●● ●●●●●●●● ● ● ● ● ● ● ● ● ●● ●● ● ●● ●● ● ● ● ● ● ● ● ●●●● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ●●●●●● ● ● ● ●● ● ● ●● ● ● ● ●●●●● ●●● Pseudogenes count ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ●●● ●●●● ● ●●●●● ● ●● ●● ● ● ● ● ● ● ●●●●● ● ● ●●●● ●● ● ● ● ● ● ● ● ●●● ●●●● ● ●● ● ●● ● ●● ● ● ● ●●● ● ●● ● ●● ● ●● ● ● ●●● ●● ●●● ● ●●●●●● ● ● ● ● ● ● ●● ● ● ●●●●●●●● ●● ●●●●●●● ● ● ●● ●● ● ● ●●● ●●● ●●●●●●● ●●●● ●● ● ● ● ●●●●● ●● ●● ● ● ●●● ● ●● ● ● ● ●● ●●●●●● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ●●● ●● ●●●●●●●●●●●● ●● ●●●● ● ● ●●● ●●●● ●●●●●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ●● ●●●●●●●●●●●●●●●●●● ●●●●● ● ● ●●● ●● ●● ● ● ● ● ●●● ●●●● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ●●● ● 0 ● ● ●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●● ●● ●●●● ●●●●● ● ● ●

0.6 0.8 1.0 1.2 1.4 0.6 0.8 1.0 1.2 1.4 Genome size Genome size bioRxiv preprint doi: https://doi.org/10.1101/2020.07.17.209767; this version posted July 18, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Other types of Glycosphingolipid O-glycan biosynthesis Mucin type Mannose type Lipopolysaccharide Nucleotide biosynthesis - biosynthesis Glycan Biosynthesis globo and isoglobo series Various types of O-glycan biosynthesis O-glycan biosynthesis Metabolism and Metabolism N-glycan biosynthesis Metabolism of N-Glycan Streptomycin Cofactors and Vitamins biosynthesis biosynthesis Glycosphingolipid Neomycin, kanamycin and biosynthesis - GPI - anchor gentamicin biosynthesis lacto and neolacto series biosynthesis Biosynthesis of Other glycan ansamycins Puromycin [B] Phytochemical [B] Glycosyltransferases degradation biosynthesis compounds Starch and sucrose metabolism Purine metabolism Thiamine Glycosaminoglycan metabolism biosynthesis - heparan sulfate / heparin

Glycosphingolipid biosynthesis - Ascorbate and Vitamin B6 Galactose aldarate metabolism metabolism ganglio series metabolism Flavonoid Glycosaminoglycan biosynthesis Caffeine Folate biosynthesis - beta-Alanine keratan sulfate metabolism biosynthesis Phenylpropanoid alpha-Linolenic acid metabolism Riboflavin biosynthesis metabolism Glycosaminoglycan Glycosaminoglycan metabolism biosynthesis - Pyrimidine metabolism degradation Polyketide sugar Pentose and glucuronate chondroitin sulfate / unit biosynthesis interconversions Isoflavonoid dermatan sulfate biosynthesis [B] Proteoglycans Biosynthesis of vancomycin Tyrosine metabolism group antibiotics

Amino sugar and nucleotide sugar metabolism Pentose phosphate metabolism Phenylalanine , tyrosine and Biosynthesis of 12-, 14- and tryptophan biosynthesis Glycerophospholipid 16-membered macrolides metabolism Fructose and mannose metabolism Phenylalanine Linoleic acid One carbon pool Arachidonic acid Acarbose and metabolism by folate Anthocyanin metabolism metabolism validamycin biosynthesis biosynthesis Photosynthesis - Isoquinoline alkaloid Retinol Biosynthesis of Novobiocin biosynthesis metabolism unsaturated fatty acids antenna proteins Biosynthesis of [B] Lipids biosynthesis Other Secondary Metabolites Glycerolipid Amino Acid [B] Lipids biosynthesis metabolism Carbohydrate Acridone alkaloid proteins Metabolism biosynthesis Ether lipid Metabolism Staurosporine biosynthesis Lipid metabolism Photosynthesis Phenazine Taurine and Sphingolipid Sulfur biosynthesis metabolism metabolism hypotaurine Flavone and Metabolism Inositol phosphate metabolism flavonol Carotenoid metabolism Glycolysis / Biosynthesis of Indole alkaloid biosynthesis [B] Photosynthesis biosynthesis biosynthesis Gluconeogenesis proteins siderophore group Fatty acid Fatty acid nonribosomal degradation peptides biosynthesis Energy Tryptophan Peptidoglycan Metabolism metabolism biosynthesis Ubiquinone and other terpenoid-quinone biosynthesis Phosphonate and Histidine metabolism Pyruvate phosphinate metabolism metabolism D-Alanine Benzoxazinone Metabolism of Biosynthesis of type II metabolism biosynthesis Terpenoids and Polyketides polyketide products Valine, leucine and Pantothenate and CoA isoleucine biosynthesis biosynthesis Limonene and Tetracycline pinene degradation biosynthesis Biosynthesis of type II polyketide backbone Monoterpenoid Synthesis and biosynthesis degradation of Biosynthesis of Selenocompound Ketone bodies Glycine, serine and Porphyrin and enediyne antibiotics metabolism threonine metabolism chlorophyll metabolism Zeatin Fatty acid elongation Butanoate in mitochondria Sesquiterpenoid and biosynthesis metabolism Clavulanic acid triterpenoid biosynthesis biosynthesis Cutin, suberine and Propanoate Lysine metabolism Valine, leucine and Urea cycle wax biosynthesis degradation isoleucine degradation Insect hormone Reductive biosynthesis Chloroalkane and chloroalkene degradation carboxylate cycle Lysine (CO2 fixation) biosynthesis Terpenoid backbone Lipoic acid biosynthesis metabolism Caprolactam degradation Carbon C5-Branched fixation dibasic acid Steroid biosynthesis metabolism

Primary bile acid Methane biosynthesis Citrate cycle D-Arginine and Atrazine metabolism (TCA cycle) Alanine , aspartate and D-ornithine degradation glutamate metabolism metabolism Glyoxylate and Nitrogen dicarboxylate Monobactam metabolism metabolism biosynthesis Steroid hormone Glucosinolate biosynthesis Cystein and methionine biosynthesis metabolism Aminobenzoate Cyanoamino acid degradation Naphthalene metabolism degradation Arginine biosynthesis Dioxin degradation Arginine and proline Chlorocyclohexane and Carbapenem metabolism Indole diterpene Xylene chlorobenzene degradation alkaloid biosynthesis degradation biosynthesis Bisphenol Diterpenoid Fluorobenzoate degradation biosynthesis degradation Glutathione metabolism Tropane, piperidine and pyridine alkaloid biosynthesis Styrene Nicotinate and Secondary bile acid Oxidative degradation nicotinamide biosynthesis Benzoate phosphorylation Penicillin and degradation cephalosporin Biotin metabolism Brassinosteroid metabolism biosynthesis biosynthesis D-Glutamine and Steroid Nitrotoluene D-glutamate degradation degradation beta-Lactam metabolism Toluene Ethylbenzene resistance degradation degradation DDT degradation Polycyclic aromatic hydrocarbon degradation Metabolism of [B] Cytochrome P450 Furfural degradation Other Amino Acid Xenobiotics Biodegradation Metabolism of xenobiotics Drug metabolism Drug metabolism and Metabolism by cytochrom P450 - cytochrom P450 - other enzymes 01100 5/11/17 (c) Kanehisa Laboratories