Functional stability despite high taxonomic turnover characterizes the Ulva microbiome across a 2,000 km salinity gradient

17 The green seaw eed Ul va depends on its associated bact eria for morphogenesis and is an 18 important model to study algal-bacterial interactions. Ul va -associated bact eria exhibit 19 high turnover across environmental gradients, leading t o the h ypothesis that bacteria 20 contribute to the acclimation potential of the host. Y et little is known about the variation 21 in the functional profile of Ulva-associat ed bacteria in relation to environmental 22 changes. T o test which microbial fu nctions shift alongside a strong en vironmental 23 gradient, w e analysed microbial communities of 91 Ul va samples across a 2,000 km 24 Atlantic–Baltic Sea salinity gr adient using metagenomic sequencing. Metabolic 25 reconstruction of 639 metagenome-assembled genomes rev ealed widespread potential 26 for carbon, sulphur , nitrogen, and vitamin metabolism, including amino acid and vitamin 27 B bios ynthesis. While salinity explained 70% of taxonomic variation, it only accounted 28 for 17% of functional v ariation, indicating extensiv e fun ctional stability . The limit ed 29 variation was attributed to typical high-salin ity bacteria exhibiting enrichment in genes 30 for thiamine, pyridoxal, and betaine biosynthesis. These metabolic modules likely 31 contribute to oxidativ e stress mitigation, cellular osmotic homeostasis, and membrane 32 stabilization in r esponse to salinity variations. Our results emphasise the importance of 33 functional profiling to understand the seaw eed holobiont and its collecti ve r esponse to 34 environmental change. 35 36 37 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 3

38 Microbial communities interact with eukaryotic hosts in e very ecosystem, yet our 39 understanding of how (variation in) microbial metabolisms impact host performance is 40 very l imited. Seaweeds, or macroalg ae, are a group of important coastal inhabitants. As 41 ecosystem engineers and primary p roducers, they are an essential sour ce of food and 42 they provid e shelt er and habitat f or di verse marine organisms 1 . Health, reproduction, 43 and dev elopment of the macroalgal host — and indirectl y their ecosystem functioning 44 — are strongly dependent on the associat ed symbionts. Beyond the ex change of ke y 45 nutrients, vitamins, and secondary metabolites, the microbial biofilm forms a physical 46 and chemical barrier acting as a “second skin ” that prot ects the host and modulat es its 47 int eractions with the en vironment 2–4 . Beneficial bacteria can, f or ex ample, shield the 48 host against pathogens 5 , stimulate algal growth 6 , or mitigate the ad verse effects of 49 environmental pollution 7 . 50 Species of the gr een seaweed Ulva are particularl y w ell-studied and ar e considered a 51 model system t o study algal-bact erial interactions 8 . The r elation between Ulv a an d its 52 symbionts, t ogether often referred to as a holobiont, is so indisp ensable that the 53 seaw eed fails to develop its typical leaf- or tube-lik e morphology in the absence of 54 particular bacteria 9,10 . Morphological de velopment is induced b y chemical compounds 55 produced by bacteria that trigger cell wall de v elopment, rhizoid formation, and cell 56 division 11 . Initial studies id entified a combination of two complementary strains (a 57 Roseo varius and a Maribacter strain) that were necessary for full morphogenesis in Ulva 58 mutabilis . Since then, multiple strains ha ve been identified that ha ve the same capacity 59 12 . This functional r edundancy across taxa is important for the host in relation to shifts 60 in bacterial communities. 61 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 4 Seaweed-associated bacterial communities are f ar from static. Although a small core 62 community can sometimes be identified, the bacteria form a highl y d ynamic community 63 tha t fluctuates through time and space 13,14 . Abiotic f act ors, such as temperatur e, light, 64 and salinity influence the community composition, as does the host itself 15–17 . P rior 65 research has mainly focused on dyn amics in the tax onomic composition of the 66 communities r ather than the functional potential that the community harbours. 67 T axonomic profiling is based on 16S rRNA amplicon sequencing that in recent y ears has 68 become wid el y accessible due to increased speed and reduced costs. Genome-wide 69 studies or metagenomic sequencing that gi ve information on the functional potential of 70 the bact erial communities in comparison are more expensi v e and the sample size in 71 those studies is oft en limit ed. Metagenomic functional profiling in, for e xample, 72 Sargassum spp. 18,19 , P yropia haitanensis 20 , and Nereocystis luetkeana 21 highlighted the 73 importance of the bacterial biofilm in t erms of vitamin production, poly saccharide 74 degr adation, and nutrient cy cling. Howev er , these studies w er e based o n a relati v ely 75 small number of replicat es (between 3–7 samples) and theref ore do not capture 76 differences across en vironments. T o understand how the functional potential of a 77 bacterial community r esponds to en vironmental change and t o ev aluate the potential 78 consequences for the seaw eed host, it is essential to expand functional profiling on a 79 broader scale across di v erse environmental gradients 22 . 80 Ulva species are known for their broad salinity tolerance, thriving in en vironments 81 ranging from fresh w ater habitats t o highly saline conditions. The Baltic Sea is a 82 particularly interesting area to study Ulva , as it hos ts more than 15 species distributed 83 across its distincti ve salinity gradient 23 . This r elati v ely stable gradient stretches across 84 more than 2,000 km of distance, transitioning from near fresh water conditions in the 85 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 5 innermost parts t owards fully marine conditions on the North Sea side 24 . While the 86 distribution of certain Ulva speci es is confined to higher salinity le vels, other species lik e 87 U. int estinalis and U . linza can be found throughout the entire range. Pre vious studies 88 showed that salinity also strongl y structur es the taxonomic composition of Ul va -89 associat ed bacterial communities 25 . The question remains , however , whether these 90 taxonomic shifts in r esponse to salinity ar e reflect ed by corresponding changes in 91 functional profile of the same bacterial communities, or if functional traits are 92 conserv ed across the environmental gradient 26 . 93 T o address these knowledge gaps and enhance our understanding of the Ulva holobiont 94 model s ystem, we investigated the taxonomic and functional composition of Ulva -95 associat ed bacteria across the Balti c–Atlantic salinity gradient using metagenomic 96 sequencing. Th e aim of this stud y w as tw ofold: (i) t o provide a compreh ensi ve ov erview 97 of the functional potential within th e Ulva bacteriome based on a large number of 98 samples (n=91), and (ii ) to assess if the functional potential of the microbial community 99 was stable across an en vironmental gradient. W e expected that the functional repertoire 100 of Ulva-asso ciat ed bacteria would be rich in carbon utilisation and nutrient cycling 101 path way s, as these functions ar e traditionally associated with sea weed microbiomes 102 20,21 . Gi ven that salinity has been identified as a major dri ving force of global bact erial 103 di versity and community structure 27 , we also h ypothesized pronounced ch anges in the 104 functional profile of the Ul va -ass ociated bac ter ial community , partic ularly in functions 105 related to osmopro tection. 106 107 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 6

and discussion 108 Ulva -associ ated bacteria utilize host-deri v ed carbon and sulphur 109 Rich in organ ic carbon, oxygen, and nutrients, the algal surf ace provides an ideal habitat 110 for the growth of microbes 28,29 . The microbial community r eciprocat es by supplying the 111 algal host with phosphate, nitrogen, and vitamins in r eturn 2,30 . Such an exchange of k ey 112 nutrients is an essential f eatur e of the Ulva holobiont. W e used metagenomic sequencing 113 to comprehensi vely describe the functional profile of Ul va -associated bacteria, obtaining 114 a large dataset of metagenome-assembled genomes (MAGs) that were annotat ed with 115 the Car bohydrate-Active EnZ ymes database (CAZy) 31 and K yoto Encyclopedia of Genes 116 and Genomes (KEGG) 32,33 . Across 91 Ulva microbiome samples (T able S1) and 639 117 MAGs (T able S2 ), we identified a total of 6,525 K O (KEGG Orthology) terms and 399 118 KEGG modules (T able S3). MAGs with an estimated completeness of >90% contained on 119 av erage 59 KEGG modules (ranging from 33-94 modules). Apart from basic cellular 120 metabolic pathw ays (e.g., fatty acid biosynthesis, RNA polymerases, DN A polymerases, 121 ribosome, F-type ATP ases, cyt ochro me C oxidase, etc.), the Ul va -as sociated bac ter ial 122 communities contained a range of functions that ar e clearly linked t o the association 123 with its host. This included the wid espread potential of bacteria t o utili ze host-deriv ed 124 carbon and sulphur . 125 All 639 MA Gs were able t o utilize organic carbon as th eir primary energy source 126 through, e.g., complex carbon degradation, gl y colysi s (KEGG modules 127 M00001+M00002), and the tricarboxylic acid (TCA) cy cle (M00009) (Fig. 1a). Likewise, 128 many bact eria were capable of utilizing sulphur compounds produced by the host. 129 Sulphur metabolism r elated gen es in Ulva-associated communities wer e predominantly 130 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 7 invol ved with assimilatory sulphat e reduction (M00176), leading t o the formation of 131 sulphite and, ultimately , sulphide (Fig. 1a). In addition, 137 MAGs were able to oxidize 132 thiosulphate (the r esult of the incomplet e oxidation of sulphides) through sulphur-133 oxidizing prot eins Sox AB and SoxXYZ that together f orm the Sox complex (M00595) (Fig. 134 1a). Due to the presence of sulphate esters in cell w all polysaccharides, Ulva species 135 ha ve a relativ el y h igh sulphur content 34,35 and our findings support the hypothesis that 136 the widespread pre v alence of sulphonates in marine algae contribut es to the high 137 abundance of sulphonate-degr ading bacteria in marine h abitats 36 . 138 Other pre valent sulphur cy cling enzymes in Ulva -as sociated bac ter ia were related to 139 dimethylsulfoniopropionat e (DMSP) metabolism, an organosulphur compound 140 produced by Ulva 37,38 . DMSP has a wide range of ecophy siological functions and in the 141 model species Ulva mutabilis it is known to mediat e the int eraction with the symbiont 142 Roseov ar ius sp. MS2. This symbiont releases morphogenetic compounds that stimulate 143 algal morphogenesis and growth 39 . Interestingly , although the bacterial strain is 144 attracted by the release of DMSP and rapidly takes up the compound, its growth is not 145 affected b y DMSP . Instead, DMSP is conv erted int o dimethylsulphide (DMS) and 146 methanethiol (MeSH). In our dataset, the capacity t o conv ert DMSP to DMS was almost 147 exclusi vely restricted to a few gener a within the Rhizobiaceae (the undescribed genus 148 JAALLB01) and Rhodobacteraceae (predominantly Jannaschia , Ros e ov ari u s , S ulfitobacter , 149 and Ta t e y a m a r i a ). Thi s observ ation aligns with previous find ings that another well-150 known morphogenesis-inducing symbiont of Ulva , Maribacter sp. MS6 (belonging to the 151 Flavobacteriaceae ), is not attracted by DMSP 39 . 152 Apart from examining carbon and sulphur metabolic functions organized in KEGG 153 modules, we also screened the Ul va -associated bacterial MAGs for the presence of 154 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 8 carbohydrate-acti ve enzymes using the CAZ y database. Th e CAZ y database categorizes 155 enzymes invol ved in d egrading, modifying, or creating gl ycosidic bonds into f amilies, 156 including gl y cosid e h ydrolases (GHs) and poly saccharid e l yases (PLs). A total of 121 157 different GH families and 26 differ ent PL families w ere id entified within Ul va -associated 158 MAGs. Our results d emonstrated that a substantial majority [97%] of the bact eria in 159 Ulva -associated communities are capable of breaking down complex polysaccharides 160 constituting the Ulva cell w all. The Ulva cell wall is mainl y composed of cellulose, 161 glucuronan, ulv an, and xyloglucan, which together mak e up 45% of the dry-w eight 162 biomass 40 . Ul van, a key component, is composed of sulph ated rhamnose, xylose, and 163 uronic acids (D-glucuronate and iduronic acid) 40 . The initial steps in the breakdown of 164 ulv an can be catalysed by ulv an l yases (sulfatases) 41 , but additional gl y coside 165 hydrolases ar e needed to obtain monomeric sugars 42 . Ulvan lyases [amongst other s 166 CAZ y family PL24, PL25, and PL37] were found in 146 MAGs (Fig. 1a), man y of which 167 were typical Ulva- a ssoc iat ed b a ct eri a, s uch a s Lewinella , Nonlabens, A lgibacter , 168 Glaciecola , and P olaribacter 12 . Pre v alent glycoside h ydrolase famili es included GH3 [553 169 MAGs], GH16 [305 MAGs], and GH5 [299 MAGs]. F amily GH3 includes, amongst others, 170 xyloglucan-sp ecific exo-β-1,4- glucosidase (EC 3.2.1 .-), xylan β-1,4-xylosidase (EC 171 3.2.1.37) , several glucanases , and several other glucosidases. Enzymes of family GH16 172 are acti ve on β-1,4 or β-1 ,3 gly cosidic bonds in v arious glucans and galactans and 173 includes xyloglu can:xyloglucosyl transferase (EC 2.4.1 .207). Members of the GH5 family 174 are mainl y acti ve on cellulose, but this family also includes high specificity 175 xyloglucanases. 176 177 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 9 Nitrogen metabolism and amino acid biosynthesis 178 Unlik e animals, seaw eeds can synth esise essential amino acid s themsel ves and relati vel y 179 little is known about the uptake of dissol ved organic nitrogen compounds by seaw eed s 180 43 . It is lik ely , howe ver , that seaw eeds and bacteria do e xchange amino acids, especially 181 when dis solved inor ganic nitrogen availab ility is low 44 . The biosynthesis of multiple 182 amino acids w as part of the core modules of the Ulva - assoc iat ed b a ct eri a ( with 0 .6 5-183 1.42% relati ve abundan ce in at least 99% of samples). Thi s in cluded the biosynthesis of 184 proline (KEGG module M00015), threonine (M00018), l ysine (M00016, M00526, 185 M00527), valine/isoleucine (M00019), serine (M00020), cy steine (M00021), 186 tryptophan (M00023), hi stidine (M00026), ornithine (M00028), leucin e (M00432), and 187 arginine (M00844). Of these core m odules, cystein e biosynthesis was d et ected in the 188 fewest number of MAGs (222 MAGs), while histidine biosynthesis was found in the 189 highest number of MAGs (443 MAGs). Lik ewise, genes necessary for gl y cine biosynthesis 190 (from threonine or serin e), alanine biosynthesis (from pyru vate) and glutamine 191 biosynthesis (from glutamate), as w ell as the conv ersion between aspartate and 192 asparagine were pre valent in the majority of the MAGs. 193 Nitrogen metabolism pot ential largely centred on the conv ersion from nitrat e and 194 nitrit e to ammonia (F ig. 1a). A large number of 174 MAGs contained the nr fAH or nirBD 195 genes that ar e essential for nitrite r eduction to ammonia, which is the final st ep in 196 dissimilat ory nitr ate reduction to ammonia. A total of 30 MAGs exhibited the genes for 197 the full dissimilatory nitrat e redu ction to ammonia, utilizing nitrate r espiration for 198 energy production (M00530). Onl y sev en MAGs, all belonging t o the Cy anobact eriia, 199 demonstrat ed complet e assimilat ory nitrate reduction capability , in volving the 200 reduction of nitrate to ammonia for biosynthesis pur poses (M00531). Never theless , the 201 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 10 first step in this process (nitrat e t o nitrite reduction mediat ed by the narB gene) was 202 identified in 53 MAGs, and the subsequent st ep (nitrite to ammonia reduction facilitated 203 by the nir A gen e) w as detected in 4 5 MAGs. The narB gene was predominantly observ ed 204 in Flavobact eriaceae and Saprospiraceae , while the nirA gene w as pre valent in 205 Akkermansiaceae and Pirellulaceae . T wo MAGs harboured all the genes necessary for the 206 denitrification process in volving the reduction of nitrate and nitrite t o nitrogen 207 (M00529), while none of the MAGs carried genes associat ed with nitrification (the 208 oxidation of ammonia to nitrate; M00528). Finally , t en MAGs from div erse families 209 showed the capacity to fix nitrogen (M00175). 210 Previous studies demonstrat ed that observ ed amino acid uptake rates in Ulva w er e 211 highest f or gl ycine and alanine and that while Ulva prefers inorganic nitrogen, the 212 organic c ompounds may als o play a significant role 45,46 . A transcriptomic stud y in 213 Laurencia dendroidea show ed that t he associated bact eria had a higher relativ e 214 contribution to amino acid metabolism than the host itself, indicating a symbiotic 215 relation 47 . Similarly , in P yropia haitanen sis , co- culti vation with a Bac illus sp. did not only 216

in an increased growth rate, but also in a downregulation of genes relat ed to the 217 biosynthesis of sev er al amino acids and other metabolit es 48 . As free amino acids are 218 known to increase primary production of algae 49,50 , it is h ypothesised that the release of 219 amino acids by bact eria serv es as one mechanism through w hich they may facilitate 220 algal growth. 221 222 223 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 11 224 Fig. 1 | Fu n ct ion al p r ofil e of the Ulva mic r obiome . ( a ) O v er v i ew o f th e m et abolic p ot ent ial of th e Ulva 225 micr ob io m e , h i ghli g h ti n g vit a mi n pr oduct i o n (p urple) , algal cell wa ll d egr ad ati on (gr e e n ) , a nd th e 226 m e t ab oli sm of car bo n (blu e ) , nitr og en (or a n g e) , an d sulph u r ( yello w ) . Th e nu mber of ide nti fi ed 227 m e t ag en o mic as se m bl e d ge n omes ( MAGs ) that ar e cap abl e p erfor m i ng t he r e a c ti ons a re di splay e d i n 228 br ackets . (b) Th e pr opor ti o n o f t he n um b er o f s e a wat er -is o l at ed (blu e) o r Ulva-i s o la te d ( g ree n ) ba c te r ia l 229 gen o m e s t ha n co nt ain e d a sp ecif ic CA Z Y f amil y ( p ol y sacc h a r ide l yas es = P L , glyc os id e h y dr o l as es = GH). 230 231 Large potential for vitamin B production by bacteria 232 Previous metagenomic studies in the red alga P yropia haitanensis 20 and brown alga 233 Nereocystis luetkeana 21 ha ve highlighted the potential of the bact erial symbionts to 234 produce vitamin B 12 (cobalamin). Our data show ed that the capacity of seaw eed-235 associat ed bacteria to biosynthesize vitamins is widespread within the Ul va microbiome 236 and extends beyond vitamin B 12 (Fig. 1a). Vitamins are essential to a w ell-functioning 237 central metabolism in both microbes and their hosts 51,52 . Algae require a combination of 238 different vitamins f or growth, but are lik el y unable to synthesise some of these organic 239 compounds themsel ves 53–55 . A stud y b y Croft et al. 56 , f or example, showed that more 240 than half of the algae studied by them required an exogenous suppl y of vitamin B 12 241 (cobalamin), 22% required vitamin B 1 ( thiamine) , and 5% required vitamin B 7 (biotin). 242 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 12 In Ulva speci es, the addition of vitamin B 12 and particularly vitamin B 6 (p yridoxal) to 243 culture medium is necessary for growth and promot es nitrogen metabolism 57,58 . This 244 vitamin auxotroph y suggests that the host is dependent on the production by its 245 microbial symbionts. Within ou r dataset, the cap acity of bacteria to produce 246 tetrahydrofolate ( a derivative of vitamin B 9 ; KEGG module M00126) w as most 247 widespr ead and was identified in 1 10 MAGs from varying families, including the 248 Alteromonadaceae , Flavobacteriaceae , Granulosicoccac eae , Saprospiraceae , and 249 Spirosomaceae (Fig. 1a). This vitamin was the only vitamin that w as part of the core 250 functions of the Ulva microbiome, being consistently pr esent in all samples with a 251 minimum relati v e abundance of 0.1%. KEGG modules relat ed to the production of 252 vitamin B 1 ( thiamine; M00127) , vitamin B 2 (riboflavin; M00125), vitamin B 5 253 (pantothenate; M00119), vitamin B 6 (pyridoxal; M00124), vitamin B 7 (biotin; M00123), 254 and vitamin B 12 (cobalamin; M00122) w er e identified as w ell (F ig. 1a), although not as 255 core functions. This could be attribut ed to either the absence of MAGs with the capacity 256 to produce these vitamins in all samples or their pr esen ce in samples but not with high 257 abundance. 258 259 The Ulva microbiome is defined b y its ability t o degrade the host 260 Our metagenomic d ataset provided a thorough o verview of the functional profile of the 261 Ulva -associated microbiome. Next, we aimed to det ermine the functions contributing to 262 a "typical" Ulva microbiome, by contrasting taxa isolat ed from Ul va with those isolated 263 from sea water . In t otal, we selected 152 MAGs from our metagenomic dataset, 264 representing 33 different genera (T able S4). W e then searched for publicly a v ailable 265 genomes of bact eria from the same genera but isolated from sea water (71 genomes) 266 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 13 (T able S4). Subsequently , w e conducted a comparativ e analy sis based on odds ratios to 267 identify pot ential enrichments of specific KEGG Orthology (K O) terms or Carboh ydr ate-268 Acti ve enZ ymes (CAZ ys) in bacteria from the same genus collected from Ulva ver sus 269 sea water . 270 Our results suggest that a def ining aspect of an Ulva microbiome is its ability to utilize 271 and break down the host organism, shown by the significant enrichmen t of CAZ ys (incl. 272 ulv an l yases, α-L-rhamnosidases, and rhamnogalacturonan α-L-rhamnoh ydrolase) that 273 specificall y target and break down Ulva’s polysaccharides and cell wall components 274 (e.g., ul van, iduronic acid, cellulose, xyloglucan) (Fig. 1b), as well as ABC transport ers 275 that facilitat e the extracellular uptake of small monosaccharides lik e fructose ( frcBCA ; 276 KEGG K O K10552, K10553 & K10554), α-glu coside ( aglEFGK ; K10232, K10233 & 277 K10234), and rhamnose ( rha SPQ T ; K1055 9 & K10560) r esulting from the degr adation 278 of the cell w all polysaccharides. It is clear that Ul va -associated bacteria do not only live 279 on the algal tissue, they li ve of i t a s w ell . 280 281 High tax onomic turnov er and fun ctional stability across salinity gradient 282 Subsequently , we examin ed the shifts in both the taxonomic and functional composition 283 of Ulva-asso ciat ed bacteria across a salinity gradient tra versing four distinct salinity 284 regions: sp anning from the horohalinicum (5–8 PSU), through the mesohaline (8–18 285 PSU) and the polyhaline (18 –30 PSU), to the euhaline (30–35 PSU) (Fig. 2a). The 286 Atlantic–Baltic salinity gradient exp lained more of the observ ed variation in the 287 taxonomic composition of Ul va -associated bacterial communities (p = 0.0001 , R 2 = 0.70) 288 than of the observed v ariation in the functional composition of the same communities (p 289 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 14 = 0.0006, R 2 = 0.17) (Fig. 2b, 2c). Pairwise comparisons, for example, showed that the 290 taxonomic composition of Ulva -associated bacteri a d iffer ed between all salinity regions 291 (horohalinicum vs mesohaline vs polyhaline vs euhaline, p<0.01 for all comparisons; 292 pairwise Adonis t est). On the other hand, the functional gene profile of bact erial 293 communities in the mesohaline, pol yhaline, and euhaline w ere not significantly different 294 from each other (p>0.05 f or all comparisons, pairwise Adonis t est), as only the 295 horohalinicum differed from the two high er salinity regions (horohalinicum vs 296 polyhaline, p=0.04; horohalinicum vs euhaline, p=0.04, pairwise Adonis test). 297 Nutrient concentrations, temperatu re and oxygen concentrations had little to no eff ect 298 on taxonomic turnov er (NO x , p = 0.04, R 2 =0.07; PO 4 , p = 0.08, R 2 =0.05; temperatu re, p = 299 0.01, R 2 =0.10; oxygen, p = 0.19, R 2 =0.04) and did not affect functional composition (NO x , 300 p = 0.90 , R 2 =0.002; PO 4 , p = 0.22, R 2 =0.03; temperature, p = 0.15, R 2 =0.04; oxygen, p = 301 0.84, R 2 =0.004). 302 Indicati ve of a high taxonomic turnover , 294 MAGs changed in relativ e abundance across 303 the salinity grad ient (p<0.01, LinDA linear regression) (F ig. 3a) (T able S2), of which 126 304 MAGs decreased with s alinity and 168 MAGs increased with salinity . Several MAGs 305 belonging to Dokdonia [MAG082, MAG518], Leucot hrix [MAG022, MAG360], and 306 Litor imonas [MAG193, MAG149] for example incr eased with salinity , while 307 Alteraurantiacibacter [MAG014, MAG020, MAG578], R ubripirellula [MAG422], and 308 Blastomonas [MAG197] decreased with salinity (Fig. 3b). 309 The taxonomic turnover was larger than the observed functional change, implying that 310 multiple taxa across the salinity gradient were able to perform similar functions. F ig. 4 311 gi ves a complete overview of the pr esence or absence of KEGG modules in the 312 metagenomic assembled genomes of 26 taxa that are characteristic to a specific salinity 313 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 15 region ( horohalinicum 5–8 PSU, mesohaline 8–18 PSU , polyhaline 18–30 PSU, or 314 euhaline 30–36 PSU). P antothenate (vitamin B 5 ), for example, can be produced at low to 315 medium salinity (5–18 PSU) by an unknown Saprospirac eae (MAG591), but at medium 316 to high salinity (18–36 PSU) this function in the community was taken over b y 317 Robiginitomaculum and another unknown Saprospiraceae (MAG149) (Fig. 4). Similarly , 318 proline biosynthesis at low salinity can be conducted by Alteraurantiac ibac ter and 319 Rubripir ellula (5–8 PSU), at medium salinity by Litorimonas A (8–18 PSU), and w as 320 gradually taken over by Neomegalonema (18–30 PSU), and Le ucothr ix and Litorimonas 321 at high salinity (30–36 PSU). In addition, most MAGs shared a set of cor e genes that ar e 322 necessary for essential functions, in cluding nu cleotide metabolism, fatty acid 323 biosynthesis, energy metabolism (F-type ATPases, succinate deh ydrogenases, et c.), and 324 genes encoding structural modules (e.g., ribosomes, RNA polymerase, and DNA 325 polymerase) (Fig. 4). 326 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 16 327 Fig. 2 | T axon om i c and fun cti onal patt e r ns a c r oss s al in ity. (a) Ge ogra phic d istri bu tio n of s ampl i n g 328 sit es ( n=63) al o n g t he Atl a n t ic –Ba l tic S e a g rad ie nt w h er e a t ot al of 9 1 Ulva sensu lato sa mples w e r e 329 c o l l e c t e d . T h e f o u r m aj o r s al i n i t y r e g i o n s ( h o r o h al i n i c u m , m e s o h a l i n e , p o l yh a l i n e , a n d eu h a l i n e ) h a v e 330 b e e n i n dic a t ed on t h e m ap a nd r i vers a r e p roj e c ted i n blue . P CA pl ots sh o w th e t axo nom ic (b ) a n d 331 fu nctio n al (c) co mpo s it ion o f Ulva -a ss o ci ate d ba c te ri a l communi t ie s a cr o s s s a lin i ty. Th e o r d ina ti o n in t h e 332 ta xo no m i c co mpo s it ion is M AG- b as ed (m et age n o me- ass em bled ge no m e ) a n d th e func tio nal co mpo s it ion 333 i s K O- ba se d ( K E G G O r t ho l og ie s ) . T he co n t ou r line s ( smo o th su r fa c e line s ) we r e fi t te d to t he o rd ina ti on 334 p l o t s b a se d o n t h e c o r re l a t i o n w i t h sa l i n i t y . C o l o u r r e p re se n t s s a l i n i t y i n a l l p l o t s. 335 336 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 17 337 Fig. 3 | Abundance of metagenome-assembled genomes across salinity. (a) P h y lo g en et ic tr ee o f t h e 294 338 me ta ge nome - a ssem b le d gen o me s ( MA G s ) th at s ign i fi c ant l y di f fe r e d ( p <0 . 0 1 ) in r e la tive a b und an c e 339 acros s t he s a l i nit y gr adi en t. Th e out e r ri n g repr es e nts t h e l og 2F old Ch an g e (Li nD A li n ear r egr ess i o n 340 model) , wit h th e M AGs i n yello w-gr e en t ha t a r e e n r ich ed i n l ow s al i n i t y a n d t h e MA Gs th at a r e e nrich ed i n 341 high s ali n it y i n pu r ple- blu e . Th e c o l ours i n the p h yl o ge n et i c t r e e re p re sent th e ba cte ria l p hy la ( t he l eg en d 342 is ord er ed i n cloc kwis e or der of a pp e a r anc e ) . (b) Rea d a bun dan c e o f t he 2 8 m o s t di f fe r e n ti a l ly ab un dan t 343 MAGs acros s s al i nit y. Th e x- ax is d ispl a y s s a li n it y (P S U ) and t he y - axi s CL R-t ra n sfor m ed r ead a bu n danc e. 344 Col o ur s r epr es e nt s a li n it y . Cur v es wer e fit t ed wi th a g en er aliz ed l i ne a r mo d e l (GL M) usi n g th e R p acka ge 345 ggplot2 . Sh a d e d ar e as r epr es e nt t he 0 . 95 c o nfid enc e i nt e r val . 346 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 18 347 Fig . 4 | T ax on om i c tur no v er a nd fu n ctio nal stab ility . H e at m a p d e pic t ing t he pr es e nce (d a r k-g r ee n) or 348 abs enc e (li ght - blu e ) of KEGG mod ules i n 2 6 m et age n ome - ass e mbl e d ge no mes ( MA Gs ) th a t ar e 349 char act eris tic t o a sp ecif ic s ali ni t y r e g i o n (hor o h al i n icum 5– 8 P S U , meso h alin e 8 – 1 8 PS U , po l y h a l i n e 18 –350 30 PSU , or eu h a li n e 30 – 35 P SU). 351 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 19 Osm oregulation drives functional var iation 352 Despite low er functional turnover , we also identified 23 KEGG modules that differed in 353 abundance across the salinity en vir onment (p<0.05, LinDA linear regr ession) (F ig. 5). 354 These modules are lik ely in volv ed in the osmoregulation of bacterial cells, but may also 355 affect the osmoregulation and acclimation potential of the host. One of the best-known 356 strategies in bacteria and eukaryotes alike is the accumulation of low molecular w eight 357 compounds, su ch as sugars and amino acids, that act as osmoprotectants t o maintain 358 osmotic homeostasis and turgor pressure. Oth er strategies included the stabilization of 359 cell membranes and mitigating oxidativ e stress. 360 Salinity-induced osmotic changes trigger the o verproduction of r eacti ve oxygen species 361 (ROS), causing oxidati ve stress 59 and damaging membrane lipids, proteins, nu cleic 362 acids, and chloroplasts 60 . Bact eria t hat were abundant in high salinity areas more often 363 contained the genes needed t o produce factor F 420 (KEGG module M00378, p=0.0004), 364 thiamine (vitamin B 1 ; KEGG module M00127, p<0 .0001) and p yridoxal-5P (the acti ve 365 form of B 6 ; M00124, p=0.001). Se veral antioxidant mech anisms ar e F 420 -dependent 61 366 and cofactor F 420 is known to help bact eria combat oxidati ve str ess 62 . In t errestrial 367 plants, both vitamin B 1 and B 6 are known to alleviat e salinity stress, for example in 368 Arabidopsis t haliana 63 and in milk thistle 64 , due to the stimulation of antioxidant 369 enzyme activity and proline cont ent . Similarly , gene exp ression of vitamin B 1 is 370 upregulated in phytoplankton during salt str ess and oxid ati ve stress is t hought to 371 function as a stress signalling molecule 65,66 . In addition to pro tection from oxidative 372 stress, pyridoxal-5P also incr eases photosynthetic pigment in wheat 67 , and it facilitates 373 growth through the reduction of eth ylene accumulation that usually occurs under 374 salinity stress in rice 68 . As algae are likely not always able to produce their own 375 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 20 vitamins 53 , it is possible that thiamine and p yrodixal-5P produced by bacteri a also aid 376 the Ulva cells against oxidativ e rad icals that are more pre valent under high salinity 377 levels. 378 Bacteria that were abundant at high salinity also more oft en contained the genes needed 379 to produce betaine (M00555, p = 0.03), an osmolyte that acts as one of the pref err ed 380 compatible solutes in the majority of prokaryotes 69,70 . It is known to accumulat e in 381 bacterial cells with incr easing salinity 71 , and betaine transport ers are enriched in 382 bacterial communities originating from marine habitats compared to fr esh water 383 environments 72 . Although betaine plays a more important role in higher plants than in 384 algae as osmoprotectant 73 , a transcript omic study show ed that three choline 385 deh ydrogenase genes (in vol ved in the conv ersion from choline to betaine) w ere 386 upregulated in Ulva compr essa during a reco very period after hyposaline stress 74 . W e 387 also found an enri chment at high salinity of bact erial genes necessary for the produ ction 388 of the es sential amino acid phenylalanine, which is known to accumulate in plants and 389 algae under high salinity conditions 75,76 and increases salinity tolerance in maize 77 . 390 Genes associated with other well-known osmolytes and osmoprot ectants, such as 391 ectoine and tr ehalose, were compar ativ el y less pre valent in our d ataset. How ever , it is 392 important to k eep in mind that these findings are solel y deriv ed from gene cont ent 393 analyses and do not reflect gene expression patterns. Future transcriptomic studies 394 could shed more light on the up- or downregulation of specific genes b y Ulva and its 395 associat ed bacteria in response to changing salinity conditions. 396 Another w ell-known strategy of bacteria to cope with cell turgor pressu re is to alter the 397 membrane composition through changes in fatty acids or phospholipid s 78 . The cell 398 membrane separates the cell’s interior from the external en vironment and is therefor e 399 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 21 the first structure to encounter the effects of fluctuating salinity and osmotic stress. The 400 disruption of membranes aff ects man y processes such as transport of compounds, 401 enzyme activities and si gnal transduction. It is therefore important t o maintain the 402 correct fluidity of the lipid bila yer 79 . W e found that many typ ical h igh-salinity MAGs 403 wer e associated with the ability to synthesize ubiquinone (M00117, p= 0.02), wh ich 404 agrees with the findings by Dupont et al. 72 on pelagic bact erial communities. 405 Ubiquinone (also called coenzyme Q) is a membrane-stabilizing isopr enoid and the 406 accumulation of this compound increases salt t olerance in bact eria, especiall y in the 407 thin-walled gr am-negativ e bacteria 80,81 . Ubiquinone alters the phy sicochemical 408 properties of the membrane by incr easing the lipid packing and density . This results in 409 reduced membrane permeability (i.e., a slower r elease of small solutes) and in cr eases 410 the strength of the membrane (i.e., resistance to cell rupture) 82 . Con versely , we found 411 that the module for phosphatid ylcholine biosynthesis was enriched in low salinity 412 conditions (M00091, p=0.002). Phosphatidylcholin e (PC) is a membrane-forming 413 phospholipid that is synthesized from Phosphatid yl ethanolamine (PE) and studies 414 found that PE levels increased in salt-adapted cells 83 . It has been not ed before that 415 while PE is the more common phospholipid in bacteria, gram-negativ e bacteria with 416 high proportions of unsatu rated fatty acids often contain add itional PC to maintain 417 stable bilayers 84 . Indeed, in our dataset, th e ability to con vert PE to PC w as mainl y 418 found in low salinity enrich ed Sphingomonadales (e.g., Alteraurantiacibacter , 419 Erythrobacter ) and Rhodobact erales (e.g., Jannaschia, Pseudorhodobact er ). Membranes 420 lacking PC are more fluid, hav e a higher permeability for small molecules and ar e more 421 sensitiv e to osmotic ch anges 85 . Both quinone and phosphatid ylcholine seem to stabilize 422 the membrane to withstand changes in tu rgor pressur e and maintain osmotic balance. 423 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 22 Bacteria-mediat ed acclimation 424 Our char acterization of the metabolic functions of a typical Ul va -microbiome, highlig hts 425 the large metabolic potential inherent to bacterial-algal symbiosis. The holobiont 426 concept, whi ch r egard s the seaw eed host and its associat ed microbes as an int egr ated 427 functional unit, is essential w hen studying the physiological response of sea weeds to 428 environmental change. This work demonstrat ed that both the taxonomic and functional 429 composition of Ulva-associated bacterial communities change across a 2,000 km salinity 430 gradient. While Ulva -associat ed bacterial taxa displa yed high taxonomic turnover across 431 salinity , se veral Ulva species were able to colonize the entire gr adient. This indicates 432 that Ul va -bacter ia have a smaller salinity-bas ed niche than the host and are les s tolerant 433 to changes in salinity conditions (i.e. limit ed acclimation potential). The high turnover of 434 microbial taxa is accompanied b y functional redundancy , wh ere guilds of taxa along the 435 entire environmental gr adient can perform crucial functions. These functions, includ ing 436 amino acid and vitamin B production, are pot entially important to the seaw eed host. 437 Alongside functional r edundan cy , w e identified distinct functional modules exhibiting 438 enrichment in either low or high salinity areas. These modules ar e likel y in vol ved in 439 mitigating oxidativ e stress, maintai ning cellular osmotic homeostasis, and stabilizing 440 cell membranes. Ulva depends on its microbiome for morphological de velopment and 441 growth — lik ewise, the forementioned bact erial acclimation mechanisms may play a 442 role in host metabolism and acclimation. In light of bacteria-mediated acclimation, 443 future laborator y experiments — involving the inoculation of s eaweed cultures with 444 targeted microbial communities — will be necessary to inv estigate whether bact eria can 445 indeed facilitate the acclimation of Ulva species to changes in salinity . 446 447 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 23 448 Fig . 5 | F u n ct ion al c h ang es i n the Ulva -a ss o ci a t e d ba c t er ial c om mun itie s a c r o s s s ali nit y . (a ) 449 log 2 F oldCh a n ge v al ues ( ± s t and ar d e r r or) of th e KEGG m o d ules t h a t s i gni fic a ntl y (p< 0.05 ) d iff er e d i n 450 a b un dan c e wi t h sa l i ni ty . (b ) T a x o no m ic co mpos it io n (or d er l e v el) of th e me t age no me -a ss e m bl e d 451 gen o m e s ( M A G s) t h at h ar bour ed t h e id e nt i fi e d KEGG m o dul es . Th e r el a ti v e a b u nd a nce is b as ed on r e ad 452 coun ts of th e or d ers acr o s s th e e n tir e dat as et. 453 454

455 Sample collection 456 Samples of Ulva sensu lato individuals (n=91) w ere collected during June–August 2020 457 in the Baltic Sea area (T able S1). Of each indi vidual, a tissue sample w as collected t o 458 molecularly identify the host species and a sw ab sample for microbiome an alyses was 459 gener ated by rubbing f or 30 s on the tissue. Sterilized disposable glov es and st erilized 460 equipment wer e used throughout the sampling procedure to minimize contamination. 461 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 24 All samples w er e stored in a portable fr eezer (-20 °C) until transferred to -80 °C in the 462 laboratory . 463 In total, 63 sampling sites w er e visited along a salinity gradi ent in the Baltic Sea and 464 adjacent areas such as the Kattegat, Skagerrak, and the east ern North Sea (Fig. 1a). Th e 465 salinity r anged from 5.1 to 34.3 PSU and is pr esented in the figur es in this study either 466 on a continuous scale (0–35 PSU) or in salinity zones defined according to the V enice 467 classification sy st em (5–8 = horohalinicum, 8–18 = mesohaline, 18–30 = polyhaline, and 468 30–35 = euhaline) 86 . In addition, w ater temperature (°C), oxygen le vels (mg L -1 ), and 469 nutrients concentrations (NO 3 - , NO 2 - , silicate, PO 4 3- in µmol L -1 ) were measured at eac h 470 site (T able S1). 471 Molecular identification, based on the tuf A marker , confirmed the samples in our dataset 472 represented sev en different species: Blidingia minima (Nägeli ex Kützing) K ylin (n=8), 473 Ulva compr essa Linnae us (n=10) , Ulva fenestrata Postels & Rup recht (n=8), Ul va 474 int estinalis Linnaeus ( n=29), Ulva lacinulata (Kützing) Wittrock (n=10), Ulv a linza 475 Linnaeus (n=20), and Ulva torta (Mert ens) T revisan (n=6) [see v an der Loos et al. 25 and 476 Steinhagen et al. 23 for detailed molecular methods and additional r esults concerning 477 Ulva d i versity in the Baltic r egion]. Throughout this stud y , “ Ulva ” r ef er s t o Ulva sensu 478 lato (including Blidingia ). 479 480 DNA extraction and metagenomic sequencing 481 T otal microbial DNA of the sw ab samples w as extracted with the Qiagen DNeasy mini kit 482 following the manufactur er’s protocol, with the addition of a bead beating step before 483 lysis using zirconium oxide beads (RETCH Mixer mill MM400; 5 minutes at 30 Hz). 484 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 25 Quantity and quality of the DNA extracts were v erified with Qubit (Lif e T echnologies, 485 Grand Island, USA) and NanoDrop (Thermo Scientific, Wilmington, USA). DNA extracts 486 wer e sent to Novogene (Cambridge, United Kingdom) for libr ary pr epar ation and 487 metagenomic sequencing on an Illumina No vaSeq 6000 (150 bp paired-end). A neg ativ e 488 DNA extraction control and a positi ve control (A T CC microbial standard MSA-1002) 489 wer e included. A t otal of 4,297,091,260 reads were generated (35,622,356–78,544,930 490 reads per sample). The sequences are archiv ed at SRA (BioProject PRJNA1040445). 491 492 Bioinformatics and statistical anal yses 493 The metagenomic sequen cing data was processed with the ATLA S Snakemake workflow 494 87 , which integrates quality control, assembly , genomic binning, and annotation. In short, 495 quality control w as perf ormed using the B BT ools suite 88 . This includes removal of PCR 496 duplicates and adapters, trimming and filtering of r eads based on qualit y and length, 497 and compressing the raw data files. Host sequences were remov ed based on an a v ailable 498 Ulva r efer ence genome (B ioProject PR JEB25750) 89 . The de novo metagenome assembly 499 was done using MEGAHIT v1.0 90 . Next, metagenome-assembled genomes (MAGs) were 500 predicted with binning tools MetaBAT 2 91 and MaxBin 2.0 92 . Binning results w er e 501 aggregat ed with D AS T ool 93 . Quality assessment of the r esulting MAGs was performed 502 with CheckM 94 . MAGs were der eplicated across samples with dR ep 95 , and taxonomic 503 classification was performed with G TDB-Tk 96 . Finally , genes w er e predicted with 504 Prodigal 97 , red undant genes w ere clustered with lin clust 98 , and annotat ed with 505 eggNOG-mapper 99 . This resulted in a classification of the genes following the 506 Carbohydrate-Activ e EnZ ymes database (CAZ y) 31 and K yoto Encyclopedia of Genes and 507 Genomes (KEGG) 32,33 databases. 508 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 26 The effect of salinity , NO x , and PO 4 on bacterial community composition at both MAG 509 (taxonomic profile) and K O (KEGG Orthology; functional p rofile) le vels was assessed 510 using the en vfit function from the v egan package with 9999 permutations 100 . 511 Multivari ate comparisons with 9999 permutations w er e subsequently conduct ed with 512 pairwise Adonis among the d iffer ent salinity zones 101 . Functional differ ences across 513 salinity w ere visualised with a PCA and smooth surface lines were fitted to the 514 ordination with the ordisurf function (v egan package) based on the correlation with 515 salinity 100 . A LinD A linear regression was used to identify which MAGs and which KEGG 516 modules (a set of genes with a specific reaction within a metabolic p athwa y) 517 significantly changed across salinity and nutrient concentrations 102 . P-values were 518 Benjamini & Hochberg corr ect ed. Gi ven the compositional natur e of the data, r ead 519 abundance v alues w er e tr ansformed with the centric log-ratios (CLR) prior t o the 520 analyses 103 . 521 T o gain a deeper understanding of what defines the metabolic potential of Ulva 522 microbiomes, we compared the genomes of bacterial taxa isolated from Ulva to those 523 isolated from seaw ater . In total, we selected 152 MAGs from our metagenomic dataset, 524 representing 33 different genera. W e then searched for publicl y available genomes of 525 bacteria from the same genera that wer e isolated from sea w ater (resulting in a set of 71 526 genomes) (T able S4 ). These originated from a v ariety of geographical locations and 527 habitats (e.g., surface w ater , deep sea, hydrothermal sy st ems, and oceanic gyr es). 528 Subsequently , we conducted a comparativ e analy sis based on odds ratios t o identify 529 pot ential enrichments of specific K O t erms or CAZ y f amilies within bacteria of the same 530 genus collected from Ulva versus seawat er . F or each K O term and CAZ y family , the odds 531 ratio was calculat ed as the number of discordant genome pairs in f avour of the Ulva 532 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 27 (ter m/family present in the Ulva bacterial genome, but not present in the sea water 533 bacterial genome) di vided by the number of discordant pairs in f avour of sea water 534 (term/family present in the sea wat er bacterial genome, but not present in the Ul va 535 bacterial genome). An offset of 0.5 was added to the number of discordant pairs to 536 prev ent dividing by zero. As multiple genomes were a vailable per genus, pairs were 537 randomly assigned 1000 times (permutations) and odds ratios w ere calculated for each 538 permutation. The median odds ratio was r etained. A t erm/family more frequent in Ulva 539 bacterial genomes results in an odds r atio larger than one. The opposite results in an 540 odds ratio smaller than one. 541 All statistical tests were performed in R 104 and data wer e visualised using the ggplot2 542 105 and phyloseq 106 packages. 543 544 Data availability statement 545 Raw whole-genome sequ ence r eads and r elated metadata are deposited in the SRA 546 (BioProject PR JNA1040445). 547 548

549 The research leading t o the results present ed in this publication w as carried out with 550 infrastructur e funded by the FW O P hD F ellowsh ip fundamental r esearch (3F020119), 551 the EMBR C Belgium (F WO project I001621N), and the F ormas-funded ‘ A manual for the 552 use of sustainable marine resources’ project (Grant no. 2022-00331). W e would like to 553 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 28 thank Samanta Hoffmann for her assistan ce during field work and Nadja Stär ck f or 554 assistance with nutri ent analy ses. WS is funded through the FW O grant nr 1252821N. 555 556 Author contributions 557 L.M.L .: Conceptualization, Methodology , F ormal analy sis, Data Cu ration, Writing - 558 Original Draft, V isu alization. S.ST .: Investigation, W riting, F unding acquisition - Re view 559 & Editing. W. S . : Resources, W riting - R eview & Editing. F. W . : W riting - R e view & Editing. 560 S.D.: Methodology , R esources, Writing - Review & Editing. A.W .: W riting - Review & 561 Editing, Supervision. O .D.C.: Writin g - Review & Editing, Supervision, F unding 562 acquisition. 563 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 29

564 1. P e s s arrodona , A . et al. Glob al seawe ed p r oduc t i vity . Sci . Adv . 8 , ea b n2465 (20 22) . 565 2. Eg an , S. et a l . The s ea wee d hol obio n t: U nderst anding s e awee d - b ac t e ria int e r acti ons . FE MS 566 Microbiol . R ev . 37 , 462–476 (2013) . 567 3. van der Loo s, L . M., Eri kss on, B. K. & F al c ão Sa lle s, J . Th e macr oa lg al holobi ont in a c hanging 568 se a . T ren d s Mic ro biol. 27 , 635 – 650 (2019 ) . 569 4. W ahl, M. , Go ec k e, F . , Lab e s , A., D o b r e tso v, S. & W einb e r ger , F . The s ec ond skin: E col ogical r ol e 570 of epibi o t ic bio fil ms o n marine o r g a ni sm s . Fro n t. Micro bio l. 3 , 1 – 21 (2012 ). 571 5. Li, J . et a l. Ba c t e rial c o n t r olled mit ig a ti on of dysbio si s in a seawe ed di s ea se. ISME J. 16 , 378–3 87 572 (2022). 573 6. G e min, M . et a l. Gr o wth -p r om ot i ng bac teria f o r the g r ee n se aw ee d U lva c lat hra ta. Aq uac . R e s. 574 50 , 374 1–3748 (2019). 575 7. Riquelme , C., R oj a s , A . , Flores , V . & C orre a, J . A. Epiph y tic ba c t eri a in a copp er -enri ched 576 en vir onme nt in nor the rn Chil e . M ar . P ol l ut. B ull. 34 , 816 –820 (199 7). 577 8. W i char d, T . et al . The green seaw eed Ul v a : A model s yst em t o study mo r pho gene si s. F r ont. 578 Pla n t Sc i. 6 , 1–8 (20 15 ). 579 9. Pr ova soli, L. E ff e c t of pl a nt hormon e s on Ulva. Bi ol. B ull . 114, 375–384 (1958 ). 580 10. Spoern er , M . , Wichard, T ., Ba chh ube r , T . , Str atmann, J . & O er t el, W . Gr owth a nd t h allus 581 morphog e ne si s of Ul va m utabili s (C h l or op h yt a ) d e p e nd s on a c o m bi na t i on o f t w o b a ct e r i al 582 s pe ci e s e xcr eting r e gul at o ry f a c t ors . J . Phycol. 48 , 1433– 1447 (20 12). 583 11. W eiss , A ., C o s ta, R. & Wic ha r d, T . Mo rph og ene s i s o f U l va m ut ab i l i s (Chl or oph y t a ) induc ed by 584 Mar i bac t er speci e s ( Ba c t eroide t e s, F la vo ba ct e riac e ae) . Bo t . Ma r . 60 , 1 97–206 (20 17). 585 12. G ruen e be r g , J., E ng el en, A. H. , Co st a , R. & Wicha r d, T . Ma croalgal morpho gene si s i nduced by 586 w at erb orne c ompound s and b acteria in c oa s t a l se aw a t e r . PLo S ON E 11 , 1– 22 (201 6). 587 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 30 13. van der Loo s, L . M., De Wil de, C., Wil lem s, A., De Cler ck , O . & S t e inhage n, S. Th e c ultivat e d s e a 588 let tuc e ( Ulv a ) micr obi o me: Suc ce s sion al a nd s ea sonal dy namic s . A qua c u lt u re 585, 74 0692 589 (2024). 590 14. D avis , K. M. et al . S ucce ssional dynam ic s of th e c ultivated k elp micr obio me . J. P hyco l . 59 , 53 8–591 551 (202 3). 592 15. D ü seda u , L. et al. E levat e d t empera t ur e -i nduc ed epi micr obi ome shi ft s in a n in v a s i ve s eaweed 593 G rac ilaria v er micu lop h yl la . M i cr oo r gan is ms 11 , 599 (20 23 ). 594 16. Minich, J. J . et al . E l e va t e d t e m pe r a t u r e dr i ves k e l p m i c r o b io m e d y s b io si s, w h i le e le va t e d 595 c arbon dio xid e induc e s w at e r mic r obiom e di sruption . PL o S O N E 13 , 1–23 ( 2 018 ). 596 17. P aix , B. et a l . S y ner gi st i c eff e ct s o f t e mpe r a tu r e an d l igh t a ff e c t th e r el a ti on ship be tw ee n T aon i a 597 at o mari a and it s e pi bac t e r ia l c o mmunity : a c ontr oll ed c ondi t io n s study . E n v ir on. M i c r o b i ol . 23 , 598 6777–6 797 (2021 ). 599 18. D a i, C. & W ang , S. Th e s truc t u r e and func t io n of the S ar gas su m f us if o r me mic r obi ome und er 600 diff erent c o ndi tion s . J. Mar . Sc i. En g. 10 , 1 401 (202 2). 601 19. G l a s l , B. e t al. Mic r obia l s u rf ac e bi o f i lm resp ond s t o t he gr ow th- r ep r oducti on -s en esc e nce c y c le 602 of th e domina n t c oral r e ef mac r o alg a e Sarga s s u m s p p . Lif e 11 , (202 1). 603 20. W ang , J . , T ang , X., Mo , Z. & Mao , Y . Me t a genome -a s s e mbl ed gen ome s from Pyrop ia hai t a nen s i s 604 micr obio me pr ovide in sight s int o th e po ten ti a l met a bol ic func tion s t o t h e se aw ee d. Fr o nt . 605 Microbiol . 13 , 1 –14 (2022 ). 606 21. W eig el , B. L. , Miranda, K. K., F og arty , E. C ., W a tson, A. R . & P fi s t er , C . A. Fu nc tional i ns ig hts int o 607 t h e k e lp mic r ob iom e from met ag enome - as sembl ed gen ome s. mSys t e m s 7 , e014 2 2-21 (2022 ). 608 22. Ra th , K. M., Fi erer , N., Murph y , D . V . & Ro us k , J . Li nking bacteria l comm unity comp os i t i on t o s o il 609 s a lini ty a long e n vir onme nt al gr adie n ts . IS M E J . 13 , 836–846 (2019). 610 23. St ein ha g e n, S ., Ho ffmann , S ., Pa via, H. & T ot h , G . B . Molec ul ar ide nt i fi ca ti on of t h e ubiquit ou s 611 gr een a lg a e Ul v a r ev e al s high biodive r sity , cryptic ity , a nd in vasive spec ie s in t he A tl a n t i c -Bal tic 612 Sea r e gio n. Al ga l R e s. 73 , 103132 (2023 ). 613 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 31 24. R eu sch, T . B. H. et a l . The B altic Se a a s a t ime ma chine f or t h e futu r e c o a s t al oce a n . Sc i. Adv . 4 , 614 (2018). 615 25. van der Loo s, L . M. et al. Sali ni ty and ho s t drive Ulv a a ssoci ated bac t eri al commun itie s ac r o ss 616 t he A t l an t i c- B alt i c Se a gr ad ie n t . M ol. Ec ol . 1–18 ( 2 022) doi :1 0.1111 /mec .16462. 617 26. G r e e n, J . L. , Bohann a n, B. J. M. & Whi t ak e r , R. J. Mic r obial b i og eog r a ph y: from ta xonom y t o 618 tra i t s . Sc i e nc e 320 , 1039 –1043 ( 2 008). 619 27. Lo zupone, C. A . & K night, R . G l o bal pa t t e rns in ba c t e rial dive rsity . Proc . Na tl. Aca d . Sc i. U . S. A . 620 104, 11 436–114 40 (2007 ) . 621 28. Br o ck, T . D . & Cly ne, J . S igni fic anc e o f algal ex cr et o ry pr o duct s for gr owth o f epi li mnetic 622 bact e ria . Appl . Environ . Mi crobiol . 47 , 73 1 –734 (198 4). 623 29. Lage, O. M. & G r a ça, A. P . Bi o film s : An Ex tr a Co a t on Mac r o a lg ae . Alga e - O rg . Im m ine n t 624 Biotech nol . (201 6) d oi:10.5 772 /63053. 625 30. H ol lant s , J ., Le lia er t, F . , De Cle r c k, O . & W illems, A. What we c an lea rn f rom s u shi : A r eview on 626 s e aw e e d- b act er i a l as s o cia t i o ns. FEMS Microbiol . E col. 83 , 1 –16 (2 013) . 627 31. Cant a r el , B . L. et a l. The Ca rboh y d r a t e -A c tive EnZ yme s d ataba se (CAZ y ): An e xp ert r e sour c e f or 628 G l y c og enomic s . N u cl e i c A c i ds R es . 37 , D2 33–D238 (2009 ). 629 32. K anehi s a, M. & G oto , S . KE G G: K yo t o Encyclope dia o f Gen e s a nd Ge nome s . Nucle i c Ac ids R es . 630 28 , 27– 30 (2000 ) . 631 33. K anehi s a, M. et al. K EG G f o r li nking g eno me s t o l i f e a nd t he en vi r onm e n t . Nucle ic Ac ids R e s. 36 , 632 480–48 4 (2008). 633 34. W ahl s tr öm, N. et a l . Compo si tion a nd st r uc t ur e o f cel l w a ll ulvan s r e c ov er e d from Ulva spp. 634 along t h e S wedi sh we s t co a s t. Ca r b o h ydr . Polym. 233, 11 5852 (20 20). 635 35. Y ang , Y . et al. Ide n ti ficati on and char ac teri z at i o n of ma rine sea weed s f o r biocomp ounds 636 pr oduction . E nviron . T ec hnol. In nov . 24 , 1 01848 (2 021). 637 36. Scho lz , S. et al . Sulf oq u inovo se me t a boli s m in ma r ine algae. Bo t. M a r . 64 , 30 1 –312 (20 21). 638 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 32 37. H a n, L. , Y an g , G. - P ., Liu, C . -Y ., Jin , Y . -M. & L iu, T . Emiss i on s o f bio g eni c sulf ur c om p ounds and 639 t h eir r egul a tio n by nutri ent s during a n U lv a pr o lif e r a bl oom in the Y ellow Sea . Mar . P oll u t. B ull. 640 162, 11 1885 (202 1). 641 38. V an Al styne, K. L., K oell e r me i er , L . & Nel s on, T . A. S p atial vari a tio n in 642 dimeth yl s ul f on iopropion at e ( DMS P) p r od uc tion in Ul v a l a c t u c a (Chl or o ph y t a ) f r o m the 643 No r t h e a s t Pa c i f i c . Mar . Biol . 150, 11 27–1 135 (2007). 644 39. K es s le r , R. W ., W e i ss, A. , K ue gler , S. , Her mes, C. & Wicha r d, T . Ma croalgal–b ac teri al i n t e r a ct ion s: 645 R ole of d ime th ylsul f onio p r op ion at e in mi cr obia l g a r de ning by Ulva (Chl or op h y t a ). Mol. E c ol . 27 , 646 1808–1 819 (2018 ). 647 40. Kidg el l, J . T ., M agn uss on, M. , de N y s , R. & Gla s s on, C . R . K. Ulvan: A s y s t em at i c rev iew of 648 e xtr ac t i o n, c om po s i tion and func t i on. Al gal R e s . 39 , 10 1422 (201 9 ). 649 41. T ang , T ., Cao , S ., Zh u, B. & Li, Q. Ulvan pol ys a cc ha ride -degr a d ing enz yme s : An upd at ed a nd 650 c ompr eh e n s ive r evi ew of s our ce s cat ego ry , pr oper ty , struc ture, and appli c a tion s o f ulvan 651 ly ase s . Alg al R e s. 60 , 10 2477 (202 1) . 652 42. R eisky , L. et al. Bio chemic al char ac t eriz a t i on of an ul van ly a se fr om t he ma r i n e fla v obac t e r iu m 653 F ormo sa agari ph il a KMM 3901 T . Ap p l. M icrobi ol. Bio t ec h nol. 102, 6 987 –699 6 (20 18). 654 43. St ein ha g e n, S ., S t e dt, K ., T rigo, J . P . , Unde land, I. & P a vi a, H . A st ep t ow a r d s c los i ng the f oo d-655 w aste g a p in novel pr o t ei n s o ur c e s : P o s t - har v e s t p r o t ein boo s t o f th e sea w e ed cr o p Ulva b y 656 herr i ng pr o duc t i on tub w a t e r . Futur e Foo ds 9 , 10 0347 (20 24 ). 657 44. T yl er , A. C., M c Gl a th ery , K. J. & A nders o n, I. C. B enthi c al g ae c ontr ol sedim ent—wat e r column 658 flux e s o f or ganic and ino r ganic nitr o g en c ompound s in a tempera te la goon. Li mn o l . O ce a no gr . 659 48 , 212 5–2137 (2003). 660 45. Li, H . e t al. N it r og e n upt ak e an d ass im i la ti o n pr e f e r en c es of t he m ai n g r ee n t i de a lg a Ulva 661 prolifera in t h e Y el low S ea, Ch in a. J . A ppl . Ph ycol. 31 , 625– 635 (201 9 ). 662 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 33 46. T yl er , A. C., M c Gl a th ery , K. J. & M ack o , S. A . U p t a k e o f u r ea and a mino ac ids by the macr oa lg ae 663 U lv a lac tuca (Ch loroph yta) a n d Gr ac ilaria v ermic uloph ylla (Rhodo phyt a ). M ar . Ec ol . Pro g. Se r . 664 294, 16 1–172 (2 005). 665 47. O li veir a, L. S . D . et a l. T r a ns c r i pt o m i c a n aly s is o f t h e r e d s e a w ee d La ur e ncia de ndr oide a 666 (Florid eophy ce a e, Rhod o ph yta) an d i t s mi cr obiome . B MC Ge nomic s 13 , 1 –13 (201 2). 667 48. Xiong , Y ., Y ang , R., Su n, X., Y ang , H . & Che n , H. E ff ec t o f t h e epiphy tic ba cterium Ba c illus sp . 668 W P ySW2 on t h e met a boli sm of P yropi a h a it a nen si s . J. Ap pl. Phyc o l. 30 , 1 225–12 3 7 (2018). 669 49. Flyn n, K. J . & W righ t , C. R. N. The simulta ne ou s a ssimila t i on o f ammoni um and l- a r ginine by the 670 marine di atom Ph ae o dactyl um tricornu t um Bo hlin . J. E xp. Mar . Biol . Ec ol. 95 , 257 – 269 (198 6) . 671 50. Lina r e s , F . E ff ect o f di s solved f ree amino acids ( DF AA ) on t h e bioma s s and produc ti on of 672 micr oph y t ob enthic c ommuni t i e s . J . E x p . M ar . B i o l. E c o l. 330, 469–481 (2006) . 673 51. LeBla nc, J . G. et al . Ba c t eri a a s vit amin su pplie r s to the ir ho st: a g ut micr obi ota pers pe c tiv e . 674 Curr . O pin . Bio t e chn ol. 24 , 16 0–168 (20 13). 675 52. R ybac k, B. , Bor tf e ld -Mil ler , M . & V o rholt , J . A . Me t a bolic adap t a t i on t o vit am in aux otr op h y by 676 leaf -a s s oc i a ted bac t eri a. ISME J . 16 , 2712 –2724 (2 022). 677 53. Cr o f t, M. T ., Law r e nce , A. D . , R aux-D ee ry , E., W a r ren, M . J. & Smi th, A . G . Algae a c quir e vitamin 678 B 12 thr oug h a s ymbio tic r el a ti on ship with bac t e ria. Na tu r e 438, 90–93 (2005 ). 679 54. D r oop , M. R . Vi t amin B 12 i n marine e c olo g y . Nature 180, 1041 –1042 ( 1 957) . 680 55. Frie s, L. R equi r em ents f or o r g ani c sub s t a nces in seaw eed s . Bot . Mar . 16 , 19–3 1 (1 973). 681 56. Cr o f t, M. T ., W a r ren, M. J. & Smith , A . G. Algae nee d thei r vit a min s . Euk a r yot. Cell 5 , 1175– 1183 682 (2006). 683 57. Mohsen , A. F ., Khale a f a, A. F ., Ha shem , M . A . & Metw alli, A . E ff ec t o f s ome vi t amin s on gr owth , 684 r e pr o duction , a mino a cid, fa t and sug ar c ont e n t s in Ulv a fasc ia t a Delil e. Bo t. M ar . 17 , (1974). 685 58. N a s r , A . H . & Bek hee t, I . A . The eff e ct o f s ome or g an ic mi cr onut r i e n t s o n s ome ma rine a lg ae 686 fr om Al e x a ndria . Hydrobi ol ogia 34 , 295 – 304 (1969 ). 687 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 34 59. Jahnk e , L. S . & Whi t e , A . L. Long -t e rm hyposa line and h y pe r sa line str e s s e s pr o duc e d i s tinc t 688 an tio x ida n t re spon se s in t h e marin e a lg a D un aliell a t er t i olec t a . J . Pla n t Physi ol. 160, 1193– 1202 689 (2003). 690 60. G a r c ía - Capa r r ó s, P . , Ha sanuzz ama n, M. & Lao, M. T . O x ida tive s tr e s s and a nt ioxida nt d ef en s e in 691 plan t s und er salini ty . in R e activ e Oxyg en, Nitr o g en a nd Sul fur S peci es i n Pla n ts (ed s. 692 H a san uz z aman , M., Fo t opoul o s, V . , Naha r , K. & Fujit a , M . ) 291–309 (W i ley , 2019) . 693 doi:10.1 002 /97811194 68677.c h12. 694 61. G urumur th y , M. et al . A nove l F 420 - d ep en dent a nt i -o xida nt mecha nis m p r o t e ct s M yc oba ct eriu m 695 t u bercul o s i s again st o xida t i ve str e ss and bact e ric idal a gent s. Mol . Mic r o bi ol. 87 , 7 44– 755 696 (2013). 697 62. G rinter , R . & Gr e eni ng , C . C o f ac t or F 420 : a n expande d vie w of it s di str ib u t i on, bio s y n th esi s a nd 698 r o le s in bac t eri a and ar ch ae a . FE M S Microbi ol . R ev . 45 , fuab021 (2021 ). 699 63. Rapala -K o z i k, M ., W ola k, N ., K uj da, M. & Bana s, A. K. The up r eg ul a tion o f t hi amin e (vit amin B 1 ) 700 bios ynthe si s in A ra b i do ps i s t ha l i a na s ee d lings unde r s alt an d o s m o t ic s tr e s s c ondi tio ns i s 701 m e d i a t e d b y ab s cis ic a ci d a t t he e ar l y s t ag e s o f t h is s tr e s s r es pons e . BM C Pl ant Bi ol . 12 , 2 702 (2012). 703 64. Mosa vikia , A. A. , Mo sa v i , S. G., S egha t ol e s la mi, M . & B ar ada r a n, R . Chi t o san n a no particle and 704 pyrido x ine see d primi ng impr ove s t ol er a nce t o s a lini ty i n milk t hi s tle seed ling. No t. Bo t. Hor ti 705 Agr o b ot . Cluj- Nap oc a 48 , 2 21–233 (2020 ). 706 65. F aa r g hl , A . A. T hiamin e and py r i do xine all evia te oxidat ive dam age b y c o pper s tres s i n gr een 707 alg a Chlorell a v ul g aris . Egypt. J . M i cro biol . 47 , 97– 110 (201 2) . 708 66. F ern, L . L. , A bidin, A. A . Z. & Y u so f , Z . N . B . Upr e gul a t ion o f thiamin e (vit amin B 1 ) b i os y nthe s i s 709 g ene upo n s t re s s applic ation i n An ab aen a s p . and N ann oc hlorop si s ocul a t a . J . P lan t B i ot e c h n ol. 710 44 , 462 –471 (20 17). 711 67. Liu, R. et al . The eff e ct s o f e x og enou s pyr ido xal-5-ph o sphate o n se edli ng gr owth a nd 712 deve lopme n t o f w he a t und er salt str e s s . Cereal R e s. Com mu n. 47 , 44 2–454 (2 019 ) . 713 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 35 68. H us sain , S . et al . Pyri do xal 5kiDB-ph o s ph at e enhanc e s th e gr owth a nd mo rpho -ph y s io lo gic a l 714 char ac teri s tic s o f r i ce c ul tiva r s by mitig a ti ng the eth y len e a ccu mulation un de r sal i nity str e s s . 715 Pla n t Ph ys i ol . Bioc he m. 154, 78 2–795 (20 20). 716 69. McP arla nd , E. L ., Al e xande r , H. & Jo hns o n, W . M. The o s mol y te tie s t hat bind: Ge nomi c insigh ts 717 in to s ynthe si s an d br e ak down of o r g anic osmoly t e s in ma r i n e mic r o be s. F r on t. M ar . Sc i . 8 , 718 689306 (2021). 719 70. Slea t o r , R. D. & Hill, C. Ba cterial o s moa da pt ation: the r ol e o f o smolyte s in b acteria l s tr e s s a nd 720 virulenc e. F EMS Mic r ob iol . R e v . 26 , 49– 7 1 (2002). 721 71. H u, X. et al . Sal t t ole r a nce mecha ni s m o f a h ydr oc a rbon -deg r a ding str ain : S alt t ole r anc e 722 mediat ed by a ccum ulat ed be t a i ne in cell s . J. Haz ar d. M at er . 392 , 122 326 (202 0) . 723 72. D u po n t, C. L . et al . F unctio na l tr a deo ffs u nderpin sali ni t y -d r i ven d ive r gen ce i n micr obi al 724 c o m m u ni t y c om p os i t io n. PLoS ON E 9 , (2 014). 725 73. Kir s t, G . O. Sa linity t ol erance o f euk a r yoti c ma r i ne algae. Ann u. R ev . Pla n t Ph ysiol. P la n t M ol. 726 Biol. 40 , 21–5 3 (1989). 727 74. Xing , Q. et a l. Compa r a t ive tran scrip t om e a nalysi s pr ovide s i n s ig hts int o r e spon se of Ulva 728 co m p re s s a t o fl uc t u at i ng salini ty c ondi t i o ns . J. Ph ycol. 57 , 1295– 1308 (202 1). 729 75. H at t a b , Z. N. A., Al -Aje el, S . A . & Kaaby , E. A. E. E ff ec t of salini ty s t r e ss on Cap sicum a nnu um 730 c al lu s gr owth, reg e nera tio n and c al lu s co n tent of cap saic in, phenylal a nine , pr ol in e and ascorbic 731 aci d. J. Li f e Sc i. 10 , (2015) . 732 76. D i t t a mi, S . M . et al . I nt e gr ative an aly s i s o f met a b oli t e and tr a n s c r ip t ab undanc e d uring t h e 733 s ho rt - t erm re s po n s e t o sal i ne and o xida t i ve str e ss in the b r ow n alga E c t ocarpu s sili c ulosu s . 734 Pla n t Ce ll En v iro n. 34 , 629–64 2 (2011). 735 77. Zahr a, N ., Ra z a , Z. A . & M ahmo od, S . E ff e ct of sali ni ty st re ss o n va r i ou s g r o wth an d 736 ph y siol og ic al attr i bu t e s o f t wo c o nt r a stin g ma iz e g eno t y pe s. Br a z. Arch . Biol . T ec hnol. 63 , 737 e20200 072 (2020 ). 738 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 36 78. O sh one, R. et a l. G enomi c, tran scriptomic , and pr o t e omic a pp r oache s t owa r d s u n ders t and ing 739 t he m o l e cu l ar m e ch a n is ms o f s a lt t o l e r an ce i n Fr a nkia str ain s i sol at ed fr om C a su arina t r e e s. 740 BMC Gen omic s 18 , 633 (2017 ). 741 79. Los, D . A. & Murat a, N . Membr an e flu idit y a nd its r o le s in th e pe r c eption o f e n vir o nment a l 742 si g n a l s . Bi oc him. B io p h ys. Acta BBA - Bi o memb r . 1666, 142–15 7 (2004) . 743 80. Sévi n, D . C. , S t ä hlin, J . N., P ol lak , G . R. , K u ehne, A. & Sa uer , U . Gl obal metaboli c re s ponse s to 744 s alt s tr es s i n f ift e e n s pe c i e s . PL OS ON E 11 , e0 148888 (2016). 745 81. Sévi n, D . C. & Sa u er , U . Ubiq uinon e ac cu mulat i on imp r ov e s o smotic - s t r e ss t ole r a nc e in 746 Es c heric hia coli . Nat . Che m. Biol . 10 , 26 6 –272 (20 14) . 747 82. Agmo Hernánd e z, V . , Eriks s o n, E. K. & Edw ar d s, K. Ubiqui n one -10 al t e r s mec ha nic al pr ope r tie s 748 and inc r ea s e s s tabi lity o f pho spholi pid m embr an e s . Bi ochi m. Biop h ys . Acta BBA - Bio mem br . 749 1848, 2 233–224 3 (2015). 750 83. Šajbi dor , J. E ff e c t of some en vir onme n tal f actors on th e co n tent and comp o sition o f mic r obial 751 membr ane lipid s. Cri t. R ev . Bi o t ec hnol . 17 , 87– 103 (199 7). 752 84. G ol d fine, H . Bac t eri al me mbr ane s an d li pid pac king theory . J. Lipid R e s. 25 , 1501–1 5 07 (1984). 753 85. G e iger , O., L óp e z-Lara, I . M. & Sohle nk am p, C. Pho sph a tidy lcholi n e bio s yn th e si s an d func tion in 754 bact e ria . B ioc him. Biop h ys. Acta BBA - Mo l. Cell Biol . L ipid s 1831, 503– 513 (201 3 ). 755 86. Alve s, A . S . et al. S p a tial di s tri bution o f s u btida l meiob e n t ho s a long e s tuari ne gr adi en t s in two 756 s ou ther n E ur o pe an e stuari e s (Portug al ). J. M a r . Bi ol. As soc. U . K . 89 , 15 29 –1540 (2 009). 757 87. Kiese r , S. , B r own, J., Z dob nov, E. M . , T rajk ov ski, M . & Mc Cue, L . A. A TL AS: a Snak e mak e 758 w orkflow f or as semb ly , a nnotation, a nd g en omic b inning of me t agen ome seq uen c e dat a. BM C 759 Bioin f or matic s 21 , 257 (2 020) . 760 88. Bus hn e ll, B . BBT ools h ttps: //s our c e for ge. net/p r ojec ts / b bmap /. (2019 ). 761 89. D e Cl er ck , O. et al . In s ig ht s in t o t h e evol u t i on of mul tic ellula r i t y f r om the s ea le t tu c e g en ome . 762 Curr . Biol . 28 , 292 1-2933. e5 (2018) . 763 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 37 90. Li, D . et al. MEGAH IT v 1 .0: A f a st and sca l able met a g e nom e a s s e mble r drive n by a dvanc ed 764 methodolog ie s a nd community p r a ct ice s . Meth od s 102, 3–11 (2016 ). 765 91. K ang , D . D . et al. Me t a BA T 2: an ada p t i ve bi nning a lgorithm f or robu st and effic ie n t g e nome 766 r e c o n s t r uc tion fr om me t age nome a ssem blie s. Peer J 7 , e7 359 (201 9 ). 767 92. W u, Y . - W ., Simmo n s , B . A . & S in g e r , S . W . Ma xBin 2.0: an au t omat e d b inning algori thm t o 768 r e c over g en ome s f r om mul t ip le me t ag en omic da t a se t s . Bi oin f or ma tics 32 , 605–6 0 7 (2016). 769 93. Siebe r , C. M . K. et al. Re c ove r y of ge nom es fr om me t ag e no me s via a dereplic atio n, a g gr ega tio n 770 and s c oring s trat eg y . N at . Mic r ob iol . 3 , 8 36–843 (2018). 771 94. P arks, D. H., Im el f ort, M ., Sk e nn er t on , C. T ., Hug enho l t z, P . & T yson, G. W . Check M : as se ssi ng 772 t h e quali t y of mic r o bi al genom es r ec overed from isol at e s, single cel l s, and metage nomes . 773 Ge n o m e Re s. 25 , 1043 –1055 ( 2 015). 774 95. O lm, M. R ., Br own , C. T ., B r ooks , B. & Ban fie ld, J. F . dRep: a t ool for fa s t and acc ur at e gen omic 775 c ompa r i son s t hat ena bl e s impr oved gen ome r e c ov er y fr om metagen ome s thr o ugh de-776 r e plica tion. ISME J . 11 , 2 864 –286 8 (2017 ). 777 96. P arks, D. H. et al . A s t a nda r diz ed bac t e r i a l t a x on om y base d on genom e phylog e n y sub s t a n t i ally 778 r e vis es t he t r e e of li f e. N a t. B iot e c h nol . 36 , 996–100 4 (2018). 779 97. H ya tt, D . et al. P r odig al : p r ok aryo t i c g e n e r ec o gni tion a nd tr a n s l ation ini ti at i on s i t e 780 identifi cation. BMC Bi oi nform atics 11 , 11 9 (2010 ). 781 98. St ein eg g er , M . & S öding , J . C lu s t e ring hug e pr o t e in se quenc e s e ts in l inea r time. N a t. C om mu n. 782 9 , 2542 (2018). 783 99. H ue rt a -C e pa s , J . et al. F a s t genome - w ide func tio nal anno ta tio n thr o ugh or th olo gy a ssignme nt 784 by eg gNO G-mapp er . M ol. Biol . E vo l. 34 , 2 115–212 2 (2017). 785 100. O k s an en , J. et al . V eg a n: c o mmuni ty ecol ogy pac k age. R pa ck a g e vers io n 2.5 -7. (2 02 0). 786 101. Mart in e z Arbizu, P . pai rw i s eAd on i s : Pai r wise mul t il evel c ompa r i son u sing adoni s. R pa ck a g e 787 version 0.4 . h tt p s: // g ith ub.com/pma rtin ez arbizu/ p a irwi se Adoni s . (2020) . 788 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint 38 102. Zhou, H., He, K. , Che n, J . & Z hang , X. Li nD A: linea r mode l s f o r d i ff e r e n tial a bundan c e ana lysi s of 789 m i c r o b i om e c om p os it i o na l da t a. G e n om e Bi ol. 23 , 95 (2022 ). 790 103. G l oor , G. B ., Mac kla im, J . M ., Pawlow s k y- Glahn, V . & Ego z cue , J. J. Micr obi ome datase t s are 791 c ompo sition al : and thi s is n ot opti ona l. Front. Mic robi ol. 8 , 2224 (2017 ). 792 104. R Core T eam. R: A langu a g e a n d en vir on men t for s t a t i stical c omputing. R Found a t ion f or 793 St ati st i c al Computi ng. (2 020) . 794 105. W i ckha m, H. G gplo t2: El eg a n t Gr ap hic s for Da t a A nal ysi s. Http s : // G g p lo t2. Tid yve rse .Org . 795 (Springer -V erlag Ne w Y or k , 2016 ). 796 106. McMur die , P . J. & Holm e s, S. Ph ylos eq: a n R pac k age f or r ep r oduci ble i nt er a c tive analy s i s and 797 gr aphic s o f micr obiom e ce n s u s d at a. PLo S O NE 8 , (2013) . 798 799 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted June 20, 2024. ; https://doi.org/10.1101/2024.06.20.599874doi: bioRxiv preprint