Features of the Opportunistic Behaviour of the Marine Bacterium Marinobacter algicola in the Microalga Ostreococcus tauri Phycosphere

Although interactions between microalgae and bacteria are observed in both natural environment and the laboratory, the modalities of coexistence of bacteria inside microalgae phycospheres in laboratory cultures are mostly unknown. Here, we focused on well-controlled cultures of the model green picoalga Ostreococcus tauri and the most abundant member of its phycosphere, Marinobacter algicola. The prevalence of M. algicola in O. tauri cultures raises questions about how this bacterium maintains itself under laboratory conditions in the microalga culture. The results showed that M. algicola did not promote O. tauri growth in the absence of vitamin B12 while M. algicola depended on O. tauri to grow in synthetic medium, most likely to obtain organic carbon sources provided by the microalgae. M. algicola grew on a range of lipids, including triacylglycerols that are known to be produced by O. tauri in culture during abiotic stress. Genomic screening revealed the absence of genes of two particular modes of quorum-sensing in Marinobacter genomes which refutes the idea that these bacterial communication systems operate in this genus. To date, the ‘opportunistic’ behaviour of M. algicola in the laboratory is limited to several phytoplanktonic species including Chlorophyta such as O. tauri. This would indicate a preferential occurrence of M. algicola in association with these specific microalgae under optimum laboratory conditions.


Introduction
Phytoplankton and bacterial dynamics are closely linked in coastal marine environments [1,2]. Numerous findings suggest that these specific relationships result from a long coexistence in the oceans for more than 200 million years [3][4][5]. Some bacterial strains are frequently isolated from natural blooms and cultivated in the laboratory [6][7][8]. Phytoplankton exudates can be important carbon substrates for bacteria, especially in early phytoplankton bloom conditions [9,10] although other carbon sources might also be important for bacterial growth [9]. In turn, it is clear now that heterotrophic bacteria can improve involved in such interactions with experiments aimed at supporting the genomic data. Then, given the ability of Marinobacter species to degrade lipids [31] and the diversity of fatty acids and triacylglycerols (TAGs) produced by O. tauri [35], the capacity of M. algicola OT to use lipids for growth was tested. We again developed a double approach to combine bioinformatics screening of Marinobacter genomes and experiments on M. algicola OT to find insights of QS capacity in Marinobacter. Lastly, a phylogenetic reconstruction including the M. algicola strain from the O. tauri RCC4221 culture and all available Marinobacter 16S sequences was first performed to detect whether there was any evidence of a common evolutionary trend between Marinobacter strains found in microalgae cultures and if they share similar phylogenetic traits.

Phylogenetic Analysis
To comment on the phylogenetic position of our M. algicola OT strain with respect to the other strains, a molecular phylogeny of 16S rRNA gene sequences was performed using sequences from different Marinobacter species. The dataset comprises 56 nucleotide sequences from GenBank including the M. algicola OT sequence (MT023716), for a total of 1415 positions after multiple alignment using the ClustalW algorithm in MEGA X version 10.1.7 software (available at https://www.megasoftware.net/, accessed on 9 August 2021) [36]. A tree was constructed using a GTR + I (invariable sites proportion) + G (discrete gamma distribution) model selected using the 'Find Best DNA Models' option in MEGA X version 10.1.7 software (available at https://www.megasoftware.net/, accessed on 9 August 2021) [36] and the maximum likelihood (ML) method using 100 bootstrap replicates (BP). All accession numbers are indicated in the tree (Figure 1). Particular attention was focused on the M. algicola clade with a comparison of sequences and the identification of specifically shared nucleic acid characteristics.

Transmission Electron Microscopy
We used transmission electron microscopy as routinely performed in the laboratory [37], to check the possibility of a close physical interaction between microalgae and bacteria. A pellet of five hundred million microalgae cells in exponential growth phase was prefixed in L1 culture medium with glutaraldehyde 1% (#16320, Electron Microscopy Sciences, Hatfield, PA, USA) for 30 min at room temperature. Then, the cells were centrifuged at 2500× g (Beckman Avanti™ centrifuge J-20XP, Beckman Coulter, Fullerton, CA, USA) for 25 min at 8 • C. Pellets were then resuspended in 40 µL of 1% low melting point agar (#A9414, Sigma, St. Louis, MO, USA) in L1 culture medium. After solidification, the plug was fixed for 2 h at 4 • C with 2.5% glutaraldehyde in one volume of 0.4 M sodium cacodylate buffer (#12300, Electron Microscopy Sciences, Hatfield, PA, USA) and two volumes of culture medium. The plug was then washed three times for 15 min in 0.2 M sodium cacodylate buffer, postfixed in 2% osmium tetroxide (#19150, Electron Microscopy Sciences, Hatfield, PA, USA) for 1 h and finally washed three times again as previously described. Samples were dehydrated in successive baths with increasing ethanol concentrations (70%, 95% and 100%) and propylene oxide (#O4332, Fisher Scientifics, Hampton, NH, USA). Samples were progressively impregnated and embedded in Epon 812 (#14120, Electron Microscopy Sciences, Hatfield, PA, USA). Thin sections were prepared with an ultramicrotome (Leica Ultracut R, Leica Biosystems, Nussloch, Germany) and stained with 2% uranyl acetate in 50% ethanol (ETOH) solution and 0.5% lead citrate in aqueous solution before examination on a transmission electron microscope (Hitachi H-7500, Hitachi, Tokyo, Japan).   Ostreococcus tauri strain RCC4221 was isolated in 1994 from the North-West Mediter-raneanThau lagoon [14] and maintained in the laboratory (cultures and cryopreservation). The O. tauri RCC4221 strain was grown in liquid medium in 50 mL aerated flasks (Sarstedt) at 20 • C under white light at approximately 100 µmol photons·m −2 ·s −1 using an LD 12:12 photoperiod. Cell growth was followed by measuring the cell concentrations in small aliquots (50 µL) of culture every two or three days with a flow cytometer (CytoFlex, Beckman Coulter, Fullerton, CA, USA). O. tauri cells were detected using the red fluorescence emission (FL3 acquisition at 670 nm) of chlorophyll pigments.

M. algicola OT Culture
The M. algicola OT strain, used for the experiments in this study, is the same that the one initially isolated from O. tauri RCC4221 in 2016 [25]. More precisely, numerous M. algicola OTUs with partial 16S sequence (obtained by Illumina sequencing) were evidenced in this previous study [25]. After cultivation of the O. tauri RCC4221 phycosphere in solid medium, only one M. algicola strain could be successfully isolated, the M. algicola OT strain used in this study. Cell growth was followed either by optical density (OD) or by measuring the cell concentrations with a flow cytometer as above. In this latter case, for enumeration of bacteria, small aliquots (50 µL) of culture were used for measurement by flow cytometry every two or three days. Bacterial nucleic acids were labelled with SYBR ® Green I (#50513, Lonza, Basel, Switzerland) and detected by green fluorescence (FL1 acquisition at 530 nm) [38].

O. tauri RCC4221 and M. algicola OT Coculture Experiments
O. tauri RCC4221 cultures used for coculture experiments were not axenic despite the use of antibiotic treatments. At the lowest possible bacterial contamination (approximately 3%) at the stationary phase, the O. tauri RCC4221 culture was diluted to 10 6 cells·mL −1 in ESAW-F/2 medium and distributed into several 50 mL sterile flasks. To test the effects of M. algicola OT on O. tauri RCC4221, different final bacterial concentrations (5 × 10 6 and 2 × 10 7 M. algicola OT cells·mL −1 ) were added to the O. tauri culture RCC4221. A control (without the addition of M. algicola OT to the O. tauri RCC4221 culture) allowed us to visualise the growth pattern of the few remaining bacteria (fewer than 3 × 10 5 cells·mL −1 ) after axenisation was not totally efficient. M. algicola OT cells were distinguished easily from other bacterial communities in O. tauri RCC4221 cultures owing to the fluorescence signal for M. algicola OT after SYBR Green labelling being lower than those for other bacteria. The same experiment was carried out in a medium without vitamin B 12 . As before, different final concentrations (5 × 10 6 and 2 × 10 7 M. algicola OT cells·mL −1 ) were added to the O. tauri culture. All coculture experiments were carried out in triplicates experiments and were performed for a total duration of 30 days. triarachidin (TAG 20:0/20:0/20:0) and hexadecyl palmitate (wax ester). Lipids were from Tokyo Chemical Industry Co. LTD, Tokyo, Japan. The strain was grown in marine broth (MB) overnight at 30 • C and resuspended at an optical density at 600 nm (OD 600nm ) of 0.1 in MB minimum [40]. Substrates were added at 0.1% (w/v) to the cell suspension in glass tubes and incubated at 30 • C under agitation at 150 rpm. Growth was monitored at 20, 44, 68 and 164 h by reading the OD 600nm .

In Silico Genomic Screening of O. tauri Phycosphere Sequences and Complete Marinobacter Genomes
Whole-genome sequencing data of 19 O. tauri mutation accumulation line sets were used for genomic screening. These 19 O. tauri non-axenic culture lines (containing O. tauri and all bacterial communities from its phycosphere) were the result of a previous extinctiondilution experiment of 378 days, which corresponded to approximately 512 generations per line with 27 points of extinction dilution to one cell [41]. These lines were shown to exclusively exhibit M. algicola cultivable species at the end point [25]. To detect whether the genomes of the remaining bacteria present in the 19 non-axenic O. tauri line cultures possessed genes of interest, we aligned the raw reads against the target gene sequences (detailed in the following sections) using Burrows-Wheeler Aligner (BWA, http://bio-bwa. sourceforge.net, accessed on 26 July 2018) with standard parameters [42] and treated the resulting bam files with SAMtools (http://www.htslib.org, accessed on 26 July 2018) to obtain the coverage of these genes [43].
We performed BLASTP of translated sequences of interest against the reference proteomes of the 21 available and complete reference genomes and proteomes of Marinobacter species, including the M. algicola DG893 strain, the closest strain to date to M. algicola OT, which were also used for screening (all accession numbers are noticed in Table 2).

Searching for Methionine Synthases Genes
We searched for evidence of methionine synthesis by M. algicola by screening the different genomic datasets.

Searching for Lipid Degradation Genes
We searched evidence for lipid degradation genes. We focused on 2 genes already identified as important for lipid degradation [31]: fadB (Fatty acid degradation alpha subunit, TIGR02437) and fadA (3-ketoacyl-CoA thiolase, TIGR02445) from the fatty oxidation complex.

Evaluation of M. algicola OT Quorum-Sensing Capacities
We investigated in this study AI-1 (AHL-driven) and AI-2 (Furanosyl diester boratedriven) modes of QS. Functional experiments with different types of biosensors were conducted to evaluate the QS capacities of M. algicola OT through AI-1 and AI-2 production, using protocols based on previously published research conducted in our group [50]. Briefly, M. algicola OT culture supernatants were tested using Pseudomonas putida and Escherichia coli-based biosensors. P. putida F117 (pRK-C12; Kmr; ppuI:mpt) was used to detect longchain (>8 carbons in the acyl side chain) N-acyl homoserine lactones (AHLs) [51][52][53], and E. coli MT102 (pJBA132) was used to evaluate the emission of short-chain (≤8 carbons in the acyl side chain) AHLs [53,54]. A mutant strain of Vibrio harveyi, BB170, was cultivated to detect the presence of AI-2 in the culture supernatants [55,56]. Protocols are precisely described in Text S1 (Supplementary Materials).

Phylogenetic Position of M. algicola OT in the Global Marinobacter Species Tree
A phylogenetic reconstruction was performed to more precisely describe the M. algicola OT phylogenetic position and to determine if all Marinobacter strains found in phytoplankton cultures tend to form a cluster. M. algicola OT, isolated from a culture of O. tauri RCC4221 [25], belongs to the M. algicola phylogenetic group (BP = 99) (black arrow, Figure 1A). This clade also includes M. algicola strains from the dinoflagellates Alexandrium AJ294359, Gymnodinium AY258116, AY258110 and Lingulodinium DQ486497, the chlorophytes Ostreococcus tauri and Dunaliella KJ573108 and the haptophytes Emiliana EF140750 and Coccolithus EF140753. All sequences from this clade share a diagnostic nucleotide: a T is present at position 762 in all sequences ( Figure 1B) while in all other sequences, a C is present at this position. While all sequences belonging to this clade share 97.4% sequence identity (1378 identical nucleotide sites over the 1415 base pairs in the total alignment), it is not possible to state if M. algicola strains found in dinoflagellate cultures are more closely related to each other than to the other strains given the relatively low variability level inside the clade. Transmission electron microscopy observations of O. tauri RCC4221 cultures revealed numerous bacterial cells that exhibited a morphology similar to M. algicola DG893 ( Figure 1C) [56]. The bacterial cells and the microalgae did not physically interact ( Figure 1C).

Phylogenetic Features of M. algicola Relative to Other Marinobacter Species Found in Phytoplankton Cultures
The closest phylogenetically related Marinobacter species to M. algicola is M. salsuginis which appeared as a sister group of M. algicola (BP = 47, considered weak support, Figure 1A) with two diagnostic characters: A at position 618 (instead of the G in other sequences) and T at position 688 (instead of the C in other sequences) (personnal observation). Other Marinobacter strains observed in phytoplanktonic cultures, such as M. adhaerens known to be associated with diatoms, do not belong to the M. algicola and M. salsuginis clades ( Figure 1A). Marinobacter sp. OTB1 (grey arrow, Figure 1A), did not cluster in the M. algicola clade but appeared to be related to a more distant lineage (M. alkaliphilus). The 16S sequences from Marinobacter sp. OTB1 and M. algicola OT strains exhibited 44 differences over the 829 homologous sites, representing 5.3% of the differences over these partial 16S sequences, which made these two strains evolutionarily distant from each other.

Effect of M. algicola OT on O. tauri RCC4221 Growth in the Absence of Vitamin B 12
Coculture experiments were compared after inoculation of M. algicola OT at different concentrations and in presence and absence of vitamin B 12 . The absence of vitamin B 12 in the culture medium affected the growth curve of O. tauri RCC4221. In the presence of vitamin B 12 (complete ESAW-F/2 medium), the maximal cell abundance was 2.5-fold higher than that in the vitamin B 12 -depleted medium. Moreover, the decay phase was much faster in the absence of vitamin B 12 ( Figure 2). Interestingly, when M. algicola OT was added to the cultures (at 5 × 10 6 and 2 × 10 7 cells·mL −1 in ESAW-F/2 complete medium), O. tauri was maintained at the highest density from day 26 to the end of the experiment ( Figure 2). However, the presence of M. algicola OT in the culture of O. tauri RCC4221 in vitamin B 12 -depleted medium did not restore the growth pattern of the algae observed in the presence of vitamin B 12 . As an evidence of the bacterial capacity to synthesise methionine (if genes are transcribed and translated), we retrieved that metH and metE methionine synthase genes from several O. tauri mutation accumulation lines containing reads from bacterial communities from its phycosphere ( Table 1) and also from in the genome of M. algicola DG893, the closest sequenced strain to M. algicola OT ( Table 2).
be associated with diatoms, do not belong to the M. algicola and M. salsuginis clades (Figure 1A). Marinobacter sp. OTB1 (grey arrow, Figure 1A), did not cluster in the M. algicola clade but appeared to be related to a more distant lineage (M. alkaliphilus). The 16S sequences from Marinobacter sp. OTB1 and M. algicola OT strains exhibited 44 differences over the 829 homologous sites, representing 5.3% of the differences over these partial 16S sequences, which made these two strains evolutionarily distant from each other.

Effect of M. algicola OT on O. tauri RCC4221 Growth in the Absence of Vitamin B12
Coculture experiments were compared after inoculation of M. algicola OT at different concentrations and in presence and absence of vitamin B12. The absence of vitamin B12 in the culture medium affected the growth curve of O. tauri RCC4221. In the presence of vitamin B12 (complete ESAW-F/2 medium), the maximal cell abundance was 2.5-fold higher than that in the vitamin B12-depleted medium. Moreover, the decay phase was much faster in the absence of vitamin B12 (Figure 2). Interestingly, when M. algicola OT was added to the cultures (at 5 × 10 6 and 2 × 10 7 cells.mL −1 in ESAW-F/2 complete medium), O. tauri was maintained at the highest density from day 26 to the end of the experiment (Figure 2). However, the presence of M. algicola OT in the culture of O. tauri RCC4221 in vitamin B12-depleted medium did not restore the growth pattern of the algae observed in the presence of vitamin B12. As an evidence of the bacterial capacity to synthesise methionine (if genes are transcribed and translated), we retrieved that metH and metE methionine synthase genes from several O. tauri mutation accumulation lines containing reads from bacterial communities from its phycosphere ( Table 1) and also from in the genome of M. algicola DG893, the closest sequenced strain to M. algicola OT ( Table 2).

M. algicola OT Growth Behaviour in Coculture with O. tauri RCC4221
A bacterial monoculture experiment revealed that M. algicola OT did not grow in ESAW-F/2 medium contrary to what was observed in the positive control culture in MB medium ( Figure 3A). However, M. algicola OT could grow in ESAW vitamin B 12 -depleted medium in coculture with O. tauri RCC4221 (ESAW minus B 12 , Figure 3B). M. algicola OT started to grow at day 8 and reached its maximal cell abundance at day 16, while O. tauri RCC4221 growth began at day 3 to reach maximal cell abundance at day 8 and then declined until complete cell disappearance at day 21 ( Figure 3B). The M. algicola exponential growth phase was strongly delayed compared to the O. tauri one (10 days for M. algicola OT versus 1 day for O. tauri, Figure 3B). The two species exhibited shifted growth curves.

M. algicola OT Growth on Lipids
For each compound tested, the M. algicola OT growth (until OD = 1.34 for the highest, Figure 4) was obtained at four different tested times, ranging from 20 to 164 h (Figure 4). Figure 4 shows that M. algicola OT was able to grow on four TAGs (tricaprylin, tripalmitin, triolein and triarachidin), a wax ester (hexadecyl palmitate) and a fatty acid (palmitic acid). These results indicated that M. algicola OT was able to assimilate the tested lipids. However, it was not possible to draw any conclusions from these data regarding the efficiency of the assimilation of the different lipids as they were nearly insoluble in water and their assimilation rate depended on the area of the water-lipid interface. The ability of M. algicola OT to assimilate a variety of lipids was also suggested by the occurrence of lipids and degradation genes in the M. algicola DG893 genome reference (if transcribed and translated). The two genes fadA and fadB target genes coding for proteins involved in lipid degradation, were found with relatively high coverage in genomics data of O. tauri phycosphere cultures (Table 3). Complete screening of available Marinobacter genomes also revealed that these two genes were retrieved in the M. algicola DB893 genome reference (Table 2)  growth phase was strongly delayed compared to the O. tauri one (10 days for M. algicola OT versus 1 day for O. tauri, Figure 3B). The two species exhibited shifted growth curves.  translated). The two genes fadA and fadB target genes coding for proteins involved in lipid degradation, were found with relatively high coverage in genomics data of O. tauri phycosphere cultures (Table 3). Complete screening of available Marinobacter genomes also revealed that these two genes were retrieved in the M. algicola DB893 genome reference ( Table 2) (with p-values = 0 after BLASTP searching). The occurrence of fadB and fadA indicates the probable presence of a fatty acid oxidation pathway in M. algicola DG893, and thus probably in M. algicola OT.   [41], all the cultivable bacteria in solid media were determined to be M. algicola OT MT023716 [25]. Here, the most abundant lipid degradation genes found from the genomic screening of the culture data are presented.    [41], all the cultivable bacteria in solid media were determined to be M. algicola OT MT023716 [25]. Here, the most abundant lipid degradation genes found from the genomic screening of the culture data are presented. We evaluated the production of AHLs by M. algicola OT using sensor strains. The M. algicola OT strain did not produce type 1 ( Figure S1A) or type 2 ( Figure S1B) auto-inducers. However, genes coding for proteins involved in QS, mainly luxI, luxS and hdtS, were present with relatively high coverage in O. tauri phycosphere cultures. These genes belong to taxonomically distinct bacteria (Table S1) and were all present in O. tauri RCC4221 culture at very low abundance rate [25]. However, complete screening of available Marinobacter genomes did not identify any homologues of the luxI, luxS, luxM, vanM, ainS or hdtS QS genes involved in AI synthesis. Nevertheless, LuxR receptor homologues were retrieved in 13 out of 21 Marinobacter genomes, including M. algicola DG893 (Table S2). INTERPRO and SMART analysis of the Marinobacter LuxR protein sequences confirmed that all of them harbour the typical LuxR DNA-binding helix-turn-helix (HTH) domain at their C-terminal region. In the N-terminal region, eight sequences presented a signal transduction response regulator domain (IPR001789), three presented a Malt-like TPR region (IPR041617) and two presented N terminal domains that were not affiliated with any known type of protein domain. However, none of them presented the AI-binding domain (IPR005143), which was confirmed by a COACH-based analysis and alignment of LuxR sequences with the Vibrio fischeri and Agrobacterium LuxR sequences (personnal observation).

Contrasting Behaviours of Marinobacter Strains in Microalgae Cultures
Bacterial growth can be substantially favoured because of organic nutrients in the surrounding environment. During blooms in Mediterranean coast, Trombetta and collaborators noticed dominant positive relationships between bacteria and phytoplankton [7] and it is known that bacterial cells can satisfy part of their own carbon demand by consuming phytoplankton exudates [57,58]. In coculture with O. tauri, organic matter quantity strongly depends on the growth stage of the microalgae, being particularly low until the middle of the exponential phase and high once the O. tauri cells reach the stationary phase. From our results, we hypothesise that bacteria largely benefit from O. tauri dead cells directly or from the exudates of accumulating living cells during exponential and decline phases. The behaviours of O. tauri RCC4221 and M. algicola in coculture contrast. Ostreococcus tauri is a known B 12 auxotroph [21]. The maximal cell abundance of O. tauri was 2.5-fold lower in the ESAW medium without vitamin B 12 than in the ESAW-F/2 complete medium in the control cultures, thus confirming that O. tauri RCC4221 needs vitamin B 12 for optimal growth. The presence of vitamin B 12 biosynthetic genes in many Marinobacter species suggested that M. algicola OT could provide vitamin B 12 to O. tauri RCC4221. However, here, we did not demonstrate any impact of M. algicola OT in favour of O. tauri RCC4221 growth in coculture when the culture medium was depleted in vitamin B 12 . From our experiments, M. algicola OT could not provide vitamin B 12 to O. tauri RCC4221. In contrast, the decline in O. tauri cells obviously matched the beginning of M. algicola OT growth, arguing that bacteria largely benefit from O. tauri (directly from dead cells or from the exudates of accumulating living cells) to sustain their growth. However, O. tauri was maintained at the highest density from day 21 to the end of the experiment when the initial density of M. algicola OT was twice that of the microalgae. This finding could indicate that O. tauri RCC4221 survived longer in the presence of the bacteria, probably benefiting from some decomposition of organic matter during the decline phase. Nevertheless, in a recent study, another Marinobacter behaviour, one that favoured O. tauri growth in culture, was evidenced [59]. In that study, the occurrence of a Marinobacter sp. OTB1 strain was reported in cultures of the O. tauri RCC745 strain, phylogenetically close to our O. tauri RCC4221 strain. The authors showed that Marinobacter sp. OTB1 favours the O. tauri RCC745 growth. They showed that in presence of Marinobacter sp. OTB1, the growth of O. tauri was stimulated following the addition of a precursor component in vitamin B 1 -limited cultures. In contrast, no growth stimulation was observed when antibiotics were added to prevent bacterial growth [59], therefore demonstrating the beneficial presence of bacteria for microalgae growth. In our case, even in the presence of vitamin B 12 , the presence of M. algicola OT did not enhance O. tauri RCC4221 growth in culture. It can be a little paradoxical that phylogenetically closed O. tauri strains (RCC4221 and RCC745) harbouring Marinobacter strains in their laboratory culture phycospheres, are not impacted the same way depending on the Marinobacter strain. In a previous study, the growth of two Prochlorococcus strains (MIT9313 and MED4) was differentially affected when they were cultivated with the same collection of heterotrophic bacteria [60]. The authors highlighted that the kind of interaction was dependent on the phylogenetic signature. From this point of view, closely related bacteria (whose 16S gene sequences differed by 1-2%) show more similar effects in cultures [60]. As pointed out previously from our phylogenetic reconstruction, M. algicola OT and Marinobacter OTB1 are two independent species that are not closely related, showing 5.3% differences in a 16S partial alignment (corresponding to the length of the 16S partial sequence of Marinobacter OTB1), and thus probably exhibit different metabolisms in microalgae cultures. Last, in our phylogenetic reconstruction, the clustering of the green microalgae, dinoflagellate and haptophyte Marinobacter clade (M. algicola) and diatom-associated Marinobacter clade (M. adhaerens-M. manganoxydans-M. flavimaris) together with M. salsuginis provided a highly bootstrap support (BP = 83). Contrary to all Marinobacter species mentioned above, which include bacteria with neutral or positive effects on microalgae to our knowledge, the M. salsuginis BS2 strain is the only species that does not seem to exhibit positive behaviour towards microalgae. The M. salsuginis BS2 strain was discovered from environmental sediments only two years after M. algicola was discovered. Although it appears closely phylogenetically related to M. algicola in our phylogenetic reconstruction, the M. salsuginis BS2 is an algicidal bacterium showing killing activity towards the harmful bloom-forming dinoflagellate Noctiluca scintillans [10].

M. algicola OT Metabolises Different Lipids
The growth defect of M. algicola OT in ESAW-F/2 without O. tauri RCC4221 was most likely due to the absence of a carbon and energy source in this medium. O. tauri RCC4221 is known to produce a wide range of lipids including TAGs that represent approximately 15% of the total lipids under nutrient-replete conditions and that accumulate up to 10 times more in nitrogen-depleted conditions [35]. Many Marinobacter species are known to degrade a large variety of hydrophobic substrates, including TAGs, wax esters, fatty acids, fatty alcohols and alkanes [31,61]. The ability of M. algicola OT MT023716 to grow in the presence of O. tauri RCC4221 in ESAW-F/2 medium, while it did not grow alone, suggested that the bacterium benefited from some exudates of O. tauri. From our experiments, M. algicola OT grew in the presence of different lipids, including four different TAGs. We can hypothesise that M. algicola OT might take advantage of TAGs produced by O. tauri, which are some of its major lipids [35]. In M. algicola OT and O. tauri coculture experiments, the co-occurrence of the onsets of the growth of the bacterium with the stationary phase of the algae is consistent with a hypothetical release of lipids or other substrates by O. tauri during the stationary phase, possibly by cell lysis. Even if it does not reveal they are actively transcribed and translated, the presence of different genes involved in different fatty acid degradation pathways in M. algicola DG893 would be a strong indicator to corroborate this hypothesis. In addition, M. algicola OT might also degrade starch from the microalga if an alpha-amylase (glycosidase) present in the M. algicola DG893 genome (personal observation) was also actively synthesised. Further investigations will be needed to directly assess the growth of M. algicola OT on O. tauri lipidic extracts and determine which microalga lipids are particularly metabolised.

M. algicola OT Does Not Produce AI-1 or AI-2 QS Autoinducers
The capacity of Marinobacter species to communicate by QS should be an essential feature of the genus. M. hydrocarbonoclasticus, an oil-degrading bacterium, forms biofilms at the substrate-water interface [62], when carbon sources such as lipids are condensed in nonaqueous phases in marine waters. Producing biofilms is a particular bacterial activity that is commonly regulated by QS in numerous bacterial families. Our first global screening for genes coding for proteins involved in QS found many QS genes involved in AI synthesis in the global phycosphere of O. tauri RCC4221 laboratory cultures. However, M. algicola and all other Marinobacter species screened do not possess any of the classical QS genes. Contrary to what we expected given the high prevalence of M. algicola OT in O. tauri RCC4221 cultures, the M. algicola OT MT023716 strain did not produce either AI-1 or AI-2 according to our various detection tests. In contrast, a Marinobacter strain recovered from snow particles in a marine environment (Marinobacter sp. PCOB-2 HP36, AJ000647) produce AHLs [34]. In our phylogenetic reconstruction, Marinobacter sp. PCOB-2 was not related to the M. algicola clade but rather to the diatom-associated Marinobacter clade, which included M. adhaerens known to live closely attached to diatoms. However, our genomic screening found no evidence of the QS genes involved in AI-1 and AI-2 synthesis in any Marinobacter genome available, including M. adhaerens. Nevertheless, other types of QS modes are possible in Gram-negative bacteria, but their study requires complex and in-depth multidisciplinary studies that are out of the context of this current work. In addition, except in a study from Gram and coworkers [34], experimental evidence of QS capacity in any Marinobacter strain has never been demonstrated. False positives are regularly observed in this kind of experiment [63], and thus, we hypothesise that this could probably be the case in the Gram and coworkers study unless new experiments demonstrate otherwise. Meanwhile, the encoding of a Lux-R receptor (assuming it is correctly translated and functional) in the majority of the Marinobacter genomes available, including M. algicola DG893, suggests that M. algicola senses AI from other bacterial communities and benefits from them to synchronise its own activities. However, for the thirteen retrieved M. algicola LuxR protein sequences, in silico protein analyses revealed the lack of an AI-binding domain (and of other features such as conserved amino acids involved in AHL binding and the presence of InterPro domains) in the LuxR protein sequences, thus eliminating this kind of 'cheater' behaviour [64]. In our study, we more specifically focused on AI-1 (AHL-driven) and AI-2 (furanosyl diester borate-driven) modes of QS. However, other types of QS have been depicted in the literature [65]. As their study requires intense experimental work, we did not consider including them in the context of this study. Nevertheless, their inclusion in future research could provide interesting insights into Marinobacter physiology, especially in the context of microalgae-bacteria interactions [32].

M. algicola OT Is a Commensal Species in O. tauri Cultures and Is Widespread in the Natural Environment
It appears clear that the colonisation of different phycospheres by Marinobacter sp. strains under laboratory conditions not only occurs in dinoflagellate cultures in which it was initially described [56] but can be extended to coccolithophorid, diatom and Chlorophyta microalgae cultures, including cultures with O. tauri RCC4221. From our partial 18S phylogenetic reconstruction, these strains all cluster together. It seems that bacterial species that best behave throughout rapid host growth have been selected from the diversity of bacteria observed in phytoplanktonic cultures, as previously suggested [12]. We wonder if this kind of selection over generations is a general phenomenon that also extends to other microalgae cultures to result in M. algicola OT MT023716 predominance in O. tauri RCC4221 cultures. The M. algicola OT growth phase matched the decline in O. tauri cells, suggesting that the bacteria fed on the exudates or dead cells of the algae. A study by Trombetta and collaborators during bloom period in mediterranean coast [7] revealed evidence of positive interactions between phytoplankton and bacteria that suggest that vitamin-synthesising bacteria would provide vitamins to phytoplankton in exchange for organic carbon [19,66,67]. Our results show that without an adequate carbon source, M. algicola is difficult to cultivate and grows particularly well in phytoplanktonic cultures in which it probably finds additional nutrients provided by living and/or dead algal cells. Given the relatively low variability level inside the M. algicola clade, it is not possible to state if M. algicola strains found in dinoflagellate cultures are more closely related to each other than to the other M. algicola strains inside the clade, which would indicate whether the bacteria are specific to some kinds of microalgae. In addition, a recent study reported the occurrence of the Marinobacter sp. OTB1 strain [59] in cultures of O. tauri RCC745. Another recent study suggests that algal-associated bacterial communities are controlled by algal hosts, culture conditions and the initial inoculum composition of the microalgae phycosphere [68].
The range of lipid compounds allowing M. algicola growth that were highlighted in this study could be the key elements explaining the occurrence of this bacterial species in so many different microalgae laboratory cultures. Interestingly, the majority of Marinobacter strains found in diverse Chlorophyta, Dinoflagellates and Haptophyte cultures are all in the M. algicola clade in our phylogenetic reconstruction, demonstrating a particular acclimation of this opportunistic species in different phycospheres in laboratory cultures. In the presence of organic carbon, Marinobacter can grow rapidly, out-competing other bacteria in enrichment cultures [69] and easily dominates other communities in natural marine aggregates [70]. This r-strategist (or copiotrophic) behaviour renders them very easy to cultivate, even compared with other heterotrophic marine bacteria [71]. Marinobacter species probably use other kinds of communication system to optimise their growth, and these bacteria might form biofilms, a central mechanism in the natural environment [72], to colonise any substrates that are not well dissolved [61], including phytoplanktonic blooms. Preliminary exploration of public datasets from Tara Oceans reveals probable colocalisation of M. algicola and O. tauri in coastal environments. It is highly probable that O. tauri, as a bloom-forming phytoplanktonic species, can be an important nutritive interface for numerous bacterial strains in the natural environment.

Conclusions
The persistence of M. algicola OT MT023716 bacteria in O. tauri RCC4221 cultures would result from opportunistic behaviour. Our experiments suggest that the M. algicola OT MT023716 strain needs O. tauri microalgae to grow in synthetic medium. Although neither genomic evidence nor experimental work argued in favour of QS activity by M. algicola OT MT023716, the strain grows at high density on a wide range of lipids, including triglycerides, known to be produced by O. tauri and numerous marine microalgae in culture during abiotic stress. Marinobacter algicola strains are frequently found in different microalgae cultures in the laboratory, which reveals their plasticity in colonising different laboratory culture phycospheres. Understanding why this colonisation is limited to certain phytoplanktonic species in the laboratory will bring new insights into the optimum activity of M. algicola in cultures, and perhaps also in marine environments.