SORTEOS / LOTERIAS

Poseidonocella sp. on Nannochloropsis sp.

Following the studies of Giraldo et al., (submitted) on bacterial-induced biofilm formation in Nannochloropsis sp. cultivated in closed bioreactors, we tested if bacteria isolated from S. robusta could induce or reduce biofilm formation in the same system. The high throughput biofilm assay established by Giraldo et al., allowed us to test up to 20 different combinations of bacteria and bacteria spent medium co-inoculated with Nannochloropsis sp. (the experimental design is shown in figure 36). One-way ANOVA with Bonferroni´s correction for multiple comparison was used to test the significance of each treatment compared to the control sample (Nannochloropsis sp. alone).

Our experiment confirmed that Poisedonocella sp. is inducing biofilm formation when co- cultivated with Nannochloropsis sp. (p=0.025) (Figure 40). Roseovarius sp., Maribacter sp. and Croceibacter sp. are not inducing significant biofilm formation when they are added to the wells as single strains, but their combination with with Poseidonocella sp. results in a significant increse of the absorbance of the biofilm, which reaches its maximum in presence of all strains (OD630=1.2-1.3, p < 0.0001). Interestingly, the most significant effects are seen when bacterial exudates are added to Poseidonocella sp. and Nannochloropsis sp. co- cultivation.

Glucose and glycerol seem to have a big impact on biofilm formation: Nannochloropsis sp. inoculated with f/2 modified medium with glucose and glycerol forms biofilm even without inoculation of bacteria (p<0.0001). The same effect is seen when Poisedonocella sp. is added to the well plate.

Since our Nannochloropsis sp. cultures are not axenic, it could be that bacteria are taking advantage of the income of fresh carbon resources and they are overgrowing, causing an increase production of biofilm.

However, the effect of residue glucose and glycerol on biofilm formation in co-cultivation of bacteria from S. robusta, Poseidonocella sp. and Nannochloropsis can be exluded, since the spent medium of these bacteria is not triggering any significant biofilm formation (Figure 40).

Figure 40: Biofilm formation of Nannochloropsis sp. in presence of different bacteria: red indicates

controls with no bacteria, orange indicates Poseidonocella sp. only, blue Roseovarius sp. treatment, light blue

Poseidonocella sp. + Roseovarius sp., green Maribacter sp. treatment, light green Poseidonocella sp. + Maribacter sp, purple Croceibacter sp. treatments and light purple Poseidonocella sp. + Croceibacter sp.

(experimental design Fig 3). Each treatment was made in triplicates. Data were statistically evaluated with a one-way ANOVA, α=0.05, Bonferroni´s correction for multiple comparisons. Asterisks refer to multiple comparisons to Nannochloropsis sp. control sample.Poseidonocella sp. (P), Roseovarius sp. (R), Maribacter

sp. (M), Croceibacter (C). 0.01 and 0.1 are the OD600 values for inoculated bacteria. Med stands for bacterial

6.3. Discussion

Chemically mediated interactions between microalgae and bacteria have been underestimated for a long time, but in the last fifteen years many studies have proved their importance from an ecological and an industrial point of view (Ramanan et al., 2016). A better understanding of microalgae-bacteria interactions could help enhance the production and the quality of microalgal cultivation (Fuentes et al., 2016), as well as the development of new technologies for several industrial applications (Rizwan et al., 2018). Moreover, these inter-kingdom communications are a key to decipher biofilm formation, both from an ecological (Van Colen et al., 2014; Wahl et al., 2012) and industrial perspective (Berner, Heimann, and Sheehan, 2015). Many studies focused on ways to reduce biofilm onset in order to improve microalgal production in photobioreactors (Zeriouh et al., 2017), while other works concentrate their efforts on finding ways to produce stable biofilms, for example to improve wastewater treatment (Gonçalves, Pires and Simões, 2017). The understanding of bacteria-algae interactions can also lead to the development of synthetic microbial communities that are enhancing microalgal production. Microbiome engineering is a relatively novel concept that is attracting growing attention not only from human health research (Foo et al., 2017), plant science (Quiza et al., 2015) and agriculture (Bender, Wagg and van der Heijden, 2016), but also from aquaculture (Charrier et al., 2017). Some studies have tested the effect of indigenous bacteria from wastewater on microalgae growth (Toyama et al., 2018), while other have employed growth-promoting bacteria from plant rhizosphere to improve microalgal productivity in confined environments (Gonzalez and Bashan, 2000).

Following the observation of chapter 4, we decided to test the effect of these bacteria isolated from a natural sample on a different, commercially relevant species belonging to the class of Stramenopile, the oleaginous microalga Nannochloropsis sp. This work was done in collaboration with Proviron © (Antwerp, Belgium), a chemical company that since then years has started to invest in microalgal biomass production to create high quality feed for aquaculture. Recently, Giraldo et al. (submitted) have found that Poseidonocella sp., a gram negative alphaproteobacterium from the Rhodobacteraceae family and phylogenetically close to the Roseobacter clade (Romanenko et al., 2012), is inducing biofilm formation in closed bioreactors for Nannochloropsis sp. culturing. We therefore tested different bacterial combinations to explore changes in microalgal growth, endometabolome and biofilm

formation in order to understand if an engineered microbiome could enhance biomass productivity while reducing biofouling side effects.

In co-cultivation experiments, we observed different effect of bacteria on Nannochloropsis growth (Figure 37). The presence of Poseidonocella sp. and Roseovarius sp. do not affect Nannochloropsis sp. biomass increase until seven days of cultivation. On the other hand, Maribacter sp. significantly supports algal growth already 3 days after inoculation. This effect is reduced when Maribacter sp. is co-inoculated with Poseidonocella sp.: in this treatment the growth is significantly increased only at the end of the seven-day co- cultivation, but the cell density reaches its maximum if compared to all the other conditions. The effects on the non-associated alga are thus different to those on algae with which the bacteria co-exist in nature (Chapter 3). Many studies confirmed that bacteria are able to both support microalgal growth (Seymour et al., 2017), for example by releasing auxins (Amin et al., 2015), and reduce it (Van Tol et al., 2016). Some of them may exhibit a two-faced behaviour, first promoting and then inibithing algal growth, until becoming pathogenic (Seyedsaymdost et al., 2011). Our work demonstrates that this effect is highly species- specific and it is dependant on both the organisms that are interacting.

To understand if bacteria also have an impact on Nannochloropsis sp. endometabolome, we analyzed the different co-cultivations using a GC-MS-based untargeted endometabolomics approach.

Amino acids were the most affected metabolites in Nannochloropsis sp. (Figure 39), with many of them being downregulated in the presence of bacterial strains, especially glycine, valine, leucine, isoleucine and tyrosine. Interestingly, glutamate and proline were upregulated in presence of Poseidonocella sp. in combination with one of the two bacteria from S. robusta. Moreover, Maribacter sp. displayed a particular effect on intermediates of the proline cycle: in its presence, pyrrole-2-carboxylic acid is accumulated while pyrroline hydroxycarboxylic acid is decreased. Several studies showed that bacteria have a significant impact on microalgal amino acid metabolism, inducing the production of specific compounds to sustain their own growth (Amin et al., 2015; Chen et al., 2017a; Paul, Mausz, and Pohnert, 2013). This interaction mechanism could be of particular interest for applied purposes, in regard to nutritional value of algae for animal and human consumption (Rebolloso-Fuentes et al., 2001). The increase of proline and intermediates of its biosynthetic pathways may suggest a response of the alga to stress, since proline is involved in active response towards toxicity in plants and in some algae (Szabados and Savouré, 2010).

The hypothesis of amino acids breakdown under stress conditions is confirmed by the presence of several degradation byproducts. Spermidine, a polyamine that is a catabolic product of amino acids (especially arginine, Hildebrandt et al., 2015) and putrescine pathway (Tabor and Tabor, 1976) and involved in abiotic stress response in plants (Gupta, Dey, and Gupta, 2013), was higher in Nannochloropsis sp. when Maribacter sp. was present (Figure 39). Since putrescine was higher in non-treated Nannochloropsis sp. cultures, these results show a species-specific stress response of the microalga towards bacteria that involves polyamines. Remarkably, the presence of Poseidonocella sp. is increasing the concentration of aminobutyric acid (Figure 39a), another product of amino acids transformation. Among the aminobutyric acids family, γ-aminobutyric acid (GABA) is a fundamental neurotransmitter in mammals and humans (Watanabe et al., 2002) and an important defensive molecule in plants (Gupta et al., 2013) that has been found also in several microalgae (Brown, 1991). A study by Pal et al. (2013) correlated the increase of GABA concentration in Nannochloropsis oculata to osmotic stress response adaptation. This finding hints to a different effect of Poseidonocella sp. on amino acids metabolism and microalgae signaling.

Sterol and fatty acids, two of the main products from Nannochloropsis sp., were only slightly influenced by bacteria. Cholesterols levels are higher in microalgal monoculture samples, while Maribacter sp. seems to have an enhancing effect on sterols and fatty acid biosynthesis. However, our experimental setup did not show a big influence of bacteria on lipid metabolism. We see clearly that the modulation of growth also goes ahead with the modulation of the cellular metabolism. As demonstrated earlier bacteria have a most pronounced activity on the amino acid content of the algae, confirming the role of these metabolites in cross kingdom nitrogen shuttling (Williams, 2007; Amin et al., 2015). It is important to note that in aquaculture applications not only the growth promoting effect of bacteria but also their putative effect on the concentration of the targeted value product has to be considered. The unbiased metabolomic survey presented here demonstrates that different pathways including amino acid metabolism, lipid and sterol metabolism and saccharide metabolism are influenced by bacteria and most likely cascading effects on targeted value products have to be considered. This consideration is confirmed by the finding that the commercially important α-tocopherol is down regulated in presence of bacteria (Figure 39). Nannochloropsis is known to produce Vitamin E in relative high amount depending on nutrient availability and growth stage (Durmaz, 2007), and studies on different

microalgae (Vidoudez and Pohnert, 2012; Mausz and Pohnert, 2015) have shown that its concentration increases during stationary phase as a response to oxidative stress.

All these observations suggest a broad metabolic rewiring of microalgae in presence of bacteria, particularly their stress response.

Finally, bacteria were tested in bioassay to investigate their contribution to biofilm formation of Nannochloropsis sp (Figure 40). Although biofilm formation mechanism and dynamics in marine environments are still a subject of intense study, the mostly general accepted theory is that bacteria are initiating biofilm formation by releasing exopolymeric substances (EPS) in the so-called surface conditioning phase (Lorite et al., 2011). EPS are usually composed of exopolysaccharides (40-95%), protein (1-60%), nucleic acids (1-10%) and lipids (1-40%) (Davey and O’toole, 2000; Flemming and Wingender, 2001). These substances create a protecting and resource-rich environment for both bacteria and microalgae (Xiao and Zheng, 2016) and shape multifaceted communities (Bohórquez et al., 2017) that have an important and broad ecological role (Van Colen et al., 2014).

Interestingly, Maribacter sp. (or its spent medium) did not significantly increase biofilm formation when co-cultivated with Nannochloropsis sp., while Roseovarius sp. induced biofilm development only at high cell concentration (Figure 40). However, when these bacteria (or their spent media) were added to microalgal cultures in the presence of Poseidonocella sp., biofilm formation was significantly higher as compared to Nannochloropsis sp. control cultures and the effect was even larger when compared to the presence of Poseidonocella sp. alone. The same promoting effect was seen when only saline medium, enriched in glucose and glycerol, was added to a co-cultivation of Nannochloropsis sp. and Poseidonocella sp. Glycerol has been used to enhance biomass productivity of Nannochloropsis sp. in mixotrophic conditions (Poddar, Sen, and Martin, 2018), while glucose has been recently employed to increase productivity in mixotroph cultivation of Chlorella vulgaris on biofilms (Roostaei, Zhang, Gopalakrishnan, and Ochocki, 2018; Ye et al., 2018). Carbon sources (especially glucose) are important to sustain biofilm formation in various bacterial species, such as gram-positive bacteria from the Bacillaces order (Mhatre, Monterrosa, and Kovács, 2014). Dissolved sugars appear therefore to be a strong trigger for biofilm formation and have to be precisely monitored in closed bioreactor cultures of Nannochloropsis sp. in order to prevent unwanted growth of thick biofilm. Further experiments on bacterial exudates, both with fractionation assays and metabolomics approach, will give us more information about putative signaling molecules that are sustaining microalgae growth and biofilm formation. This study already provides new

instruments to control culturing conditions in closed photobioreactors but reducing the manipulation from introduction of bacteria to the application of a pure growth or production stimulator might be an attractive alternative, increasing biomass productivity and reducing contaminations and biofilm formation.

Our study evaluates the opportunity to use bacteria isolated from a natural microalgal species (Seminavis robusta) to influence the growth, the metabolome and the biofilm formation of a different, commercially relevant microalgal species (Nannochloropsis sp.) in an attempt to design a productive engineered microbiome by microbial transfer. These results demonstrate that further efforts are necessary to study bacteria-algae interactions, using both indigenous and transplanted bacteria, as well as employing –omics approaches (metabolomics, transcriptomics, proteomics) and real-time test in production sites, in order to better understand bacteria-algae interactions and employ them to develop new technological applications.