H. ostrearia is a microalga complex to grow under laboratory conditions, and even after years of development, it is still a challenge to propose a protocol that allows it to grow without any bacteria [12, 13, 14]. Indeed, the tests previously performed in our laboratory showed that, even with all the nutrients needed, a high amount of antibiotics or a long exposure to them tended to kill the microalgae culture itself. With this complexity in mind, all the manipulations and analyses on H. ostrearia must consider the bacterial compartment. Therefore to better understand H. ostrearia and to ease future research, this study was focused on the bacteria present in the culture and the variation in their population under 2 conditions. The objective was to find a protocol to cultivate H. ostrearia with a known bacterial population that does not change at each sampling (organisms and proportions known), in other words, to ensure the reproducibility of the cultures. Four cultures were grown with the usual laboratory conditions for H. ostrearia, but a treatment by antibiotics was added to 2 of them 24h before the DNA extraction. Two control regions were also used: each DNA sample was sequenced by 16S for the V1V3 region and the V3V4 one. Thus the comparison between the bacterial populations of cultures under different culture conditions will allow us to determine the method with the greatest potential for reproducibility. Furthermore, it has been shown before that the selection of the 16S region is of great importance for the identification of bacteria [15]. To select the region to be retained in the analysis, a known bacterial community (Mock samples) was sequenced with the H. ostrearia samples. The reliable method, and therefore the one chosen, would be the one that identifies all expected bacteria.
We therefore first determined which one is the best 16S region to identify the bacteria in the samples. The results obtained for the V1V3 region with the mock community showed its weakness to spot the Gammaproteobacteria. Indeed, none of the bacteria belonging to this phylum has been found with this method although they were present in the sample (Fig. 2). A small proportion of it was also classified as Oxyphotobacteria (Chloroplast) which was not in the community in the beginning. However, since it represented less than 1% of the sample, this result was not considered. This misidentification may be due to a stochastic noise known on the Illumina MiSeq platform when the DNA input is below a critical threshold [16]. In contrast, the V3V4 sequencing provided the expected results, and only the bacteria genus present in the mock community were found. The only possible criticism concerns the relative abundance, the observed values being slightly different from those expected (Fig. 2). But, even if this relative abundance varied greatly depending on the genus considered, there was no one taxon that overwhelmingly dominated the others. The V3V4 sequencing still seems to well detect the genera present in a diverse bacterial community and give a good overview of their abundance in the samples. Therefore we decided to only focus on the V3V4 16S region, as for four other marine diatoms, where a high amount of Gammaproteobacteria was observed [17]. If H. ostrearia is not a special case, the Gammaproteobacteria should be observed in the samples too and so needed to be detected by our 16S region.
Next, thanks to the chosen region, we determined which culture condition allows us to maintain a stable bacterial population at each sampling. When comparing the V3V4-N1 and V3V4-N2 samples, it was obvious that the bacterial composition was not similar between the two of them. Indeed, when we examined the general profile of the classes presented in the samples, many variations were observed. V3V4-N1 seemed to have much less Alphaproteobacteria than V3V4-N2, with a variation almost equal to 10 points (Fig. 3, Table 2). Same phenomena was observed with the Clostridia class with a variation around 4 points. There was also the presence of two classes, Oxyphotobacteria and Deltaproteobacteria (1.63%), which were noticed only in V3V4-N1. Both were not very present in the corresponding N2 sample (< 1%); these variations were also more apparent when focusing on the orders (Fig. 4). Even the orders observed in a class with the same relative abundance differed greatly depending on the sample. The composition of Gammaproteobacteria was the best example of it: n the V3V4-N1 sample, the Oceanospirillales was the most prominent order in this class with a relative abundance > 10%, whereas in the V3V4-N2 sample, this percentage represented the total of the relative abundance of the three most present orders: Alteromonadales, Cellvibrionales and Oceanospirillales. No variation as high as these was observed between the V3V4-T samples, and only small oneswere noticed ( < = 1 point, Figs. 3 and 4). Thus, an antibiotic treatment helps to recreate the same bacterial population at each culture. Nevertheless, more samples are needed to really confirm this hypothesis in a large range of conditions.
Therefore V3V4-T samples were used to describe the bacteria found within H. ostrearia cultures. But, during the taxonomic association of the sequences, the first thing noticed was the high proportion of bacteria in the microalgae cultures. In fact, the sequences associated with Chloroplast represented around 1% of the samples (Table 2). If we take the hypothesis that this organelle would only come from H. ostrearia (V3V4-T-Neg is negative, Supplementary Data), this indicates that in the best case at least 99% of the samples are bacterial sequences. This high percentage of bacteria could be still problematic to study the microalgae DNA. In addition, with all results combined, more than 30 orders were found and they appeared to originate from H. ostrearia cultures (Fig. 4). The three most abundant phyla described here (Fig. 3) have also been found in other studies of microalgae, including diatoms [17, 18]. Proteobacteria, specifically the Alphaproteobacteria, was the majority of the bacteria found in a study of microalgae’s bacterial consortium, including from the diatom Thalassiosira pseudonana [18]. Previously, a study showed a high proportion of Gammaproteobacteria in diatom cultures which tended to increase with the culture time [17]. This might explain why they accounted for about a quarter of V3V4-T samples, since the H. ostrearia samples used here have been isolated in 2018 and have been grown discontinuously since. In this same study, Bacteroidetes were also present in the cultures, as in H. ostrearia ones. Litterature on the effect of these bacteria on microalgae even led us to consider the hypothesis that the presence of these bacteria was not a mere coincidence [12, 13]. Indeed, here are examples of the three most abundant classes that were present in all H. ostrearia samples (~ 47 − 6%, Fig. 4): Rhodobacteriales is a marine bacterium suspected of having a growth enhancement properties for microalgae [19], Rhizobiales displaced the same properties but for cultures of the green microalgae Chlorella [20], and Flavobacteriales showed a growth inhibitory effect to the microalgae Nannochloropsis sp. CCAP211/78 [21]. But their specific effect on H. ostrearia still needs to be described. For the other orders, especially for the ones with a small abundance (< 1%), it could be possible that the V3V4 region underestimates their presence, or that they were indeed relatively rare in our samples. Again, a larger quantity of samples would help us to confirm this hypothesis, especially if it is composed of isolated from other geographical areas. In the end, it is necessary to remember that the high presence of bacteria remains a problem to study H. ostrearia. It is therefore necessary to look for another antibiotic treatment more specific to the bacteria found in this study in order to increase the proportion of this microalgae in our sequencing data (and thus in our cultures). But, as mentioned earlier, the antibiotic treatment does not have to kill all the bacteria because some appear to be necessary for the growth and survival of H. ostrearia.