A novel alphaproteobacterial ectosymbiont promotes the growth of the hydrocarbon-rich green alga Botryococcus braunii

Botryococcus braunii is a colony-forming green alga that accumulates large amounts of liquid hydrocarbons within the colony. The utilization of B. braunii for biofuel production is however hindered by its low biomass productivity. Here we describe a novel bacterial ectosymbiont (BOTRYCO-2) that confers higher biomass productivity to B. braunii. 16S rDNA analysis indicated that the sequence of BOTRYCO-2 shows low similarity (<90%) to cultured bacterial species and located BOTRYCO-2 within a phylogenetic lineage consisting of uncultured alphaproteobacterial clones. Fluorescence in situ hybridization (FISH) studies and transmission electric microscopy indicated that BOTRYCO-2 is closely associated with B. braunii colonies. Interestingly, FISH analysis of a water bloom sample also found BOTRYCO-2 bacteria in close association with cyanobacterium Microcystis aeruginosa colonies, suggesting that BOTRYCO-2 relatives have high affinity to phytoplankton colonies. A PCR survey of algal bloom samples revealed that the BOTRYCO-2 lineage is commonly found in Microcystis associated blooms. Growth experiments indicated that B. braunii Ba10 can grow faster and has a higher biomass (1.8-fold) and hydrocarbon (1.5-fold) yield in the presence of BOTRYCO-2. Additionally, BOTRYCO-2 conferred a higher biomass yield to BOT-22, one of the fastest growing strains of B. braunii. We propose the species name ‘Candidatus Phycosocius bacilliformis’ for BOTRYCO-2.


Figure S4
pH curves of the cultures indicated in Fig. 5. Bars indicate the standard error of four biological replicates. A significant difference between the two cultures was found only at day 28 (indicated by *) (homoscedastic one-tailed t-test, P = 0.0027). pH was measured using a LAQUAtwin pH meter (Horiba, Kyoto, Japan).

Figure S8
Fluorescent image of B. braunii Ba10 − stained with DAPI. Spotted blue signals were visible in the nuclei of Ba10 − , whereas bacterial signals were absent in the image.
Note that the extracellular matrix was slightly stained with DAPI, which can be discriminated from true signals by its lower fluorescent intensity.

Figure S10
The full version of the NJ tree depicted in Fig. 3. Bootstrap values for each node (>50%) are indicated. Higher order taxa compressed in Fig. 3 are indicated with *.

Supplementary
Results of BOTRYCO-2 specific PCR detection using the primer pair 27F/Ba1R. c Bloom samples including >10 5 cell mL -1 of M. aeruginosa were analysed. d Samples or strains whose seqences are indicated in the alignment of Supplementary Fig. S2. a DNA polymorphism within the 27F/Ba1R amplified fragment. ND, not determined. e B. braunii was collected using a plankton net and a centrifuged pellet of B. braunii cells was used for PCR. M. aeruginosa was not detected in the sample under a microscope.

Establishment of an axenic culture of B. braunii Ba10
An axenic culture of B. braunii Ba10 (Ba10 − ) was established from the original culture of Ba10 (with coexisting bacteria). Ba10 was cultured in 20 mL of liquid AF-6 medium in a 25-ml test tube for 2 months with ampicillin (final concentration, 50 μg L −1 ) under the same temperature, light and CO 2 conditions as in the culture experiments described in the main text. Next, a small colony (approximately 0.5 mm) was picked using a micropipette in a sterile chamber and was washed several times by repeated transfer to the fresh liquid AF-6 in a microtube using the micropipette.
Finally, the colony was transferred to the liquid AF-6 medium and was incubated for 3 months under the same culture conditions as described above. Axenicity of the culture was confirmed by microscopic observation of the cells with DAPI staining ( Supplementary Fig. S9) and negative results for PCR detection of 16S rDNA using the bacterial universal primer pair 27F/1492R (50) and the specific primer pair 27F/Ba1R (Fig. S2). The axenicity of Ba10 − was further confirmed by the experiment in Fig. 5, showing that no bacterial signal was found in the GF/C filtrates of Ba10 − at every sampling point.
Cycling parameters comprised an initial denaturation step of 94 °C for 3 min, followed by 40 cycles of 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 2 min. PCR products were purified using ExoSAP-IT ® † (USB Corp., Cleveland, OH) and were sequenced in both directions using a BigDye Terminator v1.1 Cycle Sequencing Kit (Life Technologies, Carlsbad, CA) with an Applied Biosystems 3130 Genetic Analyzer (Life Technologies). In addition to the primers 27F and 1492R, primers 533F (Weisburg et al, 1991) and 536R (Mummey & Stahl, 2004) were used for sequencing. To obtain the 5′ and 3′ ends of the PCR-amplified segment, the same PCR reactions were performed using the high fidelity Taq enzyme Pyrobest (Takara, Shiga, Japan), and the amplicons were cloned using the Zero Blunt ® TOPO ® PCR Cloning for sequencing kit (Life Technologies). Using the 27F primer, the 5′ regions of >20 clones were partially sequenced to confirm that only a single sequence was recovered from each culture (i.e., to confirm that only one bacterial species/strain was present in each culture). To avoid a potential PCR error, five clones were sequenced and compared with the sequence obtained by the direct sequencing experiment.
For phylogenetic analyses, 250 16S rDNA sequences similar to those of the Ba10 symbiont were obtained from the NCBI DNA databank on the basis of the result of a BLAST search. Sequences representing related alphaproteobacterial species were also retrieved and included in the phylogenetic analyses. On the basis of a preliminary phylogenetic analysis, several sequences in phylogenetically defined clades that are distantly related to BOTRYCO-2 were excluded from subsequent analyses to minimize the number of OTUs while maintaining the recognized overall phylogeny of Alphaproteobacteria. In addition, all but one OTU showing an identical sequence were removed. An initial alignment was obtained using Clustal X version 1.8 (Thompson et al, 1997). This was checked visually and corrected manually, followed by the removal of ambiguous regions, yielding a final dataset consisting of 1,440 nucleotide sites (including gaps) of 113 OTUs.
The detailed methods for the NJ tree reconstruction and its statistical analyses are described (57). Prior to the ML tree reconstruction, all gaps were removed from the alignment and three OTUs that showed the same sequence as the other OTUs were removed, yielding a dataset consisting of 1,206 nucleotide sites of 110 OTUs. An ML tree reconstruction using RAxML was performed at Cipress Science Gateway (Miller et al, 2010) under the default settings, except that bootstrap analysis was performed with 1,000 replicates.