Establishing and validating axenic cultures of the microalga Haematococcus lacustris (Chlorophyceae)

ABSTRACT Although many isolation techniques for algaeaxenic cultures of algae are known, successful isolation approaches are species and contaminant-specific. The commercially important alga Haematococcus lacustris has been intensively studied due to its natural production of the high-value carotenoid astaxanthin, yet there are no reports of axenic H. lacustris cultures. In this article, we describe the successful isolation of axenic H. lacustris, originally isolated in non-axenic form from its natural habitat and adopted to be used daily in vitro, in our laboratory. We verify the absence of bacteria and identify the bacterial communities in the non-axenic algal culture. using PCR amplification with selected universal primers for 16S rRNA gene amplification in the gDNA. Amplification of the 16S rRNA gene of the non-axenic culture showed the presence of bacteria with high identity to Massilia, Blastococcus and Deinococcus species. The bacteria which were identified in our strain further expand our knowledge of bacterial communities on algae, and because we established axenic H. lacustris culture, testing their effect on growth or astaxanthin accumulation in controlled co-cultures is now available. The isolation technique described herein can be applied to other green algal species to eliminate bacteria, and the selected primers can be used to verify the axenic nature of other green algal cultures. The resulting axenic culture is required for molecular genetics research and will be very valuable for establishing heterotrophic growth in laboratory or industrial bioreactors.


Introduction
Microalgae are photosynthetic microorganisms, which are naturally found in fresh and marine water environments, inhabiting a vast range of ecosystems. Under specific environmental conditions, certain types of algal species can synthesize various valuable biological materials, such as proteins, storage lipids, antioxidants, neuroprotective products, therapeutic proteins, pigments, polysaccharides, antibiotics and vitamins. As an example, approximately 15,000 novel compounds synthesized by marine algae were chemically determined between 1960 and 2000 (Cardozo et al., 2007). However, only very few of these algae have been commercialized (Mobin & Alam, 2017), for instance Chlorella and Arthrospira as food supplements (Durand-Chastel, 1980;Soong, 1980;Tamiya, 1957), Dunaliella as a source of β-carotene (Borowitzka & Borowitzka, 1989) and, later, Haematococcus as a source of the highly valuable carotenoid astaxanthin (Lorenz & Cysewski, 2000).
In general, the establishment of the axenic state of microalgal cultures remains an unsolved issue because of the diverse interactions of microalgae and their associated bacteria (Fuentes et al., 2016;Ramanan et al., 2016;Sena et al., 2011), requiring species-specific treatments (Agostini, Lopes, Muxagata, & Macedo, 2019), such as physical filtration (Carney & Lane, 2014), bio/chemical treatments and other common methods, which are emphasized and summarized in these articles (Carney & Lane, 2014;Sena et al., 2011;Singh, Mantri, Reddy, & Jha, 2011;Zhu, Jiang, & Fa, 2020). The need for new and efficient methods to combat contamination (microbial association) was described by Lian, Wijffels, Smidt, & Sipkema (2018). However, it is important to stress the point that not all coexisting bacteria are regarded as unwanted contaminants (occasionally entering the culture). Recent studies (Lian et al., 2018;Padmaperuma, Kapoore, Gilmour, & Vaidyanathan, 2018) show that their presence benefits algal growth. In fact, the phycosphere, or the region immediately surrounding a phytoplankton cell, is enriched in organic molecules exuded by the cell into the surrounding water and contains bacteria that exchange metabolites and infochemicals with the alga (Seymour, Amin, Raina, & Stocker, 2017). There is evidence that algae shape the bacterial community surrounding the algal cell in the phycosphere (Kimbrel et al., 2019); nevertheless, it is still necessary to obtain axenic microalgal cultures of target species to study algal biochemistry, physiology and molecular biology. The three advantages of obtaining or isolating axenic microalgae are (i) the ability to study distinct physiological characters such as carbon assimilation, biomass production, etc., independent of associated bacteria or in controlled alga-bacteria co-culture (Fuentes et al., 2016), (ii) molecular characterization studies like whole genome sequencing and other high-throughput sequencing require highly purified algal genomic materials free of any bacterial genome contamination and (iii) some commercial applications of microalgae such as heterotrophic cultivation demand axenic algal cultures (Barros et al., 2019;Cooper & Smith, 2015;Pienkos & Darzins, 2009;Ryan Georgianna & Mayfield, 2012;Stephens et al., 2010).
Our study organism, the unicellular freshwater green microalga Haematococcus lacustris -formally known as Haematococcus pluvialis (Nakada & Ota, 2016) -is under intensive study, due to its naturally high productivity of the valuable keto-carotenoid metabolite astaxanthin (Borowitzka, 2013;Shah, Liang, Cheng, & Daroch, 2016), a powerful antioxidant mainly used as nutritional additive (Ambati, Phang, Ravi, & Aswathanarayana, 2014). Astaxanthin is a bright red-coloured secondary carotenoid with a wide range of applications in the cosmetic, food, feed, aquaculture, pharmaceutical and nutraceutical industries due to its extremely high capability of free radical scavenging (Guerin, Huntley, & Olaizola, 2003;Turck et al., 2020). The commercial production of astaxanthin from H. lacustris is mainly affected by the host-specific parasitic infection caused by the fungus Paraphysoderma sedebokerense (Asatryan, Boussiba, & Zarka, 2019;Hoffman et al., 2008), while the bacteria coexisting with H. lacustris do not interfere with the production process. There are previous reports that these coexisting bacteria can produce vitamin B 12 , which is required for the growth of H. lacustris (Pringsheim, 1966;Provasoli, Carlucci, & Stewart, 1974); in general, this vitamin is not synthesized by the algae (Croft et al., 2006), and in the laboratory growth of the algae, the vitamin B 12 is supplemented in some culture media. Moreover, the complex carbohydrate-rich cell wall of H. lacustris (Hagen, Siegmund, & Braune, 2002) can be very attractive to different contaminants.
The life cycle of H. lacustris is complicated, and five different morphologies can be distinguished at the different growth stages, including gametospores (microzooids), bi-flagellated oval zoospore, immotile round green aplanospore, daughter cell coenobia (mother cell) and round astaxanthin-rich red resting cyst; the vegetative stage includes both flagellated and immotile cells, while the time span of each morphology is strain dependent (Boussiba, 2000;Kobayashi, Kurimura, Kakizono, Nishio, & Tsuji, 1997;Lee & Ding, 1994). The complex carbohydrate-rich cell wall of H. lacustris (Hagen et al., 2002) can be very attractive to different bacteria, and the empty mother cell wall left after completing the consecutive division cell-cycle (Reinecke, Castillo-Flores, Boussiba, & Zarka, 2018) can provide nutrients to different bacterial communities. The phycosphere bacteria of H. lacustris NIES-144 were analysed at different growth stages, and at the genus level, bacteria belonging to Blastomonas, Acidovorax, Flavobacterium, Achromobacter and an uncultured phylotype in the family Chitinophagaceae (EU104210_g) were identified (Lee et al., 2019). In this study, the auxin-producing symbiotic bacterium Achromobacter (strain CBA4603) increased cell density, biomass and chlorophyll concentrations of H. lacustris when co-cultured at a relatively high final concentration (OD 600 = 0.3). In another study, the 16SrRNA gene-based metagenome of the phycosphere of natural H. lacustris, from the White Sea coastal rockpools, has revealed the following bacteria families: Comamonadaceae, Cytophagaceae, Xanthomonadaceae, Acetobacteraceae, Rhodobacteraceae, and Rhodocyclaceae (Kublanovskaya, Solovchenko, Fedorenko, Chekanov, & Lobakova, 2020).
Studying H. lacustris interrelations with its bacterial community, as well as obtaining an axenic algal culture free of bacteria are of great interest scientifically and commercially. Isolation of axenic strains of microalgae can be done by various ways, such as micro-picking, frequent or continuous sub-culturing, ultrasonication, lysozyme treatment and antibiotic treatment (Yim & Lee, 2004;Choi, Bae, Ahn, & Oh, 2008;Gasulla & Barreno, 2010;Sena et al., 2011;Cho et al., 2013;). The most common method used for obtaining axenic cultures is antibiotic/s treatment because of its simplicity. This study aimed to remove the bacteria associated with the H. lacustris cell surface to establish an axenic strain of the alga; we thus applied antibiotic treatment method and established an appropriate molecular method to prove axenic status in cultures.

Organism
Haematococcus lacustris (formally H. pluvialis) Flot. 1844 em. Wille 1903 (SCCAP K-0084) was obtained from the Scandinavian Culture Collection of Algae and Protozoa (SCCAP). It was used daily in vitro in our laboratory for years and used as the initial culture for the isolation of bacteria-free algal culture, as well as for the identification of H. lacustris cell surface-associated bacterial communities.

Isolation of axenic culture
A routinely used H. lacustris stationary culture (7 days old) was washed two times with sterile water and resuspended in fresh mBG 11 liquid medium. The culture was diluted tenfold and divided into 25 ml flasks (10 ml culture). In each flask, one type of antibiotic-gentamicin (25 µg/ml), streptomycin (2.5 µg/ml) or amoxycillin (2.5 µg/ml) -was added, aiming to obtain axenic algal culture. The cultures were grown for 6-7 days as described above. From each culture, 1 ml diluted culture (~10 2 -10 3 cells) was spread on a solid-enriched CGM medium, supplemented with antibiotic, at a concentration previously used in the liquid culture. Enriched CGM is an organic nutrient-enriched medium, which we found to be very efficient for visualizing any bacterial presence. After plating on Petri dishes, the plates were transferred to algal growth incubator (see above). After 1-2 weeks, growing non-contaminated colonies were streaked on fresh mBG 11 agar plate with the corresponding antibiotics. Grown colonies were upscaled to 100 ml volume culture in mBG 11 liquid media (changed once per week for ca 2 months). One putative axenic monoclonal culture, which we named AA, was used here for in-depth axenicity studies. The isolated axenic and the routine, i.e., non-axenic, H. lacustris cultures were further inoculated onto nutrient-rich microbial media: Nutrient agar (Acumedia, Lansing, USA), MacConkey agar (Difco, Detroit, USA), LB agar (Difco, Detroit, USA), Hi Crom Escherichia coli agar (HiMedia, Mumbai, India) and BGM, all containing 1.5% agar, to check for growth of any bacterial colony. Cultures were kept at 37°C for 4 days. Once axenic cultures were verified, H. lacustris colonies were transferred to and maintained on mBG 11 plates and were also inoculated in liquid mBG 11 medium and grown as described above. For further characterization of the axenic cultures, vitamin B 12 requirement for growth was also tested. Cultures were divided into two flasks: control (without vitamin B 12 ) and treatment -with vitamin B 12 (5 μg ml -1 ). After 20 days of incubation, the number of cells in each treatment was counted in a counting chamber. Tukey's post hoc test in ANOVA was calculated for the growth of axenic cells in different media in IBM® SPSS® statistical software (IBM SPSS25).

Scanning electron microscopy
The axenicity of the isolated algal culture was further tested by observing it under a scanning electron microscope. The axenic algal cells were grown in mBG 11 medium, and cells were harvested by centrifugation at 16,000 g for 5 min, washed with Dulbecco's Phosphate-Buffered Saline (DPBS, Gibco, ThermoFisher scientific), pH 7.4 and fixed with 2.5% (v/v) glutaraldehyde (Sigma-Aldrich, Rehovot, Israel), at 4°C overnight. The fixed sample was gently rinsed twice with PBS, followed by post-fixation in 1% (v/v) osmium tetroxide (Sigma-Aldrich, Rehovot, Israel). Then sample was washed with DPBS before gradient dehydration in ethanol (50, 70, 90 and 100%) under air, keeping each dehydration step for 15 min at room temperature. Then tertiary butanol was used to collect samples, which were placed on round 0.5 mm transparent glass slides, for permanent fixation before drying and mounting on metallic stubs. The preparation was coated with carbon using Emitech K575X sputter coater with Peltier cooled sputtering head and turbo pump to deposit thin layer of carbon. The acceleration voltage was fixed at 3.0 kV. High-resolution scanning electron microscope (HR-SEM) images were obtained on a Jeol JSM-7400 F electron microscope.

Algal biomass harvesting and gDNA extraction
gDNA extraction of putative axenic and non-axenic cultures was done using Qiagen DNeasy Plant Mini Kit (Hilden, Germany), according to the manufacturer's instructions. About 50 mg (wet weight) of 6-day-old cultures in mBG 11 medium were collected, and three big (2.5 mm) and two small (0.5 mm) iron beads were used to homogenize the samples by impact and friction (Retsch MM400 Mixer Mill) for four rounds of 40 s, 25 Hz.

PCR amplification of 16S rRNA gene
Bacterial 16S rDNA (~1500 bp long), which is universal in Domain Bacteria and highly conserved within existing organism of the same genus and species, has similarities to algal chloroplast DNA. However, sufficient nucleotide variation in the 5' region (~500 bp long) between bacterial species of the same genus allows to differentiate algae from bacteria as well as bacteria to the genus and species level. We used this information as a strategy/tool for our cloning experiment. More specifically, to check whether the alga is free from bacteria, we amplified the 16S/18S rDNA gene with universal primers. We examined common conserved regions of prokaryotic (16S) and eukaryotic (18S) rRNA sequences using 16S/18S universal primers. One pair of 16S/18S candidate primers, U519 and U1390, was carefully selected from probeBase online resources, considering a moderate amplicon size for later rRNA amplicon sequence origin identification. Therefore, eukaryotic 18S and prokaryotic 16S rRNA genes in alga were amplified simultaneously using the universal primers U519 and U1390, and the subsequent sorting of pyrosequenced reads revealed some distinctive nucleotide communities in the amplicon of the sample. A 16S rRNA gene "universal" primer pair that share consecutive conserve sites in Bacteria, Archaea and Eukarya (Klindworth et al., 2012), S-*-Univ-0519-a-S-18 (23_U519F_Fr -5'-CAGCMGCCGCGGTAATWC-3') and S-*-Univ-1390-a-A-18 (24_Uni1390R_Rw -5'-GACGGGCGGTGTGTACAA-3'), was used. The 5' ends of these U519F and Uni1390R primers correspond to positions 519 and 1390 of the Escherichia fergusonii 16S rRNA gene, respectively, with an amplicon size of 889 bp. Algal chloroplast DNA, which has similarities to bacterial 16S rDNA, in the selected primer set has three fewer nucleotides (886 bp). With this primer pair, an ~1077 bp product was derived from the ribosomal RNA gene of all Haematococcus samples (e.g., GenBank: MF992170.1). Amplification was carried out using proofreading Q5® High-Fidelity (HF) DNA polymerase (NEB). The reagents used for PCR experiments were prepared using ultra-pure water provided with the kits. The final concentrations of the reagents for the PCR were as follows: 0.02 units/µl Q5 HF DNA polymerases, 1x Q5 HF buffer, 1x GC Enhancer, 0.2 mM each dNTP, and 0.5 µM each primer. The PCR conditions were (i) 30 s at 98°C, (ii) 35 cycles of 10 s at 98°C, 30 s at 50-60°C, and 45 s at 72°C and (iii) 7 min at 72°C. Gradient PCR was performed in a volume of 20 µl with 2 µl of putative axenic culture, AA (22 ng/µl), non-axenic culture, which we named RN (16 ng/µl), and positive control bacteria template -Enterococcus (0.3 ng/µl) gDNA. PCR was performed in a ProFlex PCR System (Applied Biosystems, USA), and the amplified products -stained with ethidium bromide -were visualized under UV light (Bio-Rad) after electrophoresis on 1.5% agarose gel in 1x TAE (Tris-acetate-EDTA) buffer.

Product cloning into pJET1.2 blunt Vector
CloneJET PCR Cloning Kit (Thermo Scientific, Waltham, US) was used to clone PCR-amplified "Lower" and "Upper" fragments into pJET1.2 blunted vectors. This approach allowed us to strictly determine diversity of "Lower" fragment "community" and therefore evaluate culture purity; it also allowed us to identify bacterial prototypes in non-axenic culture. The heat shock transformation method (Froger & Hall, 2007) was applied to E. coli HST08 Stellar Competent Cells to obtain fragment carrying lines. Cultures grown in liquid LB medium containing 50 µg ml -1 (w/v) ampicillin raised from the streaks of transformed single colonies were used to analyse colonies for inserts by PCR, using CloneJET PCR Cloning Kit primers (pJet F and pJet R) for amplification and product size evaluation on agarose gel as a first step. After inserted fragment verification (digestion with restriction enzymes and comparison to sizes of in silico digestion fragments), nine colonies from each clone (lower or upper fragment from putative axenic and non-axenic cultures) were processed for mini plasmid preparation (Presto™ Mini Plasmid Kit (GeneAid, New Taipei, Taiwan)).

DNA sequence analyses
The plasmids were sequenced at the Genomic Analysis unit of Ben-Gurion University of the Negev, Israel, using an ABI 3500 series sequencer with BigDye Terminator v1.1 Cycle Sequencing Kit. Ta = 60°C and CloneJET kit primers were used to sequence the inserts, in both upstream and downstream directions. The sequence data obtained using both upstream and downstream sequencing for the partial 16S rRNA gene were edited with BioEdit Sequence Alignment Editor. Consensus sequences were created, and their sequence similarities were identified in NCBI GenBank database with BLASTn (Basic Local Alignment Search Tool). The sequences have been submitted to the NCBI Genbank database under the accession numbers MW338707.1-MW338708.1 (upper and lower fragments of axenic culture before cloning), MW332196-MW332203 (lower band sequences from axenic culture after cloning, AA_LBC), MW338709-MW338722 (upper band sequences from axenic -AA_UBC -and non-axenic cultures -RN_UBC -after cloning) and MW345832-MW345838; MW375031-MW375032 (lower band sequences from non-axenic culture after cloning, RN_LBC).

Phylogenetic tree analysis
Sequences in NCBI that showed high nucleotide similarities with our cloned sequences in BLASTn (Altschup, Gish, Miller, Myers, & Lipman, 1990) were used for phylogenetic tree construction. Additionally, we used H. pluvialis 16S ribosomal RNA (HQ317401.1) and H. lacustris 18S ribosomal RNA gene partial sequences (DQ009774.1) as outgroup and ingroup controls, respectively (Fig 6). Saccharomyces cerevisiae 18S rRNA gene partial sequence (NG 063315.1) was used as an external control (to other eukaryotic species) to show 18S rRNA sequence similarities (Fig 6) between algae and fungi 18S rRNA sequences. H. lacustris 18S rRNA partial sequence (DQ009774.1) was used as an outgroup control for 16S rRNA sequences in phylogenetic tree in Fig 7. Concomitantly, H. pluvialis 16S (chloroplast) and Escherichia coli 16S ribosomal RNA partial sequences were used as ingroup control 1 for axenic (AA_UBC) and ingroup control 2 for nonaxenic (RN_UBC) cloned sequences, respectively. The multiple sequence alignments were performed by ClustalW algorithm with default settings (Larkin et al., 2007) in MEGA 7.0.26 (Kumar, Stecher, & Tamura, 2016). Sequence extensions were manually trimmed, and the alignment output file was used in further steps. Maximum Likelihood statistical method was used for phylogeny tree reconstruction. To test the phylogeny, we used Bootstrap method with 500 replications (Felsenstein, 1985). Only the tree with the highest log likelihood is shown in both figures (Figs 6-7). The Jukes & Cantor (1969) model was selected as the best model (in both trees) according to the DNA models search function in MEGA7.0.26. All positions containing gaps and missing data were completely deleted. We used nearest-neighbour-interchange as the best option for the tree searching strategy. The percentage of trees in which the associated taxa clustered together (known as Bootstrap support values) is shown next to the branches.
Vertical lines represent branch statistics\frequency, and horizontal lines display branch length (substitutions per sequence site).

DNA gel electropherogram analysis
Image Lab Version 6.1.0 build 7 Standard Edition (2020, Bio-Rad Laboratories) was used to estimate amplified bands, approximate molecular weight and relative quantification (Supplementary data S1 and S2).

Axenic culture establishment
Bacteria associated with the surface of our normal algal cultures were not evident when algae were plated on LB agar, i.e., when we inoculated algae on solid LB in the dark at 30-37°C for 2 days, no bacterial colonies were visible to the naked eye or under the microscope. In contrast, bacterial colonies were observed when BGM fungal medium (Asatryan et al., 2019) was used, making this medium appropriate for visualization of bacterial presence. However, on this sucrose-containing medium, algal cells had shrunken chloroplasts lying along the cell wall. Replacing the sucrose in the medium with glucose, i.e., using enriched CGM (Gutman et al., 2011) instead of BGM, resulted in vivid green algal cells and yet allowed the visualization of bacterial contamination. Among the three antibiotics used, gentamicin (25 µg ml -1 ) and streptomycin (25 µg ml -1 ) were not good for the algae as the cells died during incubation, whereas amoxycillin (2.5 µg ml -1 ), a broad-spectrum antibiotic, was efficient in inhibiting bacterial growth without any toxic effects on the algae. After repeated sub-culturing on solid mBG 11 medium (upscaling from a single colony) for more than 3 months, a putative bacteria-free algal culture (named AA) was obtained. The isolated putative monoclonal axenic H. lacustris AA cells, which were also plated onto the nutrient-rich microbial media Nutrient agar, MacConkey agar, LB agar, Hi Crom E. coli Agar and BGM, showed no bacterial growth even after 4 days of incubation at 37°C, indicating the axenic nature of the isolated algal culture. An additional test for bacterial presence was conducted by the addition of organic carbon sources in the form of sucrose and glucose (5 g l) to the liquid algal growth medium. These carbon sources do not support the growth of the wildtype alga (Waissman-Levy et al., 2019) but can boost bacterial growth if present. After 2 weeks of inoculation, with both carbon sources, the cultures attained the same cell concentration as that of the control culture, and microscopic observation showed no bacteria in the cultures (data not shown).
The putative axenic H. lacustris AA monoclonal culture was selected for further characterization. It was routinely inoculated in mBG11 algal growth medium. Since it has previously been shown that Haematococcus requires vitamin B 12 for its growth (Provasoli et al., 1974), and in our non-axenic cultures it might be provided by bacteria on algal cells, we tested the effect of adding vitamin B 12 to the medium on growth of the putative axenic line. The AA culture grows well and has a green vivid colour, both in the presence or absence of vitamin B 12 (Fig 1A), reaching a cell density of ~2 × 10 6 cells/ml at the stationary phase ( Fig 1B). Tukey's test applied as a post hoc test to identify the differences in the axenic culture cell density with and without vitamin B 12 found no significant difference (p = 0.945) between these treatments. We thus conclude that vitamin B 12 is not needed for the growth of the axenic culture.

Axenicity examination by microscopic observations
The presence of any bacterial growth in algal culture was further examined by observing the culture under scanning electron microscope. SEM observations at 1600x magnification clearly showed both coccus and bacillus bacteria in the non-axenic (named RN) algal culture samples (Fig 2A). Bacteria were observed on individual cells, but more frequently, they were seen on the mother cell wall envelope. In contrast, there were no bacteria in the axenic culture AA (Fig 2B).

Molecular verification of axenicity
Bacterial selective media experiments and growth experiments in the presence of organic carbon source confirmed the absence of culturable bacteria under the specific tested trophic conditions. However, since there could be non-cultivable bacteria, we further verified the axenicity of the putative H. lacustris AA axenic culture with high sensitivity molecular tools. The16S rRNA gene, a common prokaryotic genetic marker, was used in this study to reveal foreign sequences in H. lacustris gDNA preparations. Successful amplification of both algal 18S and bacterial 16S rRNA genes was obtained with the selected universal set of primers (see Material and Methods) and Q5 HF DNA polymerase enzyme (Fig 3). Gradient PCR analysis (Ta = 50-60°C), with both gDNA templates, extracted either from the putative axenic (AA) or the non-axenic (RN) H. lacustris algal culture, resulted in the amplification of two bands which were visible in agarose gels: the upper band (UB) was ca 1100 bp in size, while the lower band (LB) was 900-1000 bp in size (Supplementary data S1). The LB was similar in size to the amplified band of the Enterococcus sp. positive control (Fig 3, lane 11). In the gDNA extracted from the non-axenic algal culture (Fig 3, lane 9), the ratio between the lower and the upper amplified bands (Supplementary data S1) was higher than the ratio of the corresponding bands amplified from the gDNA extracted from the putative axenic culture (Fig 3,  lanes 3-8).
For the initial characterization of the putative axenic H. lacustris culture, the UB and LB amplified fragments at 56°C annealing temperature were sequenced. It was found that the AA_LB obtained from the putative axenic culture gave the highest coverage and similarities with the H. lacustris chloroplast genome MG677935.1 SSU rDNA gene (Table 1), indicating the absence of bacteria and the axenicity of the isolated culture; the upper band (AA_UB) from the putative axenic culture showed high similarity (99.65%) and coverage (100%) to H. lacustris 18S rDNA gene fragment DQ009774.1 (Table 1).
For further characterization of the PCR-amplified fragments, derived from the putatively axenic and nonaxenic H. lacustris gDNA samples, we cloned and sequenced those PCR products. Both UB and LB PCR products from each putatively axenic (AA_LB and AA_UB) and non-axenic (RN_LB and RN_UB) H. lacustris gDNA were blunted and ligated separately into pJET 1.2 cloning site. The plasmids were used to transform E. coli HST08 Stellar Competent Cells. Cloned insert colonies obtained from the axenic culture were named AA_UBC1 to AA_UBC8 and AA_LBC1 to AA_LBC8, and the cloned insert colonies obtained from the non-axenic culture were named RN_UBC1 to RN_UBC6 and RN_LBC1 to RN_LBC9.  and bacterial (contamination control) gDNA. U519F and Uni1390R primers and Q5 HF DNA polymerase were used for product amplification. 1, 50 bp ladder; 2, No template negative control at 56°C; 3-8, putative axenic AA H. lacustris gDNA at Ta of 50, 52, 54, 56, 58 and 60°C, respectively; 9, non-axenic H. lacustris RN gDNA at 56°C; 10, Escheria coli Stellar HST08 competent cells at 56°C; 11, Enterococcus sp. gDNA at 56°C (positive control bacterial template); and 12 100 bp ladder. After streaking the cloned lines on selective media (Fig 4), nine transformed colonies from each construct were screened by on-colony PCR, using pJET 1.2 kit Fw and Rw primers and Q5 HF proofreading DNA polymerase. All selected colonies were positive except the ones obtained from the negative control -pJET 1.2 empty vector. Consequently, nine transformed lines from streaks on plates were grown in liquid cultures under selection for plasmid isolation for fragment length evaluation. One representative strain from each "cloned construct" is shown in Fig 5. The putative axenic upper band (AA_UBC1) (Fig 5, lane 6) and the non-axenic upper band (RN_UBC1) (Fig 5, lane 8) were 1000-1300 bp in size, while the putative axenic AA_LBC1 (Fig 5, lane 7) and the non-axenic RN_LBC1 (Fig 5, lane 9) fragments were about 1000 bp in size (Supplementary data S2). The negative control with no construct (Fig 5, lane 3) showed no amplified product.
After insertion verification, plasmid preps from the nine selected transformed lines from each construct were sent for sequencing with CloneJET PCR Cloning Kit recommended primer pair (the sequencing output was not good for all nine transformed lines of each construct; only 8 out of 9 were good for the UB from axenic culture, 6 out of 9 for the UB from non-axenic cultures and 8 out of 9 for the AA_LBC were sequenced good, and all the 9 were sequenced good for the RN_LBC). Sequence similarities and coverage obtained after processing both upstream and downstream sequences obtained from the cloned UB and LB of the putative axenic and non-axenic H. lacustris cultures are summarized in Tables 2 and 3. The UB sequences obtained from the cloned putative axenic strain were named AA_UBC1 to AA_UBC8; the UB sequences from the non-axenic strain (RN) were named RN_UBC1 to RN_UBC6. The sequence similarity analysis in NCBI GenBank showed that all sequences

Phylogenetic tree 1
To prove that our cloned sequences obtained from the upper band are of algal origin (H. lacustris 18S ribosomal RNA gene), we rooted the tree with H. pluvialis chloroplast rRNA sequence as outgroup (Fig 6). Based on the tree shape, we separated tips of the tree (tree leaves) into three clusters. Apart from these three clusters, we also included the outgroup, ingroup and an external control (see Material and Methods). Sequences containing both axenic (AA_UBC) and non-axenic (RN_UBC) cultures (Cluster 2) cluster together with H. lacustris 18S ribosomal RNA gene partial sequence (ingroup control). This cluster is grouped with Cluster 3 containing all the 18S partial sequences from different green algal species (taken from NCBI). This means that the population of cloned amplified fragments originated purely from algal 18S ribosomal RNA gene. Surprisingly, two axenic (AA_UBC2 and 7) and one non-axenic (RN_UBC4) sequences formed a separate cluster (Cluster 1), indicating that nucleotide sequences of these clones are divergent. Compared to the sequences from Cluster 2, the sequence of AA_UBC7 contains an insertion of four nucleotides, which produces D122E and a shift in the reading frame, and the sequences of AA_UBC2 and RN_UBC4 contain multiple substitutions resulting in protein truncation.

Phylogenetic tree 2
To prove that our putative axenic culture does not contain any bacteria, we built a phylogenetic tree with the sequences of our clones (axenic and non-axenic) obtained from the lower band. We identified four distinct clusters in the tree (Fig 7). H. lacustris 18S rRNA gene partial sequence (DQ009774.1) was used as an outgroup control.  All sequences from our putative axenic clones (AA_LBC) grouped in a single cluster (Cluster 4). Moreover, highly similar sequences we obtained from NCBI, i.e., Internal control 2, H. pluvialis 16S rRNA gene complete sequence chloroplast and Chlorophyceae species 16S ribosomal RNA (e.g., Dunaliella, Chlamydomonas), are grouped, together with AA_LBC clones, in Cluster 4. This supports our hypothesis that all sequences of our AA_LBC clones are identical and have high similarity to algal chloroplast sequences.
The sequences of our putative non-axenic (RN_LBC) clones form three distinct clusters (Cluster 1-3). RN_LBC9 sequence forms a cluster with species of Blastococcus; RN_LBC2, 4, 5 and 8 form a cluster with Deinococcus species, and RN_LBC1, 3, 6 and 7 show high similarity to species of Massilia (Fig 7). We conclude that our non-axenic clones are associated with several bacterial species.

Discussion
In this study, we report a detailed simple protocol to obtain an axenic culture of the commercially important microalga H. lucustris, together with a molecular tool to test the axenic status of the culture. Previously, Cho et al. (2013) described a novel approach to isolate algal cells free of bacterial contaminants; however, this requires sophisticated equipment for cell sorting and is not adequate for the removal of bacteria associated with the cell surface. To our knowledge, there have been no reports of axenic cultures of H. lacustris, although this commercially important alga has been intensively studied. Here, we used simple conventional microbiology methods of antibiotic (amoxycillin) treatment, plating and manual picking of colonies to free H. lucustris culture from bacteria attached to its cell surface. Since H. lucustris is not a fast-growing alga, this simple process took a few months, as recovery and growth to high density was very slow.
Of the three antibiotics used, gentamicin, streptomycin and amoxycillin, only amoxycillin (2.5 µg ml -1 ) was effective in killing bacteria without harmful effects on the growth of the alga. According to Agostini et al. (2019), streptomycin is one of the antibiotics widely used to inhibit bacterial growth in phytoplankton cultures. In general, streptomycin when used in combination with other antibiotics (such as kanamycin, neomycin, ampicillin and gentamicin) is effective for the isolation of axenic microalgal cultures (Jichang, Song, & Lin, 2015). Similarly, Soffer & Baker (2008) and Xiang et al. (2013) also reported that streptomycin in combination with kanamycin and ampicillin helps in obtaining axenic cultures of Symbiodinium. In our study, streptomycin was not useful, as it greatly inhibited the growth of H. lacustris. Gentamycin was observed to have similar algicidal effect in our study, whereas it was reported to be effective for the axenic isolation of the marine microalga Isochrysis galbana when used in combination with other antibiotics (such as ampicillin, kanamycin, neomycin and streptomycin) (Cho et al., 2002). Both antibiotics -streptomycin and gentamicinare aminoglycoside antibiotics, which can enter the algal chloroplast and inhibit plastid protein synthesis (Schwartzbach & Schiff, 1974) by targeting the 30S ribosomal subunit (Cammarata, 1973). This explains why in our study these two antibiotics inhibited the growth of H. lacustris. In contrast, amoxycillin is a broad-spectrum beta-lactam antibiotic, which acts by binding to the penicillin-binding proteins, inhibiting the transpeptidation process, leading to the inhibition of peptidoglycan synthesis and autolysis of bacterial cell wall. This process ultimately leads to the death of the bacterial cell (Elizalde-Velázquez et al., 2016). In our study, amoxycillin was found to be effective against the bacteria and non-toxic to the H. lacustris. In agreement with our findings, showing no toxic effect of amoxycillin towards H. lacustris, González-Pleiter et al. (2013) reported that amoxycillin is toxic to the cyanobacterium Anabaena strain CPB4337 but not to the green alga Pseudokirchneriella subcapitata.
H. lacustris is a freshwater alga and its environment contains different kinds of waterborne/airborne bacterial contaminants. It is not clear whether these bacteria help the alga to survive in its natural niche by providing vital substances. However, in contrast to the previous report (Provasoli et al., 1974) on a vitamin B 12 requirement for H. lacustris growth, we did not find any significant difference in the growth of the axenic culture with or without vitamin B 12 supplement, and in both cases, it grew well to high densities. We thus conclude that H. lacustris can synthesize its own vitamin B 12 or use B 12 -independent enzymes in its biosynthetic pathways. We searched the literature for cobalamin biosynthesis in the bacterial genera found in our nonaxenic H. lacustris culture and found that in the genome of Deinococcus radiodurans, many genes of the pathway were identified, suggesting a possible partly functional pathway for cobalamin biosynthesis in Deinococcus (Makarova et al., 2001); in Massilia and Blastococcus, the search yielded no results, indicating these genera do not exhibit B12 synthesis. Interestingly, in the draft genome of H. lacustris, several contigs encoding B12 biosynthetic enzymes and the B12-dependent methionine synthase, exhibiting high similarity to Deinococcus, were identified in our non-axenic H. lacustris culture (http://aranne5.bgu.ac.il/others/BlaschkauerMihal1985. pdf). These results suggest that Deinococcus malanensis found in the non-axenic culture in the current study can synthesize B12. Though B12 supplement is not mandatory for H. lacustris, it is still possible that in nonaxenic culture, H. lacustris expresses B12-dependent methionine synthase and relies on Deinococcus for B12 supply.
The axenicity of an isolated algal culture can also be confirmed by microscopic observations of the algal culture under SEM (Yim & Lee, 2004;Cho et al., 2013;Lee et al., 2015;). Our SEM observations clearly showed bacterial presence in the non-axenic culture but no bacterial presence in the isolated strain H. lacustris AA, thus confirming the axenicity of our isolated H. lacustris AA culture.
For conclusive confirmation of the axenicity of our isolated culture, amplification of the 16S/18S rRNA gene of the isolated axenic culture with a "universal" primer pair for Bacteria, Archaea and Eukarya (see Material and Methods) was carried out. The results indicate that the isolated axenic algae culture is free of bacteria. Yet, there is a possibility that the lower band sequence obtained for the axenic culture may also contain some amplicons of bacterial origin that were not detected by band sequencing. To rectify this issue, the amplified lower and upper bands of the axenic culture and non-axenic cultures were purified, cloned and sequenced (Table 3). The results not only verified the axenicity of the putative axenic culture but also allowed us to identify the bacterial genera that harbour the non-axenic culture as Massilia, Deinococcus and Blastococcus. The phylogenetic results showed that the upper band amplified from the axenic strain and the non-axenic culture were all similar and belonged to the same H. lacustris strain. The neighbour joining method shows that the sequences obtained from the lower band amplified from the axenic culture are closely related to the chloroplast gene of the H. lacustris, whereas the sequences from the lower band amplified from the non-axenic culture show an evolutionarily close relationship to the bacterial genes (Fig 7).
As bacterial genera similar to those identified in our non-axenic alga were isolated from sites with extreme conditions, we hypothesize that the bacteria found in our H. lacustris non-axenic culture originate from the natural habitat from which the alga was first isolated. For example, D. malanensis was isolated from radiation polluted soil (Microbiol et al., 2017), D. hohokamensis and D. navajonensis strains were isolated from desert soils (Rainey et al., 2005), B. aggregatus was isolated from the surface of marble and calcareous stones (Urzì, Salamone, Schumann, Rohde, & Stackebrandt, 2004), B. colisei was isolated from limestone from the Amphitheatre (Hezbri et al., 2017) and M. niabensis and M. jejuensis were isolated from air samples (Weon et al., 2010). Since H. lacustris is an inhabitant of freshwater and small rock pools on the Island of Trutbådan, it is possible that different types of waterborne, airborne and epilithic bacteria were attached to its cell wall. Hence, these bacterial genera could be considered as members of the core microbiota of the alga but could also be the contaminants from the natural habitat of the microalga, which were strongly associated with the algal cell surface. Though the specific interactions of the different bacteria identified in the current work with H. lacustris were not characterized, we can hypothesize that M. niabensis can exert a positive effect on the growth of H. lacustris in its natural habitat, as certain species of the genus Massilia were reported to promote plant growth through their ability to solubilize recalcitrant phosphate sources in soils (Zheng, Bi, Hao, Zhou, & Yang, 2017).
Nowadays, there is an increasing interest in bacteria associated with algae, either in the phycosphere or on the surface of the cell as biotic factors to promote algal growth/astaxanthin accumulation. In laboratory-grown cultures of H. lacustris BM1 (IPPAS H-2018) (isolated from the White Sea coastal zone), Proteobacteria and Bacteroidetes were found to be the predominant phyla, while the Caulobacter genus (phylum: Proteobacteria) became abundant during astaxanthin accumulation (Chekanov, Zaytseva, Mamedov, Solovchenko, & Lobakova, 2021). This bacterium, however, was not found in the phycosphere of the algae at the site of isolation (Kublanovskaya et al., 2020), indicating that this bacterium has no physical interaction with the alga. Interestingly, the bacterial genus Achromobacter, which was found in the microbiome of another H. lacustris strain NIES 144, Lee et al. (2019) and reported to accelerate the growth of the alga, was absent in the abovementioned microbiome. Although no bacterial genus was found to be common in both microbiomes, or in common with our identified bacterial strains, in general, all three H. lacustris strains were rich with bacteria belonging to the phylum Proteobacteria. This phylum was reported to be very abundant in different environmental algal samples (Ramanan et al., 2016), and our results, together with the microbiomes of the other two H. lacustris strains, support a close association or physical interaction of Proteobacteria with the algal cell surface. Based on our results, we suggest commensal relationships between the alga and bacteria, which can acquire the "waste" organic material from empty mother cells after algal cell division.
To conclude, the addition of antibiotics to microalgal cultures to get rid of bacteria is a common method in isolating axenic cultures in laboratories; however, finding the appropriate antibiotic (not toxic to the target isolated cell) and method for axenicity verification is challenging. We found that treating the microalga H. lacustris with amoxycillin and frequent sub-culturing of the alga in amoxycillin-amended rich media provides a tool for isolating bacterial-free H. lacustris culture. Macroscopic observations (showing no bacterial colonies visible to the naked eye in different organic nutrient-rich solid media) and microscopic (both light and SEM) observations clearly indicated the axenicity of the isolated culture, while it was finally confirmed by molecular characterization of the 16/ 18S rRNA gene of the isolated H. lacustris culture. In addition, the bacterial genera associated with the nonaxenic alga were also identified. The protocol described herein for establishing axenic H. lacustris culture is simple, and it can be adopted to eliminate bacteria that are strongly attached to or embedded in the microalgal cell surface. Our findings further expand the knowledge of the bacteria inhabiting H. lacustris, and together with the ability to establish axenic culture of the alga, the ability to explore controlled co-cultures is now available. In the future, the methods of metabarcoding, which combines massamplification of short DNA sequences encoding 16S rRNA (barcode) with high-throughput sequencing (Oliveira et al., 2018), can be used to characterize multiple species in a single sample or to verify culture axenicity.

Author contributions
A. Asatryan: conceptualization, methodology, writing -original draft preparation, data curation, formal analysis and visualization; M. Gunasekaran: data curation, formal analysis and visualization; A. Zarka: resources, funding acquisition, project administration, supervision writing-reviewing and editing; S. Boussiba: resources and funding acquisition.