A marine Chlamydomonas sp. emerging as an algal model

The freshwater microalga Chlamydomonas reinhardtii, which lives in wet soil, has served for decades as a model for numerous biological processes, and many tools have been introduced for this organism. Here, we have established a stable nuclear transformation for its marine counterpart, Chlamydomonas sp. SAG25.89, by fusing specific cis‐acting elements from its Actin gene with the gene providing hygromycin resistance and using an elaborated electroporation protocol. Like C. reinhardtii, Chlamydomonas sp. has a high GC content, allowing reporter genes and selection markers to be applicable in both organisms. Chlamydomonas sp. grows purely photoautotrophically and requires ammonia as a nitrogen source because its nuclear genome lacks some of the genes required for nitrogen metabolism. Interestingly, it can grow well under both low and very high salinities (up to 50 g · L‐1) rendering it as a model for osmotolerance. We further show that Chlamydomonas sp. grows well from 15 to 28°C, but halts its growth at 32°C. The genome of Chlamydomonas sp. contains some gene homologs the expression of which is regulated according to the ambient temperatures and/or confer thermal acclimation in C. reinhardtii. Thus, knowledge of temperature acclimation can now be compared to the marine species. Furthermore, Chlamydomonas sp. can serve as a model for studying marine microbial interactions and for comparing mechanisms in freshwater and marine environments. Chlamydomonas sp. was previously shown to be immobilized rapidly by a cyclic lipopeptide secreted from the antagonistic bacterium Pseudomonas protegens PF‐5, which deflagellates C. reinhardtii.

The freshwater microalga Chlamydomonas reinhardtii, which lives in wet soil, has served for decades as a model for numerous biological processes, and many tools have been introduced for this organism. Here, we have established a stable nuclear transformation for its marine counterpart, Chlamydomonas sp. SAG25.89, by fusing specific cisacting elements from its Actin gene with the gene providing hygromycin resistance and using an elaborated electroporation protocol. Like C. reinhardtii, Chlamydomonas sp. has a high GC content, allowing reporter genes and selection markers to be applicable in both organisms. Chlamydomonas sp. grows purely photoautotrophically and requires ammonia as a nitrogen source because its nuclear genome lacks some of the genes required for nitrogen metabolism. Interestingly, it can grow well under both low and very high salinities (up to 50 g Á L -1 ) rendering it as a model for osmotolerance. We further show that Chlamydomonas sp. grows well from 15 to 28°C, but halts its growth at 32°C. The genome of Chlamydomonas sp. contains some gene homologs the expression of which is regulated according to the ambient temperatures and/or confer thermal acclimation in C. reinhardtii. Thus, knowledge of temperature acclimation can now be compared to the marine species. Furthermore, Chlamydomonas sp. can serve as a model for studying marine microbial interactions and for comparing mechanisms in freshwater and marine environments. Chlamydomonas sp. was previously shown to be immobilized rapidly by a cyclic lipopeptide secreted from the antagonistic bacterium Pseudomonas protegens PF-5, which deflagellates C. reinhardtii. qPCR, Real-time quantitative PCR; SNP, single nucleotide polymorphism; Tubulin, Tub; UTR, untranslated region Eukaryotic photosynthetic unicellular organisms, known as microalgae, contribute significantly to carbon fixation on Earth and form the basis of food webs in freshwater and marine ecosystems (Field et al. 1998). They can be found in aquatic environments, wet soil, or soil crusts, and they are exposed to ever-changing abiotic conditions such as temperature and light. In nature, they co-exist with other microorganisms, such as bacteria and fungi, which can have an impact on their survival and growth rates. To study and understand the influence of abiotic and biotic factors within the hundreds of thousands of different microalgae in detail, we need several microalgal model systems with established molecular tools that can be manipulated genetically. For decades, the unicellular green biflagellated alga Chlamydomonas reinhardtii, which lives in wet soil (reviewed in Sasso et al. 2018), has been in constant use and developed as a model to study, among others, photosynthesis, the structure and function of the flagella, as well as light-and temperature-driven processes (reviewed in Salom e and Merchant 2019). Recently, C. reinhardtii has also been used to study biotic interactions with antagonistic heterotrophic bacteria. It has been shown that specific secondary metabolites produced by Pseudomonas protegens attack the algal cells and inhibit their mobility and growth (Aiyar et al. 2017). One of these, the cyclic lipopeptide Orfamide A, causes an increase in cytosolic Ca 2+ , which leads to the deflagellation and subsequent immobilization of C. reinhardtii within minutes (Aiyar et al. 2017). Other known secondary metabolites of P. protegens PF-5 present in a bacterial extract are involved in its growth arrest and it has yet to be determined which are responsible for this step. Interestingly, the identified lipopeptide was also active in an aquatic marine Chlamydomonas species (Chlamydomonas sp. SAG25.89), which was immobilized rapidly as well upon treatment. Chlamydomonas sp. SAG25.89, verified to be Chlamydomonas sp. CCMP235 (Fig. S1a in the Supporting Information), was isolated from the Nantucket Sound (USA; https://ncma.bigelow.org/ ccmp235). This marine biflagellated green alga offers the possibility to compare the interactions and mechanisms studied in the freshwater Chlamydomonas species in the marine environment.
Marine green Chlamydomonas species have been the focus of research on other important biological questions recently. The Marine Microbial Eukaryote Transcriptome Sequencing Project (MMETSP) aimed to shed light on the functional diversity of eukaryotic life in the oceans (Keeling et al. 2014). Thereby, Chlamydomonas euryale, a marine species discovered by Lewin in 1957(Lewin 1957 in Novae Scotia, Canada, and also isolated in the Yellow Sea (https://ncma.bigelow.org/ccmp219) was one of two Chlamydomonas species to be selected in this project (BioSample: SAMN02739957; BioProject ID: PRJNA231566). This species also served as a marine green algae model for studies on fatty acid and polyketide synthesis (Kohli et al. 2016). Another marine strain, Chlamydomonas sp. UWO241, which is found in the arctic area, tolerates low temperatures, hypersalinity, and extreme shade, and revealed special trades such as an altered organization of the photosystem I (Kalra et al. 2020). In addition, Chlamydomonas sp. KNM0029C, an additional arctic strain, has been used for the co-production of biodiesel and bioethanol (Kim et al. 2020). Some Chlamydomonas species from coastal subtropical areas have also been recently sequenced, which were found to have especially abundant genes involving sulfate transport, sulfotransferase, and glutathione S-transferase activities in their genomes (Nelson et al. 2019).
To study the biotic and abiotic interactions of algae in their environment, investigate their evolution, and efficiently perform algal-based biotechnology processes, we need a variety of algal models. To date, only a few algal model organisms have emerged, in which not only their genome and/or transcriptome sequence data are available, but also whose genome can be transformed (Grossman et al. 2007, Cock and Coelho 2011, Chang et al. 2016, Salom e and Merchant 2019, Falciatore et al. 2020. For marine green algae, this includes the non-flagellated marine picoeukaryote Ostreococcus tauri (Keeling 2007, van Ooijen et al. 2012, Lozano et al. 2014, Sanchez et al. 2019, as well as the extreme halotolerant cell wall-less green alga Dunaliella salina, which lives in sea salt fields (Song et al. 2019); and the green macroalga Ulva mutabilis, which relies on biotic interactions with bacteria to shape its structure (Oertel et al. 2015. Very recently, a transformation system has also been established in some archiplastidal marine species, including Micromonas and Tetraselmis (Faktorov a et al. 2020).
Here, we aimed to set up a model for a marine Chlamydomonas as a typical cell wall containing, biflagellated unicellular alga using Chlamydomonas sp. SAG25.89. With this model, we will be able to study its biotic and abiotic interactions at a molecular level and compare them with the freshwater counterpart C. reinhardtii. We first determined the growth conditions for Chlamydomonas sp. SAG25.89 and found that it tolerates a broad range of salinities, requires ammonium as a nitrogen source, grows purely photoautotrophically, and halts its growth at 32°C. We further defined hygromycin and paromomycin as usable targets for selection markers and obtained the first sequencing data of its nuclear genome. We cloned the relevant cis-acting elements of its Actin (Act1) gene and used them to design GENETIC TOOLS IN CHLAMYDOMONAS SP. and assemble a transformation vector with a Aph7" resistance cassette. Finally, we established a transformation protocol involving a high-quality electroporation system with several variables used for stable nuclear transformation of Chlamydomonas sp. SAG25.89, taking advantage of its ability to survive at low and high salinity levels.

MATERIALS AND METHODS
Strains and culture conditions. Chlamydomonas reinhardtii SAG73.72 and Chlamydomonas sp. SAG25.89 were obtained from the "Sammlung von Algenkulturen" (SAG) of the University of G€ ottingen. Chlamydomonas euryale CCMP219 and Chlamydomonas sp. CCMP235 were obtained from the National Center for Marine Algae and Microbiota (NCMA). Culturing of C. reinhardtii was performed as described previously . Cultures of Chlamydomonas sp. and C. euryale were grown as indicated, either under LD16:08 or under LD12:12 with a light intensity of 60 lmol Á m -2 Á s -1 at 18°C unless otherwise indicated. The medium used was a modification of the 3N-BBM + V medium (Bischoff and Bold 1963), in which the nitrate has been replaced in some cases by 17 mM NaNO 2 or 17 mM NH 4 Cl and 20 mM of HEPES, pH 8, as indicated. We refer to the latter medium that was used routinely as NH 4 -BBM. In the cases when extra salt was added, NH 4 -BBM is followed by the concentration of salt added (e.g., NH 4 -BBM + 25 g Á L NaCl -1 ).
Genomic DNA extraction and library preparation for Illumina sequencing. The genomic DNA extraction was performed using NucleoSpin Plant II (MACHEREY-NAGEL) kit with an initial biomass of 2 g algal cells obtained from a 10-d-old culture of each strain (Chlamydomonas sp. SAG25.89 and C. euryale, respectively). DNA integrity and fragment lengths were verified using an Agilent 2100 Bioanalyzer system. Illumina libraries were prepared using the TruSeq DNA Nano kit according to the manufacturer's instructions. The resulting two libraries were multiplexed and sequenced in rapid mode on one lane of an Illumina HiSeq2500 platform, yielding 100 bp paired-end reads.
Genomic DNA extraction and library preparation for PacBio sequencing. The genomic DNA extraction was performed using a QIAGEN Genomic-tip 20/G kit with an initial biomass of 2 g obtained algal cells from a 10-d-old culture of each strain. DNA integrity and fragment lengths were checked using an Agilent 2100 Bioanalyzer system. Prior to sequencing, short fragments were removed using the BluePippin system. Then, long-read library preparation and sequencing were performed following the manufacturer's instructions for a PacBio RS II platform using four SMRTcells per strain.
Preliminary genome draft of Chlamydomonas euryale and Chlamydomonas sp. SAG25.89. For each strain, we generated a preliminary draft assembly using long-read data. The raw reads from four PacBio SMRTcells per strain were combined, quality controlled with FastQC v0.11.7, and finally assembled using Canu 1.5 (Koren et al. 2017) with the following parameters: correctedErrorRate = 0.06 genomeSize = 120m. No further polishing or error-correction was performed, and the Illumina data were not used in the assembly process. We assessed the quality of these two assembly drafts with QUAST v4.4 (Gurevich et al. 2013).
Phylogenetic analysis of the 18S rRNA genes of Chlamydomonas and related species. For the phylogenetic analysis of the 18S rDNA gene, we used the sequences available for various known Chlamydomonas and related species with the names given after the several reclassifications that the Chlamydomonas genera have undergone (Lewin 1957, Nakazawa et al. 2001, Pr€ oschold et al. 2001, Buchheim et al. 2003, Pocock et al. 2004, Eddie et al. 2008, Demchenko et al. 2012, Yumoto et al. 2013, Nakada and Tomita 2014, Lemieux et al. 2015, Mertens et al. 2015, Wang et al. 2016, Munakata et al. 2016, Watanabe and Lewis 2017 Appendix S1 in the Supporting Information). A Multiple Sequence Alignment (MSA) was calculated using MAFFT (Katoh and Standley 2013) using the E-INS-I algorithm and enabled sequence direction correction. Otherwise, the default values were used. The 5' and 3' ends of the sequences were trimmed, and only the region available for all species was kept.
Evolutionary analyses were performed with MEGA X (Kumar et al. 2018). The phylogenetic tree with the highest log likelihood was derived using the Maximum Likelihood method and the General Time Reversible model (Nei and Kumar 2000). The tree with the highest log likelihood was calculated. The percentage of trees, in which the associated taxa clustered together, is shown next to the branches (bootstrap value of 1000). The initial tree(s) for the heuristic search were determined automatically by applying Neighbor-Joining and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach and then selecting the topology with the superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites ([+G], 5 categories). The rate variation model allowed for some sites to be evolutionarily invariable ([+I]).
Comparison of the 18S and 26S rRNA genes. We used the preliminary draft of an assembled genome of Chlamydomonas sp. SAG25.89 and C. euryale to pinpoint and extract the rDNA loci. The PacBio reads were mapped back to the fragment using the default values of minimap2 (Li 2018) using themap-pb option. The Illumina reads were mapped using the default values of BWA-mem (Li 2013). The mapped reads were then extracted and SPAdes 3.12 (Antipov et al. 2016) was used for hybrid assembly, selecting the options --onlyassemble, --careful, and multi-k-mer sizes of 81,85,89,93, and 97. The resulting contig was used as the final rDNA.
Comparisons between Chlamydomonas sp. and C. euryale were achieved by pairwise alignment between the contigs using MAFFT (Katoh and Standley 2013) with the G-INS-I algorithm. Comparisons of the 18S and 26S rDNA from Chlamydomonas sp. CCMP235 were performed with the same procedure but using the sequences found in the NCBI database (Accession no. DQ009754 and DQ015721).
Analysis of the poly(A) signal. The transcriptome of Chlamydomonas euryale from the iMicrobiome project (Accession number: MMETSP0063) was used together with an in-house script (Appendix S2 in the Supporting Information). The analysis was performed in three steps. In the first step, the mRNAs containing a poly(A) tail with a length equal or longer than one A were included and then clustered according to the poly(A) tail length. In the second step, the transcriptome was shuffled three times and the procedure from the first step was repeated. This second step makes it possible to determine the cutoff length of the poly(A) tail to avoid false positives for the dataset. In the third step, the script was executed again specifying the minimum poly(A) tail length (six adenines in this case) and the 100 bp preceding the poly (A) tail were extracted. As a control, the 100 bp dataset was shuffled. MEME suite (Bailey et al. 2009) 4.11.1 was then used to determine motifs in the last 100 bp preceding the poly(A) tail. MEME (Bailey and Elkan 1994) and DREME (Bailey 2011) were run with the options shown in Table S1 in the Supporting Information. The motifs detected were used for further analysis. With the predicted motifs from MEME and DREME, CentriMo (Bailey and Machanick 2012) was run 56 to see whether there was any positional enrichment of the motifs in the sequence.
Evaluation of the different genes from Chlamydomonas euryale. To determine gene positions in the draft genome of Chlamydomonas euryale, we used tBLASTn with the default parameters. The C. reinhardtii protein sequences were used as query and the genome draft of C. euryale as the target. It is generally accepted that an identity of 30-40% in long protein alignments (100 aa or longer) is significant to consider two proteins as homologs, while 20-35% is considered the twilight zone and structural information is needed to infer homology (Rost 1999, Pearson 2013. We set the cutoff value at 40% identity as the strains are reported to be related. The second criterion used was the minimum query coverage. We binned the results in four groups, higher or equal to 75%, between 50 and 75%, between 20 and 50%, and lower than 20%. For all the candidates, a fragment including 1,000 bp upstream and downstream from the hit region was used. The DNAseq data were mapped to the fragment using BWA-mem (Illumina reads) and Minimap2 (PacBio). The mapped reads were then filtered and qualitytrimmed using Trimmomatic (Bolger et al. 2014) with the parameters LEADING:24; TRAILING:24; SLIDINGWIN-DOW:4,28; MINLEN:81. The resulting reads were used to reassemble the fragment using SPAdes 3.12 (Antipov et al. 2016) with the--only-assemble and--careful options enabled and the k-mer sizes of 81, 85, 89, 93, and 97. Afterward, the RNAseq data were mapped using HISAT2 v2.1 (Kim et al. 2019) to the curated genomic fragment, and the intron/ exon borders were established.
Evaluation of the different genes from Chlamydomonas sp. SAG25.89. For the curation of the genomic DNA of Chlamydomonas sp. SAG25.89, we used the same methods as described for C. euryale. Since no RNAseq data are available for Chlamydomonas sp. SAG25.89, we calculated the theoretical gene model using NCBI SPlign (Kapustin et al. 2008) by aligning the predicted protein for C. euryale to the curated genomic fragments of Chlamydomonas sp. SAG25.89. To calculate the domain structure of the proteins, SMART (Kapustin et al. 2008) was used.
Agarose plug preparation and restriction analysis for Southern blots Agarose plug protocol (adapted from Sambrook andRussell 2001, Pai et al. 2018). For each plug, 5 x 10 8 cells of either Chlamydomonas sp. SAG25.89 wild type or two of the transgenic lines were harvested from 10-d-old cultures. Cells were centrifuged for 10 min at 4000g, washed twice with distilled water, and resuspended in 500 lL of distilled water. The final volume of each suspension (about 600 lL to 650 lL) was transferred to a 1.5 mL Eppendorf tube, prewarmed for 5 min at 42°C, mixed with 500 lL of 2% low melting point agarose (SeaKem â GTG Agarose) at 42°C, and poured immediately into the plug molds (190 lL Á plug -1 ) giving a final cell amount of roughly 10 8 cells Á plug -1 equivalent to about 10 lg gDNA Á plug -1 . The plugs were let to solidify at 4°C for 15 min.
Afterward, the plugs were transferred to a 50 mL Falcon tube containing 10 mL of TLB buffer (100 mM NaCl, 10 mM Tris-HCl, 25 mM EDTA, 0.5% [w/v] SDS, pH 8.0). RNAse A (20 µg Á mL -1 ) was added freshly just before use and incubated for 30 min at 37°C without shaking. Thereafter, 1 mg Á mL -1 of Proteinase K was added, and the plugs were incubated overnight at 50°C without shaking. The supernatant was poured off and the plugs were washed twice for 1 h with 10 mL Agarose embedded DNA wash buffer (20 mM Tris-HCl, 50 mM EDTA; pH 7.2). The plugs were then stored in Agarose embedded DNA wash buffer at 4°C until further use.
Individual plugs were transferred to 2 mL Eppendorf tubes and washed twice for 15 min at 4°C with 200 lL 1X New England Biololabs (NEB) CutSmart buffer; then, 200 lL of fresh 1X NEB CutSmart buffer was added and incubated at 4°C overnight. The 1X NEB CutSmart buffer was removed, and 100 lL of fresh 1X NEB CutSmart buffer was added. For each Eppendorf tube, 20 U of the desired restriction enzyme was added, flicked gently to evenly distribute it, and was incubated overnight at 37°C.
Southern blot analysis. The previously prepared agarose plugs were cut in four pieces. One piece was loaded into a well of a 0.5% Megabase certified agarose (Bio-Rad), sealed with 1% low melting point agarose, and run for 1.5-2 h at 3.5V/cm.
As probes, 300 bp long fragments of the Act1 gene from Chlamydomonas sp. SAG25.89 and the Aph7" CDS from the plasmid were synthesized using the Roche PCR DIG Probe Synthesis Kit according to the manufacturer instructions. As template for the probe synthesis 10 pg of purified SfiI-restricted fragment of pDCF4 and 10 pg of purified PCR-product from the Act1 gene were used. The sequences of the primers used to synthesize the probes and to amplify the fragment from the gDNA (in the case of Act1) were as follows: Hybridization and detection were done following the "DIG Application Manual for Filter Hybridization" from Roche using Roche's Anti-Digoxigenin-AP Fab fragments and AppliedBiosystems' Tropix â Ready-to-use CDP-Star as substrate.
pH tolerance. Cells were grown in medium NH 4 -BBM as described before with the following modifications. The buffer of the medium was replaced by a combination of MES, HEPES, and TRIS, each at a final concentration of 20 mM, and the pH was adjusted to the desired value using HCl and NaOH, respectively. The cells were grown for 10 days, and the number of cells was counted at the time of inoculation time and after 10 d. The resulting growth rate was calculated.
Salinity tolerance. Cells were grown in the NH 4 -BBM medium as described before with the following modifications. Additional NaCl was added to the medium, the concentration of which varied between 0 and 50 g Á L -1 in intervals of 10 g Á L -1 . The cells were grown for 10 d, and the number of cells was counted at inoculation time and final time. The resulting growth rate was calculated.
Quantitative reverse transcriptase PCR (RT-qPCR). Cells were grown in a 12:12 h light:dark cycle as described above and harvested at LD 3-4. The RNA extraction was performed using the Qiagen RNeasy Plant Mini kit and following the manufacturer's instructions with an initial biomass of 100 mg algal cells obtained from a 10-d-old culture of each strain. The RT-qPCR was performed using QIAGEN OneStep RT-PCR Kit according to the manufacturer's instructions, using 7.8 µL of the extracted RNA in a 20 µL reaction volume. The primers used to amplify the Tubulin genes were Tub-Rv, CTCCAGGTCCATCAGGATGGCACGAGGAAC, Tub1-Fw, ATTGACCCCACCGGCACCTACCATGGCGCT and Tub2-Fw, ATCGACCCCACCGGTACCTACCATGGCGCC. The primers used to amplify the Act1 gene were Act1-Fw, CCACACGTTTTTCAACGAGCTGCGCGTGGC and Act1-Rv, GGATGCCGACATACATTGCTGGGACGTTGAAGG. The cycling parameters were a two-step cycle as described in the kit.
Electroporation using NEPA21 (Yamano et al. 2013). A 7-to 10-d-old culture (50 mL at 1-5 x 10 7 cells Á mL -1 ) of Chlamydomonas sp., SAG25.89 was pelleted at 4000g and resuspended twice in sterile milliQ water to wash out all the salts of the medium and lower the conductivity. The cells were pelleted again and resuspended in either NH 4 -BBM + 40 mM sucrose (replicates 1 to 3) or in MAX Efficiency TM Transformation Reagent for Algae (Invitrogen, A24229; replicate 4) to a final concentration of 10 8 cells Á mL -1 . For each transformation, an aliquot of 500 µL of cell suspension was transferred to a 1 mL Eppendorf tube and was supplemented with the DNA to be transformed to a final concentration of 10 ng Á µL -1 . To be able to proceed with the electroporation, the impedance of the sample must range between 400 and 500 O. To achieve that value, an initial volume of 90 µL of the cell mixture was applied to the electroporation cuvette and the volume was increased in steps of 20 µL until the right impedance was achieved. The volumes needed oscillated between 90 µL and 400 µL. Detailed parameters for the transformation are listed in Table 1. After electroporation, the cells were resuspended in 10 mL NH 4 -BBM including 40 mM sucrose, transferred to a 15 mL Falcon tube, and incubated under dim light (20 µmol) overnight on a rotary shaker at 23°C. The next day, cells were pelleted, resuspended in 100 µL NH 4 -BBM, and plated onto an NH 4 -BBM agar plate containing 200 µg Á mL -1 hygromycin. It should be noted that it is very important to use plates without additional NaCl as high concentrations of NaCl cancel the effect of hygromycin. Transgenic lines grew after four to six weeks.

RESULTS AND DISCUSSION
Phylogeny of the marine green algae Chlamydomonas sp. SAG25.89 and Chlamydomonas euryale. To establish a model that would allow the study of biotic interactions between marine motile green algae and other microorganisms and in which the mechanisms and mode of action could be compared with their freshwater counterparts, we evaluated the previously used marine biflagellated alga Chlamydomonas sp. SAG25.89 (Aiyar et al. 2017). Its close relative C. euryale was also considered due to the availability of its transcriptome (iMicrobiome assembly MMETSP0063; Keeling et al. 2014). Chlamydomonas sp. SAG25.89 is smaller (3-4 µm of diameter) than C. reinhardtii SAG73.72 (10 µm of diameter) but similar in shape (Fig. 1a). A pyrenoid-like structure can be seen (red arrow in Fig. 1a, middle and right panel). Chlamydomonas sp. SAG25.89 corresponds to the strain CCMP235, which was verified with a pairwise alignment of their 18S and 26S rRNA genes, which completely matched the known sequences of Chlamydomonas sp. CCMP235 (Accession no. DQ009754 and DQ015721, respectively; Fig. S1a). Chlamydomonas sp. SAG25.89 is closely related to C. euryale (only one nucleotide difference in the 18S rRNA gene and eight in the entire 26S rRNA gene; Fig. 1b and S1b) and other marine Chlamydomonas species from the Moewusinia clade (Watanabe and Lewis 2017). It should be noted that the Moewusinia clade is not solely a marine clade (i.e., C. moewusii is a freshwater/soil microalga) but rather contains populations adapted to marine environments, which cluster within this clade. The marine species of the Moewusinia clade are relatively distant from C. reinhardtii, which belongs to the Reinhardtinia clade (Lemieux et al. 2015); in total, eight Insertions/Deletions (InDels) and 132 nucleotide substitutions accumulate in the 18S rRNA gene alone comparing Chlamydomonas sp. SAG25.89 and C. reinhardtii (Aiyar et al. 2017). The close relation of Chlamydomonas sp. SAG25.89 and C. euryale offers the possibility to use the existing transcriptome data from C. euryale to facilitate the elucidation of interesting genes from Chlamydomonas sp. SAG25.89. Although we used C. euryale for comparison (see   (Lemieux et al. 2015). Sequences are presented in Appendix S1. SNP: single nucleotide polymorphism; InDel: Insertions and/or Deletions. below), we focused primarily on Chlamydomonas sp. SAG25.89 as C. euryale was not as motile and, therefore, could not be used in the Orfamide A motility tests to study certain biotic interactions (Aiyar et al. 2017).
We also included other microalgae that are either (i) relevant from an industrial/pharmaceutical point of view (e.g., Haematococcus pluvialis), (ii) are used as marine unicellular model organisms (e.g., Ostreococcus tauri, Micromonas pusilla), or (iii) for their presence in hypersalinic waters (D. salina) to provide an overview of the relatedness of these models to Chlamydomonas reinhardtii and Chlamydomonas sp. SAG25.89. We used the most distant clade as the outgroup (O. tauri and M. pusilla).
Environmental factors influencing the growth of Chlamydomonas sp. SAG25.89. As mentioned in the Introduction, Chlamydomonas sp. SAG25.89 was found in the waters of Nantucket Sound (USA) and C. euryale in the Yellow Sea (China). First, we analyzed their growth rates under different salt concentrations, where a concentration of 30-32 g Á L NaCl -1 (513-547 mM) is typical for the sites where both were isolated (Lanbu et al. 1986, Conant, 2006. We used C. reinhardtii for comparison as a typical freshwater alga, which can only tolerate with difficulties low salinities up to about 10 g Á L NaCl -1 (171 mM; Fig. 2a), although it is capable to adapt to these low concentrations (specifically 200 mM NaCl, about 11.7 g Á L NaCl -1 ) and to recover its normal growth rate after several generations (Perrineau et al. 2014). Chlamydomonas euryale, on the other hand, shows optimal growth between 10 to 30 g Á L NaCl -1 . These concentrations are also within the tolerance range of Chlamydomonas sp. SAG25.89 (Fig. 2a). However, Chlamydomonas sp. SAG25.89 has a broader salinity tolerance, which ranges from very high salt concentrations (up to 50 g Á L NaCl -1 ; 856 mM) to freshwater medium (0.025 g Á L NaCl -1 , close to 0.43 mM) in which it also grows efficiently. Therefore, this alga is extremely flexible regarding the salt concentration, which was very advantageous for the establishment of a transformation protocol (see below). This also allows for the opportunity to use Chlamydomonas sp. SAG25.89 as a model for studies in osmotolerance. In terms of growth at different pH values, both marine strains (Fig. S2a in the Supporting Information), as well as C. reinhardtii, tolerate a broad spectrum ranging from pH 5 to pH 9 (Messerli et al. 2005).
We also tested whether Chlamydomonas sp. SAG25.89 could grow on acetate like C. reinhardtii in mixotrophic growth (light-dark cycle) as well as under constant darkness (heterotrophic growth; Fig. S2b). There was no difference in mixotrophic growth rates with acetate compared to autotrophic growth in a light:dark cycle (16:8 h light:dark). Under heterotrophic growth conditions, Chlamydomonas sp. SAG25.89 was not able to grow. Thus, Chlamydomonas sp. SAG25.89 cannot use acetate as a carbon source, so far due to unknown reasons, in contrast to C. reinhardtii.
We assessed different nitrogen sources for the growth of Chlamydomonas sp. SAG25.89, such as nitrate, nitrite, and ammonia. Interestingly, it could only grow on ammonia, the reduced nitrogen source (Fig. 2b). Using the long-read PacBio sequencing data (see Methods), we assembled a preliminary draft of the genome of Chlamydomonas sp. SAG25.89 and also of the genome of C. euryale. We assembled 699,427 (C. euryale) and 450,484 (Chlamydomonas sp. SAG25.89) reads per strain with a mean read length of 6,560 bp and 7,782 bp into draft assemblies comprising 745 and 403 contigs (4,392 and 2,417 unitigs) comprising total assembly lengths of 109 Mb and 96 Mb (135 Mb and 114 Mb if based on the unitigs), respectively. The largest contigs reached a size of 845 kbp (C. euryale) and 1,8 Mbp (Chlamydomonas sp. SAG25.89). Even without additional polishing steps, these drafts were of sufficient quality to assess the general genome structure and synteny of specific gene clusters in this study.
Using the genome draft of Chlamydomonas sp. SAG25.89, we were able to analyze potential reasons for the lack of growth in oxidized nitrogen sources. Notably, we could not find any gene sequences encoding neither the nitrate reductase (Nia1), nor the nitrite reductase (Nii1), nor the transporters for nitrate and nitrite in the plasma membrane (Fig. 2c,  Fig. S3 in the Supporting Information). In contrast, the gene sequences encoding the ammonium transporters (Amt family) and glutamine synthetases (Gln family) were present. In C. reinhardtii, randomly occurring single nucleotide mutations in the Nii1 gene lead to a loss of its activity. Although the gene is still present, these algal strains cannot use nitrate  (Voytsekh et al. 2008. The presence of several Met and Gly residues after the second RRM domain, as present in the ELAV-like family (CELF) protein C3 (Zhao et al. 2004) or Musashi , is indicated in yellow. RRM, RNA recognition domain; KH, K homology; WW, protein interaction domain.    ). To investigate further whether the marine Chlamydomonas species may have lost some of the genes involved in nitrogen metabolism, we used publicly available genome and transcriptome data from all the available Chlamydomonas strains. We also assembled a preliminary draft of the C. euryale genome as mentioned above, which was used in conjunction. Interestingly, we were unable to find any signs of the usual nitrate gene cluster (Fernandez and Galvan 2008) in either Chlamydomonas sp. or C. euryale. Chlamydomonas sp., C. euryale and some other species lacked all or some of the genes encoding nitrate and nitrite transporters as well as Nia1 and Nii1 (Fig. S3). In the marine model picoalga Ostreococcus tauri and in Micromonas pusilla, however, the nitrate gene clusters are present (Sanz-Luque et al. 2015). These data suggest that some marine green Chlamydomonas algae may have lost these genes completely.
We were also interested in finding out whether the growth of Chlamydomonas sp. SAG25.89 changed at different temperatures. The average temperature in the ocean is approximately 17°C, ranging from 27 to 30°C near the equator to below the freezing point (À2°C) within the polar region (Banfalvi 2016). At Nantucket Sound, where Chlamydomonas sp. SAG25.89 was isolated, the average temperatures are 23°C in summer and 10°C in winter (Conant, 2006). We chose a range of 15°C, 18°C, 23°C, 28°C, and 32°C to cover the average spectrum and to examine at which temperature the cells may become heat stressed. Chlamydomonas sp. SAG25.89 grew well and at a similar rate in a light:dark cycle of 16:8 h from 15°C to 28°C with a slight preference for 23°C (Fig. 2d). It reaches its stationary phase after 12 d before it starts its decline phase (Fig. 2d). At 32°C, Chlamydomonas sp. SAG25.89 cannot grow; this temperature is obviously already in the heat stress range for this alga. These data suggest that the heat stress temperature is lower for Chlamydomonas sp. SAG25.89 compared to C. reinhardtii where it is around 39°C (Tanaka et al. 2000(Tanaka et al. , R€ utgers et al. 2017. In Chlamydomonas reinhardtii, some molecular components that mediate thermal acclimation at temperatures between 18°C and 28°C are known (Zhao et al. 2004, Voytsekh et al. 2008). These include the C1 and C3 subunits of the RNAbinding protein CHLAMY1, the 5'-3' endoribonuclease Xrn1 as well as the RNA-binding protein Musashi, which is present in various splicing variants. These components are either expressed at different levels in cells grown between 18°C and 28°C or their phosphorylation pattern changes (in case of C1). C1, C3, and XRN1 are also known members of the circadian clock in C. reinhardtii (Iliev et al. 2006, Dathe et al. 2012. The knock-down, knock-out (C3 and Xrn1), or overexpression (Musashi) of some of these components alter the algal temperature-dependent growth rates, suggesting that they confer thermal acclimation in addition to their clock function . The genes encoding these subunits seem to be conserved in Chlamydomonas sp. SAG25.89 (Fig. 2e); their typical domains are conserved well enough to be recognized by detection software, but so far we have no functional data in Chlamydomonas sp. SAG25.89. Thus, the C1 subunit carries two of three KH domains responsible for RNA binding and the WW domain (protein interaction domain) at its C-terminus ( Fig. 2e; Zhao et al. 2004). In the case of C3, all domains are conserved. Three RRM (RNA recognition domains) are present and the second and third are separated by a Metrich spacer (Zhao et al. 2004). Musashi, having several splicing variants in C. reinhardtii, resembles some of them with three RRM domains ). This provides a well-defined basis for further molecular studies and potential involvement in thermal acclimation in the marine Chlamydomonas sp. SAG25.89.
Characterization of Chlamydomonas sp. SAG25.89 antibiotic resistances and cis-acting elements as basis for transformation. Our main goal was to establish a transformation protocol for Chlamydomonas sp. SAG25.89, which would allow for the introduction of diverse reporter genes in the future and performance of functional studies on genes of this marine alga (e.g., by gene silencing or genome editing). For this purpose, several important features had to be evaluated. First, it was necessary to find a selection marker. Therefore, common antibiotics were checked at different concentrations (Table 2). Although Chlamydomonas sp. SAG25.89 showed growth with kanamycin, streptomycin, and neomycin up to a relatively high concentration (50-200 µg Á mL -1 ); so far for unknown reasons, in contrast to C. reinhardtii, it could not grow at concentrations of paromomycin and hygromycin equal or higher than 50 µg Á mL -1 . We chose the hygromycin conferring resistance gene Aph7'' as the selection marker for the vector because the established aequorin reporter for C. reinhardtii (Aiyar et al. 2017) is based on this resistance and we want to use parts of this vector also for Chlamydomonas sp. SAG25.89 in the future. The vector has to contain the right GC content as well as codon-compatible sequences so that it can be properly expressed by the host. Most Chlamydomonas species have a high GC content (>60%) including C. euryale and Chlamydomonas sp. SAG25.89 (Table S2 in the Supporting Information) and their freshwater counterpart C. reinhardtii ). Since the transcriptome of C. euryale was available and this marine species is related closely to Chlamydomonas sp. SAG25.89, we used it to calculate the codon usage. The codon usage of C. euryale and C. reinhardtii is very similar in most cases; both share the same preferred codons (Fig. 3a). Because C. euryale and Chlamydomonas sp. SAG25.89 are even closer relatives than C. euryale and C. reinhardtii (Fig. 1b), we expect that this also accounts for Chlamydomonas sp. SAG25.89. This offers the unique opportunity to use genes established already for C. reinhardtii in Chlamydomonas sp. SAG25.89 as well.
For the successful transformation of Chlamydomonas sp. SAG25.89, the coding sequence of the Aph7'' gene conferring resistance to hygromycin was used, as established for C. reinhardtii in the vector pHyg4 (Berthold et al. 2002). It contains the native sequence of the Aminoglycoside phosphotransferase (Aph7") gene from Streptomyces hygroscopicus, which is also GC-rich (70.8%), has a similar codon usage to C. reinhardtii, and contains an additional stop codon at the 3' terminus with the most frequent stop codon of C. reinhardtii to guarantee the proper stop of transcription (Berthold et al. 2002).
Moreover, we evaluated the polyadenylation signal present before the start of the poly(A) tail in the transcriptome of Chlamydomonas euryale. The preferent poly(A) signal found was TG(C/T)AA (Fig. 3b), which matches the most dominant poly(A) signal present in~50% of the genes of C. reinhardtii (TGTAA; Zhao et al. 2014, Bell et al. 2016. The TG (C/T)AA poly(A) signal is also present in the b- TGC  TGT  GAC  GAT  GAA  GAG  TTC  TTT  GGA  GGC  GGG  GGT  CAC  CAT  ATA  ATC  ATT  AAA  AAG  CTA  CTC  CTG  CTT  TTA  TTG  ATG  AAC  AAT  CCA  CCC  CCG  CCT  CAA  CAG  AGA  AGG  CGA  CGC  CGG  CGT  AGC  AGT  TCA  TCC  TCG  TCT  ACA  ACC  ACG  ACT  GTA  GTC  GTG  GTT  TGG  TAC  TAT  TAA  TAG  Tubulin (Tub1 and Tub2) and the Act1 genes of Chlamydomonas sp. SAG25.89 that were selected for further analysis.
Members of the cytoskeleton, such as tubulin and actin, are usually expressed strongly, often in a constitutive manner, as they are essential for many cellular processes and/or flagellar structure. Thus, we looked for the Tub1, Tub2, and Act1 genes in Chlamydomonas sp. SAG25.89. Our aim was to take the cis-acting elements including the promoter as well as the 5'-and 3'-UTRs from one of these genes to ensure good expression levels of the hygromycin .89 Act1 promoter and its 5'-and 3'-untranslated regions (UTRs) presented in gray, the Aph7'' coding sequence (purple), and the plasmid replication and selection system for E. coli with the origin of replication (yellow), the Kanamycin resistance (green), and the Ampicillin promoter (white). The enzymes used to linearize the plasmid for transformation in Chlamydomonas sp. SAG25.89 are depicted in red and the qPCR primers used for bulk screening are depicted in blue. (B) Overview of the transformation protocol. (C) Pictures of transgenic lines and controls for the transformation of Chlamydomonas sp. SAG25.89 using pDCF4 on media with (+Hyg) and without (-Hyg) hygromycin. Colonies were picked and transferred to a new plate as soon as they were visible to avoid merging. The pictures were taken after further incubation of the original plate once the (already picked) colonies were big enough to be seen in the pictures. (D) Verification of nuclear integration of the Aph7" cassette using the qPCR primers for bulk screening. A fragment of 113 bp can be seen for the positive control with vector DNA (pDCF4) as well as for the transgenic lines (TL) but is missing in wild type (WT) and in the no template control (NT). (E) Southern Blot analysis with an Act1 probe in WT, TL2, and TL4. Genomic DNAs were restricted with the indicated enzymes as described in Methods. (F) Southern Blot analysis with an Aph" probe in WT, TL2, and TL4. Genomic DNAs were restricted with the indicated enzymes as described in Methods. (E, F). The used membrane was the same for both probes; it was stripped in between and rehybridized. selection marker Aph7'' to create a highly efficient vector. In C. reinhardtii, there are two b-Tubulin genes (Tub1 and Tub2; Youngblom et al. 1984), which are located in chromosome 12~0.8 Mbp apart (Cre12.g542250 and Cre12.g549550, respectively; Fig. 4a). In Chlamydomonas sp. SAG25.89, there are also two b-Tub genes, but they are separated only by 255 bp (Fig. 4a), making it challenging to select and distinguish the promoters and UTRs that do not overlap with the other gene, which could lead to problems with expression. The Act1 gene was present in both organisms as a single copy gene. The structure of the exons/introns varied slightly (Fig. 4b). To determine whether the Tubulin genes were expressed at different levels than Act1, we performed RT-qPCR. We found that Act1 is expressed at a slightly higher rate than Tub1 and Tub2 (Fig. 4c). For all reasons mentioned above, we focused on Act1.
Establishment of a transformation vector and protocol for Chlamydomonas sp. SAG25.89 stable nuclear transformation. We constructed the transformation vector pDCF4 (Fig. 5a) for amplification in E. coli and expression in Chlamydomonas sp. SAG25.89 (see sequence in Fig. S4 in the Supporting Information). In addition to E. coli specific features, it contains 666 bp of the Chlamydomonas sp. SAG25.89 Act1 promoter and 5'-UTR upstream of the Aph7'' coding sequence and 1,386 bp of the potential Act1 3'-UTR downstream, including all possible polyadenylation signals (Fig. 5a). The vector was transformed in three independent experiments either as a circular plasmid or linearized with the restriction enzyme AgeI, which cuts within the Kanamycin resistance gene for E. coli. Moreover, the vector was restricted with SfiI, which has two restriction sites that precisely flank the Aph7" cassette (Fig. 5a), and the resulting Aph7" cassette fragment was also transformed. For the transformation, an electroporation protocol using the NEPA 21 electroporator was established, which allows for the adjustment of numerous parameters necessary for optimizing the transformation of algal cells (Yamano et al. 2013). The transformation buffer must lack NaCl for the electroporation process to be successful. We took advantage of the possibility to keep Chlamydomonas sp. SAG25.89 without additional NaCl (Fig. 5b). First, cells were cultured in NH 4 -BBM + 25 g Á L NaCl -1 (see Methods). To ensure a working transformation, the cells had to be washed thoroughly in distilled water to remove all traces of salt; otherwise, it would have been impossible to achieve the correct impedance for electroporating. One poring pulse with positive voltage was required to open the pores, followed by five square wave transfer pulses to translocate the DNA within the cell (Table 1). It took approximately six to eight weeks for successfully transformed cells to form colonies on hygromycin NH 4 -BBM agar plates (Fig. 5c).
All transformations were performed at the same ratio of 10 ng DNA per 10 5 algal cells. The linearized vector achieved a significantly higher yield than circularized plasmid DNA (Table 3). The results with the linearized plasmid using either AgeI or SfiI were similar (Table 3).The transgenic lines were also grown successively on medium containing hygromycin and, thereafter, on medium lacking hygromycin and again on medium with hygromycin. The growth was consistent in all cases.
In addition to hygromycin resistance, the integration of the cassette into the genome of Chlamydomonas sp. SAG25.89 was verified through PCR using some of the transgenic lines with primers amplifying 113 bp of the 3' terminus of the Aph7'' coding sequence (Fig. 5a) after at least two rounds of re-streaking the colonies before use in colony PCR (Fig. 5d).
Moreover, Southern blots were performed to check for the integration of the cassette into the genome (Fig. 5). As a positive control, we used the Act1 gene (Fig. 5e). The genomic DNAs of wild type and the transgenic lines 2 and 4 were restricted with StyI, MseI, NcoI, and PstI and probed with labeled Act1 (see Methods). The calculated sizes of 1.64 kbp for StyI, 3.5 kbp for MseI, 4.5 kbp for PstI, and 5.2 kbp for NcoI were visible in all cases (Fig. 5e). In transgenic line 2, a second higher kbp band was also visible in case of PstI, which is likely due to partial restriction as it does not appear in wild type and in transgenic line 4. The size of this band is in accordance with the size (8.4 kbp) of a skipped PstI restriction site in the genome. Also, in case of MseI, a second higher and weaker band was visible in all cases in accordance with a skipped site. The genomic DNAs of wild type and the mentioned two transgenic lines were also used for labeling with the Aph7" probe (see Methods). All enzymes used (StyI, MseI, NcoI, and PstI) do not cut within the Aph7" or Act1 probe regions. As expected, no signals were visible in wild type with the Aph7" probe ( Fig. 5f, left panel). In transgenic line 2 (Fig. 5f, middle panel), two fragments hybridized in case of the NcoI digest, and more than two for StyI, where partial digests may have occurred. These data indicate that the Aph7" cassette was integrated more than once in the genome of transgenic line 2. In transgenic line 4 (Fig. 5f, right panel), a single integration event of the Aph7" cassette Linearized versions of replica 1 to 3 represent pDCF4 cut with AgeI; in the case of replica 4, pDCF4 was cut with SfiI. Details about the amount of DNA and used cell numbers are described in Methods. 66 seems to have taken place as one major band is visible in all cases. These data corroborate that the Aph7" cassette is being integrated into the genome by the established transformation method.
In summary, the vector pDCF4, along with the established transformation protocol, can now be used as a basis for the genetic manipulation of this marine green algal species. As outlined above, reporters used for Chlamydomonas reinhardtii should be easy to transfer to Chlamydomonas sp. SAG25.89 based on the high GC content (Table S2) and similar codon usage, at least in the closely related C. euryale (Fig. 3a). Thereby, the recently developed MoClo toolkit enabling synthetic biology in C. reinhardtii (Crozet et al. 2018) can serve as an efficient basis for Chlamydomonas sp. SAG25.89.
A long-term plan is to improve the genome sequences of Chlamydomonas sp. SAG25.89 by additional Oxford Nanopore sequencing, which yields larger fragments and, thus, ensures an efficient and more contiguous assembly. Nanopore sequencing should provide reads that extend beyond most of the genomic repeats, allowing an assembly with good contiguity and avoiding chimeras between chromosomes due to long stretches of repetitive and nearly identical sequences. Furthermore, once the contiguity and potential misassemblies have been resolved, the Illumina data can be used to polish it, resulting in a high-quality, high-contiguity genome. CONCLUSIONS We developed an efficient electroporation-based transformation protocol for the marine Chlamydomonas sp. SAG25.89, which can now serve as a basis for studies on biotic and abiotic interactions. The data show that the exponential growth of this marine protist stops at 32°C while it grows well in a range from 15°C to 28°C. Some components that confer thermal acclimation in C. reinhardtii are conserved in Chlamydomonas sp. SAG25.89, allowing comparative studies in the future. Due to the high GC content in both C. reinhardtii and Chlamydomonas sp. SAG25.89, the numerous established codonadapted GC-rich reporters for C. reinhardtii (Salom e and Merchant 2019) are easily transferable to the marine species. The alga will be also amenable to genetic approaches including silencing or gene editing. Our studies also revealed unexpected insights on the physiology (osmotolerance) and the genome of Chlamydomonas sp. SAG25.89 such as the likely loss of some genes of nitrogen metabolism or the translocation of the two Tub genes.
We thank Ivonne G€ orlich and Marco Groth from the Core Facility DNA Sequencing of the Leibniz Institute on Aging and the Fritz Lipmann Institute in Jena for their help with DNA sequencing; Elke Martina Jung for help with the light microscopy black-white picture; and Melvin Schubert and Anxhela Rredhi for helping in the initial set-up of the Southern Blot system. We thank Jan Petersen and Severin Sasso for helpful comments on the manuscript and James Brooks for proofreading. Our work was supported by a JSMC (Jena School for Microbial Communication) fellowship awarded to D.C.F., financed at first by the German Research Foundation (DFG), followed by the Carl Zeiss Foundation. M.Mi. was further supported by the DFG-financed CRC 1127-2 ChemBioSys -239748522. M.Ma. D.D., and M.F. were supported by the DFG "iDIV" and the Ministry of Thuringia "DigLeben" funds and M.H. by the DFG-funded CRC 1076 AquaDiva. In addition, M.Mi. and M.Ma. were supported by the DFG cluster of excellence "Balance of the Microverse". Open access funding enabled and organized by Projekt DEAL.

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web site: Figure S1. Full alignment of the 45S pre-rRNA cluster of Chlamydomonas sp. SAG25.89 with the 18S and 26S rRNA genes of Chlamydomonas sp. CCMP235 (A) and differences of the 45S pre-rRNA cluster compared to C. euryale (B).    Table S1. Parameters used to run CENTRIMO and DREME. Appendix S1. Multiple sequence alignment of the 18S rDNA (FASTA file; Fig. 1b).
Appendix S2. In-house script for processing the transcriptome and selecting mRNAs containing the poly(A) tail. GENETIC TOOLS IN CHLAMYDOMONAS SP.