NUCLEAR MONOPLOIDY AND ASEXUAL PROPAGATION OF NANNOCHLOROPSIS OCEANICA (EUSTIGMATOPHYCEAE) AS REVEALED BY ITS GENOME SEQUENCE¹

N. oceanica and its culturing. The original Nannochloropsis sp. was provided by the Key Laboratory of Mariculture of Chinese Ministry of Education affiliated with Ocean University of China. Seawater near Qingdao shore was filtrated with a membrane with pores of 0.4 μm in diameter, autoclaved at 121°C for 30 min, and used for preparing f/2 medium (pH 7.8, salinity 30) (Guillard and Ryther 1962, Guillard 1975). The solid medium was prepared by melting 1.2% (w/w) agar into f/2. The alga was cultured at 23 ± 1°C and a continuous irradiation of 70 μmol photons · s⁻¹· m⁻² at a rhythm of 12:12 light:dark (L:D) dark. The alga was purified twice by streaking the cells (1 × 10³ cells · mL^{− 1} in liquid medium) onto the solid plate medium containing 100 μg · mL⁻¹ ampicillin. When algal colonies appeared in 13 d, a single colony was inoculated into 5 mL of liquid f/2 medium containing 200 μg · mL⁻¹ ampicillin. The culture was then amplified into 100 mL and verified as being bacteria free by evenly inoculating the amplified cells (1 × 10³ cells · mL^{− 1} adjusted with autoclaved liquid medium) onto solid medium without antibiotics. The purified line was identified as N. oceanica with the phylogenetic analysis of its 18S rDNA, partial rbcL gene, and internal transcribed spacer (ITS) region sequences (HQ201712–HQ201714).

DNA extraction. Algal cells were collected at the exponential growth phase with centrifugation (3k30, Sigma, Osterode, Germany). The cell pellet was washed three times with polysaccharide elimination buffer (Wang et al. 2005). DNA was extracted with cetyl trimethylammonium bromide (CTAB) method described previously (Chen et al. 2001). Conventional phenol extraction was applied to purify the DNA with the quality of DNA checked by digesting with Hind III, EcoR V, and EcoR I and separating in 0.8% agarose gel.

Library construction and sequencing. The DNA libraries were constructed with the Paired-ends Library Preparation Kit (Illumina Inc., San Diego, CA, USA) following manufacturer’s instructions. Two libraries with insert sizes of 500 bp and 2 kb, respectively, were constructed. Methods for DNA manipulation, including formation of single-molecule arrays, cluster growth, and paired-end sequencing, were performed following the standard protocols available from Illumina Inc. and on Illumina GA sequencer as were described by Boodhun et al. (2008) and widely practiced at Beijing Genomics Institute (BGI) at Shenzhen. The base-calling pipeline (SolexaPipeline-v1.7) (Sequencing Control Software, Illumina Inc.) was used to process the raw fluorescent images and call sequences.

De novo assembling. Those paired-end reads with low quality were filtered ahead of assembling. The filtered reads also included those with low complexity, which contained either a preset proportion of nondetermined or low-quality bases or a preset length of overlap between adapter and reads, the duplicated and the contaminated. The high-quality reads were assembled into scaffolds with SOAPdenovo (Li et al. 2010). To remove the possible flaws, the quality reads were assigned back to the assembly with BWA software (http://bio-bwa.sourceforge.net/index.shtml) with the suspected boxes (<the length of a read) replaced with the sequence of the reads accounting for ≥80% of the total (>20 in number).

Polymorphic nucleotides searching. Polymorphic nucleotides were searched by aligning the qualified reads (with a quality score of >20) to the assembly with SOAP2 (http://soap.genomics.org.cn/soapaligner.html). Once all reads were assigned and the nucleotides leveraging across reads and the gaps in the assembly were removed, the numbers of the most abundant or major (n₁) and the second most abundant or minor (n₂) nucleotides at each polymorphic site were counted according to the number of reads in which they were located. Of the sites covered by >10 nucleotides (n₁+ n₂, n₂>2), the heterozygous ones covered by either >10 n₂ or 3–9 n₂ but within the expectation of 1n₁:1n₂ (P > 0.01) were considered as polymorphic.

A further correction of the sequence assembly was made at this step by replacing a minor nucleotide in the assembly with the major one when the number of reads at which the major nucleotide located was >10.

Gene prediction and function annotation. Based on the final assembly, messenger RNAs were predicted using hidden Markov model (HMM)–based GeneMark-ES version 2.3a (Ter-Hovhannisyan et al. 2008) with those encoding peptides shorter than 30 amino acid residues in length discarded. The yielded mRNAs were annotated by aligning with the known functional genes in the protein knowledgebase SwissProt and its computer annotated supplement TrEMBL (http://www.uniprot.org/downloads, release 56.1), GO (http://www.geneontology.org/GO.downloads.database.shtml), KEGG (ftp://ftp.genome.jp/pub/kegg/, release 46), and National Center for Biotechnology Information (NCBI)–HR (ftp://ftp.ncbi.nih.gov/blast/db/, 20100313) databases under the criteria of e-value ≤1 e⁻⁵ and identity ≥40. The rRNA and tRNA were identified by referring to rRNA (Pruesse et al. 2007) and tRNA (http://gtrnadb.ucsc.edu/) databases with rRNAmmer (Lagesen et al. 2007) and tRNAscan-SE, respectively (Lowe and Eddy 1997). Other noncoding RNAs, including miRNA, sRNA, and snRNA, were identified by referring to Rfam database (Griffiths-Jones et al. 2005) with Infernal software (http://infernal.janelia.org/) with default parameters. Transposons were determined by either referring to Repbase transposable elements library (Jurka et al. 2005) or aligning the genome sequence to the curated transposable element related proteins with RepeatMasker and RepeatProteinMasker (http://www.RepeatMasker.org/), respectively. Tandem repeats were predicted using TRF (tandem repeat finder) software (Benson 1999).

To remove the polymorphic nucleotides within the repeated region, the predicted repeats were searched against the assembly. Once a polymorphic nucleotide was identified within a match (identity >95% and length >50 bp), it was deleted from the polymorphic nucleotides list.

Interesting proteins searching. The amino acid sequences of the interesting proteins well defined previously were downloaded and used as the queries searching the predicted proteins. Among the matched (e ≤ 1 e⁻⁵, identity≥40), the best match was considered as the ortholog of a queried protein.

Genome size. Paired-end sequencing reads (73 bp at one end and 75 bp at the other) of the fragments of two libraries (0.5 kb and 2 kb fragments, respectively) generated >4 Gb raw sequence in total. From each quality, 73 bp reads of 0.5 kb library, 17mers were framed out with the number (depth) and the abundance (proportion to the total) of each kind calculated. The abundance of 17mers in different depths should be in a Poisson distribution; however, such a distribution may change due to diverse causes, such as sequencing errors, genomic heterozygosity, and chromosomal fragment repeat. It was found that the abundance of 17mers with depths ranging from seven to 67 was in Poisson distribution, and the depth of the most abundant 17mers was ∼37. Accordingly, the genome size was estimated to be ∼30 Mb. The raw sequence covered 139 times of the estimated genome of N. oceanica. The raw reads were assembled into 886 scaffolds with a total length of 29.5 Mb and a scaffold N₅₀ value of 294.5 kb (Table 1). Raw sequences have been deposited into the NCBI Sequence Read Archive with accession number SRA027386.

Polymorphic nucleotides. In total, 1,360 polymorphic nucleotides were identified in the genome of N. oceanica (diverse repeats not included). The density of polymorphic nucleotide was one in ∼22.06 kb if they were evenly distributed, obviously lower than that in other genomes. In the scattering plot (Fig. 1, A and B), it was found that the ratios of the major to minor (≥10) nucleotides each polymorphic locus obviously deviated from 1:1, the ratio expected in diploid organisms. A clue could be drawn here that N. oceanica is a monoploid alga, namely, its rareness of polymorphic nucleotides and inequality of major and minor nucleotides at each polymorphic locus (n₂ ≥ 10). The polymorphic nucleotides detected may originate from natural mutations and clonal expansion in continuous culture.

Functional genes and metabolic network. From the assembly, 11,129 mRNAs (genes) were predicted. The length of predicted genes was 1,345 bp on average. The protein-encoding regions accounted for ∼50.69% of the total scaffold length. The assembly and the predicted genes have been deposited at DNA Data Bank of Japan (DDBJ)/European Molecular Biology Laboratory (EMBL)/GenBank under the accession AEUM00000000. The version described in this article is the first version AEUM01000000. In total, 6,639 predicted genes were annotated by searching the proteins in diverse databases, accounting for 59.65% of the total (Table 2). Of the annotated 3,303 genes (29.68% of the total) matched with those deposited in KEGG database, each was hit by at least one function (Table 3) defined in diverse metabolic pathways like glycolysis, fatty acid biosynthesis, and photosynthesis. The annotated included also six miRNAs, three snRNAs, and 86 tRNAs; and one 18S, one 28S, and seven 8S RNAs.

It was unexpected that a certain portion of predicted genes were assigned functions involved in human diseases including cancers and immune, infectious, and metabolic disorders (Table 3). It was obvious that these assignments were not reasonable; these functions absolutely are not the characteristics of N. oceanica, a single-cell microalga. It was also noted that these genes were assigned multiple functions.

A protein should function in a species in which its function is defined; however, it may have an ancient origin in evolution and function differentially, across species. We found that those proteins of N. oceanica assigned functions involved in human diseases also function in genetic information processing and metabolism and cellular processes in most cases, indicating that their function can be traced back to the early stage of evolution. Domazet-Lošo and Tautz (2010) found two strong booming time points of cancer proteins, one at the origin of cellular structure and the other around the formation of multicellular organisms. These proteins are involved in cancer formation in humans but perform more general functions of maintaining genome stability (caretakers) and cellular signaling and growth processes (gatekeepers) in evolutionarily primary organisms. We searched the predicted proteins of N. oceanica using well-defined cancer genes (Futreal et al. 2004) as queries and found 45 orthologs (P < 3.00 e⁻²⁵). The alignments of these proteins with those of mammals showed the existence of homologous domain(s), although the sequences themselves are diverse in length. The variation in sequence may have allowed the proteins to function differentially among species. Rather than functioning in cancers, these genes should function as the maintainers of cellular stability and processors of genetic information in N. oceanica. The biological characteristics of a species should be integrated into the annotation and function assignment of its genes.

Asexual reproduction. The nuclear ploidy and reproductive strategy of N. oceanica remained as puzzles in the past. It is critical to the breeding of this alga to know its reproductive strategy; mutation and elite clone expansion could be very effective ways of breeding if it is asexual. Meiosis is absolutely required for sexual reproduction. Nearly all sexual eukaryotes undergo meiosis with which some unique characteristics are associated (e.g., recombination, synapsis). As proposed by Schurko and Logsdon (2008), the presence of multiple genes required specifically for meiosis in a genome, in particular, those for recombination, formation of synaptonemal complexes, and chromosome cohesion, is a positive indication that the organism is capable of meiosis and, implicitly, sexual reproduction. In contrast, the absence of the majority, or all, of these genes would be consistent with asexuality. To verify the lack of meiosis in this alga, the meiosis-specific proteins well defined by Ramesh et al. (2005) and Malik et al. (2008), including Spo11, Hop1, Mnd1, Dmc1, Msh4, Msh5, Mer3, and Rec8 identified in diverse eukaryotes, were downloaded and used as queries to search the predicted proteins of N. oceanica. It was determined that these proteins tightly matched only two predicted proteins, Noc_GME3248 and Noc_GME7986 (0 < P < 7.00 e⁻²⁹), the latter involved in DNA mismatch repairing. We believe that N. oceanica is not capable of performing meiosis using near-universal meiotic machinery and accordingly reproducing asexually.

Biosynthesis of long-chain PUFAs. It is known that N. oceanica synthesizes rich EPA. In total, 280 lipid metabolism functions hit the predicted proteins; however, the pathway of long-chain PUFA synthesis was not clearly revealed. To elucidate the mechanism of PUFA synthesis, we downloaded the protein sequences of diverse desaturases and elongase involved in the biosynthesis of long-chain PUFA, which were well identified previously (Hashimoto et al. 2008) and used as queries to align with the predicted proteins. Three desaturase-encoding genes (Noc_GME1570, Noc_GME8161, and Noc_GME7482, P < 2.00 e⁻⁵⁷) and one elongase-encoding gene (Noc_GME10858, P = 3.00 e⁻⁷²) were identified. The matched genes included also a few other genes; however, their P values were >e⁻²⁰, and accordingly, they should not be involved in long-chain PUFA. According to Hashimoto et al. (2008), the genes we identified should involve in the biosynthesis of fatty acids with a carbon chain length >18C. The synthesis of long-chain PUFA by N. oceanica is certain; the genes of two types of key enzyme encoding genes existed in its genome.