Cytochrome c oxidase subunit I gene in Thalassiosira nordenskioeldii strains inhabiting in cold and warm sea waters

Abstract

From the biota beneath the sea ice in Lake Saroma, which is adjacent to Sea of Okhotsk, a diatom culture of Saroma 16 was isolated. Strutted processes and a labiate process in Saroma 16 were characteristic of those in Thalassiosira nordenskioeldii. Similarity search analysis showed that the 826-bp rbcL-3P region sequence of this strain was 100% identical to multiple sequences registered as T. nordenskioeldii in a public database. The 4305-bp PCR-amplified mitochondrial cytochrome c oxidase subunit I (COI) gene (COI)-5P region of Saroma 16 included a 1060-bp open reading frame (ORF), which was interrupted by 934-bp and 2311-bp introns that included frame-shifted ORFs encoding reverse-transcriptase (RTase)-like proteins. Previous reports showed that a strain of the same species, CNS00052, originating from the East China Sea included no introns in the COI, whereas North Atlantic Ocean strains of the same species, such as CCMP992, CCMP993, and CCMP997, included a 2.3-kb intron in the same position as Saroma 16.

1. Introduction

Marine diatoms are the most important food source for aquatic animals. Some diatoms survive winter clinging under sea ice, where the water temperature is almost at freezing point, and only a limited level of air is available. The marine diatom Thalassiosira nordenskioeldii is one such species. This algae is observed in cold sea water regions^1–4) or warm water region in colder seasons.^5–10) T. nordenskioeldii which originated from the southern North Sea was reported to grow at −3 ℃.¹⁾ A strain of the same species that was isolated from Lake Saroma facing the Sea of Okhotsk, grew at −1.8 ℃ but did not above 16.8 ℃.⁵⁾ In the present study, we isolated cultures of diatoms from ice biota from Lake Saroma and determined a partial nucleotide sequence of the cytochrome c oxidase subunit I (COI) gene (COI) from the mitochondrial genome in a strain of T. nordenskioeldii, and compared this with the obtained sequences from different strains of the same species, which originated from cold and warm sea water around the world.

2. Materials and methods

2.1. Alga.

Depths of sea ice and water temperatures were checked in every 250 m from 500 m to 1750 m west-northwest from the lake edge (44.1196°N, 143.969°E) in front of Aquaculture Fishery Cooperative of Lake Saroma (Supplementary Fig. 1), and sea ice blocks of 20 × 20 cm in width were excised and collected at spots 985 m or 1240 m in the same direction from the lake edge in February 28 to March 1, 2018. Sea water under the collected ice was also recovered and transferred to the laboratory. Temperatures were measured using a conductivity temperature depth profiler, AAQ-125-CAD (JFE Advantech Co., Nishinomiya, Japan). Sea water temperatures at 985 m and 1240 m were −1.156 ℃ and −1.136 ℃, respectively, which were calculated from data recorded at 1000 m and 1250 m. Ice was thawed with two volumes of filtered sea water (< 2 ℃). Single colonies were collected using a Pasteur pipet and inoculated into Gillard’s f/10 marine enriched seawater to establish a culture, Saroma 16. The strain was cultured in a 12 h light (2.7 µmol m⁻² s⁻¹): 12 h dark photoperiod at 4 ℃.

2.2. Microscopy.

After cultured cells were observed using a light microscope, some cells were treated with alkaline detergent, washed with distilled water, embedded in Mount Media (Fujifilm Wako, Osaka) and observed using a light microscope again. Bright-field images were photographed using a CX23 Biological Microscope (Olympus, Tokyo) equipped with digital camera EOS M10 (Canon, Tokyo). For scanning electron microscopy, specimens after cleaning were coated with gold particles in an Ion Coater IB-3 (Eiko, Tokyo), with an ionization current of 5 mA for 3 min. Scanning electron microscopic images were taken using a Tabletop Microscope TM3000 (Hitachi, Tokyo).

2.3. PCR, nucleotide sequencing and alignment of sequences.

Genomic DNA was isolated from Saroma 16 in accordance with a standard phenol:chloroform extraction protocol.¹¹⁾ In order to amplify the rbcL-3P region from this strain, PCR amplification was performed using isolated genomic DNA as a template and forward (rbcL3P-CfD: 5′-CCR TTY ATG CGT TGG AGA GA-3′) and reverse (rbcL3P-DPrbcL7: 5′-AAR CAA CCT TGT GTA AGT CT-3′) primers.¹²⁾ An accession number of LC781032 was assigned to the rbcL sequence. In order to amplify the COI-5P region from this strain, PCR amplification was performed using Saroma 16 genomic DNA as a template, forward (52F: 5′-TGA CTT TTT TCA ACA AAT CAC AAA GAC AT-3′) and reverse (1167R: 5′-ATG TGC AAC CAC ATA ATA AGT ATC GTG-3′) primers and Blend Taq polymerase (Toyobo, Tokyo). Primers were designed in reference to degenerate primers pC1 and pB1.¹³⁾ The 3′ ends of 52F and 1167R anneal to the same genome positions as those of pC1 and pB1, respectively. PCR cycles were as follows; one cycle of 94 ℃ for 2 minutes, 5 cycles of 94 ℃ for 30 seconds, 45 ℃ for 30 seconds and 72 ℃ for 90 seconds, 35 cycles of 94 ℃ for 30 seconds, 46.5 ℃ for 90 seconds and 72 ℃ for 90 seconds, and one cycle of 72 ℃ for 10 minutes. A 4.3-kb product was recovered. Amplified DNA band was recovered from agarose gels using Wizard SV Gel and PCR Clean-Up System (Promega, Madison) and was subject to direct nucleotide sequencing using either forward or reverse primers. After trimming away end sequences with low phred quality scores, adjacent sequences were connected to generate a 4305-bp consensus sequence. An accession number of LC762311 was assigned to the resultant sequence. In reference to a previous report,¹³⁾ related amino acid sequences encoded in intronic ORFs were retrieved from the database and aligned with the sequences of Saroma 16 using MAFFT version 7.511 online. After a default alignment, part of alignment was adjusted manually to find motifs for RTase and maturase. Phylogenetic trees were constructed using aligned sequences with the neighbor joining method with bootstrap values using phylo.io.1.0.0. Considering the synonymous nucleotide sequence substitutions in COI ORFs in related strains, a phylogenetic tree using nucleotide sequences was constructed instead of using amino acid sequences.

3. Results and discussion

Figures 1A–D show light microscopic images of Saroma 16. Several chloroplasts were observed within the cell (Fig. 1A). Organic threads sprouted radially from the frustule shoulder. Figure 1B shows a girdle view. The cell was octagonal and a distinctive depression was observed over the central regions of frustules. An organic connecting thread extended to an adjacent cell, and the length of this thread was almost in the same as in that of the pervalvar axis. Figures 1C and D show a valve view of a cleaned cell. Strutted processes in the valve shoulder were arrayed in a ring, on which a single labiate process was also found. The base of the labiate process was distinguished from those of strutted processes in the ring (Figs. 1C and D). Figures 1E–H show SEM images of Saroma 16. Figure 1E shows an external view of the frustule. Consistent with the depression observed in Fig. 1B, an electron-dense region was observed in the central part of the frustule (Fig. 1E). The central strutted process was adjacent to a central areola. Corresponding to the distinctive base projection of a marginal process, which was observed using light microscopy (Figs. 1C and D), a single labiate process was observed in a somewhat upright direction, whereas strutted processes projected in rather horizontal directions (Fig. 1E). The inside view of the labiate process was distinctive from those of strutted processes (Fig. 1F). As shown at a higher magnification (Fig. 1G), the central strutted process projected upright. The tips of the labiate process possessed openings (Figs. 1G and H). These observations were consistent with the morphological features of T. nordenskioeldii,¹⁴⁾ and distinct from those of related species (Supplementary Fig. 2).

Figure 1.

Light- and scanning electron-microscopic images of Saroma 16. (A) Light microscopic image of the frustule view before cleaning of the cell with an alkaline detergent. Asterisks and cp indicate the organic threads and chloroplast, respectively. (B) Girdle view of the cell before cleaning. An asterisk and cp are as in panel A, and cvs indicate concavities in the valves center. (C and D) Two images of a frustule view for a cleaned cell that was photographed in the same field with different foci. A smaller black arrow indicates a central areola (annulus), and black arrowheads and a larger black arrow indicate marginal strutted and a labiate processes, respectively. Panels A to D use the same magnifications and a scale bar (5 µm) is representatively shown in D. (E) SEM image of the frustule outside in cleaned cell. A dotted white arrow indicates a central areola. White arrowheads and smaller and larger white solid arrows are as the black ones in panels C and D. (F) Frustule inside view. An asterisk shows a girdle, and a larger solid arrow is as in panel E. (G and H) Enlarged frustules outside and inside, respectively. Bars in E, G and H represent 2 µm. Panels E and F are in the same magnifications.

Blastn analysis of the nucleotide sequence of the 826-bp rbcL-3P region of Saroma 16 detected several 100%-identical sequences registered as T. nordenskioeldii. This result was consistent with the cytological analysis above. Using Saroma 16 genomic DNA as a template and primers 52F and 1167R, further PCR amplification was performed. A 4.3-kb band was recovered, and direct sequencing of this fragment was carried out. Forward and reverse strand sequences were connected to generate a 4305-bp sequence. Alignment of this sequence with that of CCMP992 cDNA (AB020229.1)¹⁵⁾ revealed that the Saroma 16 sequence included a 1060-bp of COI ORF region (Fig. 2) and 934-bp (Supplementary Fig. 3) and 2311-bp (Fig. 3) introns, which were inserted after the first bases of the valine codon at the 70th residue from the N-terminal end (V₇₀) and the alanine codon at the 289th residue from the N-terminal end (A₂₈₉), respectively. Though a COI sequence in a Lake Saroma-derived T. nordenskioeldii, was reported previously,⁶⁾ this sequence was 100% identical to, and shorter than, that obtained in this study, and the culture analyzed previously was no longer available. The Saroma 16-COI ORFs sequence presented in this study possessed nine synonymous nucleotide substitutions in eight codons as compared with those of the CCMP992 sequence.¹³⁾

Figure 2.

Location of group II introns insertion shown over the aligned COI amino acid sequences of various organisms including Thalassiosira nordenskioeldii. Organisms, introns and GenBank accession numbers are as follows. Marchantia paleacea (Ma.cox1-i-2, M68929.1), Podospora anserina (Po.cox1-i-1 and Po.cox1-i-4, X55026.1), Allomyces macrogynus (Al.cox1-i-3, U41288.1), Saccharomyces cerevisiae (Sa.cox1-i-1 and Sa.cox1-i-2, V00694.1), Kluyveromyces lactis (Kl.cox1-i-1, X57546.1), Pylaiella littoralis (Py.cox1-i-1, Py.cox1-i-2 and Py.cox1-i-3, Z72500.1), Neurospora crassa (Ne.cox1-i-1, X14669.1), Diacronema lutheri (Di.cox1-i-1, AF045691.1), T. nordenskioeldii CCMP992 (Th_1.cox1-i-1, AB038235.1) and T. nordenskioeldii Saroma16 (Th_2.cox1-i-1 and Th_2.cox1-i-2, LC762311, this study). Note that the second and third introns of P. anserina are not shown. T. nordenskioeldii amino acid residue positions are shown on the left side with reference to CNS00052. Abbreviated amino acid sequences are shown within parentheses. Asterisks show positions of the intron-insertion in the corresponding nucleotide sequences. An asterisk in the three-letter symbol, for example, “A*sp”, indicates that after the first base of aspartate codon, an intron is inserted, whereas one letter symbols, for example “D”, indicate that no intron is inserted in the corresponding codon.

Figure 3.

Alignment of COI intronic amino acid sequences of various organisms. Conserved domains of reverse transcriptase (RT-I to VII) are underlined, and that of maturase are boxed. The Saroma 16 sequence includes two frameshifts (+1) and (−1), whereas the CCMP922 sequence possesses only one frameshift (−1). Positions of frameshifts are shown with vertical closed arrows. Putative initiation codon for Thalassiosira nordenskioeldii CCMP992 is shown with an open arrow adjacent to “(ORF start2)”. Abbreviations for intronic amino acid sequences listed on the left of the alignment are as follows; the first intron of Saccharomyces cerevisiae (Sa.1), the first intron of Diacronema lutheri (Di.1), the second intron of Pylaiella littoralis (Py.2), the first intron of T. nordenskioeldii CCMP1096 (Th.1a), the first intron of T. nordenskioeldii CCMP992 (Th.1b) and the second intron of T. nordenskioeldii Saroma 16 (Th.2c). Tryptophan is encoded by either TGG or TGA codons (“transl_table=4” genetic code). Note that TGA codon is assigned to termination codon in the standard genetic code. Numbers in parentheses and at the right end show abbreviated amino acid sequences and amino acid positions from the N-terminal end, respectively.

The Saroma 16 intron at A₂₈₉ (Th_2.cox1-i-2) contained a ORF for a 686-residue amino acid sequence for RTase, which was connected with a maturase. As in CCMP992,¹³⁾ the Saroma 16 RTase coding region included a (−1) frameshift at the third base for codon L₄₃₅ (Fig. 3). Notably, the Saroma 16 RTase coding sequence possessed an additional (+1) frameshift as an A-insertion in front of the I₄₁₃ codon (Fig. 3). These findings suggested that the CCMP992 intronic enzyme may consist of 449 residues of the amino acid sequence including only motifs of RT-I to RT-IV, whereas that of Saroma 16 may possess 686 residues including all the motifs of RTI to RTVII and maturase, in which a 23-residue sequence was substituted in the region between RT-IV and RT-V. These results suggested that RTase in the CCMP992 intron (Th_1.cox1-i-1) might be null function, whereas that of Saroma 16 (Th_2.cox1-i-2) might retain a certain level of function when the H-N-H motif is complemented. Phylogenetic analyses of RTase, maturase, and COI suggested very close relationships between Saroma 16 and CCMP992 (Fig. 4), compared with the other sequences used here.

Figure 4.

Phylogenetic trees based on amino acid sequences of reverse transcriptase (A) and maturase (B) and nucleotide sequences of COI (C). Amino acid sequences of reverse transcriptase range from N-terminus of RT-I domain to C-terminus of RT-VII domain shown in Fig. 3, those of maturases surrounded by a box in Fig. 3, and nucleotide sequences of COI corresponding to amino acid sequences shown in Fig. 3 were aligned using MAFFT, and phylogenetic trees were obtained using the neighbor joining method and shown using phylo.io., with default parameters. Organisms, introns and accession numbers are as in Fig. 2. An intron of Thalassiosira nordenskioeldii CCMP1096 is referred to as cox1-i-1 (AB037974). Numbers at the bases of branches are bootstrap values above 50.

The 934-bp Saroma 16 intron (Th_2.cox1-i-1) after the first base of the V₇₀ codon (Supplementary Fig. 3) was in the same position as Marchantia paleacea cox1-i-2 and Saccharomyces cerevisiae cox1-i-1 (Fig. 2). Existence of an intron after V₇₀ in various organisms suggests that this intron insertion position is very stable across organisms. We performed Blastx analysis using the 934-bp Saroma 16-COI intron sequence (Th_2.cox1-i-1). Frames 1 and 3 of the predicted amino acid sequences encoded on this intron hit two regions of Psammoneis japonicum RTase (PjRTase) sequence (YP_009495510.1) ranging from 84th to 103th residues and from 5th to 57th residues, respectively (Supplementary Fig. 3). Manually aligned sequences of Saroma 16-frame 2 and PjRTase also shared some level of identity in a span between above-mentioned two regions. These results suggested that the 934-bp Th_2.cox1-i-1 was present in a common ancestor of T. nordenskioeldii and was lost during evolution giving rise to the Atlantic strains.

As shown in Fig. 5, invasion of intron(s) into COI is observed only in cold water culturable T. nordenskioeldii strains from the Atlantic rim (CCMP992, CCMP993, CCMP997 and CCMP1096)¹⁶⁾ and Pacific rim (Saroma 16), whereas no COI intron was detected in a warm water culturable T. nordenskioeldii strain of East China Sea (CNS00052).¹⁷⁾ Generation of introns in COI may affect the level of translation of COI, which is a rate-limiting enzyme for respiration.¹⁸⁾ The COI intron in cold water culturable strains might be explained by global ocean circulation, which is referred to as thermohaline circulation.¹⁹⁾ Deep sea water on the earth is estimated to be replaced every 200 to 1800 years.²⁰⁾ In the middle of the Pacific Ocean, there is a water stream-crosspoint in which surface sea water proceeds west and deep sea water moves up north. One speculations is as follows. If the T. nordenskioeldii strains discussed in the present study was derived from a common ancestor originating from this point, this ancestor, which would have no COI intron, flowed from the East China Sea to the Sea of Okhotsk in a cold and salty deep current, influenced by abyssal circulation from middle of the Pacific Ocean to the Chishima Islands,²¹⁾ to obtain two introns, and some population of the descendant independently went west to the Atlantic Ocean in a warm shallow current,¹⁹⁾ moving north to the Arctic Sea past Norway, to lose one intron, and made an anti-clockwise turn to the North American seashore into a cold and salty deep current.¹⁹⁾ The intron gains and losses might be associated with cold adaptation. Because we cannot exclude the possibility that the distribution of cold water-culturable strains and the presence of the intron might be coincidentally observed only in strains mentioned in the present study, further accumulation of COI sequences in related strains will be important in the future. More physiological analyses of various strains in this species, including CCMP993, CCMP997, or CCMP1096 in the future might also provide important information for elucidation of the above-mentioned hypothesis. Because cold adaptations in cultures of T. pseudonana have been reported so far,^22,23) investigation of several species in the genus Thalassiosira should be widely conducted.

Figure 5.

COI ORF sequences of Thalassiosira nordenskioeldii from various regions around the world. Cold water culturable strains CCMP992, CCMP993, CCMP997 and CCMP1096 originated at 42.8833° N 69.6833° W, 43° N 69° W, 69.6667° N 18.9667° E and 76.25° N 82.55° W, respectively (https://ncma.bigelow.org/, accessed October 27, 2023). Longitude of Tromso, collection site for CCMP997, was corrected from the original datum represented in the URL. According to Andersen et al. (1997),¹⁶⁾ the locality of CCMP992 was station 834A, 42°52′ N 69°41′ W. COI sequences for CCMP992, CCMP993, CCMP997, CCMP1096 and Saroma 16 were interrupted with an approximately 2.3-kb intron after the first base of the alanine codon 289, whereas Saroma 16 also included an additional 934-bp intron after the first base of the valine codon 70. In contrast to the above-mentioned strains, a warm water strain CNS00052, which was isolated at 31.95° N 125.05° E, did not contain an intron in the COI sequence. Filled red circles indicate origins of strains and horizontal lines illustrate the open reading frames encoding the COI amino acid sequence. Positions of amino acid residues are in reference to CNS00052. The locality of this strain was reported previously.⁷⁾ Although the nucleotide sequences of CCMP992 and Saroma 16, which were mentioned in Fig. 2, were deposited in DDBJ with accession numbers AB038235 and LC762311, respectively, those of CCMP993 and CCMP997 were not reported so far. The sizes of introns in CCMP993 and CCMP997 were estimated from PCR products reported previously. A T. nordenskioeldii strain reported previously,⁵⁾ was also referred to as Saroma 16 because of the same locality. Description of latitude and longtitude is quoted from references in the table.

Non-standard abbreviation list

COI

cytochrome c oxidase subunit I

COI

COI gene

ORF

open reading frame

PCR

polymerase chain reaction

RTase

reverse transcriptase

SEM

scanning electron microscope