The dinoflagellates are an interesting model for eukaryotic evolutionary studies, due to their extraordinary genomic characteristics. Dinoflagellate chromosomes remain permanently condensed during the entire cell life cycle, their nuclear membranes remain intact during mitosis, and they lack nucleosomes and typical histones(Dodge 1966, Hackett et al. 2004, Moreno Diaz de la Espina et al. 2005, Lin et al. 2010). Moreover, dinoflagellates contain modified nuclear DNA; for example, 5-hydroxymethyluraci replaces 12-70% of the nuclear DNA’s thymine, while 5-methylcytosine replaces some cytosine(Lin 2011). Dinoflagellates possess a sizable quantity of DNA, ranging from 1.5 to 225 pg per cell (LaJeunesse et al.2005). In addition, dinoflagellates’ gene regulation mechanisms,such as alternative splicing and post-transcriptional regulation, differ substantially from those of typical eukaryotes (Brunelle and Van Dolah 2011, Zhang et al.2011). In particular, studies have shown spliced leader(SL) trans-splicing to be a common mRNA processing mechanism in the dinoflagellate nuclear genes (Lidie and Van Dolah 2007, Zhang and Lin 2008, 2009, Zhang et al. 2009, 2011, Lin et al. 2010), whereas most eukaryotic mRNA editing employs cis-splicing in processing. In general, this mRNA processing using SL trans-splicing appends a short RNA fragment, such as a SL RNA, to the 5′-untranslated region (UTR) of transcribed pre-mRNA(Pouchkina-Stantcheva and Tunnacliffe 2005, Zhang et al. 2007). Researchers have identified SL trans-splicing in other eukaryotic organisms, including trypanosomes,euglena, nematodes, platyhelminthes, rotifers, a tunicate (Ciona intestinalis), and so on (Murphy et al. 1986, Krause and Hirsh 1987, Tessier et al. 1991, Davis et al. 1994, Pouchkina-Stantcheva and Tunnacliffe 2005).

Recently, Lin and colleagues (Zhang et al. 2007, Zhang and Lin 2009) have studied dinoflagellate SL (dinoSL) trans-splicing extensively, and identified a short, conserved dinoSL sequence, 5′-DCC GTA GCC ATT TTG GCT CAA G-3′ (D = U, A, or G), comparing 5′-UTR sequences from cDNA libraries (Zhang et al. 2007). They pointed out that dinoflagellate nuclear encoded transcripts mostly have dinoSL sequences at the 5′-UTR end (Zhang et al. 2007, 2009). With this distinct characteristic, the authors can isolate dinoflagellate genes from environmental samples by using the dinoSL sequence as a marker, or dinoflagellate-specific primer, on the SL (Zhang and Lin 2008). However, Bachvaroff and Place (2008), after determining 47 genes of dinoflagellate Amphidinium carterae, examined those having cDNAs for dinoSL sequences and detected this dinoSL in only about two-thirds of all examined transcripts (i.e., approximately one-third failed to show trans-splicing). Following this study, Zhang and Lin (2009) re-investigated the genes lacking dinoSL, which Bachvaroff and Place (2008) had suggested might be “not trans-spliced,” successfully detected the dinoSL at the 5′-ends of their transcripts, and reinstated the postulate that dinoSL is widespread among dinoflagellate nuclear-encoded transcripts. Taking these previous findings into consideration, the presence of dinoSL in the majority of dinoflagellate transcripts remains controversial. To determine the expressed sequence tag (EST) sequencings of other dinoflagellates and other strains from different geographical regions requires further studies.

In the present study, we investigated the dinoSL sequence using our EST databases that comprised a naked dinoflagellate, Cochlodinium polykrikoides, and an armored dinoflagellate, Prorocentrum minimum. To determine these EST sequences, we employed 454 sequencing (a pyrosequencing system of 454 Life Sciences, Roche, Branford, CT, USA). Researchers consider C. polykrikoides and P. minimum to be harmful algal bloom species. In particular, P. minimum can produce the potent diarrhetic shellfish poisoning, which is one of the major types of illness that result from harmful algal blooms.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium,provided the original work is properly cited.

We obtained the two dinoflagellate strains, C. polykri-koides (CP-1) and P. minimum (D-127), from the National Fisheries Research and Development Institute (NFRDI) and the Korea Marine Microalgae Culture Center (KMMCC), respectively. Cultures of both strains were grown in f/2 medium, at 20°C, following a 12 : 12 h light : dark cycle. We harvested the cells at various growth phases, using exponential growth phase cultures for various stress treatments, including heat shock, cold shock, exposure to metals, and UV. The Cochlodinium and Prorocentrum cells were then harvested via centrifugation at 3,000 rpm for 10 min. We immediately diluted all harvested cells with ten volumes of TRIzol (Invitrogen, Carlsbad, CA, USA), froze them in liquid nitrogen, and stored them at -80°C until we extracted their RNA.

To isolate the total RNA from these harvested cells, we used the TRIzol method (Invitrogen), according to the manufacturer’s instructions. After physically breaking the cells via freeze-thawing in liquid nitrogen, we homogenized them using zirconium beads (diameter 0.1 mm) and a Mini-Beadbeater (BioSpec Products Inc., Bartlesville, OK, USA). We measured each RNA sample’s concentration and purity using a DU730 life science UV-Vis spectrophotometer (Beckman Coulter, Fullerton, CA, USA) and verified the RNA’s integrity via electrophoresis on agarose gels.

First, we pooled a variety of total RNAs from various conditions (e.g., heat shock, cold shock, toxic chemical exposure, and different life stages) into a single tube, which we then subjected to EST sequencing via a GS-FLX Titanium instrument (454 Life Sciences, Roche), assembling the EST sequences having 95% similarity levels with one another. Next, we separately characterized contigs and singletons of each EST data set by means of BLAST-X comparisons, using public domain databases. This process allowed us to treat EST sequences with E-values over 1.0E-05 as “No Hit,” as they probably belonged to UTRs.

DinoSL sequence searches

Finally, we investigated EST sequences having more than 100 bp of 5′-UTR for the SL sequence. In addition, we constructed local nucleotide databases of our Cochlodinium and Prorocentrum EST data, using BioEdit

version 5.0.6 (Hall 1999), and used them for local BLAST searches. To analyze nuclear encoded transcripts (or EST sequences) that had the dinoSL sequence, we used Genetyx version 7.0 software (Genetyx Corp., Tokyo, Japan).

In the present study, we determined the large-scale EST sequences of two dinoflagellates, C. polykrikoides (93 K sequence reads, 31 Mb) and P. minimum (773 K sequence reads, 291 Mb). From our GS-FLX sequencing, we identified 3,173 contigs and 21,521 singletons in Cochlodinium cDNA and 21,120 contigs and 125,540 singletons in Prorocentrum cDNA (Table 1). BLAST-X searches showed many cDNA sequences contained 5′-UTR sequences. For example, we identified P. minimum ESTs more than 100 bp of 5′-UTR in sequence length at 1,491 contigs and 414 singletons. Upon comparing these 5′-UTR sequences of P. minimum ESTs and a conserved dinoSL sequence (5′-DCC GTA GCC ATT TTG GCT CAA G-3′), we identified eight dinoSL sequences belonging to ribosomal protein, 40S ribosomal protein, 60S ribosomal protein, cAMP-dependent protein kinase regulatory subunits, and acidic ribosomal protein (Table 2, Fig. 1). On the other hand, we only detected one dinoSL sequence belonging to the 60S ribosomal protein in the C. polykrikoides ESTs.

In addition, we used BLAST searches to investigate dinoSL-containing transcripts in our local nucleotide database. Through this analysis, we detected 55 dinoSL sequence-containing ESTs (17 contigs, 38 singletons) from the P. minimum EST data. Using BLAST searches, we analyzed all sequences in the GenBank database, listing the closest matched genes in Table 2. Of these, we could annotate 3 out of 38 singleton-ESTs (8%) and 9 out of 17 contig-ESTs (53%). The identified genes included ribosomal protein, acidic ribosomal protein, cAMP-dependent protein kinase, conserved hypothetical protein, heat shock protein 70, imm downregulate 23 protein, and unknown proteins in P. minimum, and 60S ribosomal protein in C. polykrikoides. Our data showed that transcript numbers containing dinoSL sequence were lower than the total reads of EST data. In particular, we detected only one dinoSL sequence from C. polykrikoides ESTs. These results resembled those in the study by Bachvaroff and Place (2008), in which they detected the dinoSL sequence in Amphidinium carterae EST data, but it was not universal. These findings are incompatible with those of Zhang and Lin (2009), which showed that the dinoSL sequence in the 5′-UTR has a wide distribution among dinoflagellate nuclear-encoded transcripts. With the present and previous data, we could not conclude that dinoflagellates’ nuclear-gene transcripts most commonly include the dinoSL sequence, because we identified relatively few gene transcripts containing the dinoSL sequence from large-scale ESTs of C. polykrikoides and P. minimum.

To our knowledge, the dinoSL sequence is added to the 5′-end of dinoflagellate gene transcripts. For investigating whether all or parts of dinoflagellate nuclear gene transcripts contain dinoSL sequence, researchers should retain intact 5′-ends of the genes. In addition, to detect the dinoSL-containing transcripts, studies should amplify transcripts entirely by means of reverse transcriptase.

However, many dinoflagellates contain inhibitors that will strongly inhibit either reverse transcriptase or Taq DNA polymerase activity (Zhang and Lin 2009). Problems such as these may explain why few nuclear gene transcripts contain the dinoSL sequence, in both the previous data (Zhang et al. 2007, Bachvaroff and Place 2008) and in our EST data. On the other hand, Bachvaroff and Place (2008) showed that the SL trans-splicing of dinoflagellates nuclear genes correlates with expression level, suggesting that the high-expression-level gene is more likely to be SL trans-spliced. By surveying the dinoflagellate gene transcripts that contain the dinoSL sequence, researchers have identified numerous genes involved in the various cell functions (Table 3). Interestingly, all of the annotated genes in Table 3 play important roles in cells’ biological processes and have high expression levels within these cells. In view of the summarized data, we consider that finding genes containing the dinoSL sequence might be much easier using high-expression-level genes rather than using low-expression genes. For example, studies have found ribosomal protein genes containing the dinoSL sequence in almost all dinoflagellates (except Noctiluca scintillans). Reportedly, proliferating cell nuclear antigen (PCNA) contains the dinoSL sequence throughout the phylum (Zhang et al. 2007). Both ribosomal protein and PCNA are expressed throughout the cell cycle, and at high expression levels, as well.

This study investigated the dinoSL sequence location by surveying reported dinoSL-containing gene transcripts and our EST data (Table 3). Having detected the dinoSL sequence in C. polykrikoides and P. minimum nuclear gene transcripts (Fig. 1), we found the dinoSL sequence location ranged from 52 to102 bp upstream of

the start codon (ATG). Moreover, additional summarized data (Table 3) showed that the dinoSL sequence’s major locations ranged from 40 to 160 bp upstream of ATG. Only in a few genes, and particularly in unknown function genes, did the dinoSL sequence occur more than 170 bp upstream of ATG. Perhaps the SL trans-splicing process mostly tends to append the dinoSL sequence to the short nucleotides (< 170 bp) upstream of the start codon.