Potential impact of stress activated retrotransposons on genome evolution in a marine diatom

Abstract

Background: Transposable elements (TEs) are mobile DNA sequences present in the genomes of most organisms. They have been extensively studied in animals, fungi, and plants, and have been shown to have important functions in genome dynamics and species evolution. Recent genomic data can now enlarge the identification and study of TEs to other branches of the eukaryotic tree of life. Diatoms, which belong to the heterokont group, are unicellular eukaryotic algae responsible for around 40% of marine primary productivity. The genomes of a centric diatom, Thalassiosira pseudonana, and a pennate diatom, Phaeodactylum tricornutum, that likely diverged around 90 Mya, have recently become available.

Results: In the present work, we establish that LTR retrotransposons (LTR-RTs) are the most abundant TEs inhabiting these genomes, with a much higher presence in the P. tricornutum genome. We show that the LTR-RTs found in diatoms form two new phylogenetic lineages that appear to be diatom specific and are also found in environmental samples taken from different oceans. Comparative expression analysis in P. tricornutum cells cultured under 16 different conditions demonstrate high levels of transcriptional activity of LTR retrotransposons in response to nitrate limitation and upon exposure to diatom-derived reactive aldehydes, which are known to induce stress responses and cell death. Regulatory aspects of P. tricornutum retrotransposon transcription also include the occurrence of nitrate limitation sensitive cis-regulatory components within LTR elements and cytosine methylation dynamics. Differential insertion patterns in different P. tricornutum accessions isolated from around the world infer the role of LTR-RTs in generating intraspecific genetic variability.

Conclusion: Based on these findings we propose that LTR-RTs may have been important for promoting genome rearrangements in diatoms.

Figure 1

Composition of the TE complements in the P. tricornutum and T. pseudonana genomes. (A and B) Pie chart representing the relative abundance of different TEs to the P. tricornutum (A) and T. pseudonana(B) TE complements. (C) Histogram representing percent genome coverage across the diatom TE complements. (D) Pie chart representing the relative contribution of the different CoDi groups to the P. tricornutum LTR-RT complement.

Figure 2

Phylogenetic tree showing the relationships between the CoDis and other Ty1/copia-like elements. This tree uses the RT domains from Ty3 and gypsy as outgoup and was constructed with the NJ method with the MEGA4 software [54]. The bootstrap values were calculated over 1,000 iterations and bootstrap scores over 70% are shown.

Figure 3

Phylogenetic tree showing the relationships between CoDis and other LTR-RT and retroviral lineages. The bootstrap values were calculated over 1,000 iterations and are indicated for two basal nodes. The tree was constructed with the NJ method using the SplitsTree4 software [55]. Species abbreviations: P. mt (Pseudonitzschia multistriata); P. m (Pseudonitzschia multiseries); F. c (Fragilariopsis cylindrus).

Figure 4

Abundance of CoDi-encoding ESTs in different conditions. (A) EST frequencies of the P. tricornutum CoDi elements listed in Additional file 1 within the 16 P. tricornutum cDNA libraries described and available at http://www.biologie.ens.fr/diatomics/EST3/. CoDi7 group does not have any EST support. (B) EST frequencies of the T. pseudonana CoDi elements listed in Additional file 1 within the 7 T. pseudonana cDNA libraries described and available at http://www.biologie.ens.fr/diatomics/EST3/. (A and B) Letter p indicates statistically-supported a (Pearson's Chi squares p = 0.0000) higher EST frequency of a CoDi group in this condition respect to the original library (non-stressed).

Figure 5

Regulation of Blackbeard expression. (A) Effect of nitrate limitation on the expression of the pLTRbkb-GUS-FcpA construct in transgenic P. tricornutum cells. Data represent the average with standard error from seven independent cultures after two weeks nitrate limitation (50 μM NO3-) compared to standard growth medium (882 μM NO3-). (B) Verification of Blackbeard transcriptional activation by semi-quantitative RT PCR in the cultures used for McrPCR. (C) McrPCR on Blackbeard and H4 and RPS controls using DNA extracted from P. tricornutum cells grown under normal and nitrate-limited conditions.

Figure 6

Sequence Specific Amplified Polymorphism analysis of Bkb in 13 P. tricornutum accessions. Each amplified insertion is revealed as a band on a sequencing gel and genomic DNA from the different accessions produces a characteristic fingerprint of bands.

Figure 7

Analysis of the Blackbeard locus. (A) Schematic representation of the primer pairs used to perform PCR at the Blackbeard locus. Primer pairs are embedded within ovals and dashed lines indicate the projection of the Bkb locus found in haplotype b to its native target site on haplotype a. (B) Haplotype analysis by PCR to assess the presence/absence of the Blackbeard insertion in ten P. tricornutum accessions. Haplotypes a and b respectively refer to the absence and presence of the Blackbeard insertion.

Figure 8

Schematic representation of the PtC25 and PtC75 recombinant loci. LTR-RT of the CoDi5.3 (orange) and CoDi2.3 (blue) groups are drawn with their LTRs (flanking arrows). Gene family 1 (green) and gene family 2 (purple) and other genes (grey) are drawn as arrows. Gene family 1 is further distinguished by red and/or blue bar on top and similar colors indicate similar sequences (see Additional file 4). Black or grey boxes with identical numbers indicate similar intergenic regions. Grey parallelograms project large duplicated regions from chromosome to chromosome. The blue parallelogram indicates the high similarity between the PtC25 and PtC75 elements. We indicate a 30 bp gap found in the CoDi5.3 segment flanking PtC75. We also indicate that the PtC25-associated CoDi5.3 entity contains a 5' truncated LTR which starts precisely where the gap described on chromosome 31 ends, further consolidating the historical link between these two loci. Bd 31.35 indicates a scaffold that could not be successfully mapped during P. tricornutum genome assembly.

Figure 9

Distribution of the GOS RT sequences. (A) Size of dataset in megabases (Mb) for each filter across the different geographic positions examined. Numbers indicate the number of RT hits for each filter. (B) Frequency of RT hits across the different filters.

Figure 10

Phylogenetic tree showing the relationships between the reverse transcriptase domains from the CAMERA database, retroviruses, and LTR retrotransposons. The tree was constructed with the NJ method using the SplitsTree4 software [59]. The bootstrap values were calculated over 1,000 iterations and are indicated for two basal nodes. GOS sequences are labeled by a two- letter code indicating their geographic provenance: Caribbean Sea (CA), Eastern Tropical Pacific (ET), Galapagos Islands (GI), Indian Ocean (IO), North American East Coast (NA), Polynesia Archipelagos (PA), Sargasso Sea (SS); followed by a number indicating filter size: 0.1-0.8 (1), 0.22-0.8 (2), 0.8-3.0 (3), 3.0-20.0 (4). These labels appear with blue background. Species abbreviations: P. mt (Pseudonitzschia multistriata); P. m (Pseudonitzschia multiseries); F. c (Fragilariopsis cylindrus); C. r (Chlamydomonas Reinhardtii); O. t (Ostreococcus tauri); A.a (Aureococcus anophagefferens).