Recent mobility of plastid encoded group II introns and twintrons in five strains of the unicellular red alga Porphyridium

Porphyridium purpureum strains and plastid genomes

Four Porphyridium purpureum strains, SAG 1380-1a, SAG 1380-1b, SAG 1380-1d (obtained from the Culture Collection of Algae, Göttingen University) and CCMP 1328 (obtained from the National Center for Marine Algae and Microbiota, East Boothbay, ME) were grown under sterile conditions on Artificial Sea Water (Jones, Speer & Kuyr, 1963) at 25 °C, under continuous light (100 µmol photons m⁻² s⁻¹) on a rotary shaker at 100 rpm (Innova 43; New Brunswick Eppendorf, Enfield, Connecticut, USA). Cells were pelleted via centrifugation and DNA was extracted from ca. 100 mg of material with the DNeasy Plant Mini Kit (Qiagen) following the manufacturer’s protocol. Sequencing libraries were prepared for each strain using the Nextera DNA Sample Preparation Kit (Illumina Inc., San Diego, California, USA) and sequenced on an Illumina MiSeq sequencer using a 300-cycle (150 × 150 paired-end) MiSeq Reagent Kit v2 (Illumina, Inc.). Sequencing reads were quality and adapter trimmed (Q limit cutoff = 0.05) and overlapping pairs were merged at the 3′ end using the CLC Genomics Workbench 6.5.1 (CLC Bio, Aarhus, Denmark).

Mapping, polymorphism detection and analysis

The reads from each of the four newly-sequenced strains above were mapped to the existing P. purpureum plastid reference genome (strain NIES 2140; Tajima et al., 2014) with a stringency of 90% sequence identity over a 90% read length fraction using the CLC Genomics Workbench (CLC Bio, Aarhus, DK). SNPs were called using the Genomics Workbench 6.5.1 quality-based variant detection (≥10× base coverage, quality score >30 and ≥50% frequency required to be called). An uncorrected distance phylogeny was constructed using a matrix of DNA polymorphisms detected between the five plastid genomes with the program MEGA6.06 (Tamura et al., 2013; 100 bootstrap replicates).

Group II intron and IEP identification

Novel GII introns in the plastid genomes of the four P. purpureum strains were identified by aligning de novo assembled (using the CLC Genomics Workbench v.6.5.1 de novo assembler) plastid contigs from each strain to the NIES 2140 reference. Multiple large (>50bp) insertions were identified in our de novo contigs with respect to the reference, and were annotated as putative introns. We then mapped the corresponding raw short read data to these contigs and manually inspected the mapping for assembly artifacts. Intron encoded proteins (IEPs) were identified within the putative introns by ORF detection using the bacterial/plastidic genetic code. The four domains that constitute an IEP (i.e., reverse transcriptase [RT], maturase [X], DNA-binding [D], and endonuclease [En] Mohr, Perlman & Lambowitz, 1993; San Filippo & Lambowitz, 2000) were identified by sequence alignment using ClustalX (Larkin et al., 2007) to known IEPs of the prokaryote CL1/CL2 group and to those from the Rhodophyta, Viridiplantae, Cryptophyta, Euglenozoa, and stramenopiles (listed in Table S1) obtained from NCBI and the Group II intron database (Dai et al., 2003; http://webapps2.ucalgary.ca/~groupii/, accessed Sept. 2014). To examine the phylogeny of these mobile elements, the IEP peptide sequences were aligned with the RT-domain alignment of Toro & Martínez-Abarca (2013) and maximum likelihood phylogenies were inferred under the WAG amino acid substitution model with 100 bootstrap replicates using MEGA6.06. The GII intron/IEP sequences described here are accessible using NCBI accession numbers KKJ826367 to KKJ826395 and the P. purpureum plastid genome under NC_023133 (Tajima et al., 2014).

Intron structure and evolution

Intron secondary structures were predicted using sequence alignment, manual domain identification, and automatic structure conformation in comparison with previously predicted structures of group IIB introns using the Mfold web server (Zuker, 2003; Table S1). A detailed secondary structure model was generated based on the rpoC1 intron and mat1d IEP (Fig. 1). This was then used as a guide to predict draft structures using PseudoViewer3 (Byun & Han, 2009) for all other GII introns. A domain alignment was then performed against the GII intron structures derived from the cryptophyte Rhodomonas salina (Maier et al., 1995; Khan et al., 2007) using ClustalX2.1, and a maximum-likelihood phylogeny was generated using intronic nucleotide sequence data under the GTR +I + Γ model with 100 bootstrap replicates using MEGA6.06 (Tajima et al., 2014). Prior to this, the IEPs or IEP remnants were removed to avoid potential long-branch attraction artifacts. Additionally, conserved motifs within the basal DI, DIV, DV and DVI domains (Table S2) were used as a BLASTN (Altschul et al., 1990) query to the five aligned plastid genomes to identify additional group II intron structures present in all strains (and thus not identified via length heterogeneity upon initial assessment).

The twintrons present in the P. purpureum plastid genome were aligned and compared to the other introns to allow identification of the outer and inner introns, exon binding sites, to describe their secondary structures, and potentially to understand their mode of origin.

Intron-encoded proteins

Intron-encoded proteins present at the same insertion site are nearly identical among the strains (98.9–100% amino acid identity), except for the mat1b IEP in strain NIES 2140 which has an apparent truncation of 27 amino acids due to a premature stop-codon. All nine IEPs contain two fully conserved reverse transcriptase (RT) and maturase (X) domains (Fig. S3), whereas four of the five elements present in all five P. purpureum strains (mat1a, 1b, 1c, 1e) are either truncated or have completely lost the DNA-binding (D) and endonuclease (EN) domains responsible for conferring mobility (Simon, Kelchner & Zimmerly, 2009). These latter GII introns thus appear to have lost mobility, and exhibit vertical inheritance. Additionally, mat1a and mat1b lack the YADD motif crucial for reverse transcriptase activity at the active site (Fig. S3; Moran et al., 1995). The remaining five GII introns encoding mat1d, mat1f[a,b,c], mat1g, mat1h, mat1i are distributed in lineage-specific patterns on the P. purpureum phylogeny (Fig. 2A) and likely remain mobile because they retain all functional domains (Fig. S1). Thereforre, we show here for the first time examples of recent intron mobility and putative stability; the latter being represented by plastid-encoded IEPs that lack a functional endonuclease domain due to mutation and/or sequence degeneration.

Phylogenetic analysis using the IEP peptide alignment shows that seven of the nine P. purpureum IEPs form a monophyletic clade that is sister to cryptophyte plastid IEPs, the cyanobacterial CL2B clade, and Euglenozoa plastids (Fig. 4 and Fig. S4). The mat1a and mat1b IEPs, derived from group II introns found to lack typical secondary structure, create a paraphyletic assemblage within the cryptophytes (mat1a) or group outside of the CL2B clade (mat1b). This tree, in association with Fig. 3, illustrates the shared ancestry and subsequent co-evolution of seven IEPs as well as their associated GII intron structures.

Group IIB intron secondary structure

Self-splicing group II introns are dependent on a conserved secondary and tertiary RNA structure. These autocatalytic genetic elements are composed of six distinct double-helical domains (DI to DVI) that radiate from a central wheel with each domain having a specific activity (Lambowitz & Zimmerly, 2011). As illustrated by the rpoC1 GII intron that contains mat1d (Fig. 1), the introns studied here have group IIB intron secondary structures following this model. Annotated sequence alignments and draft secondary structures for the remaining introns are presented in the supplementary information (Figs. S5–S16 (note that no intronic sequence data were removed to simplify folding)). As expected, the P. purpureum IEPs are located in the domain IV (DIV) loop, which is integral for ribozyme activity. DIVa (the maturase binding site exclusive of the IEP (see Fig. S23)) and DV contain conserved regions (96 ± 4% identity), whereas DVI is highly variable (37 ± 17% identity; length range 44–162bp; see Fig. S17).

The bulged AC nucleotide pair illustrated within DV of Fig. 1 is in agreement with the model of Toor et al. (2008) and Keating et al. (2010), however the possibility exists that (as in the remaining introns (Figs. S5–S16)) the unpaired nucleotides can be shifted downstream to create a CG bulge. The DVI domain contains a conserved, bulged adenosine that serves as a nucleophile during lariat generation upon splicing (Peebles et al., 1987; Robart et al., 2014), however most P. purpreum group II intron models described here maintain an additional unpaired guanine in an AG bulge. The effect this has on the splicing reaction remains unknown. Structural analysis reveals a novel and unusual bipartite DIII domain configuration, because it can be represented by either a canonical stem/loop structure, or as two individual stems (Figs. S5–S11 (see inset DIII)), or as two individual stems only (Figs. S12–S16). The DIII domain contributes an adenosine pair to a base stack that serves to reinforce DV opposite the catalytic site, and stabilizes the entire structure (Robart et al., 2014). Modification of this domain in the P. purpureum group II intron structures that have lost mobility may reflect the lack of an IEP and thus the need for reinforcement.

Group II intron RNAs self-splice via base-pairing interactions between exon-binding sites (EBS1 & EBS2) on the ribozyme and intron-binding sites (IBS1 & IBS2) at the 5′ exon region (Lambowitz & Zimmerly, 2011). Despite a common origin, the P. purpureum introns that encode an IEP appear to have a highly variable EBS (Fig. S18) perhaps explaining their ability to spread to novel sites in these plastid genomes. Each EBS/IBS pairing is uniquely associated with an intron/IEP combination, and complementarity between both is present. EBS1 and/or EBS2 were not identified for the mat1a, mat1b, and mat1c introns. Interestingly, EBS1 is located at the same site in the nucleotide alignment, whereas the EBS2 position is variable due to length heterogeneity between introns. Understanding how variation in these binding sites affects the ability of group II introns to self-splice and bind target DNA is paramount for ‘targetron’ development (Enyeart et al., 2014) and application of these mobile elements to biotechnology.

Finally, sequence alignment of the P. purpureum introns described here with the five Rhodomonas salina introns presented in Khan & Archibald (2008) (Fig. S17) demonstrates that the domain organization and secondary structure of these elements in both species are similar. We were thus able to derive amended secondary structures for the cryptophyte models proposed by Maier et al. (1995) and Khan & Archibald (2008) using P. purpureum as a guide. In doing so, we identified a cryptophyte domain IVa similar to that in P. purpureum that contains the IEP and has modified domains DII and DIII (e.g., Fig. S19). We propose that the non-canonical features described by Khan & Archibald (2008) in R. salina and H. andersenii (i.e., domain insertions, ORF relocation, absence of internal splicing) can be explained by degeneration of the endonuclease domain between the protein C-terminus and domain IVa. Amended structures for the remaining cryptophyte introns are presented herein (Figs. S19–S23).

Red algal twintrons

Introns nested within other introns (or twintrons) were first reported in the Euglena gracilis plastid (Copertino & Hallick, 1991). Since then, group II/III twintrons have been reported at multiple sites in complete Euglenozoa plastid genomes (E. gracilis and Monomorphina aenigmatica; Pombert et al., 2012) and from the plastid genomes of the cryptophytes Rhodomonas salina and Hemiselmis andersenii (Maier et al., 1995; Khan et al., 2007 (however see discussion, above)). Twintrons have also been described in the prokaryotes Thermosynechococcus elongatus, a thermophilic cyanobacterium (Mohr, Ghanem & Lambowitz, 2010) and in Methanosarcina acetivorans, an archaebacterium (Dai & Zimmerly, 2003). Here we provide the first description of twintrons in rhodophyte plastid genomes, and the first known report of an inner intron (mat1f) found nested within two different outer introns (while also inserted in a third gene). The plastid genomes of three P. purpureum strains each contain two twintrons encoding mat1fa and mat1fb (Figs. 2A and 2B) that are bounded by different outer introns inserted in the rpoC2 and atpI genes, respectively. Two strains contain a copy of the inner intron/IEP inserted singly within the atpB gene as mat1fc. Alignment of the outer and inner twintron regions together with the other introns shows that the two different twintrons have very similar structures (Fig. S1) Despite partial sequence similarity (78.2% sequence identity in pairwise comparisons), the two outer introns have similar IEP remnants. The IEPs are truncated at the same site, likely due to a partial protein deletion. Approximately 130 nt and 555 nt, respectively, remain in the 5′ and 3′ regions of the former IEP in the external introns. Presumably, the later insertion of the inner intron happened at the same binding site (85 nt further downstream from the excision site). Our analyses show that the closely related outer introns int.b (atpI) and int.b (rpoC2; Fig. 3) in P. purpureum retain IEP remnants that have been truncated in the same region due to inner intron insertion at the same DIV target site (Fig. S17). Of future interest is to study the splicing of these red algal twintrons to confirm that excision occurs in consecutive steps as in other plastid twintrons (Copertino, Shigeoka & Hallick, 1992).