The plastid genome of some eustigmatophyte algae harbours a bacteria-derived six-gene cluster for biosynthesis of a novel secondary metabolite

Abstract

Acquisition of genes by plastid genomes (plastomes) via horizontal gene transfer (HGT) seems to be a rare phenomenon. Here, we report an interesting case of HGT revealed by sequencing the plastomes of the eustigmatophyte algae Monodopsis sp. MarTras21 and Vischeria sp. CAUP Q 202. These plastomes proved to harbour a unique cluster of six genes, most probably acquired from a bacterium of the phylum Bacteroidetes, with homologues in various bacteria, typically organized in a conserved uncharacterized putative operon. Sequence analyses of the six proteins encoded by the operon yielded the following annotation for them: (i) a novel family without discernible homologues; (ii) a new family within the superfamily of metallo-dependent hydrolases; (iii) a novel subgroup of the UbiA superfamily of prenyl transferases; (iv) a new clade within the sugar phosphate cyclase superfamily; (v) a new family within the xylose isomerase-like superfamily; and (vi) a hydrolase for a phosphate moiety-containing substrate. We suggest that the operon encodes enzymes of a pathway synthesizing an isoprenoid-cyclitol-derived compound, possibly an antimicrobial or other protective substance. To the best of our knowledge, this is the first report of an expansion of the metabolic capacity of a plastid mediated by HGT into the plastid genome.

Keywords: Eustigmatophyceae; UbiA superfamily; horizontal gene transfer; plastid genome; secondary metabolism; sugar phosphate cyclase superfamily.

Figure 1.

Schematic phylogeny of eustigmatophytes showing the position of species with sequenced plastid genomes. The taxa with plastomes sequenced in this study are highlighted in bold. For the genera Nannochloropsis and Microchloropsis, only the type species are shown for simplicity, although plastomes have been sequenced for a number of other closely related species or strains. The topology of the tree reflects a robustly resolved phylogeny based on the 18S rRNA gene [14]. Evolutionary events impacting the gene content of eustigmatophyte plastid genomes are mapped onto the tree based on the most parsimonious interpretation of the pattern of the gene presence/absence (provided in the electronic supplementary material, table S1).

Figure 2.

Gene maps of the newly sequenced plastid genomes of Vischeria sp. CAUP Q 202 and Monodopsis sp. MarTras21. Genes are shown as blocks facing inside if transcribed in the clockwise direction or facing outside if transcribed in the counter-clockwise direction. The assignment of the genes into different functional categories is indicated by their different colours. The plot in the inner circle shows the GC content, with the thin grey line marking 50%. Red arrows point to the novel six-gene cluster acquired by HGT. Note the ebo operon (eboA to eboF genes) delimited in both genomes with the red bar.

Figure 3.

The ebo operon, a novel six-gene operon shared between the plastid genomes of some eustigmatophytes and bacteria. The figure shows examples of various arrangements of the ebo operon. Genes (or predicted ORFs) are displayed as boxes with the pointed side indicating the 3' end of the coding sequence (i.e. the direction of transcription). Accession numbers of protein sequences encoded by the ebo genes included in the figure are provided in the electronic supplementary material, table S2. Boxes in white correspond to genes that flank clusters of ebo genes in the genomes but are not considered to be a part of the ebo operon itself. Boxes in grey correspond to non-ebo genes presumably inserted into the original ebo operon. Double slashes indicate intervening genomic sequences of a varying length (sometimes including a break in the genome sequence assembly). Note that in some species an extra copy of one of the ebo genes exists in the genome, detached from the ebo operon as such.

Figure 4.

Phylogenetic analysis of EboC protein sequences. The ML trees were inferred from an alignment of 224 amino acid positions using RAxML and the LG4X+Γ substitution model. Bootstrap support values are shown at branches when more than 75%. Five groups of species are distinguished by a different background: Bacteroidetes, violet; Cyanobacteria, blue-green; Eustigmatophyceae, red; Leptospira, light green; other bacteria, blue. Taxa are divided into three conveniently defined groups according to the type of ebo operon they possess (see the vertical bars on the right side). Note that some species assigned in the figure to the ABCDEF-type of the operon secondarily deviate from this presumed ancestral form for the given major taxa (see §3.2).

Figure 5.

The origin of the novel operon in the eustigmatophyte plastid genomes can be traced to a bacterium from the phylum Bacteroidetes. The phylogenetic tree displayed was inferred with the ML method and the GTR+Γ substitution model from a concatenated alignment of protein sequences of all six ebo genes (1838 aligned amino acids). Bootstrap support values are shown at branches when more than 75%. The tree is arbitrarily rooted on a maximally supported branch separating Cyanobacteria and other taxa included in the analysis. For simplicity, monophyletic groups comprising exclusively representatives of the same genus were collapsed and are shown as triangles, with the two side lengths proportional to the distances to closest and furthest leaves. Note that sequences from the eustigmatophytes Vischeria sp. CAUP Q 202 and Monodopsis sp. MarTras21 branch among members of the phylum Bacteroidetes, possibly specifically related to the ebo operon from Sporocytophaga myxococcoides. Note also that the cluster of sequences from the genus Leptospira (phylum Spirochaetes) is firmly nested among Bacteroidetes, indicating an inter-phylum HGT event.

Figure 6.

Cluster analysis of sequences of the UbiA superfamily showing that EboC belongs to a novel separate group within the superfamily. Each dot represents a sequence and each grey line shows a PSI-BLAST comparison of two sequences, with darker lines indicating higher similarity (lower E-values). The cluster of EboC homologues is coloured bright pink, with Monodopsis and Vischeria sequences highlighted in yellow. Other members of the UbiA superfamily include COQ2 (red), UbiA (magenta), UbiAD1 (green), MenA (yellow green), homogentisate prenyltransferases (HPT, HST and HGGT in different pink colours), chlorophyll synthase (blue), COX10 (dark blue), DGGGP synthase (dark green), DPPR synthase (cyan), archaeal UbiA homologues (orange), AuaA homologues (wine colour) and Af homologues (dark grey).

Figure 7.

EboD represents a new lineage within the SPC superfamily. The phylogenetic tree displayed was inferred with the ML method using the LG+Γ substitution model and a protein sequence of alignment of 147 amino acids. For simplicity, most monophyletic branches of the same putative enzymatic activity with more than two members were collapsed as in figure 5, bootstrap support values are shown only when more than 75%. The tree was rooted at a position suggested by a previously published analysis using glycerol dehydrogenase (CglD) from Escherichia coli as an outgroup [52]. Main subgroups of the superfamily characterized by different enzymatic activities are indicated. Note that most proteins in the tree have not been characterized biochemically, so the functional assignment suggested by the tree topology has to be taken with caution (for example, the clade including confirmed EVS enzymes would apparently be considered as putative DHQS without the actual data to the contrary). AminoDHQS, aminodehydroquinate synthase; DDGS, desmethyl-4-deoxygadusol synthase; DDGS?, a clade of uncharacterized proteins related to DDGS, but distinctly different to possibly exhibit a different enzymatic activity; DHQS, 3-dehydroquinate synthase; DOIS, 2-deoxy-scyllo-inosose synthase; EEVS, 2-epi-5-epi-valiolone synthase; EVS, 2-epi-valiolone synthase.

Figure 8.

Multiple gains of the eboF gene by eukaryotes via HGT from bacteria. The phylogenetic tree displayed was inferred with the ML method and the LG4X+Γ substitution model from an alignment of EboF protein sequences from eukaryotes and their most similar homologues from bacteria (380 aligned amino acids). Bootstrap support values are shown at branches when more than 75%. Eukaryotic taxa are printed in bold (except the sequence from the liverwort Frullania sp., which is a likely bacterial contamination, see §3.5). The two branches crossed by a double slash were shortened to half of their actual length to make the figure more compact. Accession numbers of all the sequences used to infer the tree are provided in the electronic supplementary material, table S3.