DNA variation in the phenotypically-diverse brown alga Saccharina japonica

Abstract

Background: Saccharina japonica (Areschoug) Lane, Mayes, Druehl et Saunders is an economically important and highly morphologically variable brown alga inhabiting the northwest Pacific marine waters. On the basis of nuclear (ITS), plastid (rbcLS) and mitochondrial (COI) DNA sequence data, we have analyzed the genetic composition of typical Saccharina japonica (TYP) and its two common morphological varieties, known as the "longipes" (LON) and "shallow-water" (SHA) forms seeking to clarify their taxonomical status and to evaluate the possibility of cryptic species within S. japonica.

Results: The data show that the TYP and LON forms are very similar genetically in spite of drastic differences in morphology, life history traits, and ecological preferences. Both, however, are genetically quite different from the SHA form. The two Saccharina lineages are distinguished by 109 fixed single nucleotide differences as well as by seven fixed length polymorphisms (based on a 4,286 bp concatenated dataset that includes three gene regions). The GenBank database reveals a close affinity of the TYP and LON forms to S. japonica and the SHA form to S. cichorioides. The three gene markers used in the present work have different sensitivity for the algal species identification. COI gene was the most discriminant gene marker. However, we have detected instances of interspecific COI recombination reflecting putative historical hybridization events between distantly related algal lineages. The recombinant sequences show highly contrasted level of divergence in the 5'- and 3'- regions of the gene, leading to significantly different tree topologies depending on the gene segment (5'- or 3'-) used for tree reconstruction. Consequently, the 5'-COI "barcoding" region (~ 650 bp) can be misleading for identification purposes, at least in the case of algal species that might have experienced historical hybridization events.

Conclusion: Taking into account the potential roles of phenotypic plasticity in evolution, we conclude that the TYP and LON forms represent examples of algae phenotypic diversification that enables successful adaptation to contrasting shallow- and deep-water marine environments, while the SHA form is very similar to S. cichorioides and should be considered a different species. Practical applications for algal management and conservation are briefly considered.

Figure 1

Nucleotide substitution sites in the rbc LS, COI , and ITS gene regions of the Saccharina japonica morphological forms. The numbers above the top sequence represent the position of segregating sites and the start of a deletion or insertion (see also Additional file 4). Nucleotides are numbered from the beginning of our sequence. Dots indicate the same nucleotide as the reference sequence. The hyphens represent deleted nucleotides. ▴ denotes a deletion; † denotes the absence of a deletion; ▼ denotes an insertion; ‡ denotes the absence of an insertion. The sample abbreviations are in the section "Methods". COR: S. coriacea. The underlined nucleotides in bold face show fixed differences between COR and the other sequences.

Figure 2

Sliding window plot of nucleotide divergence for three gene regions (rbc LS, COI , and ITS) between the TYP and SHA forms of Saccharina japonica. Window sizes are 100 nucleotides with one-nucleotide increments. A schematic representation of the regions is displayed at bottom. The thick black line marks the COI region investigated by McDevit and Saunders [43].

Figure 3

Phylogenetic trees of the rbc LS, COI, ITS , and concatenate sequences (4,286 bp total) ofSaccharina japonicamorphological forms. The topology of trees obtained with Maximum likelihood and Neighbor-joining methods were congruent. Numbers at the nodes are bootstrap percent support values based on 1,000 replications in Maximum likelihood analysis. Saccharina latissima was excluded from this analysis due to recombination in the COI gene (see text). Other comments as in Figure 1.

Figure 4

Sliding-window plots of divergence along the COI gene region between S. latissima and other laminarialean algae: S. japonica (A), the SHA form (B), S. coriacea (C), and S. angustata (D). Window sizes are 50 nucleotides with ten-nucleotide increments. The thick black lines at the bottom mark the COI regions investigated by McDevit and Saunders [43] and Silberfeld with colleagues [2].

Figure 5

Phylogenetic trees based on different fragments of the COI coding region: (A) 5’- COI , (B) 3’- COI , and (C) full COI region. Representative sequences of the orders Laminariales (L) and Ectocarpales (E) included in these trees are marked by vertical lines. Red algae COI sequences of Callithamnion pikeanum and Hypnea nidulans (GenBank accession numbers EU194965 and FJ694907, respectively) are used as outgroups. Note the changed position of S. latissima (in bold) depending on the COI region used for the tree. The S. latissima COI sequence denoted as “S” is from [2]. See Additional file 3 for GenBank accession numbers. Other comments as in Table 1 and Figure 3.

Figure 6

Schematic representation of recombination events in the COI gene region from Saccharina latissima (A) and Cystophora retorta (B). The parental sequences are from Petalonia fascia and the SHA form (for S. latissima ); and Bifurcaria bifurcata and Phyllospora comosa (for C. retorta ).