Pleistocene Speciation of Sister Taxa in a North Pacific Clade of Brooding Sea Stars

1

Taxonomic Appendix

The most recent monographic treatment of Leptasterias (Fisher 1930) proposed numerous species, subspecies and forms within the six-rayed subgenusHexasterias. However, more recent taxonomic usage has tended to follow the suggestion of Chia (1966), who applied the name L. hexactis to all six-rayed sea stars in the Puget Sound region. Molecular genetic analysis of over 1200 six-rayed Leptasterias collected from the intertidal and shallow subtidal zones throughout the range of the subgenus in North America from Shemya Is., AK (52o44.7′ N/ 174o 7.0′ E) to San Simeon, CA (35o38.4′ N/ 121o12.0′ W) by PCR-RFLP survey of the mitochondrial DNA putative control region and flanking sequences (Foltz 1998) and sequencing (Hrincevich et al. 2000, Flowers and Foltz 2001) identified five molecular lineages that corresponded to five species or species complexes: L. aequalis, L. alaskensis, L. aleutica/L. camtschatica, L. hexactis and L. leptodoma. Three species complexes (L. aequalis, L. aleutica/L. camtschatica and L. hexactis) each showed genetic (allozymic, DNA sequence) and morphological heterogeneity suggestive of specific or subspecific differentiation (Table 2). Although an extended taxonomic treatment of these three complexes is beyond the scope of the present paper, two sympatrically distributed sister groups (L. aleutica/L. camtschatica, L. aequalis A/L. aequalis D) show genetic and morphological differences and likely correspond to conventional species under morphological, phylogenetic and/or genealogical concordance criteria for species delimitation. Two other sister groups (L. hexactis C/L. hexactis G, L. aequalis B/L. aequalis K) also show genetic and morphological differences, but because they are parapatrically distributed along the North American Pacific Coast (Table 2), determining their specific or subspecific status is problematic (e.g., Avise 2000). L. hexactis Cand L. aequalis Beach show micro-geographic (< 20 km scale) morphological variation in the Puget Sound region (Foltz and Flowers 2007), including variation in characters used by Fisher (1930) in his taxonomic revision of the genus. Some of this morphological variation is associated with environmental variation (protected versus exposed coastline), and thus could represent adaptive differences produced and maintained either by phenotypic plasticity or by natural selection. The morphological differences observed between sister groups L. hexactis C/L. hexactis G and L. aequalis B/L. aequalis K (primarily marked differences in the average number of actinal spines per ray, see Table 2) might represent similar morphological variation on a broader geographical scale. However, for L. hexactis C and L. hexactis G there is independent genetic evidence that the two lineages have been separated for a long period. In particular, the two lineages show allozyme differences averaged over 14 loci (Hrincevich and Foltz 1996) that are slightly greater than observed between L. aequalis A and L. aequalis B, a pair of sympatric non-sister species that are reproductively isolated as well as morphologically distinct (Foltz 1997, Foltz and Flowers 2007).

Figure S1. Approximate ranges along the Pacific coast of North America, from Attu Island, Alaska to San Simeon, California, of four pairs of putative sister species considered in the present study. Distributional data are based on molecular data from Foltz et al. (1996a, 1996b), Foltz (1998) and Flowers and Foltz (2001), rather than taken from Fisher (1930). L. hexactis C, L. aequalis A and L. aequalis B are sympatric in the Puget Sound Region.

Table S1 Effects of specimen age (range 128 years), preservation method (alcoholpreserved, fresh or formalinpreserved tissue) and size of target region(range 2501000 bp) on PCR amplification success for nuclear intron sequences. The amplification success rate from alcohol or formalinpreserved museum specimens was 80% for target regions of approximately 250 bp. Success rates declined for target regions of 1000 bp (alcoholpreserved specimens) and for regions of 5001000 bp (formalinpreserved specimens).

Sourceadeg. freedomType III Sum of SquaresMean Square Fvalue

Age of specimen20.790.393.09

Preservation method217.298.6567.25***

Size of target region111.2411.2487.45***

***P<0.001 [adjusted for multiple tests]

aAge (range 128 years) and preservation method were class variables and size was analyzed as a regressor variable, using PROC GLM in SAS v. 9.1.3.

Table S2. Models and parameter estimates from MrBayes for four data partitions, as well as various summary statistics for each partition. The tree length was 1.1612.

aMean values are shown with standard errors. "Missing data" refer to gene regions that have a sequence length of zero for one or more taxa.In descending order of importance, missing data were due to [1] a lack of sequence information for one gene region, mostly COI, in 20 specimens, [2] ragged ends at the 5' or 3' end of some regions and [3] alignment gaps in each gene region except COI. In addition, the following regions of the Elongation factor-1 α subunit intron that totaled 351 bp were completely excluded from the analysis: the two internal priming sites (Internal-2F and -3F) and two poorly-aligned regions. Previous analyses of simulated DNA sequence data (Wiens 2003) suggest that phylogenetic accuracy is only slightly affected with 27% average missing data, and that the most likely effect of missing data is to exacerbate long-branch attraction problems. In the present analysis, long branches were restricted to the outgroup taxa, which were included only to test the monophyly of the genus Leptasterias.

bBecause of the relatively large size (approximately 1 kb) of the Elongation factor-1 α subunit intron 4 amplicon in Leptasterias, DNA from formalin-fixed or other poorly-preserved specimens was amplified in three abutting segments by using two internal primers and their reverse complements, in addition to the flanking primers described by Foltz et al. (2007b):

Internal-2F5=TGTGTGCATACATGTGGGTTAGTAAAGAAACTT

Internal-3F5=AATACATAAACCTCTTTGGTCAAATGCTCAACTTTAATA

cThe two ATP synthase β subunit introns share a 140-bp repeat sequence that evolves at a slow rate compared to flanking non-repetitive sequence and that also shows evidence for gene conversion. As a result, both copies of this repeat region were excluded from some of the analyses.

dPlus flanking tRNA and rRNA regions. The unassigned region in sea star mitochondrial DNA is called the putative control region, by analogy to the unassigned region in sea urchin mitochondrial DNA, which has been shown by experimental evidence to harbor a replication origin (Mayhook et al. 1992).

eExclusion of the outgroup taxa resulted in a shorter alignment for the putative control region and flanking sequences for these analyses (Table S3), due to the removal of gap-only sites. Also, the cytochrome oxidase subunit I and elongation factor-1 α subunit intron 4 alignments used were trimmed in length compared to those used in the phylogenetic reconstruction, due to missing data in some taxa.
Table S3. Details of IMa analysis.

aupper and lower 95% interval

bin units of 105 individuals

1

Table S4. Catalog numbers, collection details and GenBank accession numbers for specimens included in Figs. 2-4.

#Cataloged and identified as Leptasterias sp.

*Previously-published sequences obtained from GenBank

1

Table S5. Catalog numbers, collection details and GenBank accession numbers for specimens included in Figs. 2-4 and Table 2.

*Previously-published sequences obtained from GenBank

Supplemental References

Chia FS (1966) Systematics of the six-rayed sea star, Leptasterias, in the vicinity of San Juan Island, Washington. Syst Zool 15:300-306

Foltz DW (1997) Hybridization frequency is negatively correlated with divergence time of mitochondrial DNA haplotypes in a sea star (Leptasterias spp.) species complex. Evolution 51:283-288

Foltz DW (1998) Distribution of intertidal Leptasterias along the Pacific North American coast: a synthesis of allozymic and mtDNA data. In: Mooi R, Telford M (eds) Proc 9th Int Echin Conf, Balkema, Rotterdam, pp 235239

Mayhook AG, Rinaldi AM, Jacobs HT (1992) Replication origins and pause sites in sea urchin mitochondrial DNA. Proc R Soc Lond 248B:85-94

Wiens JJ (2003) Missing data, incomplete taxa, and phylogenetic accuracy. Syst Biol 52:528-538