| Constancea 83, 2002 University and Jepson Herbaria P.C. Silva Festschrift |
|---|
Despite numerous studies, the phylogeny and taxonomy of much of the Laurencia Lamouroux complex still remains obscure. Previous phylogenetic analyses of the genus Osmundea based on comparative morphology and sequences of the plastid-encoded rbcL gene indicated two potentially phylogenetically informative characters. The presence or absence of secondary pit connections in the epidermis and the shape of spermatangial receptacles (urn-shaped or cup-shaped) were synapomorphies for some clades in European species. Osmundea Stackhouse has a markedly disjunct distribution, being confined to Atlantic and Mediterranean coasts of Europe and Pacific North America. The major goal of this study was to extend taxon sampling to Californian species (the “Spectabilis group”) which show character combinations that differ from any European species. In addition, we investigated whether the basionym Fucus oederi Gunnerus might be available for the taxonomically and nomenclaturally confused species Osmundea ramosissima Athanasiadis.The tribe Laurencieae, represented by Chondrophycus, Laurencia, and Osmundea, was well supported, with Chondrophycus papillosus and the type species of Laurencia, L. obtusa, positioned basally. All species currently placed in the genus Osmundea formed a monophyletic group with robust support. Within Osmundea, three clades representing different geographical areas were observed (North Pacific, Atlantic Europe, Mediterranean + Atlantic Europe), with good (91% to 100%) bootstrap support. This phylogeny is interpreted as indicating that the genus may have originated in the late Tethys. The Californian clade and the European Osmundea clades exhibit contrasting stability of characters. The Californian species showed a fixed urn-shaped spermatangial receptacle shape and showed a variable number of epidermal secondary pit connections. In contrast, among the European species, the shape of the spermatangial receptacle shows character reversals and appears to be actively evolving in Mediterranean species. The two distinct European clades of Osmundea are clearly separable by the presence or absence of secondary pit connections. This comparative biogeographical approach to a morphological/molecular phylogenetic investigation has therefore yielded results that challenge the use solely of morphological characters for future generic subdivision in Osmundea.
During the last decade, further major changes have been made. The genus Osmundea Stackhouse (1809) has been resurrected (Nam et al. 1994) for a group of species largely corresponding to the subgenus Saitoa. The section Chondrophycus has also been elevated to generic status (Garbary and Harper 1998). Both taxonomic changes were proposed on the basis of a suite of morphological characters (Nam et al. 1994; Nam and Saito 1995; Garbary and Harper 1998; Nam 1999). The genus Osmundea was defined by (1) the development of spermatangia on filaments formed by apical and epidermal cells rather than on spermatangial trichoblasts formed by axial cells, (2) tetrasporangia formed by random epidermal cells rather than particular pericentral cells (Nam et al. 1994), and (3) the formation of two pericentral cells in vegetative axial filaments (Nam 1999). Molecular studies, based on rbcL sequence data, have corroborated the morphological evidence for the separation of Osmundea from Laurencia (Nam et al. 2000). The species currently assigned to Osmundea have a markedly disjunct distribution, being confined to Atlantic and Mediterranean coasts of Europe and Pacific North America (Serio et al. 1999). In contrast, the genera Laurencia and Chondrophycus are more widely distributed, although the majority of species are found in the Southern Hemisphere (McDermid 1988).
There have been only two previous phylogenetic studies of the genus Osmundea. The morphological cladistic analysis of Garbary and Harper (1998) placed O. hybrida basal in the genus, and linked European and Californian species together in pairs such as O. truncata with O. spectabilis. Nam et al.'s (2000) study assessed the phylogenetic significance of several morphological characters in Osmundea by comparative morphological and molecular analyses, but was confined to the European species. The type of spermatangial pits, previously thought to be significant (Saito 1982), was shown to be homoplasious. Urn-shaped spermatangial pits are found only in Osmundea (Nam et al. 1994), but Nam et al. (2000) were unable to determine whether these or cup-shaped spermatangial receptacles are ancestral within the European species of Osmundea. Additional taxa were required to evaluate the level of phylogenetic informativeness of male receptacle shape. In contrast, the presence vs. absence of secondary pit connections between epidermal cells was found to be a synapomorphy for the two clades of European Osmundea species. Again, however, the phylogenetic significance of these characters could not be fully assessed because species from the North Pacific were not available for inclusion in the analyses.
The North Pacific species of Osmundea show character combinations that differ from any European species: all have urn- (pocket-) shaped spermatangial receptacles, but only Osmundea spectabilis is reported to have some (rare) secondary pit connections between the epidermal cells (Nam et al. 1994; Serio et al. 1999). It was the main purpose, therefore, of this study to widen our previous phylogenetic analysis of Osmundea to include species that are found in Pacific North America, particularly O. spectabilis, in order to clarify the phylogenetic significance and character evolution of secondary pit connections and spermatangial receptacle shape. The gene chosen for analyses was rbcL, as there is data available for a number of European species of Osmundea. In addition, previous studies have shown this gene to give good resolution of red algal relationships at the species to genus level (e.g., Freshwater et al. 1994; Hommersand et al. 1994; McIvor et al. 2001). We have sequenced eight of the fifteen Osmundea species currently recognized: O. blinksii, O. hybrida, O. osmunda, O. pinnatifida, O. ramosissima, O. spectabilis, O. splendens, and O. truncata, with most of the remaining seven [O. crispa (Hollenberg) Nam, O. maggsiana Serio, Cormaci & Furnari, O. multibulba (Dawson, Neushul & Wildman) Nam, O. pelagiensis (Cormaci, Furnari & Serio) Furnari, O. pelagosae (Schiffner) Nam, O. sinicola (Setchell & Gardner) Nam, and O. verlaquei (Cormaci, Furnari & Serio) Furnari] being of relatively restricted occurrence.
In addition to this, we wished to clarify the nomenclatural position of one of the two European species confused until recently under the name Osmundea truncata (Kützing) Nam & Maggs. The second of the sister species was referred to O. ramosissima by Nam et al. (2000), who recognized Osmundea ramosissima Athanasiadis (1996, p. 119) as a valid name for the polynomial Fucus ramosissimus ... Oeder (1766, pl. 276). Since then it has been suggested to us (P. Silva, personal communication) that if the Oeder name is not accepted as a binomial (as explained in Nam et al. 2000), then Osmundea ramosissima Athanasiadis must be considered the name of a new species rather than a new name for a previously published illegitimate name. As such, it is invalidly published since Athanasiadis (1996) did not designate a holotype and Oeder's illustration cannot serve as holotype because the taxon to which it applies can be permanently preserved. One of the objectives of the present study was therefore to determine whether a valid name could be found for this species.
Permanent microscope preparations were made from silica gel-preserved samples by soaking them overnight in dilute aniline blue in 4% Formalin-seawater, then mounting in 10% Karo corn syrup and acidifying with a drop of hydrochloric acid. The presence or absence of secondary pit connections in between outer epidermal cells was determined by skimming the surface off the specimen with a hand-held razor blade, and mounting the stained material with the epidermis uppermost in Karo corn syrup.
Potential lectotype material of Fucus oederi Gunnerus was sought in Kongelige Norske Videnskabers Selskab Museet, Trondheim (TRH) by correspondence with the algal curator.
For PCR amplification, a PTC-200 DNA Engine (MJ Research Inc.) was used. All PCR amplifications were carried out using the previously published primers rbcLFC as the forward primer, and rbcLRD as the reverse primer (Nam et al. 2000; McIvor et al. 2001), and all reactions contained 200 µM each of dATP, dCTP, dGTP and dTTP, 0.3 µM of each primer, 2.5 mM MgCl2, and 1.6 units of Taq polymerase (Bioline). The PCR cycle used was as previously indicated (Nam et al. 2000; McIvor et al. 2001).
About 1250 base pairs of the rbcL gene were amplified using rbcLFC and rbcLRD. The success of the PCR reactions was confirmed by running products on a 1% Tris-Acetic acid EDTA agarose gel, stained with ethidium bromide, and visualized under UV light. The PCR fragments for sequencing were then purified using the High Pure PCR Product Purification Kit (Roche Diagnostics Ltd. Lewes, UK), according to the manufacturer's instructions. The PCR products were then directly sequenced commercially by MWG-Biotech, Ebersberg, Germany.
Fucus oederi Gunnerus (non Esper) was validly described by Gunnerus (1772, p. 100) who considered that his species corresponded perfectly to Oeder's (1766) invalidly described taxon. In fact, he cited both Oeder's polynomial “Fucus ramosissimus, ramis vagis...” and plate (Oeder 1766, p. 7, pl. 276). Oeder's plate is therefore the type of F. oederi. Since such a type was demonstrated to be ambiguous by Nam et al. (2000), the epitype there designated (Nam et al. 2000, fig. 30) for F. ramosissimus Oeder nom. inval., is also the epitype of F. oederi. F. oederi is available for transfer to Osmundea. Osmundea ramosissima Athanasiadis nom. inval. should be considered as a synonym of that species.
Parsimony analysis (not shown) resulted in three most parsimonious trees of length 944 steps (consistency index = 0.607, retention index = 0.634, and homoplasy index = 0.393). These varied only in the positioning of the O. spectabilis samples relative to each other, and were congruent with the results from the ML analysis (Fig. 1). The tribe Laurencieae, represented by Laurencia, Chondrophycus and Osmundea, was well supported in all analyses [87-94% bootstrap proportions (BP); Bremer decay index (DI)=6]. The type species of Laurencia, L. obtusa, was placed as a sister group to Chondrophycus papillosus, although the support was weak (51% to 62% BP; DI=3), and the NJ bootstrap analysis failed to resolve this node. Both taxa were positioned basally within the Laurencieae with strong support (100% BP; DI=6). All species currently placed in the genus Osmundea formed a monophyletic group with robust (100% BP; DI=17) support in all analyses. Within Osmundea, three clades representing different geographical areas were oberved (North Pacific, Atlantic Europe, Atlantic Europe and Mediterranean), with robust (91% to 100% BP; DI=6 to 11) support.
The North Pacific species of Osmundea were robustly grouped together (100% BP support; DI=11). The three samples of O. spectabilis were resolved into a group with 100% support, and a decay index of 15. The remaining North Pacific species, O. blinksii and O. splendens, were also grouped together with 100% BP support in all analyses. MP and ML analyses placed both European groups of Osmundea together, although bootstrap support for this clade was lacking, and the decay index was low (DI=1). The grouping of O. truncata with O. oederi was well supported (91-100% BP; DI=7), although support for the positions of the remaining European species, O. pinnatifida, O. osmunda, and O. hybrida varied (77-99% BP; DI=1 to 6), and was lacking in some analyses. MP and NJ bootstrap analyses failed to resolve O. osmunda and O. hybrida as sister species, although the support from the ML analysis was good (83%).
A consensus parsimony tree (Fig. 2) with important character states (presence or absence of secondary pit connections and shape of spermatangial receptacles) mapped onto it showed a single loss of secondary pit connections, in the branch leading to the Atlantic Osmundea clade. In contrast, the distribution of urn-shaped vs. cup-shaped spermatangial receptacles showed clear evidence of character reversal, because the clade containing O. hybrida was robustly separated (Fig. 1) from the O. truncata/O. oederi clade, which shares the cup-shaped character state.
The inclusion of Osmundea species from the North East Pacific has permitted us to reassess the utility and phylogenetic significance of two important morphological characters. Nam et al. (2000) found that European species of Osmundea separated into two groups based on the presence or absence of secondary pit connections. In contrast, however, the shape of the male spermatangial receptacle was shown to be homoplasious, and two contrasting theories were put forward. It was suggested that either (a) a character reversal from urn-shaped to cup-shaped receptacles had occurred (in O. hybrida), or (b) cup-shaped spermatangial receptacles were ancestral within European species of Osmundea, and that urn-shaped spermatangial receptacles had evolved from them. The second hypothesis was favoured, although it was acknowledged that in the absence of samples from Pacific North America this hypothesis could not be fully tested.
The Californian species show different character state combinations from European species. All species have urn-shaped spermatangial receptacles, but whereas European species with this type of receptacle (O. pinnatifida, O. osmunda) lack secondary pit connections between epidermal calls, we have shown that they are present in all of the Californian species examined. Previously they had been observed only in O. spectabilis (Nam et al. 1994) but were thought to be absent in the other species (Setchell and Gardner 1924; Saito 1969). On European coasts, as found previously (Nam et al. 2000), there are two well-supported groups of Osmundea, with one clade, O. truncata +O. oederi, having secondary pit connections, and the other lacking them. It therefore seems that the presence vs. absence of secondary pit connections is a reliable indicator of phylogenetic position within the genus. Although secondary pit connections are difficult to detect, the shape and size of the epidermal cells may serve as a proxy for routine identification purposes. There has been a single change, from the ancestral condition in which secondary pit connections are present, in the O. pinnatifida clade. The reduction of their frequency in the Californian clade indicates the evolution of a further loss of this character.
The genus Chondrophycus is also heterogeneous for this character, containing species in which secondary pit connections are absent (the majority of species), present, or sporadic (Nam 1999). We can speculate that the ancestral condition in this genus, as in Osmundea, is the presence of pit connections, and that there has been an independent loss in this genus. Why there should be an evolutionary pressure for loss of secondary pit connections in the whole lineage is difficult to understand. The function of primary pit connections, let alone secondary ones, is very poorly known in red algae (Pueschel 1990). Secondary pit connections are normally formed only when cells are closely contiguous because cell fusions are involved (Maggs and Cheney 1990). Any functional role epidermal secondary pit connections may have is further obscured by the formation of secondary pit connections between the epidermal cells and the underlying medullary cells; these may be an adequate substitute.
All species of Osmundea currently found in Pacific North America have urn-shaped spermatangial receptacles, and it seems that this character is fixed within this group of species. However, in the European groups this is not the case. The shape of the spermatangial receptacles is fixed within species (Maggs and Hommersand 1993). The inclusion of Pacific North American species in the analysis has demonstrated that spermatangial receptacle shape has undergone a number of morphological changes between species in European Osmundea.
Osmundea is unique within the Laurencia complex as the only genus that shows a disjunct distribution (Serio et al. 1999), occurring on Pacific coasts of North America and Atlantic and Mediterranean coasts of Europe and North Africa. The species of Osmundea included in the present analysis formed three well-supported groups corresponding to their geographical distribution. Our analyses do not support any of the previous subdivisions of the genus Laurencia sensu lato based on morphology, except for recognition of the genus Osmundea (Nam et al. 1994) and the informal “Spectabilis group” proposed by Saito (1969) for six Californian species of Osmundea (as Laurencia), O. spectabilis, O. blinksii, O. multibulba, O. crispa, O. sinicola, and O. splendens. Neither the invalid subgenus Saitoa nor its emended section Pinnatifidae proposed by Furnari and Serio (1993a) is supported as they contain both Pacific and Atlantic species. We concur with them, however, in their conclusions regarding Saito's (1982) erroneous emendation of section Pinnatifidae.
This analysis shows that there has been a clear divergence between the Californian clade and the Atlantic and Mediterranean species. However, there is little support for the node joining the two European clades, and it therefore appears that these two groups may have diverged rapidly over a short evolutionary period of time. The high diversity of Osmundea species within the Mediterranean may be evidence that they are actively evolving. To date nine species have been recorded from the Mediterranean and the North-east Atlantic, several of which are separated by small morphological differences. On morphological grounds, both O. maggsiana and O. pelagiensis are likely to be closely related to O. hybrida, as they lack secondary pit connections and have cup-shaped spermatangial receptacles. Since both appear to be endemic to the Mediterranean (Furnari et al. 2001), with very narrow distributions, it is highly likely that these species share a recent common ancestor with O. hybrida. Another Mediterranean species, O. pelagosae, is reported to share the character combination of the Pacific North American taxa. It has secondary pit connections, and spermatangial receptacles described as urn-shaped (Furnari et al. 2001), a character combination not found in European species of Osmundea. However, the unbranched spermatangial filaments in O. pelagosae are very different from those of the Californian species (Furnari and Serio 1993b); the spermatangial receptacles appear to be more like the deepened and marginally inrolled receptacles of O. truncata. O. verlaquei, the final Mediterranean species, has secondary pit connections and slightly sunken cup-shaped spermatangial receptacles that are somewhat intermediate between cup-shaped and urn-shaped (Cormaci et al. 1994; Furnari et al. 2001), suggesting that this species is also closely related to O. truncata.
In conclusion, the Californian clade and the European Osmundea clades exhibit contrasting stability of characters. The Californian species show a fixed spermatangial receptacle shape but variation in the degree of loss of secondary pit connections within the clade. In contrast, among the European species, the shape of the spermatangial receptacle shows character reversals and appears to be actively evolving in Mediterranean species, requiring developmental studies for the elucidation of the path of evolution of this character. Unlike in the Californian clade, the two distinct European groups of Osmundea are clearly divided on the basis of the presence or absence of secondary pit connections, which is fixed within clades. This comparative biogeographical approach to a morphological/molecular phylogenetic investigation has therefore yielded results that will make it more difficult to rely solely on morphological characters in infrageneric classification of Osmundea.
The disjunct distribution of Osmundea in warm-temperate regions of the Northern Hemisphere with nine species recorded from the eastern North Atlantic Ocean and Mediterranean Sea and six species from the eastern North Pacific Ocean in Pacific North America (Serio et al. 1999; Nam et al. 2000) corresponds to a late Tethyan distribution pattern. As Hommersand (2003) states:
“For a phylogenetic line to be identified as a component of a Tethyan flora it should contain elements in the Indian Ocean and also in Western Europe and the Mediterranean Sea. More often than not related taxa also occur in the western Atlantic Ocean and Caribbean Sea and certain tropical and warm-temperate regions of the eastern Pacific along the coast of North and South America. In some instances there may be a disjunct distribution with species in Europe and warm-temperate Pacific North America that are presently absent in the western Atlantic Ocean.”
The only example of a Tethyan distribution to have been documented with molecular data is that of the red algal family Solieriaceae with representatives in the Indo-W-Pacific, Atlantic, and eastern Pacific Oceans (Fredericq et al. 1999). A basal group was identified in this study that contained genera and species found primarily in Australia that were previously placed in the Solieriaceae but which have since been transferred to the Areschougiaceae (Chiovitti et al. 1998). Examples of families having a disjunct distribution between the eastern Atlantic and the eastern Pacific Oceans include the Cystoseiraceae with Cystoseira and Halidrys (Hommersand 2003), and possibly the Fucaceae with Ascophyllum and Pelvetia in the North Atlantic and Hesperophycus, Pelvetiopsis, and Silvetia in the North Pacific (Serrão et al. 1999); Fucus, however, was shown to have a more recent boreal distribution in the North Atlantic and North Pacific Oceans. For Osmundea to fit a Tethyan distribution pattern the genus would likely have originated in the western Tethyan Ocean after effective closure of the Tethyan Seaway between 60Ma and 20Ma during Oligocene or Miocene. The apparent absence of species of Osmundea in the tropical western Atlantic and Indo-W-Pacific Oceans suggest that none of the species have adapted to the environment of the modern tropics. For example, in contrast to species of Laurencia and Chondrophycus, those of Osmundea may not have evolved secondary metabolites that would render them resistant to predation by tropical herbivorous fish (see Hommersand 1990).