| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics and Phytogeography |
Department of Biology, Duke University, Durham, North Carolina 27708 USA
Received for publication 11 February 2008. Accepted for publication 1 April 2008.
ABSTRACT
A seemingly obvious but sometimes overlooked premise of any evolutionary analysis is delineating the group of taxa under study. This is especially problematic in some bryophyte groups because of morphological simplicity and convergence. This research applies information from nucleotide sequences for eight plastid and nuclear loci to delineate a group of northern hemisphere peat moss species, the so-called Sphagnum subsecundum complex, which includes species known to be gametophytically haploid or diploid (i.e., sporophytically diploid-tetraploid). Despite the fact that S. subsecundum and several species in the complex have been attributed disjunct ranges that include all major continents, phylogenetic analyses suggest that the group is actually restricted to Europe and eastern North America. Plants from western North America, from California to Alaska, which are morphologically similar to species of the S. subsecundum complex in eastern N. America and Europe, actually belong to a different deep clade within Sphagnum section Subsecunda. One species often considered part of the S. subsecundum complex, S. contortum, likely has a reticulate history involving species in the two deepest clades within section Subsecunda. Nucleotide sequences have a strong geographic structure across the section Subsecunda, but shallow tip clades suggest repeated long-distance dispersal in the section as well.
Key Words: long-distance dispersal • northern hemisphere biogeography • Pacific rim biogeography • Sphagnum • Sphagnum subsecundum
For those who study evolutionary processes in a group of related organisms, a crucial and seemingly obvious issue is defining the focal group for the research. In studies of speciation, for example, incorrectly assuming sister group relationships might lead to the formulation of inappropriate and positively misleading hypotheses. Along those lines, the current study was motivated by the need to circumscribe a group of closely related species in the moss genus Sphagnum L. (peat mosses). This group of species, sometimes referred to as the "S. subsecundum complex," includes several haploid-diploid [gametophyte-sporophyte] and several diploid-tetraploid taxa that have been recognized as different species (e.g., Russow, 1894
; Flatberg, 2002), varieties (e.g., Hill, 1975; Eddy, 1977b; Crum, 1984) or as forms of a single polymorphic but indivisible species (e.g., Andrews, 1913). The genealogical history of species in the S. subsecundum complex will be the subject of future papers; the present one is focused on delimitation of the S. subsecundum complex itself within the broader phylogenetic context of peat moss species.
The perennial and free-living stage of the moss life cycle, which is conspicuous in nature, is the gametophyte. Gametophytes are typically haploid and at maturity produce male and female gametangia (archegonia containing eggs and antheridia containing sperm). Gametes are of course formed mitotically because the gametophytes are already haploid. Gamete fusion initiates the sporophyte generation, which in Sphagnum consists of little more than a sporangium where meiotic spores are produced (the capsule), elevated on a stalk or pseudopodium that is part of the maternal gametophyte. Only two chromosome numbers are known in Sphagnum; x = 19 and 38 (Fritsch, 1991). Interspecific hybridization has been documented between closely related taxa of Sphagnum section Acutifolia (Cronberg and Natcheva, 2002; Natcheva and Cronberg, 2007), as well as between more distantly related species from different sections of the genus (Shaw and Goffinet, 2000). Within the S. subsecundum complex, several species are reported to have haploid gametophytes (N = 19), whereas others have diploid gametophytes (N = 38) (Newton, 1993). Allopolyploid origins for the diploid species of this group have been proposed (Maass and Harvey, 1973; Eddy, 1977b; Melosik et al., 2005), and allopolyploid speciation has been documented in other groups of peat mosses (Cronberg, 1996; Såstad et al., 1999, 2000, 2001). However, little solid evidence for evaluating relationships in the group is available, and one of the problems in resolving ancestor–descendent relationships is defining the phylogenetic and geographic limits of the group itself.
Three haploid species in the S. subsecundum complex are reported from Europe and North America: S. subsecundum Nees, S. contortum Schultz, and S. platyphyllum (Lindb.) Warnst. In addition, four gametophytically diploid species that have generally been associated with those haploids have also been described: S. auriculatum Schimp. (Europe), S. lescurii Sull. (N. America), S. carolinianum Andrus (N. America), and S. inundatum Russ. (N. America and Europe). All three of the haploid species are reported to be more or less widespread throughout the northern part of the Northern Hemisphere in Asia, eastern and western North America, and Europe. Precise distributional ranges are poorly understood because of persistent disagreement about how to classify morphological forms such as species and varieties and how to sort plants into distinct morphotypes. Nevertheless, the haploid species have been reported from very broad geographic ranges. In the case of S. subsecundum, in particular, literature reports have placed this species in North and South America (Crum, 1984), Africa (Eddy, 1977a, Eddy, 1985), southeast Asia (Eddy, 1977a), New Guinea (Eddy, 1977a), and New Zealand (Sainsbury, 1955). The problem of resolving phylogenetic relationships among the species and determining ancestry of the diploids is significantly complicated if members of the S. subsecundum complex occur across such broad ranges. If the diploids are of hybrid allopolyploid origin, which haploid taxa are likely progenitors? When using molecular markers to test hypotheses about parentage, is worldwide sampling required to account for genetic diversity?
Sphagnum subsecundum and related species are classified in the section Subsecunda of the genus Sphagnum (Isoviita, 1966). The section Subsecunda is one of four major lineages within Sphagnum resolved by nuclear and plastid DNA sequences (Shaw et al., 2000, 2003). Monophyly of the section is supported by both morphological characters and molecular phylogenetic analyses. This study was undertaken to clearly delimit a "S. subsecundum complex," if such exists as a phylogenetically meaningful entity, within the broader section Subsecunda. Toward that end, a worldwide sampling of plants was included in the analyses, which are based on nucleotide sequences from the nuclear and plastid genomes. With this research, we identified the "actors" in the evolutionary play within which species of the S. subsecundum complex evolved. We discovered strong geographic structure in the section Subsecunda, but also found evidence of repeated long-distance dispersal in this spore-producing group of plants. We also documented a strong conflict between the placement of one taxon in the S. subsecundum complex (S. contortum) in different gene trees and interpret this conflict in terms of previous hybridization between members of two divergent relatively deep clades within the section.
MATERIALS AND METHODS
Taxon sampling
A total of 98 accessions were included in the phylogenetic analyses: 94 samples of species in Sphagnum section Subsecunda plus four outgroup taxa representing other sections of the genus. For outgroup taxa, we included one accession each of S. wulfianum Girg. (section Polyclada), S. teres (Schimp.) Ångström (section Squarrosa), S. squarrosum Crome (section Squarrosa), and S. portoricense Hampe (section Sphagnum). Ingroup accessions represent all major regions of the world with 32 samples from North America, 18 from Central and South America, 19 from Europe, nine from northern Asia (including Russia, China, and Japan), 13 from Africa, and three from Australia/New Zealand. The one area that could benefit from additional sampling is southern tropical Asia between China and Australia. We were unable to obtain recent collections from that region suitable for DNA extraction. It is unlikely (though not impossible) that plants from tropical Asia comprise a monophyletic group not otherwise included in our analyses, but this view needs to be tested in the future.
Most vouchers are housed in DUKE, with a limited number of other collections sampled from H, MICH, MO, and NY, as listed in Appendix 1. DNA sampling required only a small portion of a single capitulum (the cluster of branches at the tops of stems), and the remaining portions of sampled plants were placed in small packets, which were returned to the original herbarium specimen. As many specimens of Sphagnum contain interspecific mixtures, the actual stem from which DNA was obtained can be checked accordingly. DNA voucher labels with our laboratory DNA accession number are affixed to specimens so the herbarium packet can be readily identified as a voucher for this study.
DNA sequencing
We obtained nucleotide sequences from the 98 accessions for eight nuclear and plastid loci (Table 1). In addition to the nrITS region (nuclear ribosomal internal transcribed spacer, including ITS1, ITS2, and the 5.8S gene), our nuclear sequences represent mainly introns and intergenic spacers. These include two introns in the Leafy-Flo gene (LFY1 and LFY 2) and three anonymous loci (RapdA, RapdB, RapdF). Sequences from two plastid loci (trnL, trnG) were also included; both regions include introns, spacers, and short regions of coding sequences. The data set was complete in that we obtained good sequences for all loci from all accessions. Laboratory protocols for DNA extraction, PCR amplification, sequencing, and sequence editing were as described by Shaw et al. (2003).
|
). We used a subjective approach that combined event-based (Morrison, 2006) and similarity criteria (Simmons, 2004). In general, we tried to insert gaps that minimized substitutions, but also assessed the (subjective) probability of multiple short or single-base indels, vs. some level of nucleotide substitution. Gradual improvement of our alignments for Sphagnum has been ongoing as data from additional taxa have accumulated. Some such data are not included in the current study, but the accuracy of the alignment has benefited from research on the genus (e.g., Shaw, 2000; Shaw et al., 2003, 2004) that encompasses both closely related and more distant species (Simmons and Freudenstein, 2003). Although our alignment is subjective, it is explicit because it is available in the TreeBase database (see accessions later).
For each locus, and for the combined data set, the best-fit models of nucleotide substitution were determined by hierarchical likelihood ratio tests with the aid of the program MrModeltest 1.1b (Nylander, 2002) (Table 1). Bayesian analyses (Yang and Rannala, 1997) of each locus were conducted specifying the optimal substitution model for that locus using the program MrBayes 3 (Huelsenbeck and Ronquist, 2001). These analyses were conducted with three runs, each with four chains and 5 000 000 generations, using default, uniform priors. Model parameters including trees were sampled every 500th generation. The number of trees needed to reach stationarity (i.e., the burnin) in the Markov chain Monte Carlo (MCMC) algorithm was estimated by visual inspection of the plot of log likelihood score at each sampling point using Excel (Microsoft, Richmond, Washington, USA). Trees of the burnin for each run were excluded from the tree set. The combined data set was analyzed under a single substitution model (as determined above and specified in Table 1), in so-called homogeneous analyses, and with separate models specified for each locus in heterogeneous analyses. Maximum likelihood (ML) analyses (Felsenstein, 1981) were also conducted on the combined data set using the program GARLI (Zwickl, 2006) with the same substitution model as specified for the homogeneous Bayesian analyses, with 300 generations. The ML analyses were run twice, as recommended in the GARLI manual.
Branch support was assessed based on Bayesian posterior probability values plus ML and maximum parsimony (MP) bootstrap analyses (Felsenstein, 1985). Bayesian analyses were run on each locus separately as well as on the combined eight-locus matrix. MP and ML analyses bootstrap analyses were run only on the combined matrix (with and without S. contortum in the data set; described later). MP bootstrap analyses were run with 300 replicates, each with 10 random addition replicates with tree-bisection-reconnection (TBR) branch swapping. ML bootstrap analyses were run with 300 random addition replicates as described by Zwickl (2006).
Trees are presented without species identifications for two reasons. One is that geographic patterns are the focus of this paper. The second, and perhaps most important, is that the taxonomy of species in section Subsecunda is very poorly understood, and many names from regions outside the northern continents should be viewed with extreme skepticism because critical taxonomic revisions have not been completed and species delimitation in these aquatic mosses is especially difficult because of extreme phenotypic variation. For samples collected in areas outside North America and Europe, species identifications provided in Appendix 1 are those borne on the herbarium specimen. Although probably subject to future change in some cases, these names serve as "placeholders" until a worldwide revision of Sphagnum section Subsecunda is completed. Species names, locality and collector data, and GenBank accession numbers are provided for all collections in Appendix 1. The data matrix on which phylogenetic analyses were conducted is available as TreeBASE (http://www.treebase.org) accession number S2016.
RESULTS
The combined eight-locus data set included 5314 nucleotide sites. Forty-three sites were excluded from the analyses because of alignment ambiguity, so the final data set included 5271 sites. The length of each locus, the numbers of variable and parsimony-informative sites, and the optimal substitution model for each locus and the combined data set are provided in Table 1. As in previous analyses of Sphagnum phylogeny, the three anonymous nuclear loci (RapdA, RapdB, and RapdF) were highly informative. The results of Bayesian (homogeneous and heterogeneous), ML, and MP analyses were similar, and the combined trees (Figs. 1, 4) show the topology recovered from homogeneous analyses using the overall best substitution model, namely, the GTR+G model. Major lineages, the focus of this study, were identical using the different analytical methods. Any branch that was well supported using Bayesian, MP, or ML methods is shown in boldface. For the Bayesian analyses, significant support was defined as a posterior probability of 
0.95 and for the MP and ML analyses, "significant" bootstrap support was defined as percentages of 70%. Trees obtained from single locus and combined analyses under Bayesian inference, MP, and ML, with support values, are provided in online Appendices S1–S16 (see Supplemental Data with the online version of this article). A list and brief explanation of phylogenetic trees included as Appendices S1–S16 is provided as online Appendix S17.
|
|
Sister to the Pacific Rim plants is a large clade that is designated the Atlantic clade to distinguish it from the former, although it is in fact much more broadly distributed than along Atlantic regions. Three major groups are resolved in the Atlantic clade, though support is in some cases weak (Fig. 1). One, lineage A, includes a group of species that are highly distinctive morphologically: S. macrophyllum Bernh. ex Brid., S. cribrosum Lindb., S. cyclophyllum Sull & Lesq., and S. pylaesii Brid. On the basis of their unique gametophytic morphology, S. macrophyllum and S. cribrosum have previously been segregated as Sphagnum section Isocladus, and S. pylaesii has been segregated as section Hemitheca (Shaw et al., 2004). The present results support previous inferences (Shaw et al., 2004) that despite their unique morphologies they are nevertheless nested within the section Subsecunda and are closely related. This group of species is distributed throughout the eastern part of North America and include characteristic peat mosses of the Atlantic Coastal Plain. Sphagnum cribrosum is endemic to that region, whereas S. macrophyllum has been found at one disjunct location in Honduras (Crum, 1984), S. pylaesii is disjunct to a small area along the border of northern Spain and southern France, and S. cyclophyllum has been attributed to a few localities in northern South America (H. A. Crum, deceased, University of Michigan, unpublished Flora Neotropica manuscript). Recent material of the S. macrophyllum Honduras population and South American populations of S. cyclophyllum were not available for DNA sampling, but the present results corroborate previous analyses that resolve European and American populations of S. pylaesii as sister groups (boxes in Figs. 1, 4). A Japanese population of S. guwassanense Warnst., a species morphologically similar to S. cyclophyllum, is resolved without support as sister to the rest of lineage A, as in previous analyses (Shaw et al., 2004).
Lineage B within the inclusive Atlantic clade is resolved as sister to lineages C plus D, but this relationship lacks significant support from Bayesian, MP, or ML analyses (Fig. 1). Nevertheless, individual monophyly of lineages B and D are well supported. Lineage B includes samples from tropical and southern hemisphere populations from Africa, South America, and the Australia/New Zealand region, whereas lineage D is limited to South America and Africa. Lineage C is resolved without significant support as sister to lineage D. Lineage C includes samples from Europe and eastern North America, including S. subsecundum and the putative polyploid species, S. auriculatum, S. lescurii, S. inundatum, and S. carolinianum. All samples of the polyploids fall into this lineage, and samples from the New World have a sister group relationship to samples from Europe (Fig. 1). Accessions from western North America (for example, California, Alaska) that had been previously determined (as herbarium specimens) as S. inundatum, S. lescurii, or S. subsecundum were invariably resolved as members of the Pacific Rim clade. Reexamination of the voucher specimens showed that they fall within the range of morphological variation encompassed by California "S. subsecundum," which appears to be more variable than S. subsecundum of eastern North America and Europe. Evidently, lineage C within the Atlantic clade contains the "real" S. subsecundum (as defined by its European type specimen) and the polyploids generally associated with it. Interestingly, the analysis resolves European and New World samples as reciprocally monophyletic (Fig. 1; curved brackets), although only the European clade is well supported. The New World group includes two samples of S. lescurii from Belize (Central America; one was identified as S. trirameum Crum in MICH). These two collections share 18 synapomorphic substitutions that unite them in a single clade, but also differ from one another by more than 20 additional substitutions. They may represent a single dispersal to tropical America, but judging from the degree of differentiation between them, the event may well have been long ago.
All samples of S. contortum, one of the haploid species generally considered part of the S. subsecundum complex, is resolved as monophyletic (middle of Fig. 1). Its relationship to other species in the section is, however, not resolved, except that it is placed by the combined analysis nested within the inclusive Atlantic clade with strong support. Separate analyses of individual loci suggest a reason for its unresolved position in the combined reconstruction. On the basis of the nrITS data alone, S. contortum is resolved within the strongly supported Atlantic clade (Fig. 2). RapdA sequences alone, in contrast, resolve S. contortum within the Pacific Rim clade, also with strong support (Fig. 3).
|
|
Despite phylogenetic evidence for generally strong geographic structure within the section Subsecunda, many geographically cohesive clades contain disjunct accessions from other geographic regions (Fig. 4). Thus the Pacific Rim clade includes three samples of S. subsecundum from Italy and one from Newfoundland. These relationships are supported by all single-locus analyses (not shown), as well as by indel characters (discussed later). Several additional samples from Italy, not included in the final analyses presented here because of missing data due to nonamplification of one or more loci, also group with these samples within the Pacific Rim clade (A. J. Shaw, S. Boles, and B. Shaw, unpublished data). Sphagnum platyphyllum is strongly supported as a member of the Pacific Rim clade, whether plants were sampled from northwestern North America, eastern North America, or Europe. This species appears to bear a genealogical connection to species in the Pacific Rim clade, but it is widespread around the northern hemisphere. Within lineage A of the inclusive Atlantic clade, S. pylaesii is found in both eastern North America and Europe, and North American plants have a close phylogenetic relationship to samples from South America.
The two lineages (B and D) that include plants from tropical and southern hemisphere latitudes contain populations that appear to bear close genealogical relationships but are disjunct on different continents. Thus, in lineage B, samples from Colombia and Kenya are resolved as sister groups, although in this case the relationship is not significantly supported. The two samples share 23 synapomorphic substitutions, but also differ by more than 60 additional substitutions. In lineage D, two strongly supported clades each include populations that originated in Africa and South America. These populations, as with previous examples, share synapomorphic nucleotide substitutions but are also differentiated by numerous additional substitutions, suggesting sufficient time to accumulate differential mutations since any intercontinental dispersal.
Indels (insertion/deletions) were not scored for the phylogenetic analyses in order to assess relationships implied by nucleotide substitutions alone. We then examined the distributions of coded indels to determine which, if any, support inferred relationships. The determination of whether an indel represents a deletion or insertion was made with reference to outgroup sequences. A four-base insertion in the ITS locus is shared by a well-supported clade within lineage D comprised of three accessions from the Republic of Guinea, Tanzania, and Burundi (Figs. 1, 4). Two Brazilian samples that have a well-supported sister group relationship within lineage D share a five-base insertion in the cpDNA trnG locus, but another Brazilian sample, that is part of a different well-supported clade with a sample from Venezuela, also shares that insertion. In lineage B, four accessions, two from South Africa and two from Malawi, share a four-base insertion in the RapdB locus and a 25-base insertion in the cpDNA trnL locus. These four samples form a well-supported clade, but a fifth sample, from Uganda, that also resolved with strong support as part of that clade by nucleotide substitutions, does not have either insertion. All samples resolved as lineage B by nucleotide substitutions share an 18-base insertion in the RapdB locus. Similarly, all samples comprising the Pacific Rim clade share a five-base deletion in the RapdB locus. An additional sample from India, not included in the formal phylogenetic analyses because of missing data for several loci, also shares this deletion (A. J. Shaw, S. Boles, and B. Shaw, unpublished data). This record extends the distribution of the Pacific rim clade at least as far west in Asia as India. Four accessions within the Pacific Rim clade, one from Alaska, one from Japan, and two from Italy, not resolved as a monophyletic group by nucleotide substitutions alone, share a 17-base deletion in the LFY1 locus.
With regard to the primary goal of this research in defining the S. subsecundum complex for future genetic analyses at the population level, lineage C within the inclusive Atlantic clade contains samples of S. subsecundum from eastern North America and Europe and also contains all samples of the polyploid species, S. auriculatum, S. carolinianum, S. inundatum, and S. lescurii from these regions. Samples of S. subsecundum from other areas identified as that species belong to divergent lineages within the section and do not appear to be related to S. subsecundum sensu stricto. It is noteworthy that S. subsecundum s.s. was sampled from Spain in southern Europe and group with other European plants, but morphologically similar plants from Italy group with accessions belonging to the Pacific Rim clade.
DISCUSSION
Sphagnum subsecundum has been reported from a broad distribution including North and South America, Europe, northern Asia, Africa, and tropical to temperate Australasia (Warnstorf, 1911; Andrews, 1913; Eddy, 1977a; Crum, 1984; Daniels and Eddy, 1985; McQueen and Andrus, 2007). Indeed, Andrews (1913, p. 22) commented that the species "is regarded as cosmopolitan." The distributions of the other two European and North American haploid species generally considered part of the S. subsecundum complex, S. contortum and S. platyphyllum, are poorly documented because of taxonomic problems among the species, but are also generally attributed broad geographic ranges that include at least all the northern continents. Species in the complex that are reported to have diploid gametophytes, namely, S. auriculatum, S. carolinianum, S. inundatum, and S. lescurii, are also taxonomically controversial, but have been reported from eastern and western North America, Europe, and temperate Asia at least as far east as China and Japan.
Our molecular results indicate that there is strong phylogenetic structure within Sphagnum section Subsecunda and that S. subsecundum in the strict sense as typified by European specimens cited in the original description (Nees von Esenbeck, 1819) may well be restricted to Europe and eastern North America. Western North American plants identified as S. subsecundum from California to Alaska and that are indeed morphologically similar to European–eastern North American plants, are in fact more closely related to plants (also generally identified as S. subsecundum) from eastern Asia. This does not appear to be a simple case of cryptic speciation in which distinct populations within a widespread species are genetically differentiated; plants of S. subsecundum from eastern North America and Europe, vs. those from western North America around the Pacific Rim to Asia, are resolved as members of two deeply divergent clades within the section Subsecunda. Either the morphological characteristics used to diagnose S. subsecundum are homoplastic or that morphology is plesiotypic within the section Subsecunda.
The phylogenetic results indicate that taxonomic/nomenclatural changes need to be made with regard to S. subsecundum. Several species that are closely related to S. subsecundum have been described from Asia, especially Japan, including S. subobesum Warnst. and S. miyabeanum Warnst. (Warnstorf, 1911). It is not altogether clear at this time, however, where species boundaries should be drawn among the morphological forms that occur in western North America and eastern Asia, so no taxonomic changes are proposed here. Recognizing western North American plants currently named S. subsecundum as a distinct species relative to eastern North American and European plants (S. subsecundum s.s.) is necessitated by the phylogenetic results, which demonstrate that the two taxa are highly divergent within the section Subsecunda. However, there are no obvious and unambiguous morphological characters that distinguish them, so identification of plants would have to be based largely on geographic considerations. This approach carries with it the significant practical disadvantage that if in fact the eastern North American/European S. subsecundum s.s. does occur in the West occasionally, such new discoveries would be missed because western plants would be considered a different species by definition. Results presented here indicate that episodic long-distance dispersal characterizes Sphagnum section Subsecunda, so the occurrence of unexpected plants disjunct from their main ranges is not impossible and is even likely. Although we were unable to identify clear diagnostic characters to distinguish the Pacific Rim vs. Atlantic S. subsecundum, we have observed that the Pacific plants are substantially more variable in stem leaf morphology, which defines S. subsecundum. Sphagnum subsecundum s.s. of eastern North America and Europe is relatively stenotypic in stem leaf shape, size, and anatomy, whereas the western plants encompass stem leaf morphologies that are more typical of the gametophytically diploid species such as S. lescurii and S. inundatum.
The biogeographic patterns of the two other haploid species of the S. subsecundum complex, S. contortum and S. platyphyllum, diverge from that of S. subsecundum. Sphagnum platyphyllum is known from a circumboreal range, and we have in fact made recent collections from northern Europe, eastern and western North America, and Asian Siberia (specimens in DUKE). In addition, we can verify that plants conforming to the morphology of S. platyphyllum have been collected in Costa Rica (specimens in BING). In contrast to S. subsecundum, all samples of S. platyphyllum are resolved in the Pacific Rim clade of the section Subsecunda, regardless of their provenance. Accessions of S. platypyllum are not resolved as a single monophyletic group, which is surprising given the morphological distinctiveness of this species, but samples from widely disjunct localities, including eastern and western North America and northern Europe, all fall within the Pacific Rim clade.
Sphagnum contortum does, in contrast, form a monophyletic (exclusive) group, including populations from eastern North America and Scandinavia. Monophyly of the species is well supported, but its placement within the section Subsecunda is unresolved in the analysis of the combined eight-locus data set. Separate analyses of individual loci give strong evidence of reticulation in its history, with nrITS, for example, providing solid evidence of a relationship to taxa in the Pacific Rim clade, whereas RapdA sequences provide equally strong evidence of a relationship to species in the Atlantic clade. Hybridization rather than incomplete lineage sorting seems to be the most likely explanation for the incongruence among loci, and it is possible that S. contortum may have been involved in past hybridization. Sphagnum contortum and S. platyphyllum share alleles at highly polymorphic microsatellite markers (A. J. Shaw, S. Boles, and B. Shaw, unpublished data); nevertheless, additional sampling is required to resolve the ancestry of S. contortum. It is an intriguing possibility that the nonmonophyly we observed for accessions of S. platyphyllum could reflect genetic heterogeneity resulting from past hybridization with S. contortum.
Phylogenetic results presented here are critical to ongoing research aimed at resolving genealogical relationships among gametophytically haploid and diploid taxa in the S. subsecundum complex. All the diploid accessions that were included in these analyses, including all four species (S. auriculatum,S. carolinianum,S. inundatum,S. lescurii), were resolved in a single lineage ("C" in Figs. 1, 4) within the broader Atlantic clade. Only one of the three haploid species (S. contortum,S. platyphyllum,S. subsecundum) that was included in that clade is S. subsecundum. Independent data from both microsatellite markers and nucleotide sequences suggest a close relationship between the diploid species and S. subsecundum (A. J. Shaw, S. Boles, and B. Shaw, unpublished data). Phylogenetic patterns presented here provide strong support for the hypothesis that the genome(s) of the diploid species is(are) related to that of S. subsecundum and not to those of S. contortum or S. platyphyllum. Moreover, only S. subsecundum in the strict sense, from eastern North America and Europe, is implicated in the genealogical history of the diploid species. It appears that plants commonly referred to as S. subsecundum from western North America and eastern Asia are unrelated to the diploid species in eastern North America and Europe.
One of the diploid species, S. inundatum, was attributed to Alaska in the recently published Flora of North America treatment of Sphagnum by McQueen and Andrus (2007). Alaskan specimens examined during the course of the present work, including samples annotated as S. inundatum by R. E. Andrus (in BING), proved to be the western North American ‘S. subsecundum’ (i.e., accessions associated with that name from western North America; not conspecific with the "real" S. subsecundum of eastern North America and Europe). Western North American plants of ‘S. subecundum’ are more phenotypically variable than eastern North American–European plants of S. subsecundum s.s. and include stem leaf morphologies that are often associated with several of the diploid species, including both S. inundatum and S. lescurii. Both S. inundatum and the European diploid species, S. auriculatum, have been reported from Japan and China (the latter as S. denticulatum Brid.). We cannot confirm these reports at present, and based on our examination of specimens, the diploid species may be restricted to eastern North America and Europe. If one or more of the species does occur in eastern Asia, it will be worth investigating, utilizing genetic data, whether they share a common origin with morphologically similar plants in eastern North America and Europe. Unpublished DNA sequence and microsatellite data unambiguously show that S. inundatum originated independently in North America and Europe (M. Ricca, Duke University and A. J. Shaw, unpublished data). Those results are consistent with current analyses, which resolve North American and European accessions in the S. subsecundum complex as reciprocally monophyletic.
The primary purpose of this study was to delimit the S. subsecundum complex of haploid and diploid species so that more intensive genetic analyses can focus on the correct group of taxa in the appropriate geographic region, but the research has added value in elucidating biogeographic patterns in the section Subsecunda. Analyses presented here show a strong geographic structure at several nested levels within the section Subsecunda, but also suggest repeated dispersals among major continental areas. Ideally, biogeographic analyses should include three components of evidence: phylogenetic hypotheses, inferences of ancestral areas and directions of geographic movements, and age estimates (Donoghue and Smith, 2004). Obtaining age estimates for calibrating bryophyte phyogenies is problematic because this group of plants has a very poor fossil record (Krassilov and Schuster, 1984; Miller, 1984). At the species level, Wall (2005) used lineages endemic to oceanic islands of known age to estimate maximum dates for nodes in a phylogenetic analysis of the moss genus Mitthyridium H. Rob. Hartmann et al. (2006) dated the origin of the liverwort genus Bryopteris (Nees) Lindenb. based on fossils preserved in amber. The only other dating exercises that have been conducted for mosses or liverworts were based on estimates for the origins of land plants and the early divergence of the bryophyte lineages from tracheophytes. With such general and limited calibration of moss and liverwort phylogenies, age estimates for orders, families, and genera of mosses by extrapolation must be viewed with extreme caution.
Phylogenetic reconstructions of land plants suggest that mosses diverged from other land plants before the diversification of major tracheophyte lineages, and reconstructions for the mosses suggest that Sphagnum was one of, if not the, earliest diverging lineage of mosses (Shaw and Renzaglia, 2004). As such, Sphagnum appears to be an old genus of land plants. Nevertheless, age estimates for Sphagnum lineages are currently out of reach. The (relatively) early divergence of Sphagnum section Subsecunda into the Pacific Rim and Atlantic clades thus cannot be confidently dated. It is noteworthy, nonetheless, that the present investigation documents a closer relationship between western North America and eastern Asia than between eastern North America and eastern Asia. Most botanical studies, in contrast, have corroborated close phylogenetic connections between eastern North American–eastern Asian plants (Wen, 1999; Manos and Donoghue, 2001). Donoghue and Smith (2004) found relatively few close phytogeographic connections between western North America and eastern Asia, a biogeographic pattern that appears to be more common in animals than in plants (Sanmartín et al., 2001). A close phylogenetic connection was documented for Chinese and western North American collections of the fungus Flammula velutipes (Curtis) Singer, but several Argentinian accessions were also included in the same clade (Petersen and Hughes, 2007). Like peat mosses, these (and other) spore-producing fungi are likely able to disperse widely, at least on occasion. Most northern hemisphere sphagna, at the sectional and even specific level, are widespread across North America, Europe, and Asia, so any choice of outgroup(s) for the section Subsecunda phylogeny would be ambiguous with regard to ancestral range and directionality of movement, from Pacific to Atlantic or visa versa. We can only say at this point that there appears to have been an early vicariance event separating the Pacific Rim and Atlantic clades.
A surprising result of the phylogenetic analyses presented here is that three samples of S. subsecundum from Italy were resolved as part of the Pacific Rim clade rather than the Atlantic clade, which includes all other European samples of this species including two from Spain. Several indel characters also support this relationship. The three Italian collections are well nested within the Pacific Rim clade. A single specimen of S. subsecundum from Newfoundland also grouped with the Pacific Rim clade. These surprising occurrences that are exceptions to an otherwise clear pattern strongly suggest occasional long-distance dispersal. In fact, the three Italian samples do not form a monophyletic group, leaving open the question as to whether there was a single dispersal with subsequent differentiation (including homoplasy, obscuring their monophyletic origin), or multiple dispersals. The possibility of a relictual occurrence in Italy cannot be eliminated.
Within the Atlantic clade, there is an early divergence, albeit without strong support, between a lineage ("A": in Fig. 1) of plants characteristic of eastern North America and the rest of the Atlantic clade. This eastern North American lineage itself has evidence of both vicariant genetic differentiation and dispersal. Four species, each morphologically distinctive with respect to all other sphagna, as well as to each other, are represented in this lineage. Two of them have been segregated in their own monotypic or small sections of Sphagnum (Shaw et al., 2004). Together they form a well-supported monophyletic group, and although the species are widespread in eastern North America and only common there, several are found disjunctively on other continents. One, S. pylaesii (sometimes classified in the monotypic section Hemitheca), is found in eastern North America, where it occurs from North Carolina north to Newfoundland and is fairly common north of New Jersey, is also found disjunctively in a small area in southern France near the Spanish border. Previous analyses (Shaw et al., 2004) showed that the European and (North and South) American plants are reciprocally monophyletic and that the American plants contain substantially higher nucleotide polymorphism than do European plants. The difference in nucleotide diversity is evident in Fig. 4 (brackets). This pattern is consistent with a dispersal event from North America to Europe with an associated population bottleneck. The species also occurs sporadically in South America, and a lack of genetic differentiation between a sample from North vs. several from South America further suggests long-distance dispersal. Another species in lineage A, S. macrophyllum, is endemic to eastern North America but has been found once in Honduras, providing yet additional inferential evidence of dispersal.
Two lineages within the Atlantic clade are comprised of samples from tropical and southern hemisphere regions. One (lineage B) includes samples from Africa, South America, and Australia/New Zealand, whereas the other (lineage D) includes plants from South America and Africa only. Both lineages are well supported, suggesting that from temperate/boreal regions of the northern hemisphere, there were two independent range expansions southward. The present results corroborate close biogeographic relationships between South America and Africa documented in other mosses and liverworts (e.g., Heinrichs et al., 2006). Similar biogeographic relationships are also known in flowering plants, but in bryophytes, individual species often occur disjunctively on the two continents. In Sphagnum, well-supported sister group relationships occur between African and South American plants in both lineages B and D (for example, between accessions from Colombia and Kenya in lineage B and between Brazil and South Africa in lineage D) (Figs. 1, 4). These accessions are well nested within their respective lineages, suggesting relatively recent dispersal with very limited intercontinental divergence, rather than deep vicariance. Within lineage B, three accessions from Australia and New Zealand form a well-supported monophyletic group, suggesting the possibility that this region was colonized only once. Additional sampling is necessary, however, to test this hypothesis.
Conclusions
The main conclusion of this study is that the Sphagnum subsecundum complex of gametophytically haploid and diploid taxa appears to be endemic to eastern North America and Europe. Phylogenetic data are crucial for delimiting the cast of actors in the evolutionary play that has featured recurrent allopolyploidy, especially in this group of structurally simple plants where diagnostic morphological characters are few and may be subject to homoplasy. This is clearly exemplified by the fact that S. subsecundum has been reported from eastern and western North America and Europe, yet this broad geographic range of morphologically similar plants belies phylogenetic complexity. The search for ancestors to the eastern North American–European allopolyploids apparently need not involve sampling from western North America and Asia because these plants are only distantly related to the polyploids despite their morphological similarity. The phylogenetic results presented here reveal geographic structure at nested levels within the section and further provide evidence of repeated long-distance dispersal, consistent with other inferences that these spore-producing plants have the ability to spread over great distances, at least on occasion (Muñoz et al., 2004).
Appendix 1. Voucher information and GenBank accession numbers for taxa used in this study. Voucher specimens are deposited in the following herbaria: DUKE = Duke University, H = University of Helsinki, MO = Missouri Botanical Garden, MICH = University of Michigan, NY = New York Botanical Garden. Accessions included in Figs. 1–4 and in trees provided as online supplements are identified by the isolate number.
|
FOOTNOTES
1 This research was supported by NSF grant no. DEB-0515749. The authors thank P. Zhou for assistance with some analyses and the curators of MICH and NY for the loan of herbarium specimens. 
LITERATURE CITED
Andrews, A. L. 1913. Order Sphagnales. North American Flora vol. 15, part 1, 1–31.
Cronberg, N. 1996. Isozyme evidence of relationships within Sphagnum section Acutifolia (Sphagnaceae Bryopsida). Plant Systematics and Evolution 203: 41–64.[CrossRef][Web of Science]
Cronberg, N., AND R. Natcheva. 2002. Hybridization between the peat mosses, Sphagnum capillifolium and S. quinquefariumSphagnum, Bryophyta) as inferred by morphological characters and isozyme markers. Plant Systematics and Evolution 234: 53–70.[CrossRef][Web of Science]
Crum, H. A 1984. Sphagnopsida (Sphagnaceae). North American Flora, series 2, part 11, 1–180.
Daniels, R. E., AND A. Eddy. 1985. Handbook of European sphagna. Institute of Terrestrial Ecology, Her Majesty’s Stationery Office, London, UK.
Donoghue, M. J., AND S. A. Smith. 2004. Patterns in the assembly of temperate forests around the Northern Hemisphere. Philosophical Transactions of the Royal Society of London.B, Biological Sciences 359: 1633–1644.[CrossRef]
Eddy, A. 1977a. Sphagnales of tropical Asia. Bulletin of the British Museum. Natural History (Botany) 5: 359–445.
Eddy, A. 1977b. Sphagnum subsecundum agg. in Britain. Journal of Bryology 9: 309–319.[Web of Science]
Eddy, A. 1985. A revision of African Sphagnales. Bulletin of the British Museum Natural History (Botany) 12: 77–172.
Felsenstein, J. S. 1981. Evolutionary trees from DNA sequences: A maximum likelihood approach. Journal of Molecular Evolution 17: 368–376.[CrossRef][Web of Science][Medline]
Felsenstein, J. S. 1985. Confidence-limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791.[CrossRef][Web of Science]
Flatberg, K. I. 2002 The Norwegian sphagna: A field colour guide. Norges Tehnisk-naturvitenskapelige Universitet Vitenskapmuset. Rapport Botanisk Serie 2002-1, 1–44, 54 plates.
Fritsch, R. 1991. Index to bryophyte chromosome counts. Bryophytorum Bibliotheca 40: 19–20.
Hartmann, F. A., R. Wilson, S. R. Gradstein, H. Schnieder, AND J. Heinrichs. 2006. Testing hypotheses on species delimitation and disjunctions in the liverwort Bryopteris (Jungermanniopsida: Lejeuneaceae). International Journal of Plant Sciences 167: 1205–1214.[CrossRef][Web of Science]
Heinrichs, J., M. Lindner, H. Groth, J. Hentschel, K. Feldberg, C. Engel, J. J. Renker, M. von Konrat, D. G. Long, AND H. Schneider. 2006. Goodbye or welcome Gondwana? –Insights into the phylogenetic biogeography of the leafy liverwort Plagiochila with a description of Proskauera, gen. nov. (Plagiochilaceae, Jungermanniales). Plant Systematics and Evolution 258: 227–250.[CrossRef][Web of Science]
Hill, M. O. 1975. Sphagnum subsecundum Nees and S. auriculatum Schimp. in Britain. Journal of Bryology 8: 435–441.[Web of Science]
Huelsenbeck, J. P., AND F. Ronquist. 2001. MrBayes: Bayesian inference of phylogeny. Bioinformatics 17: 754–755.
Isoviita, P. 1966. Studies on Sphagnum L. I. Nomenclatural revisions of the European taxa. Annales Botanici Fennici 7: 157–162.
Krassilov, V. A., AND R. M. Schuster. 1984. In R. M. Schuster [ed.], New manual of bryology, vol. 2, 1173–1193. Hattori Botanical Laboratory, Nichinan, Japan.
Maass, W. S. G., AND M. J. Harvey. 1973. Studies on the taxonomy and distribution of Sphagnum. VII. Chromosome numbers in Sphagnum. Nova Hedwigia 24: 193–205.
Manos, P. S., AND M. J. Donoghue. 2001. Progress in northern hemisphere phytogeography: An introduction. International Journal of Plant Sciences 162: S1–S2.[CrossRef][Web of Science]
McQueen, C. B., AND R. E. Andrus. 2007. Sphagnaceae Dumortier. In Flora of North America Editorial Committee [ed.], Bryophytes: Mosses, part 1, 45–101. Flora of North America, vol. 27, Oxford University Press, New York, New York, USA.
Melosik, I., I. I. Odrzykoski, AND E. Sliwinska. 2005. Delimitation of taxa of Sphagnumsubsecunda s. l. (Musci, Sphagnaceae) based on multienzyme phenotype and cytological markers. Nova Hedwigia 80: 397–412.[CrossRef][Web of Science]
Miller, N. G. 1984. Tertiary and Quaternary fossils. In R. M. Schuster [ed.], New manual of bryology, vol. 2, 1194–1232. Hattori Botanical Laboratory, Nichinan, Japan.
Morrison, D. A. 2006. Multiple sequence alignment for phylogenetic purposes. Australian Systematic Botany 19: 479–539.[CrossRef][Web of Science]
Muñoz, J., A. M. Felicísimo, F. Cabezas, A. R. Burgaz, AND I. Martínez. 2004. Wind as a long distance dispersal vehicle in the southern hemisphere. Science 304: 1144–1147.
Natcheva, R., AND N. Cronberg. 2007. Recombination and introgression of nuclear and chloroplast genomes between the peat mosses, Sphagnum capillifolium and S. quinquefarium. Molecular Ecology 16: 811–818.[CrossRef][Medline]
Nees Von Esenbeck, C. G. D. 1819. Deutschlands flora, Abtheilung II, Cryptogamie 17. [Publisher unknown.]
Newton, M. E. 1993. Cytogenetics of Sphagnum. Advances in Bryology 5: 61–78.
Nylander, J. A. A. 2002. MrModeltest, version 1.1b.Website http://www.ebc.uu.se/systzoo/staff/nylander.html.
Petersen, R. H., AND K. W. Hughes. 2007. Some agaric distribution patterns involving Pacific landmasses and Pacific Rim. Mycoscience 48: 1–14.[CrossRef]
Russow, E. A. F. 1894. Zur Kenntnis der Subsecundum- und Cymbifolium Gruppe europäischer Torfmoose, nebst einem Anhang, enthaltend eine Aufzählung der bisher im Ostbalticum beobachteten Sphagnum-Arten und einen Schlüssel zur Bestimmung dieser Arten. Archiv Naturkunde Liv- Ehst- Kurlands 2. Serie 10: 361–527.
Sainsbury, G. O. K. 1955. A handbook of the New Zealand mosses. Royal Society of New Zealand, Wellington, New Zealand.
Sanmartín, I., H. Enghof, AND H. Ronquist. 2001. Patterns of animal dispersal, vicariance, and diversification in the holarctic. Biological Journal of the Linnean Society 73: 345–390.[CrossRef][Web of Science]
Såstad, S. M., K. I. Flatberg, AND N. Cronberg. 1999. Electrophoretic evidence supporting a theory of allopolyploid origin of the peatmoss Sphagnum jensenii. Nordic Journal of Botany 19: 355–361.[CrossRef][Web of Science]
Såstad, S. M., K. I. Flatberg, AND L. Hanssen. 2000. Origin, taxonomy and population structure of the allopolyploid peat moss Sphagnum majus. Plant Systematics and Evolution 225: 73–84.[CrossRef][Web of Science]
Såstad, S. M., H. Stenøien, K. I. Flatberg, AND S. Bakken. 2001. The narrow endemic Sphagnum troendelagicum is an allopolyploid derivative of the widespread S. balticum and S. tenellum. Systematic Botany 26: 66–74.[Web of Science]
Shaw, A. J. 2000. Phylogeny of the Sphagnopsida based on nuclear and chloroplast DNA sequences. Bryologist 103: 277–306.[CrossRef]
Shaw, A. J., C. J. Cox, AND S. B. Boles. 2003. Polarity of peatmoss (Sphagnum) evolution: Who says mosses have no roots? American Journal of Botany 90: 1777–1787.
Shaw, A. J., C. J. Cox, AND S. B. Boles. 2004. Phylogenetic relationships among Sphagnum sections: Hemitheca, Isocladus, and Subsecunda. Bryologist 107: 189–196.[CrossRef]
Shaw, A. J., AND B. B. Goffinet 2000. Molecular evidence of reticulate evolution in the peatmosses (Sphagnum), including S. ehyalinum, sp. nov. The Bryologist 103: 357–374.[CrossRef]
Shaw, A. J., AND K. S. Renzaglia. 2004. Phylogeny and diversification of bryophytes. American Journal of Botany 91: 1557–1581.
Simmons, M. P. 2004. Independence of alignment and tree search. Molecular Phylogenetics and Evolution 31: 874–879.[CrossRef][Web of Science][Medline]
Simmons, M. P., AND J. V. Freudenstein. 2003. The effects of increasing genetic distance on alignment of, and tree construction from, rDNA internal transcribed spacer sequence alignment. Molecular Phylogenetics and Evolution 26: 444–451.[CrossRef][Web of Science][Medline]
Wall, D. 2005. Origin and diversification of a tropical moss. Evolution 59: 1413–1424.[Web of Science][Medline]
Warnstorf, C. 1911. Sphagnales-Sphagnaceae (Sphagnologia Universalis). In H. G. A. Engler [ed.], Das Pflanzenreich, 1–546. Wilhelm Englemann, Leipzig, Germany.
Wen, J. 1999. Evolution of eastern Asian and eastern North American biogeographic disjunctions in flowering plants. Annual Review of Ecology and Systematics 30: 421–455.[CrossRef][Web of Science]
Yang, Z., AND B. Rannala. 1997. Bayesian phylogenetic inference using DNA sequences: A Markov chain Monte Carlo method. Molecular Biology and Evolution 14: 717–724.[Abstract]
Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence data sets under the maximum likelihood criterion. Ph.D. dissertation, University of Texas at Austin, Austin, Texas, USA.
This article has been cited by other articles:
|
|
 |
|
  M. Ricca, F. W. Beecher, S. B. Boles, E. Temsch, J. Greilhuber, E. F. Karlin, and A. J. Shaw Cytotype variation and allopolyploidy in North American species of the Sphagnum subsecundum complex (Sphagnaceae) Am. J. Botany, December 1, 2008; 95(12): 1606 - 1620. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
