Phylogeny of Hibiscus sect. Muenchhusia (Malvaceae) Based on Chloroplast rpL16 and ndhF, and Nuclear ITS and GBSSI Sequences

Hibiscus section Muenchhusia is a North American taxon with Žve species as recognized in the most recent taxonomic revision: H. coccineus, H. dasycalyx, H. grandiorus, H. laevis, and H. moscheutos. To investigate the monophyly of Hibiscus section Muenchhusia, its relationship to other Hibiscus species, and the phylogenetic relationships among its species, DNA sequence data were obtained. To investigate the placement of section Muenchhusia within Hibiscus, chloroplast ndhF gene and rpL16 intron sequences were generated and added to a recently published dataset. To investigate relationships within section Muenchhusia three DNA sequence data sets were generated: a non-coding region of the chloroplast genome (rpL16 intron), nuclear ribosomal ITS, and a nuclear gene encoding granule-bound starch synthase (GBSSI).Analyses of ndhF 1 rpL16 data indicated that section Muenchhusia is indeed monophyletic and is embedded in a clade that includes representatives of Hibiscus section Trionum sensu lato plus other genera segregated from Hibiscus (Abelmoschus, Fioria), and the tribe Malvavisceae. Within sectionMuenchhusia little to no phylogenetically informative variation was detected in the ITS or rpL16 sequences. The GBSSI data, on the other hand, provided sufŽcient information to resolve relationships among species. The species of section Muenchhusia fall into two primary clades, one consisting of H. grandiorus and H. moscheutos sensu lato, the other including H. coccineus, H. dasycalyx, and H. laevis. These phylogenetic data corroborate earlier biosystematic studies that also placed the species into these same two groups. Sequence polymorphism was observed in one accession each of H. dasycalyx and H. grandiorus and resolution of the underlying alleles indicates that gene ow has occurred fromH. moscheutos into both H. dasycalyx and H. grandiorus. Hibiscus L. sect. Muenchhusia (Heister ex Fabricium) O. Blanchard (Malvaceae), colloquially known as the Rose Mallows, is a North American taxon that includes Žve species as circumscribed in the most recent taxonomic revision (Blanchard 1976; see Table 1): Hibiscus coccineus Walter, Hibiscus dasycalyx Blake & Shiller, Hibiscus grandiorus Michaux, Hibiscus laevis Allioni, and Hibiscus moscheutos L. Within H. moscheutos two subspecies were recognized by Blanchard (1976): H. moscheutos subsp. moscheutos, and H. moscheutos subsp. lasiocarpos (Cavanilles) O. J. Blanchard. A number of additional taxa associated with H. moscheutos have been variously recognized historically either as varieties, subspecies, or distinct species. Primary among these are Hibiscus moscheutos subsp. palustris L. of the northeastern US (synonymized with H. moscheutos subsp. moscheutos by Blanchard [1976]) and Hibiscus moscheutos subsp. incanus Wendl. of the southeastern coastal plain (synonymized with H. moscheutos subsp. lasiocarpos by Blanchard [1976]). The recognition of Hibiscus sect. Muenchhusia as a group was resurrected by Blanchard (in Fryxell 1988); previously these species had been placed in the large and heterogeneous Hibiscus sect. Trionum (e.g., Hochreutiner 1900). The monophyly of sect. Muenchhusia is suggested by the overall morphological similarity of its included species, a shared chromosome number of n519 (Wise and Menzel 1971), similar ecological afŽnities (all species are primarily wetland species, either in marshes or along river, drainage ditch, or lake edges), a common growth habit being perennial herbs that grow back from a rootstock each year, and Žnally their shared geographic distribution with natural populations generally conŽned to eastern and central North America. The taxonomic work of Blanchard (1976) serves as a foundation for further systematic work on this group. The results of his study were two-fold. First, the species of sect. Muenchhusia were segregated from sect. Trionum in recognition of the clear afŽnities of these species to each other, but no clear relationship between them and sect. Trionum. Second, Blanchard (1976) collected extensively throughout the range of these species and catalogued the extensive intraspeciŽc variation present within those species. Blanchard’s (1976) work, however, did not address the question of species relationships within sect. Muenchhusia, thus the phylogeny of sect.Muenchhusia and the relationship of sect. Muenchhusia to the rest of Hibiscus remains unclear. Wise and Menzel (1971) took a biosystematic approach to understanding species relationships in sect. Muenchhusia by performing crossing studies within and between species. These analyses resulted in the recognition of two primary crossing groups, Group I consisting of H. grandiorus and H. moscheutos s.l., and Group II consisting of H. coccineus and H. laevis (H. dasycalyxwas not included in the analyses of Wise and Menzel [1971]). In all crosses between species within groups, fertile hybrids were formed. Crosses between groups, however, were generally unsuccessful in setting fruit. These data suggest that these two groups may represent two distinct natural groups, although the alternative that one group gave rise to the other with a single evolution of reproductive isolation in the derived group cannot be ruled out. Pfeil et al. (2002) used chloroplast DNA (cpDNA) 386 [Volume 29 SYSTEMATIC BOTANY TABLE 1. Plant materials used for DNA sequencing, with voucher information and GenBank accession numbers. All voucher specimens are deposited at the University of Tennessee herbarium (TENN). All sequences are newly reported except for rpL16 from H. trionumwhich was taken from GenBank. H. coccineus—Missouri Botanical Garden, R. Small s.n.; ndhF AY341395, rpL16 AY341407, ITS AY341386, GBSSI AY341420. FL, St. John’s Co., R. Small 187; GBSSI AY341421. H. dasycalyx—TX, Houston Co., O. Blanchard 249; ndhF AY341397, rpL16 AY341406, ITS AY341388, GBSSI AY341419. TX, Trinity Co., R. Klips s.n.; GBSSI AY341417, AY341418. H. grandiorus—FL, Gulf Co., R. Small 190; ndhF AY341398, rpL16 AY341400, ITS AY341389, GBSSI AY341410. FL, St. John’s Co., K. Siripun s.n.; GBSSI AY341408, AY341409. H. laevis—NE, county unknown, R. Small s.n.; ndhF AY341396, rpL16 AY341405, ITS AY341387, GBSSI AY341416. TN, Polk Co., J. Shaw s.n.; GBSSI AY341415. H. moscheutos subsp. incanus—FL, Alachua Co., R. Small 189; rpL16 AY341401, ITS AY341391, GBSSI AY341411. H. moscheutos subsp. lasiocarpos—IL, Lawrence Co., R. Small 203; rpL16 AY341403, ITS AY341392, GBSSI AY341413. H. moscheutos subsp. moscheutos—TN, Knox Co., R. Small 173; ndhF AY341399, rpL16 AY341402, ITS AY341390, GBSSI AY341412. H. moscheutos subsp. palustris—NY, Cayuga Co., R. Small 172; rpL16 AY341404, ITS AY341393, GBSSI AY341414. H. trionum—South Africa, R. Small s.n.; ndhF AY341394, rpL16 AF384612, ITS AY341385, GBSSI AY341422. sequences of the gene ndhF and the intron of the rpL16 gene to address phylogenetic relationships of the genera ascribed to the tribe Hibisceae, and the genus Hibiscus itself. Their results indicate that several genera that have been placed in Hibisceae are basal within the family or part of a large polytomy near the base of the family (e.g., Radyera, Howittia, Lagunaria) making Hibisceae a paraphyletic grade. In addition, within the large grade of the genus Hibiscus are embedded several genera that have been segregated from Hibiscus (e.g., Abelmoschus, Fioria, Macrostelia) as well as the tribes Malvavisceae and Decaschisteae. A Žnal result of the work of Pfeil et al. (2002) has been the recognition of potentially monophyletic groups within the large and heterogeneous grade of Hibiscus (including segregate genera). Some members of Hibiscus sect. Trionum s.l. were included in the analyses of Pfeil et al. (2002) including H. trionum L. (maintained in sect. Trionum) and H. striatus Cav. (segregated into sect. Striati). No members of sect. Muenchhusia were included in the analyses of Pfeil et al. (2002), so the placement of this section relative to sect. Trionum and the other sections segregated from sect. Trionum by Blanchard remains unclear. Given these taxonomic, biosystematic, and phylogenetic studies, the current work was undertaken to address three primary questions. First, is Hibiscus sect. Muenchhusia monophyletic? Second, what is the phylogenetic placement of Hibiscus sect.Muenchhusiawithin the genus Hibiscus and tribe Hibisceae? Third, what are the phylogenetic relationships among the species ofHibiscus sect.Muenchhusia? DNA sequence data from three sources were explored for this project: cpDNA, including the gene ndhF and the rpL16 intron for higher level analysis and rpL16 alone for lower level analysis; the internal transcribed spacer region of nuclear ribosomal DNA (ITS); and a nuclear encoded granulebound starch synthase gene (GBSSI). Chloroplast DNA noncoding regions and ITS sequences are often applied to systematic problems at low taxonomic levels because they generally show greater levels of sequence variation than cpDNA or nuclear rDNA genes. Lowcopy nuclear genes such as GBSSI are used less often than cpDNA or ITS sequences due the greater difŽculty in isolating and characterizing low-copy genes. Nonetheless, such genes may provide phylogenetically informative characters where other sequences such as cpDNA or ITS do not (e.g., Small et al. 1998; Sang 2002). GBSSI sequences have been applied to both species-level and higher phylogenetic questions in a variety of plant families (e.g., Mason-Gamer et al. 1998; Miller et al. 1999; Evans et al. 2000; Mason-Gamer 2001; Peralta and Spooner 2001; Walsh and Hoot 2001; Evans and Campbell 2002; Mathews et al. 2002; Smedmark et al. 2003). Most studies thus far have found GBSSI to be single-copy in diploid species although there are two copies per diploid genome in Rosaceae (Evans et al. 2000). Gene structure appears to be conserved and consists of one untranslated and 13 translated exons (van der Leij et al. 1991). Different groups of researchers have used different parts of the GBSSI gene for phylogenetic studies. For example, in Poaceae (Mason-Gamer et al. 1998; Mason-Gamer 2001) and Ipomoea (Miller et al. 1999) the 39 end of the gene (exons 8–13 or 9–11 respectively) has been sequenced, whereas in Malvaceae, Rosaceae, and Solanaceae (this paper; Evans et al. 2000; Peralta and Spooner 2001; Evans and Campbell 2002; Smedmark et al. 2003) the 59 end of the gene (exons 1–8 or 9) have been sequenced. Further, different subsets of the data have been used to address different questions. For higher level studies (e.g., genera within families) often only exon sequences are used because intron sequences are difŽcult to align (e.g., Mason-Gamer et al. 1998; Evans et al. 2000; Evans and Campbell 2002), while at lower levels (e.g., species within genera) both exon and intron sequences are used (e.g., Miller et al. 1999; Mason-Gamer 2001; Peralta and Spooner 2001; Walsh and Hoot 2001; Smedmark et al. 2003). MATERIALS AND METHODS Plant Materials. Specimens for these analyses were from wild-collected populations, botanical gardens, or USDA germplasm maintenance facilities (Table 1). All Žve species of sect. 2004] 387 SMALL: HIBISCUS SECT. MUENCHHUSIA TABLE 2. AmpliŽcation and sequencing primers for GBSSI, ITS, rpL16, and ndhF used in this study. Primer Sequence (59 to 39) Amp/Seq Reference