Genetic Variation and Species Boundaries in Calopogon (Orchidaceae)

Morphological and habitat similarities among the five species of the terrestrial orchid genus Calopogon have led to nomenclatural and taxonomic confusion. The taxa are marked by subtle character differences and little apparent reproductive isolation. Here we investigate allozyme diversity at the species level and the partitioning of genetic variation within and among species and their populations. Genetic identities are used to define species boundaries and suggest phylogenetic relationships. All five species of Calopogon maintain high levels of allozyme variation within their populations (P 5 50.0%294.4%, AP 5 2.67–3.32, He 5 0.11–0.43). Calopogon oklahomensis, an autotetraploid that appears to have undergone gene silencing at 13 of its 19 polymorphic loci, consistently had the highest genetic diversity values. Calopogon multiflorus, which has the most restricted range and rarest occurrence, had the lowest mean genetic diversity values. In C. oklahomensis, C. pallidus and C. tuberosus most of the genetic variation exists within rather than among populations (GST 5 0.037–0.085). The UPGMA phenogram generated using genetic identity data has three phenetic groups and supports designation of the taxa as separate species. The data suggest that C. oklahomensis most closely resembles the basal extant taxon within Calopogon. The genus Calopogon R. Brown (Orchidaceae) includes five showy, terrestrial species of short-lived, herbaceous perennials known as Grass Pinks (Dressler 1981; 1993). Calopogon spp. display considerable morphological similarity, often grow in similar habitats, have overlapping ranges, and overlapping periods of anthesis. Morphological similarities between taxa (eg. plant height, leaf length and width, number and color of flowers, petal length, width and shape, length and broadness of the lip, hair characters) are so great that taxonomic keys are unclear and characters distinguishing the taxa are often subtle, leading to nomenclatural and taxonomic confusion. Where the Calopogon taxa occur sympatrically there appears to be little reproductive isolation. All Calopogon species share a similar pollination system with cross-pollination by bees believed to be the primary breeding mechanism (Dressler 1981; Firmage and Cole 1988; Thien and Marcks 1972). There is evidence of self-compatibility (Thien 1973) although an insect is required for pollen transfer. Calopogon is also capable of vegetative reproduction via root sprouting (Thien and Marcks 1972). Their typically deep pink flowers produce no nectar (with the possible exception of C. oklahomensis Goldman) and offer no pollen reward. Consequently, flowers must attract pollinators through mimicry of other sympatric species that offer a pollen reward and by a form of deception categorized as pseudopollen (Dafni 1984) which refers to the resemblance of the showy yellow and orange lip hairs to a mass of pollen. When a pollinator is attracted to this display and lands on the labellum, if the insect is heavy enough, the labellum swings down and the posterior of the insect comes into contact with the sticky pollinia located on the end of the column (Thien and Marcks 1972; Firmage and Cole 1988). Pollinators are therefore varied and non-specific. Like most orchid taxa the tiny, dust-like seeds are wind dispersed (Dressler 1981). Pace (1909) reported that haploid daughter cells of C. tuberosus (L.) B.S.P. had 13 chromosomes, while separate counts of diploid sporophytic tissue revealed approximately 26 chromosomes. These findings were based on only a few counts each from haploid and diploid states. In 1972, Thien and Marcks found that C. tuberosus from northern Wisconsin had 2n 5 40 while a later report by Thien (1973) indicated that C. tuberosus, C. pallidus Chapman, C. barbatus (Walter) Ames, and C. multiflorus Lindley root tip cells possess 2n 5 42 chromosomes. Calopogon oklahomensis is a recently described species (Goldman 1995) for which a chromosome number has not been reported. Isolating mechanisms among the five congeners are habitat, seasonal separation, and to some extent, flower size, which necessitates pollinators of different size and weight (Thien 1973). The most common and widespread species, C. tuberosus, ranges from Canada to Cuba and the Bahamas and across the eastern United States. This species is comprised of two varieties, Calopogon tuberosus var. tuberosus and C. tuberosus var. simpsonii (Small) Magrath. Calopogon tuberosus var. simpsonii is found on calcareous soils in swamps and prairies in southern Florida, Bahamas and Cuba. This variety tends to have longer inflorescences that represent a larger proportion of the total plant height. Calopogon pallidus and C. barbatus are restricted to the coastal plain of the southeastern United States ranging from North Carolina to Mississippi. Calopogon multiflorus, which grows in open, moist pine flatwoods and meadows, has a more restricted range, being confined to Florida and its immediate borders. Calopogon oklahomensis occurs in mesic, acidic, sandy-loam prairies in 2004] 309 TRAPNELL ET AL.: CALOPOGON GENETIC VARIATION Arkansas, southeastern Kansas, Missouri, and eastern Oklahoma (Goldman 1995). Reliance on morphological characters for delineation of Calopogon species has led to taxonomic confusion. Examination of the underlying genetic diversity of the species and the degree of genetic divergence among taxa is a useful way to elucidate the taxonomic delineation. Knowledge of the genetic variation housed within a species is also valuable since it can provide insights into the species’ long-term survival potential. This information is essential for intelligent decisionmaking regarding conservation recommendations and policies. Genetic diversity increases the likelihood that a species can survive disease, environmental fluctuations, and natural catastrophes that invariably occur over time. The objective of this study was to determine the extent of genetic diversity within each Calopogon species and to describe the partitioning of genetic variation within and among species and their populations. Levels of genetic identity between these taxa are quantified and used to assess species boundaries and suggest phylogenetic relationships. MATERIALS AND METHODS Calopogon belongs to subfamily Epidendroideae Lindley, tribe Arethuseae Lindley, and subtribe Bletiinae Bentham. The two outgroup species included in this study, Bletia purpurea (Lamarck) A. de Candolle and Arethusa bulbosa L., were selected for their moderate taxonomic distance from Calopogon (Goldman et al. 2001). Bletia purpurea belongs to the same subtribe as Calopogon (Dressler 1981; 1993) and both outgroup genera belong to the same tribe as Calopogon. Both outgroup species also have overlapping geographic ranges with some of the Calopogon species. Mature leaf or stem tissue was randomly collected from 15 populations representing the five Calopogon species and the two outgroup species. Voucher specimens from populations from which plant material was collected are listed in Appendix 1. Forty-eight individuals from each population were collected wherever possible (Appendix 1). Leaves were wrapped in moist paper towels, placed in a sealed plastic bag, and kept chilled to prevent protein denaturation. Care was taken to avoid further trauma to the tissue from either freezing or creasing. Within 24–72 hrs of collection, the tissue was clipped into small pieces, placed in a chilled mortar, and crushed with a pestle and a pinch of sea sand to disrupt cellular compartmentalization. Stem material was crushed in liquid nitrogen with sea sand. Enzymes were extracted from the tissue with a polyvinylpyrrolidone-phosphate extraction buffer (Mitton et al. 1979). The resulting slurry containing crude protein extract was absorbed onto 4 x 6 mm wicks punched from Whatman 3 mm chromatography paper. Wicks were stored in microtest plates at 2708C until used for electrophoresis. Wicks were placed in horizontal gels composed of 10% potato starch and electrophoresis was performed. Twelve enzyme stains in four buffer systems resolved 21 putative allozyme loci. Enzymes stained and loci identified (in parentheses) for each of the four buffer systems were: 1) system 6; alcohol dehydrogenase (ADH1, ADH2) and fluorescent esterase (FE4), 2) system 7; aspartate aminotransferase (AAT2, AAT3, AAT4), diaphorase (DIA1, DIA2), menadione reductase (MNR2), 3) system 8-; fluorescent esterase (FE1, FE2), menadione reductase (MNR1), and triosephosphate isomerase (TPI1, TPI2) and 4) system 11; adenylate kinase (AK1), isocitrate dehydrogenase (IDH), malate dehydrogenase (MDH1, MDH2), 6-phosphogluconate dehydrogenase (6-PGD), phosphoglucoisomerase (PGI1), phosphoglucomutase (PGM2). All stain recipes were adapted from Soltis et al. (1983) except for diaphorase which was taken from Cheliak and Pitel (1984). Buffer system 8is a modification of buffer system 8 as described by Soltis et al. (1983). Two standard wicks from individuals of C. tuberosus and C. pallidus were placed on each gel. Banding patterns were consistent with those expected for each enzyme system (Weeden and Wendel 1989). Levels of allozyme diversity were estimated within species and for individual populations using a computer program designed by M.D. Loveless and A.F. Schnabel. Measures of genetic diversity were percent polymorphic loci, P (a species was treated as polymorphic at a locus if two alleles were detected); mean number of alleles per locus, A; mean number of alleles per polymorphic locus, AP; effective number of alleles per locus, Ae 5 1/Spi; and genetic diversity, He [5 12 Spi (Nei 1973), where pi is the frequency of the ith allele], which is the proportion of loci heterozygous per individual under Hardy-Weinberg expectations. Observed heterozygosity (Ho) was compared with Hardy-Weinberg expected heterozygosity for each polymorphic locus in each population by calculating Wright’s fixation indices (F; Wright 1922) and testing for significant deviations using X2 5 F2N(a 2 1); df 5 a(a 2 1)/2 where N is the total number of individuals analyzed and a is the number of alleles at the locus (Li and Horvitz 1953). Variation among populations was estimated using Nei’s (1973) measures of genetic diversity. Total genetic diversity (HT), mean genetic diversity within populations (HS), and mean genetic diversity among populations (DST) such that HT 5 HS 1 DST, were determined for each polymorphic locus. The proportion of genetic variation that occurs among populations (GST) was calculated for each polymorphic locus by GST 5 DST/HT (Nei 1973) and averaged across loci. Heterogeneity in allele frequencies among populations was tested by X2 5 2NGST (a 2 1); df 5 (a 2 1)(k 2 1), where N is the total number of individuals analyzed, a is the number of alleles at the locus, and k is the number of populations (Workman and Niswander 1970). Nei’s (1972) genetic identities and distances were calculated for each pair-wise combination of populations and species. A UPGMA phenogram of genetic identities was generated using NTSYS-PC (Rohlf 2000). A Mantel test of correspondence between genetic distances and geographic distances was performed (Smouse et al. 1986).