Introduction

The first author became aware of Helenium scaposum Britton in the late 1960s while working on his dissertation project, a treatment of Helenium L. sect. Tetrodus (Cass.) DC. As this taxon was not part of his work at that time, Bierner placed it on his back-burner, and there it sat for some 50 years. Recently, Bierner decided to tie up some loose ends in his research by first dealing with Helenium scaposum.

Over the years, Helenium scaposum has been recognized as an endemic restricted to Isla de la Juventud, Cuba, formerly Isla de Pinos (e.g., Rock, 1957; Prede Rodríguez et al., 2000; Martínez Redondo and Herrera Oliver, 2003; Greuter and Rodriguez, 2016). It is restricted to the southern part of the western area known as the sabana arenosa and the western part of the northern region known as the sabana grande, where it grows in white sand associated with current or former pine forests (Balátová-Tuláčková and Capote, 1985). Puentes et al. (1985), using different terminology, noted that H. scaposum was restricted to the northeastern portion of Isla de la Juventud in the area they called Distrito Arenas Blancas. Herrera Oliver et al. (1999), in their evaluation of the state of conservation in the family Asteraceae in Cuba, classified H. scaposum as low-risk, conservation-dependent. Later it was placed on the Red List of Cuban vascular plants as being in danger and located in semi-natural savannas of pine forest regions (Berazaín Iturralde et al., 2005). The only chemical work we were able to find was that of Frias et al. (1972), who examined a number of taxa from Cuba for the presence of alkaloids and saponins; none were detected in H. scaposum.

Given the location of Helenium scaposum (Isla de la Juventud), its habit (scapiform), and certain morphological features such as its root system (numerous adventitious roots originating from a short rhizome), the first author had two major questions: 1) is this plant actually a Helenium, and 2) if so, where does it fit into the genus? To aid in this quest, Bierner solicited the help of Bruce G. Baldwin, who produced the nuclear 18S–26S nrDNA internal transcribed spacer (ITS) region sequence and performed a preliminary phylogenetic analysis, and Michael J. Moore and his undergraduate students Helene Tiley, Flora Samis and Spencer Wight, who generated most of the remaining sequences in this study and performed a more detailed phylogenetic analysis.

Materials and Methods

A total of 60 Helenium accessions were included in phylogenetic analyses, representing 36 taxa in the genus (Table 1).

Total DNA of Helenium scaposum was extracted from a leaf fragment of Greuter et al. 25969 (NY) using the DNeasy Plant Mini Kit (Qiagen, Valencia, California, USA), with incubation for cell lysis of 1–2 h rather than 10 min, as indicated in the manufacturer's protocol. Total DNA of other newly sequenced Helenium taxa was isolated from field collected or herbarium specimens using either the Nucleon Phytopure kit (GE Healthcare Lifesciences, Pittsburgh, Pennsylvania, USA) or the CTAB method of Doyle and Doyle (1987), with the addition of 1% PVP-40.

For H. scaposum, the 18S–26S nrDNA ITS region was amplified by polymerase chain reaction (PCR) as indicated by Baldwin and Wessa (2000b) except for use of AccuPower PCR Premix (K-2016; Bioneer Corp., Chunbuk, Korea). For all other newly generated sequences, PCRs were performed in 12.5 µL volumes with 0.5 µL of 5 mM primer for both primers, 5–20 ng of DNA template, 0.1 µL of GoTaq (Promega, Madison, WI, USA), 6.25 µL of Failsafe Premix B (Epicentre, Madison, WI, USA), and 4.7 µL of sterile, deionized water. Reactions were run on a Bio-Rad PTC 200 thermocycler (Bio-Rad, Hercules, CA, USA) at Oberlin College. For H. scaposum, Exonuclease I and shrimp alkaline phosphatase were used to remove unincorporated nucleotides as directed in the PCR Product Pre-Sequencing Kit (70995, United States Biochemical Corp., Cleveland, Ohio, USA). For all other newly generated sequences, PCRs were cleaned in 16.5 µL reactions containing 10 U of Exonuclease I (Affymetrix, ThermoFisher Scientific, Waltham, MA, USA), 2 U of shrimp alkaline phosphatase (Affymetrix), 8 µL of PCR product, and 8.5 µL of sterile, deionized water. Sanger sequencing was conducted at the UC Berkeley DNA Sequencing Facility (Barker Hall) for H. scaposum and by Neogenomics (formerly SeqWright; Houston, TX, USA) using an ABI 3730xl automated sequencer (Applied Biosystems, ThermoFisher Scientific) for other taxa. The primer ITS5 (White et al., 1990) was used instead of ITS-I (Urbatsch et al., 2000) for sequencing. The resulting forward and reverse sequences for each reaction were trimmed and de novo assembled using default parameters of the Geneious assembler in Geneious version 7 (Biomatters, Auckland, New Zealand).

The ITS region alignment was created in Geneious using MAFFT (Katoh and Standley, 2016) with default parameters, and all sites with greater than 50% missing data were removed using Phyutility (Smith and Dunn, 2008). Maximum likelihood analyses were run using RAxML version 8.2 (Stamatakis 2014) using the GTRGAMMA model, 100 bootstrap replicates, and the following parameters: -T 2 -f a -x 12345 -p 12345. The resulting trees were visualized using FigTree version 1.4.4 (available at  https://github.com/rambaut/figtree/releases).

Results and Discussion

The ITS region sequence prepared by Baldwin from the Helenium scaposum sample was initially included in a maximum likelihood analysis with other sequences in the ITS dataset of Baldwin et al. (2002), which confirmed its position in Helenium (Fig. 1). The sequence was then included in a maximum likelihood analysis with other sequences in an ITS region dataset produced by Moore et al., which also confirmed its position in Helenium (Fig. 2; Appendix 1).

Table 1.

Voucher information and GenBank accession numbers for all accessions included in this study. Voucher and collection locality information are given only for newly submitted GenBank sequences. Herbarium codes follow Index Herbariorum, and numbers following herbarium codes are herbarium accession numbers. Specimens with no numbers after herbarium codes lack herbarium accession numbers (s.n.).

Continued.

Continued.

Continued.

Continued.

Continued.

Fig. 1.

Results of initial maximum likelihood phylogenetic analysis including Helenium scaposum and taxa across the Heliantheae alliance sampled by Baldwin et al. (2002), showing only the clade corresponding to tribe Helenieae. Bootstrap clade support values above 90% are shown at nodes. Note that the samples of Helenium scaposum (denoted by arrow) and the only other included species of Helenium (H. bigelovii) constitute a robust clade.

Fig. 2.

Phylogram showing results of a maximum likelihood analysis of Helenium and outgroup species in Helenieae. Taxa in Helenium sect. Leptopoda are in bold italics, and H. scaposum is highlighted in red and by the red arrow. Bootstrap clade support values at or above 50% are shown at nodes.

The analysis of Baldwin clearly answered question 1; this plant belongs in Helenium. The Helenium scaposum sequence associated with Helenium in a data set that included representatives of the closely related genera Balduina Nutt. and Gaillardia Foug. The data set also included representatives of all other Helenieae genera: Amblyolepis DC., Baileya Harv. & A. Gray ex Torr., Hymenoxys Cass., Marshallia Schreb., Pelucha S. Watson, Plateilema (A. Gray) Cockerell, Psathyrotes (Nutt.) A. Gray, Psilostrophe DC., Tetraneuris Greene, and Trichoptilium A. Gray (see Baldwin and Wessa, 2000a, and Baldwin et al., 2002).

Question 2 — where does Helenium scaposum fit into the genus — has been answered to some extent by the analysis of Moore et al. Baldwin's data set included only one Helenium sequence other than H. scaposum (that of H. bigelovii Torr. & A. Gray), while the data set of Moore et al. included sequences of 36 Helenium taxa plus multiple populations of many of the taxa (Table 1). Furthermore, the Moore et al. data set included representatives of all of the currently recognized sections of Helenium (Appendix 2). The North American sections are sect. Amarum Bierner, sect. Hecubaea (DC.) A. Gray, sect. Helenium, sect. Leptopoda (Nutt.) Wood, and sect. Tetrodus (Cass.) DC. (Bierner, 1972). The South American sections are sect. Cephalophora (Cav.) Hoff., and sect. Actinea (Juss.) Bierner (Bierner, 1978, 1987).

To add some perspective, Rock (1957) in his work on Helenium stated, “At this time, I would like to emphasize the apparent relationship of H. scaposum to the tetraneuranae of Rydberg and suggest that the most likely affinity of this species is with Plateilema Palmeri (A. Gray) Cockerell.” In fact, the only conspicuous morphological characteristic that might indicate such a close relationship between H. scaposum and P. palmeri is that both have scapiform capitulescences. Furthermore, the analyses of both Baldwin and Moore et al. (this study) indicate that H. scaposum and P. palmeri are well separated phylogenetically (Figs. 1 and 2; Appendix 1). In addition, a close relationship of these two taxa is ecologically and biogeographically doubtful; H. scaposum is confined to white sand associated with current or former pine forests on Isla de la Juventud, Cuba, and P. palmeri is found on fine clay-loam soils of Chihuahuan desert scrub/grasslands from Brewster County, Texas, to central Coahuila and southern Nuevo León, Mexico (Jackson et al., 2015; Jackson, 2017). Judging from biogeography alone, it seems most likely that H. scaposum fits into the “widespread, disjunct to western Cuba” pattern of Sorrie and Weakley (2001), and is more likely to be related to members of Helenium from the southeastern United States. In fact, Sorrie and Weakley (2001) state, “Western Cuba includes the provinces of Habana, Isla de Pinos (Isla de Juventud), Matanzas, and Pinar del Rio. Nearly all of the plants listed here are found as disjuncts to Cuba only. A number of the ‘widespread' taxa are confined in the United States to the Florida peninsula.” While H. scaposum was not among the taxa listed by Sorrie and Weakley (2001), it is endemic to Isla de la Juventud, and two members of Helenium from the southeastern United States, H. pinnatifidum (Nutt.) Rydb. and H. flexuosum Raf., have populations that extend into southern Florida.

Which brings us back to question 2 — where does Helenium scaposum fit into the genus Helenium? While it seems clear from Figure 2 that H. scaposum is a distinctive taxon, as indicated by its relatively long branch length (i.e., it has several unique base changes), and is associated with a large clade including, among other taxa, members of Helenium sect. Leptopoda (highlighted in the tree), the relationships within this large clade are generally poorly supported, making it impossible based on this phylogenetic analysis to make any strong claim as to a sister-taxon relationship. Having said that, a second look at biogeography and a first look at morphology and ecology seem to be warranted.

All of the taxa of Helenium sect. Leptopoda (Appendix 2) have distributions that fit within the “widespread, disjunct to western Cuba” pattern of Sorrie and Weakley (2001); however, only H. flexuosum and H. pinnatifidum have distributions that extend to far southern Florida (see Rock 1957, p. 139). With regard to habitat preference, H. scaposum grows in white sand associated with current or former pine forests, and the taxa of H. sect. Leptopoda grow in various types of sandy soils often at the edges of pine forests (Rock, 1957; Table 2), which also is in accord with the “widespread, disjunct to western Cuba” pattern of Sorrie and Weakley (2001). Morphologically, H. scaposum, while very diminutive compared to the taxa of H. sect. Leptopoda, seems to have more in common with H. drummondii H. Rock, H. pinnatifidum, and H. vernale Walter (Table 2), the three of which, according to Rock (1957), form a subgroup within sect. Leptopoda. Even though H. flexuosum has a distribution that extends into southern Florida, it seems on morphological grounds (Table 2) to be a poor candidate as a sister-taxon to H. scaposum. Knowing the chromosome number of H. scaposum could be very important for establishing a sister-taxon relationship; hence, a visit to Isla de la Juventud may be in Bierner's not-too-distant future. In the meantime, we believe that the most likely association of H. scaposum within Helenium is with taxa of H. sect. Leptopoda, in particular H. drummondii, H. pinnatifidum, and H. vernale.

We recognize that the phylogenetic tree produced by Moore et al. (Fig. 2 and Appendix 1) provides an opportunity to comment on sectional delimitations (Appendix 2) and relationships in Helenium. However, that is outside of the scope of this paper and will be the subject of future studies.

Taxonomic Treatment

Table 2.

A morphological and ecological comparison of Helenium scaposum with the taxa of Helenium sect. Leptopoda.

Continued.