Evolutionary Studies of the Gnetales

Evolutionary studies of the Gnetales

Chen Hou

Academic dissertation for the degree of Doctor of Philosophy in Plant Sys- tematics presented at Stockholm University 2016

Evolutionary studies of the Gnetales

Chen Hou

©Chen Hou Stockholm University 2016

ISBN 978-91-7649-371-7

Printed in Sweden by US AB, Stockholm 2016 Distributor: Department of Ecology, Environment and Plant Sciences, Stockholm University, Sweden

”Although relatively few people have chosen to study the Gnetales, those who have had the opportunity to work with these organisms experience a profound sense of extraordinary beauty and complexity of the evolutionary process.”

- Willian E. Friedman, 1996

Abstract

The Gnetales consist of three distinct genera, Ephedra, Gnetum and Wel- witschia with considerable divergence among them regarding morphological, ecological and molecular characters. A longstanding debate of the simi- larity between the Gnetales and angiosperms and the unresolved seed plant phylogeny intrigues plant scientists to further investigate the evolutionary history of the Gnetales. The presented projects deal with interdisciplinary questions on proteomics, chloroplast genomes, phylogenetic relationships, gross morphology and taxonomy. The thesis aims to summarize general problems encountered in previous studies, and to provide new insights and future perspectives based on the results of completed and ongoing projects.

In Ephedra, the Mediterranean species E. foeminea has been shown to be entomophilous and it possesses an important phylogenetic status as the sister of the remaining genus. Therefore, the chloroplast genome of E. foeminea was assembled and compared to that previously presented (of the anemophilous Asian species E. equisetina, nested in the core clade of Ephedra). The genome has a quadripartite structure and comprises 118 genes and 109,584 base pairs. A pairwise genome comparison was conducted between E. foeminea and E. equisetina, resulting in the detection of 2,352 variable sites, the obtained data can be used for prospective phylogenetic studies. A proteomic study was also conducted on E. foeminea along with three anemophilous Ephedra species, in order to investigate the biochemistry of the pollination drops. The results show that detected proteins in the pollination drops of Ephedra vary dramatically among species but always occur in very low amounts. The majority of the detected proteins are degradome proteins, i.e., waste products from degrading cells of the nucellus. Some secretome proteins were also found, which are putatively functional, but also these proteins occur in very low amounts. The repeatability of the proteomic studies can, however, be questioned. The sampling methods and proteomic analyses are probably problematic although some suggestions for improvement are provided. Thus I chose to continue with other projects.

In Gnetum, reconstruction of the genus phylogeny and assessments of divergence times of clades were performed using an extensive sampling of ingroup and outgroup accessions. The results show that the South American lineage separated from the remaining genus in the Late Cretaceous. The con- tinued diversification event gave rise to an African lineage and an Asian lineage. The crown age of the Asian clade, which comprises two arborescent species sister to the remaining liaonid species, was estimated to the Creta- ceous-Paleogene (K-Pg) boundary. In light of the genus phylogeny and estimated node ages, we suggest that the breakup of Gondwana influenced diversification patterns in Gnetum. Later dispersal events also contributed to

the current distribution of Gnetum, and to the phylogenetic patterns within each of the major clades. From my results, it is however clear that taxonomy and species delimitations are poorly defined, and needs to be further studied for all subclades of Gnetum. I have initiated this task by studying the Chi- nese lianoid clade of Gnetum more in depth. Eleven chloroplast genomes were generated, aligned and compared. Based on the information, four chloroplast markers were designed and applied to further resolve the species relationships with an extensive sampling. The results show, with strong support, that G. parvifolium is sister to all the remaining species of the Chinese linaoid clade. Another five lianoid species are confirmed using both morphological and molecular data, but several names are represented by type material that cannot be considered separate species. Modified keys for identification of male and female plants are presented, based on vegetative and reproductive structures. A subsequent dating analysis indicates that diversification in the Chinese lianoid Gnetum clade took place mainly in the Neogene, during which environmental changes probably facilitated diversification in the lineage.

List of Papers

This thesis is based on the following five papers, referred to in the text by their respective roman numerals.

Paper I. Hou, C. & Rydin, C. 2013. Proteome of pollination drops in Ephed- ra (Gnetales). Manuscript.

Paper II. von Aderkas, P.; Prior, N.; Gagnon, S.; Little, S.; Cross, T.; Hardie, D.; Borchers, C.; Thornburg, R.; Hou, C. & Lunny. A. 2015. Degradome and secretome of pollination drops of Ephedra. The Botanical Review 81:1- 27. DOI:10.1007/s12229-014-9147-x

Paper III. Hou, C.; Wikström N. & Rydin, C. 2015. The chloroplast genome of Ephedra foeminea (Ephedraceae, Gnetales), an entomophilous gymnosperm endemic to the Mediterranean area. Mitochondrial DNA. DOI:10.3109/19401736.2015.1122768

Paper IV. Hou, C.; Humphreys, A. M.; Thureborn, O. & Rydin. C. 2015. New insights into the evolutionary history of Gnetum (Gnetales). Taxon 64:239-253. DOI:10.12705/642.12

Paper V. Hou, C.; Wikström N.; Strijk J. S. & Rydin, C. 2016. Resolving phylogenetic relations and species delimitations in closely related gymnosperms using high-throughput NGS, Sanger sequencing and morphology. submitted

Reprints were made with permission from the publishers.

Contributions of the thesis Papers I, III, IV, V were written by CH with comments and suggestions from the co-authors, and the designs of the studies were performed by CH and CR. CH did analyses of gel electrophoresis in Paper I and Paper II. CH has generated the majority of sequences in Paper IV together with OT, all sequences in paper V, and the sequences in Paper III together with NW. CH is respon- sible for the phylogenetic analyses and dating analyses in Paper IV and Pa- per V, genome assemblies individually in Paper IV and together with NW in Paper III. CH has conducted the morphological studies in Paper V in cooper- ation with the co-authors.

Content

Introduction …………………………………………………………………...…………………………..10

Chapter 1. Pollination biology in the Gnetales

1.1 Anemophily versus entomophily ………………………………………………..……………………12

1.2 Investigations of pollination drops in gymnosperms……………………………………...…………..12

1.3 A proteomic study of pollination drops of Ephedra…………………………………….………….…13

1.4 A new proteomic study……………………………………………………………...…………...... 14

1.5 Prospective studies………………………………………………………………………….…………15

Chapter 2. Phylogenetic studies and dating analyses of the Gnetales

2.1 Phylogenetic studies in Ephedra and taxonomic implications……………………………..…………17

2.2 Dating analyses and evolutionary history of the Ephedraceae………………………………..………17

2.3 Phylogenetic studies in Gnetum and taxonomic implications………………………………...………18

2.4 Dating analyses and evolutionary history of Gnetum…………………………………………………20

2.5 Prospective studies…………………………….………………………………………………………22

Chapter 3. Chloroplast genome investigations in the Gnetales

3.1 Nuclear genomes of the Gnetales………………………………………………………………..……24

3.2 Structural differences among chloroplast genome in the Gnetales……………………………………25

3.3 A newly generated chloroplast genome of Ephedra………………………………………..…………25

3.4 Chloroplast genome investigation in Chinese lianoid Gnetum………………………………..………26

3.5 Prospective studies…………………………………………………………………………………….27

Chapter 4. Taxonomy and species delimitations in the Gnetales…………………………………...…….29

Concluding remarks……………………………………………………………………………………….30

Legend………..……………………………………………………………………………………………31

Svensk sammanfattning (Swedish summary)……………………………………………………..………32

Acknowledgement……………….…………………………………………………………………..……34

Literature cited…………………………………………………………………………………….………36

Abbreviations

1D SDS PAGE One dimensional SDS polyacrylamide gel electrophoresis AICc Corrected Akaike information criterion AICM A posterior simulation-based analogue of Akaike information criterion CIPRES Cyber-infrastructure for Phylogenetic Research CTAB Cetyl trimethylammonium bromide DDT Dichlorodiphenyltrichloroethane DNA Deoxyribonucleic acid GTR + Γ The general time reversible model with gamma distributed rates HPD Highest posterior density IRs Inverted repeats K-Pg Cretaceous-Paleogene LC-MS Liquid chromatography-mass spectrometry LR Likelihood ratio LSC Large single copy MCMC Markov Monte Carlo chains NGS Next generation sequencing nrITS Nuclear internal transcribed spacers (including ITS1, 5.8S and ITS2) PP Posterior probability RJ-MCMC Reversible-jump Markov chain Monte Carlo RNA Ribonucleic acid rRNA Ribosomal RNA SSC Small single copy TBR Tree-bisection reconnection tRNA Transfer RNA UCLN Uncorrelated lognormal relaxed clock WCSP World Checklist of Selected Plant Families

Hou, C. PhD thesis 2016

Introduction

The order Gnetales consists of three families, Ephedraceae, Gnetaceae and Welwitschiaceae. Each family is monogeneric and comprises Ephedra (Fig. 1a), Welwitschia (Fig. 1b) and Gnetum (Fig. 1c-d), respectively. The monophyly of the Gnetales, first inferred based on morphology (Hooker 1863, Crane 1985), has been repeatedly confirmed using molecular data, and Ephedra is sister to the clade that comprises Gnetum and Welwitschia (Price 1996, Bowe et al. 2000, Rydin et al. 2002, Rydin et al. 2004, Rydin & Korall 2009). Ephedra comprises more than 50 species with the life forms shrubs, climbers or small trees (Price 1996, Rydin & Korall 2009, Thureborn & Rydin 2015). The geographic distribution of the genus covers arid and semi- arid regions of northern Africa, western Europe to eastern Asia, and the Americas (Fig. 1e, Kubitzki 1990a). Some Asian species of Ephedra, such as E. equisetina, E. monosperma, E. regeliana and E. sinica are the natural source of ephedrine, an alkaloid commonly used as stimulant and treatment in traditional medicine (Huang 1998, Kitani et al. 2009, Kitani et al. 2011). Welwitschia comprises a monotypic species, W. mirabilis, a shrub distributed in the Namib Desert (Fig. 1e, Hooker 1863, Pearson 1907, Pearson 1909, Rodin 1953, Kubitzki 1990d). Gnetum comprises more than 30 species with the life forms trees, shrubs and liana (Price 1996, Hou et al. 2015). The genus inhabits tropical and subtropical forests of South America, Africa and South Asia (Fig. 1e, Kubitzki 1990c). Vegetative parts of Asian Gnetum (Fig. 1d), such as bark of G. gnemon, G. latifolium and G. montanum can provide high-quality fibres for fishing nets, threads and paper production (Markgraf 1951, Fu et al. 1999). Mature seeds of Gnetum (Fig. 1c) can be used for production of edible oil, soup and wine (Fu et al. 1999). Besides, biochemistry extracted from the Asian species of Gnetum is reported to possess pharmacological effects, e.g., anti-oxidant (Yao & Lin 2005), anti- inﬂammatory (Yao & Lin 2005, Yao et al. 2006) and anti-influenza (Liu et al. 2010).

Gnetales are characterized by decussate and opposite phyllotaxis (sometimes whorled), multiple axillary buds, ovules surrounded by one to two seed envelopes, a thin integument that protrudes and forms a micropylar tube, and a granular infratectal layer in the pollen wall. Other distinctive characters of the Gnetales are vessel elements with perforation plates evolved from a basic plan of circular bordered pits, a character that used to be considered homolo- gous with angiosperms but now believed to be independently evolved (Thompson 1918, Carlquist 1996) because the perforation plates of angiosperm vessels are derived from pits of a scalariform shape. There are many such examples; the direction of ovule development in Gnetales is acropetal, which is different from the basipetal direction in angiosperms (Endress

10 Hou, C. PhD thesis 2016

1996). Similarly, the seed envelopes that surround gnetalean ovules have long been compared with the angiosperm carpel, but the Gnetales clearly possess the gymnospermous mode of reproduction, with an exposed ovule. The Gnetales have a double fertilization mechanism, but eventually generate two zygotes rather than one embryo and one endosperm present in angiosperms (Friedman 1992, Carmichael & Friedman 1996). Therefore, the Gnetales and angiosperm are believed independently derived and similarities between the two groups are most probably superficial. Although most con- temporary scientists consider the Gnetales most closely related to conifers, the question of whether or not the Gnetales are most closely related to angiosperms remains unresolved despite many attempts to address the question (Arber & Parkin 1908, Pearson 1929, Eames 1952, Carlquist 1996, Crane 1996, Endress 1996, Price 1996, Mathews 2009, Friis et al. 2011). Seed plant phylogeny is still not resolved (Rydin et al. 2002, Rydin 2005, Mathews 2009), despite many previous efforts. Thus, all these unresolved questions make the evolutionary history of the Gnetales very intriguing.

Despite sharing a common ancestry and several uniquely derived features, Ephedra, Gnetum and Welwitschia reveal dramatic differences in both vegetative and reproductive morphology (Thompson 1912, Pearson 1929, Kubitzki 1990b, Endress 1996, Price 1996, Yang et al. 2015). For example, leaves of Ephedra are highly reduced (Fig. 1a), often scale-like, and with 2-3 parallel veins that converge at the tip; leaves of Welwitschia are large and strap-shaped (Fig. 1b), with parallel primary veins connected by Y-shaped second-order veins; leaves of Gnetum are petiolate and similar to those of many eudicots (Fig. 1e), with a venation that appears pinnate and reticulate. The female cone of Ephedra has several decussately arranged bracts (Fig. 1a), but normally only the uppermost two bracts (rarely one or three) are fertile and subtend female units; the female cone of Welwitschia have decussately arranged bracts piled in a cone axis (Fig. 1b), each of which subtends a female unit; the female strobili of Gnetum are spike-shaped, bearing several whorls of irregularly placed female units on collars (Fig. 1h). In addition to fertile male units, male cones of the three genera often comprise sterile female structures (rarely present in Ephedra). Although the varied gross morphology of the three genera provide valuable clues that have been used to place some fossils phylogenetically, e.g., Cratonia cotyleton (Rydin et al. 2003), the substantial extinct diversity and unknown ancestral state of many features make it difficult to place most fossils with certainty.

11 Hou, C. PhD thesis 2016

Chapter 1 Pollination biology in the Gnetales

1.1 Anemophily versus entomophily

Pollination biology of the Gnetales interests plant scientists since the order possesses morphologically bisexual reproductive organs (Pearson 1929, Endress 1996), which are prevalent in angiosperms but rare in gymnosperms (Friis et al. 2011). The bisexual reproductive organs are, however, function- ally unisexual. Previous efforts have been performed to reveal the pollination syndrome of the Gnetales. In Ephedra, anemophily (wind pollination) is prevailing (Rydin & Bolinder 2015) and has been documented, for example, for E. trifurca (Buchmann et al. 1989), E. distachya (Bolinder et al. 2016), E. nevadensis (Bolinder et al. 2015a, Niklas 2015). Two species, E. foeminea (= E. campylopoda) (Porsch 1910, Bolinder et al. 2016) and E. aphylla (Bino et al. 1984, Meeuse et al. 1990) have been described as entomophilous (insect-pollinated). Welwitschia mirabilis, the monotypic species of Welwitsch- ia, has also been described as entomophilous (Pearson 1907). Wetschnig and Depisch (1999) argue that anemophily is present in the species, but unimportant for reproductive success(Wetschnig & Depisch 1999). In Gnetum, entomophily is mostly likely prevailing and has been suggested for all studied species, i.e., G. gnemon (Kato & Inoue 1994, Kato et al. 1995), G. cuspidatum (Kato et al. 1995), G. luofuense (Corlett 2001). Authors of a recent pollination study of G. parvifolium argue that the species possesses both entomophily and anemophily (Gong et al. 2015). The statement is questionable, however, because the pollen grains of the species cannot be dispersed far enough from the microsporangia by wind, indicating that anemophily is unimportant in Gnetum, like in Welwitschia mirabilis.

1.2 Investigations of pollination drops in gymnosperms

Pollination drops play an essential role in pollination biology of gymnosperms. They are present in almost all gymnosperms (Strasburger 1872, Pearson 1929, Ziegler 1959, Takaso 1990), and the main function is analo- gous to that of the angiosperm stigma, i.e., to act as pollen trappers and pollen carrier to the internal parts of ovules for subsequent fertilization (Strasburger 1872, Owens et al. 1998, Nepi et al. 2009, Nepi et al. 2012). Although this main function of pollination drops has long been known, studied closely e.g., in extant and extinct conifers (beginning with Doyle 1945), details regarding pollen-pollination drop interactions are poorly understood and the role of pollination drops has probably been underestimated (Coulter et al. 2012, Little et al. 2014). A pioneer study inventories the biochemistry of pollination drops in Taxus baccata and Ephedra distachya ssp. helvetica, and demonstrates that sugars, amino acids, peptides and minerals are present in pollination drops of the two species (Ziegler 1959). With the advents of

12 Hou, C. PhD thesis 2016 promoted mass spectrometry techniques, proteomic studies have gained re- newed focus on investigating unknown functions of pollination drops in gymnosperms (Prior et al. 2013). O’Leary et al. (2004) found arabinogalactan proteins in pollination drops of Taxus × media. Poulis et al. (2005) launched a proteomic study in the pollination drops of Douglas fir (Pseudotsuga menziesii) and detected several types of proteins: invertases, galactosidases, peroxidases and xylosidases. The proteins were considered to participate in the process of pollen tube elongation and nutrition, but also probably engage in filtering out external pollen from other species as a result of heterospecific pollen selection. The initial attempts lead to extensive sur- veying of the protein profiles in pollination drops of various gymnosperm species, conifers, Chamaecyparis lawsoniana, Juniperus communis, Juni- perus oxycedrus, as well as members of the Gnetales, e.g. Welwitschia mirabilis (Wagner et al. 2007). Additional functions of pollination drops have gradually been revealed, for example, anti-bacterial and anti-viral, and defense mechanisms (Wagner et al. 2007, Coulter et al. 2012).

1.3 A proteomic study of pollination drops of Ephedra

The achievements of previous proteomic studies in pollination drops encour- aged us to survey and deepen the understanding of proteomics in pollination drops of the Gnetales. The female ovules of the Gnetales, like those of other gymnosperms, lack physical barriers for protection, e.g., against pathogens, which can be accidently engulfed from external environment during the course of pollination. In the Gnetales, only chitinase (a pathogen defense protein) had been found in pollination drops of Welwitschia mirabilis (Wagner et al. 2007), but little was known about protein profiles in drops of Ephedra and Gnetum. In addition, unlike anemophilous conifer species, proteomics might be shifted in entomophilous species of the Gnetales as the result of co-evolution with pollinators. A proteomic study of pollination drops in Ephedra (Paper I) was conducted using four species, E. foeminea, E. minuta, E. likiangensis and E. distachya. Pollination drops of E. foeminea and E. distachya were collected in Asprovalta, Greece during 2011-2012, whereas pollination drops from E. minuta and E. likiangensis were obtained in the botanical greenhouse of Stockholm University, Sweden during 2011- 2012. Daily collections from different specimens of Ephedra were manually pooled with the total amount up to 50 μl and stored at -20 . To roughly assess amounts and types of proteins, partial samples of each species were applied using gel electrophoresis. To reveal protein profiles of℃ each species, proteins were liquid-liquid extracted and analyzed using tandem mass spectrometry (see details of methods in Paper I and Paper II). The functions of detected proteins were predicted by inquiring a set of proteomic databases. The results of Paper 1 indicate that amounts and types of detected proteins are considerably reduced in pollination drops of Ephedra compared with those in conifers. The number of detected proteins vary among the four spe-

13 Hou, C. PhD thesis 2016 cies, ranging from eight (E. foeminea) to two (E. minuta), but they all occur in very low concentrations compared to what has been found in conifers. Furthermore, degradome proteins account for the majority of the detected proteins. Such proteins are waste products, and in this case are derived from intracellular tissues as a result of nucellus degradation. It probably occurs during the process of pollination and pollen chamber formation. Degradome proteins have never before been found or carefully studied in pollination drops of gymnosperms, and the reason is probably that studied conifer species do not produce a pollen chamber by degradation of the nucellus caps during the course of ovule development. Unlike degradome proteins, secretome proteins (products of apoplastic secretion) are thought to be useful for plants. For example, thaumatin-like proteins were found (Paper I) and predicted to destroy external pathogens in situ. These proteins have also been found in pollination drops of conifers (O'Leary et al. 2007, Wagner et al. 2007). In addition, protein profiles of the entomophilous E. foeminea were compared with those of the other, anemophilous, species (Paper I). The result shows that the pollination drops of E. foeminea are relatively “protein- rich”, probably an evolutionary consequence of plant-insect interactions. However, the ecological functions of identified proteins are often poorly understood and will be difficult to test in the future because all detected proteins are present in very low concentration in the pollination drops of Ephed- ra.

1.4 A new proteomic study

A subsequent proteomic study of pollination drops was conducted on seven Ephedra species, i.e., E. compacta, E. distachya, E. foeminea, E .likiangensis, E. minuta, E. monosperma and E. trifurca (Paper II). In the study, extensive efforts were made to further reveal the differences of protein profiles among different species. Analytical methods were identically applied in Papers I and II. The results of Paper II support those of Paper I that the majority of detected proteins were degradome proteins, and that the number of detected proteins differ dramatically among species, ranging from twenty (E. foeminea) to six (E. monosperma). It is, however, very surprising and worrying that the protein profiles shown in Paper I differ considerably from those in Paper II (see Table 4 in Paper I), even though the same samples were applied in the two studies. It may indicate that the results of these studies are not repeatable. Furthermore, the new proteomic study (Paper II) assesses the protein profile of E. monosperma over time. The result shows that the protein profile changes over time in the pollination samples of the species. Taken together, these findings question the robustness of the proteomic results in pollination drops of Ephedra as well as in previous proteomic studies in conifers. Pollination drops of conifers are very small and therefore routinely pooled to ensure sufficient sample size for the proteomic analyses. If temporal variation of protein components is present broadly in pollination

14 Hou, C. PhD thesis 2016 drops, it is of highest importance not to pool the samples in the field. Even if not pooled, it may be difficult to obtain results that are comparable across species if it requires that the samples are taken from ovules of the exact same developmental stage. New proteomic studies of pollination drops require robust and more careful methods for collection and analyses of proteins in gymnosperms.

1.5 Prospective studies

Pollination biology of the Gnetales has gained an increasing attention in the recent five years, not only in the field of proteomic studies of pollination drops, but also in fields of gross morphology of reproductive organs, pollen morphology as well as field observation. Studies of carbohydrates are more promising for investigations of ovule defense and plant-insect interactions than are proteomic studies. Pollination drops of the Gnetales, like nectar in angiosperms, have high sugar concentrations: as much as 25% was measured in E. distachya (Ziegler 1959), 14-16% in G. cuspidatum and 3-13% in G. gnemon (Kato et al. 1995), all of which are much higher than the 1.25% described for Pinus nigra (McWilliam 1958). The sugary pollination drops of the Gnetales can probably function both to attract and reward pollinators, at least in Ephedra (Bolinder et al. 2016). Gnetum and Welwitschia might not only use pollination drops as the attraction, in fact, scent emission has also been documented in the two genera (Kato et al. 1995, Endress 1996). A few studies (Bino et al. 1984, Kato et al. 1995) have indicated that extraovular nectar occurs in a few species of the Gnetales disregarding pollination drops (in two species, Ephedra aphylla, Bino et al. 1984, and Gnetum cuspidatum, Kato et al. 1995). A recent study reveals, however, that pollination drops secreted from female reproductive units provide the reward for pollinators in Gnetum cuspidatum, not nectar produced by extraovular nectaries (Jörgensen & Rydin 2015). In addition, Rydin & Bolinder (2015) found a correlation between pollination drop secretion and phases of the moon in the entomophilous Ephedra foeminea. The system has most likely evolved in co- evolution with the nocturnal pollinators, who utilize moonlight for efficient navigation. The studied anemophilous species, however, do not possess such a trait. It is highly unlikely that a similar system is present in Gnetum, which inhabits tropical rain forests, but the entomophilous Welwitschia, which like Ephedra exists in an open environment with low precipitation, may potentially have the same system although it has not been studied. Another field of pollination study, investigation of pollen morphology has also proven help- ful to predict the pollination syndromes of the Gnetales. Bolinder et al. (2015a) found that the pollen wall ultrastructure of E. foeminea is denser than that of other studied (anemophilous) species. It further lacks ability to create the aerodynamic microenviroments created around the female cones of anemophilous species of Ephedra, which use pollination drops to capture

15 Hou, C. PhD thesis 2016 air-borne pollen grains. Similar comparisons are made in pollen ultrastructure of Gnetum. The granular layer of the pollen wall in G. africanum is spa- cious and possesses few but large granules (Tekleva & Krassilov 2009). This stands in a sharp contrast with the pollen walls of other studied Gnetum species whose granular layers are filled with small granules (Yao et al. 2004, Tekleva & Krassilov 2009). Together with the presence of unisexual male cones, these findings indicate that G. africanum could be anemophilous rather than entomophilous (Jörgensen & Rydin 2015, Rydin & Hoorn 2016). Entomophily is probably the ancestral state in the Gnetales and anemophily has thus evolved at least once in the Gnetales (Bolinder et al. 2015a, Rydin & Bolinder 2015, Bolinder et al. 2016), perhaps twice. Additional studies are, however, needed to better understand the pollination biology of the Gnetales.

16 Hou, C. PhD thesis 2016

Chapter 2 Phylogenetic studies and dating analyses of the Gnetales

2.1 Phylogenetic studies in Ephedra and taxonomic implications

Early phylogenetic studies of Ephedra, e.g., Rydin et al. (2004), Ickert-Bond & Wojciechowski (2004), Huang et al. (2005), indicate that it is difficult to reveal relationships in the genus. The results were poorly resolved and sup- ported, and deep divergences differed among the studies. A subsequent study (Rydin & Korall 2009) applied an increased sampling of taxa (104 ingroup accessions and 100 outgroup accessions, which represent the main lineages of vascular plants) and seven molecular markers (18S, 26S, nrITS, rbcL, rpl16, rps4, trnS-trnfM). The reconstructed phylogeny of Ephedra places E. foeminea as sister to all the remaining species. Among the remaining species several Mediterranean groups are successive sisters to a monophyletic group that comprises American and Asian species.

However, several problems remain in phylogenetic reconstructions of Ephedra. Although the early divergences in the Ephedra phylogeny is resolved in Rydin & Korall (2009), it receives low statistic support, and species delimitations, especially the Mediterranean species complex, are uncertain. Besides of interspecific variation, intraspecific variation is investigated and discussed in population genetic studies of the Old World species (Qin et al. 2013, Wu et al. 2016) as well as the New World species (Loera et al. 2012, Loera et al. 2015). These studies corroborate that the interspecific variation can be as large as intraspecific variation, indicating that the bound- aries are very vague in closely related species of Ephedra. The statement is also congruent with results in studies of morphological, anatomical and his- tological characters in female reproductive units (Rydin et al. 2010), micromorphology of seeds (Ickert-Bond & Rydin 2011) and pollen (Bolinder et al. 2015b). Nevertheless, the reconstruction of phylogenies in Ephedra allows for tests of traditional classification schemes. Stapf (1889) divided Ephedra into three sections based on cone morphology, i.e., E. section Pseudobacca- tae, E. section Alatae and E. section Ascarca. The classification scheme of Ephedra, however, did not gain support using molecular data (Ickert-Bond & Wojciechowski 2004, Huang et al. 2005, Rydin & Korall 2009), indicating that previous classification of the genus is artificial. A new taxonomy of the genus is urgently needed. In addition, it will be quite interesting to test the phylogenetic position of several new species proposed in the last decade using molecular data, for example E. sumlingensis (Sharma & Singh 2015).

2.2 Dating analyses and evolutionary history of the Ephedraceae

17 Hou, C. PhD thesis 2016

Node ages in Ephedra, especially the ancestral nodes in the phylogeny, remain uncertain. The crown age of Ephedra has been previously assessed about 8-32 Ma under a strict clock (Huang & Price 2003, Won & Renner 2003). The results, however, are not convincing because the taxon sampling of the studies is quite small. A subsequent dating analysis of Ephedra (Ickert-Bond et al. 2009) applies a broad sampling with 53 accessions as the ingroup and 41 accessions as the outgroup. The study utilized ten molecular markers generated in previous phylogenetic studies, and also applied a Wel- witschia-like fossil, Cratonia cotyledon (Rydin et al. 2003), to constrain the age of the Gnetum-Welwitschia clade. The results show that the crown age of Ephedra is about 30 Ma. The estimated ages of extant species are thus much younger than discovered Ephedra-like meso- and megafossils from the Early Cretaceous (Crane 1996, Rydin et al. 2004, Rydin et al. 2006a, Rydin et al. 2006b, Wang & Zheng 2010, Yang et al. 2013, Liu & Wang 2015). These fossils are not members of the crown group Ephedra but represent extinct stem lineage(s) (Rydin et al. 2010).

Based on palynological data, the species diversity of Ephedra was shown to be elevated in the Early Cretaceous but declined in the Late Cretaceous (Crane & Lidgard 1989, Lidgard & Crane 1990). Cenozoic diversity has never been assessed in such a summarizing way, but preliminary results indicate substantial fluctuations also during this era (Bolinder and Rydin in prep.). The fluctuating pattern of biodiversity is probably influenced by both abiotic and biotic factors. For example, orogenetic movements and gradual acidification in Pliocene and Pleistocene may have facilitated the speciation process in Old World clades (Qin et al. 2013) as well as New World clades (Loera et al. 2012). Also biotic factors, such as varied pollination modes (Bolinder et al. 2015a, Rydin & Bolinder 2015, Bolinder et al. 2016) and seed-dispersal vectors (Hollander & Vander Wall 2009, Hollander et al. 2010, Loera et al. 2015) could play indispensable roles in the speciation and extinction processes of Ephedra.

2.3 Phylogenetic studies in Gnetum and taxonomic implications

The species diversity of Gnetum has been poorly understood although efforts have been made to delimit species based on gross morphology and geographic distribution. In the most recent monograph of Gnetum (Markgraf 1930), 30 species from South America, Africa and Asia were distinguished. Markgraf 1930 divided the genus into two sections G. sect. Gnemonomorphi and G. sect. Cylindrostachys, which roughly corresponds to the taxonomic framework proposed by Griffith (1859). The arborescent species were considered to constitute an ancestral lineage of Gnetum (Markgraf 1930). The phylogeny of the genus was first assessed in a modern framework in a series of studies using molecular data (Won & Renner 2003, 2005b, a, 2006). The studies revealed that Gnetum was a monophyletic group and indicated, alt-

18 Hou, C. PhD thesis 2016 hough with poor support, that the South American lineage diversifies early in the evolutionary history (Won & Renner 2006). Several questions re- mained, however, i.e., 1) the deepest splits in the genus 2) the monophyly of sections and subsections 3) the phylogenetic relationships and monophyly of closely related species.

In paper IV, phylogenetic analyses were launched using an extensive sampling of 58 accessions (representing 27 species of Gnetum) and 39 outgroup accessions representing the remaining Gnetales, other seed plants, ferns, and lycopods. Besides, a subsequent analysis was conducted using an increased taxon sampling of outgroup species of the Gnetales (16 accessions of Ephed- ra and one accession of Welwitschia) plus a conifer accession (Calocedrus sp), to root the trees. Six species that had never been sequenced before was included, i.e., G. buchholzianum, G. camporum, G. indicum, G. leptostach- yum, G. leyboldii and G. montanum. The results reveal that Gnetum section Erecta Griff. 1859 (section Gnemonomorphi Markgr. 1930) is non- monophyletic and comprises several clades. South American species (G. subsect. Araeognemones in Markgr. 1930) form a clade that is sister to the remaining Gnetum species with strong support. The African species comprise a monophyletic group (G. subsect. Micrognemones in Markgr. 1930), which separates from a monophyletic Asian group (the “core” Gnetum). The Asian clade comprises a clade of arborescent species (G. subsect. Eugnemo- nes in Markgr. 1930) and a clade of lianoid species (G. sect. Scandentia Griff. 1859, G. sect. Cylindrostachys in Markgr. 1930). The Asian lianoid clade comprises a clade of G. gnemonoides and G. raya and the remaining clade, which consists of a South East Asia clade and a Chinese clade. The results reveal that G. subsections Stipitati and Sessiles (Markgr. 1930) are polyphyletic.

Although paper IV aimed at resolving major relationships in Gnetum, the results also provide resolution among closely related species in most clades. There is one exception though; phylogenetic relationships and species delimitations in the Chinese lianoid Gnetum clade are not resolved and specimens representing several species are nested. The reasons are probably several: 1) genetic markers applied in Paper IV do not have sufficient sequence variation to resolve the phylogeny in this clade; 2) the sampling of the clade is restricted and mainly based on cultivated plants; 3) misidentification of the specimens might occur due to contradictory taxonomic conclusions in previous literatures; 4) other reasons, for example, hybridization, incomplete lineage sorting and horizontal gene transfer.

Hence, new efforts were made to investigate the phylogeny and species delimitations in the Chinese clade of Gnetum (Paper V). The study applies freshly collected material from tropical and subtropical forests of southern China (Guangdong, Guangxi, Hainan, Fujian, Yunnan and Hong Kong), as

19 Hou, C. PhD thesis 2016 well as herbarium material. To begin with, phylogenetic analyses were conducted based on eleven chloroplast genomes representing G. gnemon, G. luofuense, G. montanum, G. pendulum and G. parvifolium. Besides, we also build phylogenies using a concatenation of protein coding sequences. Pair- wise comparisons of the five chloroplast genomes were performed to reveal sequence divergences, and four targeted molecular markers i.e., matK, rpoC1, rps12 and trnF-trnV were designed and applied for Sanger sequencing. The subsequent phylogenetic reconstruction was conducted using five molecular markers (nrITS, matK, rpoC1, rps12 and trnF-trnV) and a dense taxon sampling, including 49 ingroup accessions and 16 outgroup accessions. The trees were rooted on an African species (G. africanum) according to results of paper IV. The results of Paper V indicate that Chinese lianoid Gne- tum form a monophyletic group, which is congruent with previous phylogenetic studies. The accessions of Gnetum parvifolium form a monophyletic group, which is sister to the remaining Chinese lianoid species. To further define species delimitations and make taxonomic conclusions, a morphological study was performed based on 156 herbarium sheets including 40 type sheets. The material used in Paper V represents all currently accepted Chi- nese lianoid species, and a few more with uncertain taxonomic status. Sub- sequent these investigations, new taxonomic conclusions were made based on molecular as well as morphological data. We conclude that Chinese lianoid Gnetum comprise six species, i.e., G. catasphaericum H.Shao, G. for- mosum Markgr., G. luofuense C.Y.Cheng, G. montanum Markgr., G. pendulum C.Y.Cheng and G. parvifolium (Warb.) W.C.Cheng. Gnetum giganteum H.Shao and G. gracilipes C.Y.Cheng are considered synonymous with G. pendulum. Gnetum hainanense C.Y.Cheng ex L.K.Fu, Y.F.Yu & M.G.Gilbert is synonymous with G. luofuense. The validity of G. cleistostachyum C.Y.Cheng and G. indicum (Lour.) Merr. is also discussed, both of which are considered questionable about the name application. In addition, modified taxonomic keys of Chinese lianoid Gnetum based on male and female plants are provided in Paper V.

2.4 Dating analyses and evolutionary history of Gnetum Compared with Ephedra, the evolutionary history of Gnetum is poorly understood. It is partially because the distribution of Gnetum is restricted to (sub)tropical forests, from which both mega- and microfossils are usually poorly preserved. So far two megafossils i.e., Khitania (Guo et al. 2009) and Siphonospermum (Rydin & Friis 2010) are considered to be of possible Gne- tum affinity, both discovered from the Early Cretaceous Yixian Formation, China. Unfortunately, only parts of the plants are preserved and the preserva- tion state is relatively poor, resulting in uncertain phylogenetic positions. Consequently, the fossils cannot be used as calibration points in dating analyses. In addition, Gnetum-like microfossils, i.e., small, spherical and echi- nate pollen (Yao et al. 2004, Tekleva & Krassilov 2009, Tekleva 2015) are

20 Hou, C. PhD thesis 2016 largely unknown in the fossil record (Friis et al. 2011). This stands in sharp contrast with the record of polyplicate microfossils of Ephedra and Wel- witschia, which is common in Cretaceous strata (Crane & Lidgard 1989), and in some Cenozoic localities as well (Han et al. 2016, Bolinder et al. in progress). The reason may be that the ectexine of Gnetum is thinner than that of Ephedra and Welwitschia and sensitive to taphonomical activities (Tekleva 2015). Initial efforts of dating analyses were made in Won & Renner (2006). The crown age of Gnetum was estimated to the Oligocene (ca. 26 Ma) and the Asian grown group to the Miocene (ca. 22 Ma), indicating that the vicariance, break up of old continents, exerts no impact on clade formation of Gnetum. The results of new dating analyses in Paper IV, however, indicate that these estimated ages of early divergences in Gnetum are underestimated. It is partially because methods of dating analyses have ad- vanced considerably in recent years and stratigraphic information has been recently updated, but may also be because the sample size applied in previous dating analyses (Won and Renner 2006) is limited.

A new dating analysis was, therefore, conducted in Paper IV, based on the dataset comprising 20 ingroup accessions as well as 39 outgroup accessions. To calibrate the results to absolute time, several well-preserved fossils with updated stratigraphic knowledge were applied. A strict clock was rejected and a relaxed uncorrelated clock was thus applied, using two alternative tree priors i.e., Yule process and birth-death process, of which the latter had sig- nificantly better fit to the data. The results of Paper IV date the crown age of Gnetum to the Campanian (Late Cretaceous, ca. 81 Ma, 95% highest posterior density, HPD, 64-98 Ma). The South American clade was dated to the Paleogene-Neogene boundary (22 Ma, 95% HPD: 8-39 Ma), and the Asian clade to the Cretaceous-Paleogene (K-Pg) boundary (65 Ma, 95% HPD: 48- 82 Ma). Discrepancies of the assessed ages are considerable between the results of Paper IV and the previous study in Won & Renner (2006), and it intrigues us to reconsider the underlying geographical and environmental factors that may have influenced evolution in Gnetum. Early divergence events of Gnetum appear to correlate well with the final separation of South America and Africa in the mid-Cretaceous, as assessed both from absolute dates and phylogenetic patterns. Later dispersal, e.g., among continents and islands of South East Asia, is also apparent from the results in Paper IV, and is likely to have influenced speciation and geographic distribution of extant species within a region or continent. The seed dispersal modes of Gnetum are documented to be zoochory (by birds, civet-cates, fishes and rodents) as well as hydrochory (ocean trends) (Ridley 1930, Markgraf 1951, Kubitzki 1985, Forget et al. 2002), but knowledge of seed dispersal in the genus is patchy and incomplete.

Divergence times were also estimated among Chinese lianoid species (Paper V). Normally distributed age priors (based on results in Paper IV) with en-

21 Hou, C. PhD thesis 2016 forced monophyly were used to calibrate the phylogeny to absolute time. The results reveal that the mean crown age of Chinese lianoid Gnetum is 21 Ma (HPD: 11-34 Ma) corresponding to the Oligocene to Miocene, and diversification events during the latter epoch gave rise to the major clades within the Chinese lianoid group. The crown age of G. parvifolium is 6 Ma (late Miocene to Pliocene, HPD: 2-11 Ma). The crown age of G. pendulum and G. montanum is 2 Ma (HPD: 1-4 Ma), which is slightly younger than the crown age of G. luofuense (3 Ma, HPD: 1-5 Ma) and the crown age of G. catasphaericum (4 Ma, HPD: 2-7 Ma). The diversification in the Chinese lianoid Gnetum during the Neogene coincides with a period of expansion of tropical and subtropical areas in southern China (Yao et al. 2011).

2.5 Prospective studies

Despite various challenges, plant systematic studies of the Gnetales have made considerable progresses in the current decade. Among other things, continuous efforts have been made recently to resolve the early divergences and estimate divergence ages in Ephedra (Thureborn & Rydin 2015). The new phylogenetic study applies four nuclear ribosomal (18S, 26S, nrITS, nrETS) and five chloroplast markers (rbcL, matK, rpl16, rps4 and trnS- trnfM) using an increased sampling of Mediterranean specimens. The study well resolves the relationships of E. foeminea, the other Mediterranean clades, and the remaining species, and also discusses monophyly and delimitations of Mediterranean species. In addition, ongoing dating analyses apply new understanding of the fossil pollen records (Bolinder et al. 2015b) to calibrate the phylogeny within the crown group, resulting in considerably older ages than previously estimated for the early divergence of Ephedra (Thureborn & Rydin 2015). Previous hypotheses of long distance dispersal in Ephedra (Ickert-Bond et al. (2009) may need to be revised. Prospective studies of Ephedra can focus on the biogeography to inspect the possible underlying mechanisms. Future phylogenetic studies of Gnetum should focus on the species delimitation within other (sub)clades of the genus (other than the Chinese clade), and test the taxonomic conclusions at the species and subspecies levels. One ongoing project concerns species delimitations in the paraphyletic assemblage Gnetum section Erecta Griff. 1859 (=section Gnemonomorphi Markgr. 1930) using extensive sampling from several Afri- can and arborescent Asian species. Besides, to well explain the historical diversification and biogeography of Gnetum, continuous studies are needed. They could utilize ecological niche models to address the impacts of biotic and abiotic factors on abundance and distribution of extant species.

Last but not least, speciation and extinction processes of the Gnetales seem to correlate with global climate change. Between the late Eocene and Oligo- cene, the global climate became much drier and cooler (Zachos et al. 2001). During the early phase of Miocene, the global climate became warmer again

22 Hou, C. PhD thesis 2016 before it turned into the present cooler environment (Dutton & Barron 1997, Zachos et al. 2001). The underlying mechanisms that drove the diversification of the Gnetales seem often associated with paleoclimatic fluctuations. One example from my work is diversification in the Chinese lianoid clade of Gnetum, which correlates in time with the expansion of tropical forests in South East Asia (Paper V). However, other factors are conceivably also important. Preliminary results in an ongoing project of Bolinder and Rydin, which aims to uncover the correlation between dynamic patterns of diversity in Ephedra and altered paleoclimate, indicate that also biotic factors are likely to have had a strong impact on ephedran evolution.

23 Hou, C. PhD thesis 2016

Chapter 3. Chloroplast genome investigations in the Gnetales

3.1 Nuclear genomes of the Gnetales Nuclear genome sizes among the three gnetalean genera are considerable different (Leitch & Leitch 2013): Gnetum has the smallest nuclear genome (1C values = 2.25-3.98 pg), including diploids (2n = 22) and tetraploids (2n = 44). Welwitschia mirabilis has a larger nuclear genome than Gnetum (1C = 7.2 pg) with the chromosome number 2n = 42 (Leitch & Leitch 2013). Ephedra possesses the largest and most variable nuclear genomes in the Gnetales (1C = 8.8-18.22 pg), with the chromosome numbers 2n = 14 and 2n = 28 (Leitch & Leitch 2013). The nuclear genome size of Ephedra has been recently investigated using flow cytometry and chromosome count methods (Ickert-Bond et al. 2015). In the study, the variation of genome size in the genus was shown to be even greater than previously estimated (1C = 8.09 - 38.34 pg), which in addition indicates that Ephedra may have the largest genome size among gymnosperms (i.e., 38.34 pg, 2n = 8x = 56 in Ephedra antisyphilitica (Ickert-Bond et al. 2015). The study also highlights the pre- dominance of polyploidization in Ephedra and demonstrates that whole genome duplication have facilitated speciation in the genus. The statement is corroborated by a very recent study of population genetics and ecological niche modelling among 12 Chinese Ephedra species (Wu et al. 2016). Un- fortunately, completely sequenced nuclear genomes of the Gnetales are, so far, not available.

Compared with nuclear genomes, chloroplast genomes are better understood in the Gnetales. The size of the chloroplast genome varies among the three genera according to Wu et al. (2009): Ephedra has the smallest chloroplast genome in the order (E. equisetina, 109 518 bp) and Welwitschia the largest (W. mirabilis, 118 919 bp). Gnetum is in between the two (G. parvifolium, 114 914 bp). The chloroplast genomes of the Gnetales are believed to be more compact and reduced than those of other land plants (Wu et al. 2009), except for some species of Pinaceae (e.g., Pinus koraiensis, 116 866 bp) (see Table 1 in McCoy et al. 2008). Intraspecific variation of genomic sizes in the Gnetales are poorly investigated but may occur as indicated by comparison of the two chloroplast genomes of W. mirabilis, generated by McCoy et al. (2008) and Wu et al. (2009) respectively. There is a 807 bp difference in length between the two. Regarding the structure, the chloroplast genome of Ephedra is characterized by fewer introns, short intergenic spaces and higher gene densities compared with the chloroplast genomes of Welwitschia and Gnetum (Wu et al. 2009). Moreover, the gene content shifts among the chloroplast genomes in the Gnetales. For example, the gene chlL is found in Ephedra but is deficient in the other two genera (McCoy et al. 2008, Wu et al. 2009).

24 Hou, C. PhD thesis 2016

3.2 Structural differences among chloroplast genomes in the Gnetales Compared with nuclear genomes, chloroplast genomes are better understood in the Gnetales. The size of the chloroplast genome varies among the three genera according to Wu et al. (2009): Ephedra has the smallest chloroplast genome in the order (E. equisetina, 109 518 bp) and Welwitschia the largest (W. mirabilis, 118 919 bp). Gnetum is in between the two (G. parvifolium, 114 914 bp). The chloroplast genomes of the Gnetales are believed to be more compact and reduced than those of other land plants (Wu et al. 2009), except for some species of Pinaceae (e.g., Pinus koraiensis, 116 866 bp) (see Table 1 in McCoy et al. 2008). Intraspecific variation of genomic sizes in the Gnetales are poorly investigated but may occur as indicated by comparison of the two chloroplast genomes of W. mirabilis, generated by McCoy et al. (2008) and Wu et al. (2009) respectively. There is a 807 bp difference in length between the two. Regarding the structure, the chloroplast genome of Ephedra is characterized by fewer introns, short intergenic spaces and higher gene densities compared with chloroplast genomes of Welwitschia and Gne- tum (Wu et al. 2009). Moreover, the gene content shifts among the chloroplast genomes in the Gnetales. For example, the gene chlL is found in Ephedra but is deficient in the other two genera (McCoy et al. 2008, Wu et al. 2009).

3.3 A newly generated chloroplast genome of Ephedra

Despite previous efforts mentioned above, additional chloroplast genomes are required to reveal potential interspecific variation in terms of genomic structure and gene content, as well as to assess interspecific sequence variation for prospective phylogenetic studies. In Ephedra, the knowledge of chloroplast genomes was restricted to a single Chinese species, E. equisetina (Wu et al. 2009), and little was known about other species of Ephedra. The early diversifying lineages are especially interesting, not least E. foeminea, which differs from all other species of Ephedra in several respects. It is sister to the remaining genus (Rydin & Korall 2009, Thureborn & Rydin 2015), and differs morphologically from other species of Ephedra in that it has bisexual male cones. Besides, in contrast with most species of Ephedra including E. equisetina, which are anemophilous, E. foeminea has been confirmed to be entomophilous with a distinct ecological status (for details of pollination biology in Ephedra, see chapter 1). Therefore, we investigated the chloroplast genome of this Mediterranean species (Paper III). Plant material of E. foeminea was collected in Asprovalta, Macedonia, Greece in 2012. Extracted DNA was sequenced using the Illumina Hiseq 2500 (Illumina Inc., San Die- go, CA) and generated raw-reads were assembled into a chloroplast genome using reference-guided and de novo methods. Gene assessment and annota-

25 Hou, C. PhD thesis 2016 tion were largely transferred from the known chloroplast genomes of the Gnetales but were also assessed.

Our results reveal that the chloroplast genome of E. foeminea possesses a typical quadripartite structure, 109 584 bp in length (Fig. 2), comprising two identical inverted repeats (IRs, each 20 739 bp), a large single copy region (LSC, 60 027 bp) and a small single copy region (SSC, 8 079 bp). The assembled chloroplast genome of E. foeminea is 66 bp longer than that of E. equisetina. There are 118 genes in the chloroplast genome of E. foeminea, comprising 73 protein coding genes, 37 transfer RNA genes and 8 ribosomal RNA genes. The gene density of E. foeminea is 1.076 (genes per 1000 bp), which is slightly larger than the 1.068 of E. equisetina. The overall GC content is 36.7%, i.e., approximately the same as the 36.6% in E. equisetina. The LSC, SSC and IRs have GC contents of 34.1%, 27.5%, and 42.1%, respectively. A pairwise comparison of chloroplast genomes of E. foeminea and E. equisetina (Paper III) reveals 2352 variable sites, of which 1018 are point mutations and 1334 insertions and deletions (indels). The detected sequence variation is of great importance for designing highly variable markers to well resolve species delimitation of Ephedra in prospective studies.

3.4 Chloroplast genome investigation in Chinese lianoid Gnetum Compared with Ephedra, the chloroplast genomes of Gnetum are better understood; the genomes of several species are available on Genbank (G. gnemon NC_026301, G. montanum NC_021438, G. parvifolium NC_011942, G. ula AP014923). However, all these species belong in the Asian clade. In light of the results from previous studies (Wu et al. 2009, Hsu et al. 2015), chloroplast genomes of Gnetum have been known to possess a quadripartite structures i.e., LSC, SSC and IRs (see Fig. 3). In paper V, five chloroplast genomes (G. gnemon, G. luofuense [=G. hainanense], G. montanum, G. parvifolium and G. pendulum) were assembled using the same methods as mentioned above. We found that the overall length of the chloroplast genomes vary among species, ranging from 114,405 (G. gnemon) to 115,011bp (G. parvifolium). Also intraspecific variation is present, for example, three specimens of G. parvifolium have been studied and the assembled chloroplast genomes reveal remarkable variation in length (being 114 850 bp, 114 950 bp and 115 011 bp, respectively). Our results show that overall length of chloroplast genomes cannot act as a species-specific trait in the Gnetales. The discrepancies of overall length of chloroplast genomes are probably due to the presence and absence of repeats. Thus in paper V (unpublished results), sequences of the five studied Gnetum species were analyzed in REPuter (Kurtz et al. 2001). The setting of detection followed Huang et al. (2014): hamming distance 3, minimal repeat size 30 bp and three types of repeats, i.e., forward, reverse and palindromic repeats were

26 Hou, C. PhD thesis 2016 taken into account. The results reveal that the total numbers of repeats differ slightly in the chloroplast genomes of the five species (Fig. 4a): 43 repeats were found in G. parvifolium, 42 repeats in G. pendulum, and 40 each in G. gnemon, G. luofuense (=G. hainanense) and G. montanum. The lengths of the detected repeats vary from 35 bp to 153 bp across the five species, of which direct repeats are prevailing and reverse repeats are fewest (Fig. 4a-d). Among the detected 205 repeats, the majority (121 repeats) are located in intergeneric regions, whereas the remaining 84 repeats are situated in protein coding regions (Table 1). These detected repeats are considered to play a crucial role of rearranging architectures of the chloroplast genomes (Palmer 1991). Based on previously known chloroplast genomes and those generated in paper V, gene content and order are considered consistent in Asian Gne- tum. The chloroplast genomes consist of 115 genes, of which 66 are protein coding genes, eight ribosomal RNA (rRNA) genes, and 40 genes transfer RNA (tRNA) genes. The gene psbA located in the IRb region is a pseudo- gene with a reduced size compared with the counterpart located in the con- junction of IRa and LSC regions. The 5’end exon of gene rps12 is located in LSC and two 3’d end exon ends in IRa and IRb, respectively. In addition, chloroplast genomes of the five studied Chinese Gnetum species were aligned and variable sites were detected using a SNP/variation finder pro- gram (Paper V). The results show that there are 9345 variable sites, comprising 5600 base substitutions (60%) and 3745 insertions and deletions (indels, 40%). The majority of sequence variation was detected in non-protein regions (i.e., rRNA and tRNA genes, introns and intergenic spacers) whereas the protein regions are relatively conserved. The sequence variation provides useful information regarding the possibility to design variable molecular markers for further resolving relationships among closely related species of Asian Gnetum.

3.5 Prospective studies

The investigation of chloroplast genomes provides indispensable opportuni- ties to deepen the understanding of the evolutionary history of the Gnetales in terms of genome architecture and sequence variation. Alignments and comparisons of chloroplast genomes may reveal interspecific variation as well as intraspecific variation in the Gnetales. Besides, genomic data can also be applied to investigate population genetics and biogeographic patterns (Powell et al. 1995, Weising & Gardner 1999, Provan et al. 2001, Petit et al. 2005, Mariac et al. 2014). For example, we assessed single sequence repeats (SSRs) from the five chloroplast genomes of the Chinese lianoid Gnetum clade generated in paper V (unpublished results). The detection and identification of SSRs was conducted in Phobos with the selection criteria described in Huang et al. (2014). In total, 283 single sequence repeats (SSRs) were detected (Table 2), of which 61 SSRs were detected in G. montanum, 58 in G. parvifolium, 58 in G. pendulum, 56 in G. luofuense and 50 in G. gnemon.

27 Hou, C. PhD thesis 2016

The detected SSRs are promising to be further applied for population genetic studies in Gnetum. In addition, it would be very interesting to generate the complete nuclear genomes of the Gnetales in future studies. The nuclear genomes of Gnetum and Welwitschia are much smaller than the single one known in conifers, the Norway spruce (Nystedt et al. 2013), and may therefore be easier to assemble de novo. Besides, although the mitochondrial genome of Welwitschia mirabilis has been published very recently (Guo et al. 2016), knowledge is still lacking for Ephedra and Gnetum, and more efforts can be made in this field.

28 Hou, C. PhD thesis 2016

Chapter 4 Taxonomy and species delimitation in the Gnetales

Taxonomy and species delimitation is challenging in Ephedra and Gnetum. It is partially because gross morphology does not reveal considerable inter- species variation, particularly in Ephedra as well in as some Gnetum lineages, e.g., Chinese lianoid Gnetum. In addition, intraspecific variation is also present in some species that possess a broad geographic distribution, e.g., E. distachya (Kakiuchi et al. 2011) and G. parvifolium (Huang et al. 2010), which further complicates taxonomic choices and assessments of species delimitation. In addition, vegetative parts of the Gnetales may exhibit plas- ticity as a response to the environment. This is for example the case for Chi- nese lianiod Gnetum, as plant height and leave size vary dramatically among and within species. In addition, recent hybridization clearly occurs between some wild populations; apparently resulting for example in blurred species delimitation of anemophilous Ephedra specie (Ickert-Bond & Rydin 2011), but this has not been well studied so far.

Nevertheless, several issues can be considered to better resolve taxonomy and species delimitation in the Gnetales. First, alpha-taxonomic work requires an extensive sampling of specimens, preferably reproductive material. Besides, an extensive number of morphological characters should be investigated in order to avoid misidentification. For example, to resolve the taxonomy and species delimitations of 11 putative species of the Chinese lianoid Gnetum clade (Paper V), 40 type specimens and more than 126 sheets of supporting material were reviewed, and 21 morphological characters were measured and compared (see also paragraph 2.3 above). In addition to gross morphology, there is an ongoing project on anatomical structure and histolo- gy patterns of female reproductive units of Chinese lianoid Gnetum (Hou and Rydin, unpublished). The study aims to generate anatomical and histo- logical characters, and to find relevant characters useful for wider comparative studies of the entire genus and potentially also extinct relatives. Fur- thermore, it should be noted that qualitative and quantitative morphological characters often require new and rigorous studies in the Gnetales. For example, the number of sterile ovules in male spikes of certain Asian Gnetum species was recently reassessed and showed considerably different numbers compared with the original descriptions (Jörgensen & Rydin 2015). Last but not least, to combine morphological and molecular data (as is done in Paper V) is a powerful strategy to provide better corroborated results on taxonomy and species delimitation, in particular in groups like the Gnetales where morphological investigation and species recognition are difficult.

29 Hou, C. PhD thesis 2016

Concluding remark

Biodiversity of the Gnetales has been the focus of my PhD studies, but more efforts are needed in the future. At first, a robust and complete phylogeny of the Gnetales is important, but still not fully achieved. To make further progress, informative genetic markers can be designed based on initial investigations of nuclear and chloroplast genomes. An efficient but much more costly and time-consuming approach would be to analyze entire genomes for all investigated specimens. However, morphological data is also extremely important and useful in this old group with ample extinct diversity. With comprehensive morphological datasets at hand, simultaneous analyses of living and extinct species can be made. The aim of such future studies can for example be to understand relationships of fossils to living species, to assess the evolution of features over time, and to resolve the relationships of the Gnetales to other seed plants. But it can also be to provide comprehensive assessments of alpha-taxonomy, species diagnoses and identification of new living and fossil species. Futures studies should also explore poorly understood topics such as pollination biology, ecology, biogeography and population genetics. I hope more efforts can be made to continuously explore “the beauty and complexity of the evolutionary process” in the Gnetales.

30 Hou, C. PhD thesis 2016

Legend

Fig. 1. Gross morphology and biogeographic distribution of the Gnetales - (a) mature seeds of Ephedra likiangensis (b) opposite leaves and strobili of Welwitschia mirabilis (c) Three female spikes of Gnetum gnemon with several developing seeds; (d) stems and leaves of G. gnemon; (e) geographic distribution of Ephedra (in red), Gnetum (in blue) and Welwitschia (in green). Photographs by CH.

Fig. 2. A map of the chloroplast genome of Ephedra foeminea (see also Pa- per III). Genes, which are transcribed at clockwise and counter-clockwise directions are arranged inside and outside of the genome map, respectively. Genes that possess various functions are labelled with different colors. Genes that contain introns are marked with bold black lines. The large single copy region (LSC), the small single copy region (SSC) and the inverted repeats (IR) are marked inside the genome map. Percentage of GC content is shown with a threshold line of 50%. The length of the chloroplast genome of Ephedra foeminea is shown in the middle.

Fig. 3 A map of the chloroplast genomes of G. gnemon, G. luofuense (=G. hainanense), G. montanum, G. parvifolium and G. pendulum (see also Paper V). Genes, which are transcribed at clockwise and counter-clockwise directions are arranged inside and outside of the genome map, respectively. Genes that possess various functions are labelled with different colors. Genes that contain introns are marked with bold black lines. The large single copy region (LSC), the small single copy region (SSC) and the inverted repeats (IR) are marked inside the genome map. Percentage of GC content is shown with a threshold line of 50%. The lengths of the cp genomes of the five Gnetum species are shown in the middle.

Fig. 4 Repeated sequences detection in the five cp genomes of Gnetum studied in Paper V. (a) number of the three repeat types; (b) frequency of the direct (forward) repeats by length; (c) frequency of the reverse repeats by length; (d) frequency of the palindromic repeats by length.

31 Hou, C. PhD thesis 2016

Svensk sammanfattning (Swedish Summary)

Gnetales omfattar tre olika släkten, Ephedra, Gnetum och Welwitschia, som skiljer sig mycket åt gällande utseende, ekologi och molekylära egenskaper. Långvariga vetenskapliga debatter kring likheter mellan Gnetales och blomväxter, samt rent allmänt oklara släktskapsrelationer inom fröväxter, har inspirerat forskare att fortsätta studera Gnetales evolutionära historia, decennium efter decennium. De presenterade projekten handlar om flera tvärvetenskapliga frågor, från proteinfunktion och kloroplastens genom, till släktskapsförhållanden, morfologi och taxonomi. Här sammanfattas nya rön, kvarstående problem, och insikter och framtidsutsikter, baserat på resultaten av mina slutförda och pågående projekt.

Inom släktet Ephedra har medelhavsarten E. foeminea visat sig vara insektspollinerad, till skillnad från släktet i övrigt, och arten har en viktig fylogenetisk status som syster till alla övriga Ephedra arter. Därför är det en intressant nyckelart som studerats mycket på senare år. Här presenteras dess kloroplast-genom, som jag sekvenserat och jämfört med det enda kloroplast- genom inom släktet som är känt sedan tidigare, det av den vindpollinerade, asiatiska arten E. equisetina. Genomet är indelat i fyra huvudområden och omfattar 118 gener och totalt 109 584 kvävebaspar. En parvis jämförelse mellan E. foeminea och E. equistetina visar på över 2000 variabla baspositioner, data som kan användas för framtida släktskapsstudier. Jag studerade också vilka proteiner som finns i pollinationsdropparna hos E. foeminea och tre vindpollinerade Ephedra-arter, främst i syfte att hitta försvarsprotein som skyddar reproduktionen. Resultaten visar dock att protein förekommer i mycket små mängder hos Ephedra jämfört med andra studerade gymnospermer, och i huvudsak som nedbrytningsprodukter, det vill säga rester av döende celler. Några tänkbart funktionella proteiner hittades också, men även dessa förekommer i mycket små mängder. På grund av problem med analysmetoderna, vilket gör att resultaten eventuellt kan ifrågasättas, samt den låga förekomsten av proteiner i Ephedras pollinationsdroppar, valde jag att gå vidare med andra projekt.

Inom Gnetum har jag arbetat med rekonstruktion av släktskapsrelationer och analyser om när arter och grupper av arter divergerade från varandra. Resultaten visar att den sydamerikanska linjen separerades från övriga arter i släktet under yngre krita. Fortsatt diversifiering gav upphov till en afrikansk grupp och en asiatisk grupp. Den asiatiska gruppen, som omfattar två trädformerande arter, syster till de återstående liaonida arterna, är från krita- paleogen (K-Pg) gränsen. Mot bakgrund av släktskapen och gruppernas åldrar kan man anta att uppdelningen av superkontinenten Gondwana har påverkat basala diversifieringsmönster inom Gnetum. Senare spridning har

32 Hou, C. PhD thesis 2016 också bidragit till den nuvarande utbredningen av Gnetum. Utifrån mina resultat var det dock tydligt att taxonomi och artavgränsningar är dåligt underbyggda, och måste studeras vidare för alla undergrupper inom Gnetum. Jag har påbörjat denna uppgift genom att studera den kinesiska lianoida Gnetum-gruppen mer på djupet. Elva kloroplast-genom genererades och jämfördes. Baserat på dessa utformades fyra kloroplast-markörer som sekvenserades för ytterligare ett stort antal individer. Informationen användes för att undersöka släktskapsrelationer. Resultaten visar att G. parvifolium är syster till alla övriga arter i den kinesiska lianoida gruppen. Ytterligare fem arter bekräftas baserat på både morfologiska och molekylära data, men flera namn representeras av växtmaterial som inte kan anses vara egna arter. Modifiering av tidigare nycklar för identifiering av han- och honplantor presenteras. En dateringsanalys visar att diversifiering i den kinesiska lianoid Gnetum-kladen skedde främst under neogen, då miljöförändringar ledde till utbredning av tropiska skogar i det som nu är södra Kina. Detta bör ha gynnat diversifiering inom Gnetum.

33 Hou, C. PhD thesis 2016

Acknowledgement

I would like to give my deepest gratitude to my supervisor Catarina Rydin for providing a precious opportunity to be enrolled as a PhD student. Many thanks for sharing her wisdom with me and your kind encouragement and patience. Your always-sweet smiles are really impressive, as well as being so warm-hearted to care of my daily life in Sweden. In addition, I would like to thank Ove Eriksson and Tanja Slotte for providing comments and suggestions on my PhD thesis.

I will give many thanks to Niklas Wikström, for being the opponent of my Licentiate dissertation. In addition, it was very interesting to discuss with you about phylogenetic knowledge and methods during the book exam, and learning bioinformatics from you using next generation data in my research. Besides, thanks for helping me editing and providing many suggestive comments on the Chinese lianoid Gnetum paper.

I will give many thanks to Anbar Khodabandeh for helping me in the molecular lab and answering my questions. I have really enjoyed working with you in the morphological lab, too.

I will also give many thanks to other members of our Gnetales group, at first, to my roommate Kristina Bolinder, I am so fortunate to share the office with you, it is very exciting to discuss with you about the research, and our chats during and after work brought me a lot of pleasure. Thanks to Aelys Hum- phreys and Olle Thureborn for comments and sequences during our collabo- ration with the TAXON paper, as well as to Eva Larsen. It has been unfor- gettable to work with you all in the undergraduate course and having pleasure together during journeys to conferences abroad.

I would like to thank other persons in the former plant systematic division. Thanks to Per-Ola Karis, I really enjoy the discussion during the book exam with you and your personal sense of humour is very impressive. Thanks to Sylvain Razafimandimbison, I really benefit from the book exam of plant taxonomy and enjoyed your docent presentation of angiosperms evolution. Thanks to Barbro Axelius for learning undergraduate pedagogy from you, and other previous members, Birgitta Bremer, Kent Kainulainen, Åsa Krüger, Frida Stångberg, Annika Bengtson for the “PhD-sitting” in the first year, and for making my work so pleasurable and convenient.

I would like to thank the persons who gave me a warm-hearted hosting and assisted me for the field collection and material transportation in southern China. Many thanks to Jenny Lau, Tang Chin Cheung, Laura Won from the University of Hong; thanks to Richard Corlett, Bo Pan, Jian-tao Yin, Ma-

34 Hou, C. PhD thesis 2016 reike Roeder, Daniele Cicuzza from Xishuangbanna Tropical Botanical Gar- den; thanks to Si-jin Zeng, Hai-jun Yang from South China Agricultural University; thanks to Nan Deng, Sheng-qing Shi, Ze-ping Jiang from Chi- nese Academy of Forestry; thanks to Zhu-qiu Song, Shi-xiao Luo from South China Botanical Garden; thanks to En-de Liu from Kunming Institute of Botany, thanks to Joeri Strijk, Kun-fang Cao from Guangxi University.

I would like to give my sincere gratitude to my parents that assisted me for the field collection in Hainan and their continuous encouragement and guid- ance in my daily life. At the end, thanks to treasured Chinese friends in Stockholm, Xiong-zhuo Tang, Xiao Wang, Kun Wang, Si-mei Yu, Li-min Ma, Ning Sun, Liqun Yao, Yun-po Zhao and Yan Wang for bringing a lot of happiness in my daily life.

35 Hou, C. PhD thesis 2016

Literature cited

Arber, E. N. & J. Parkin. 1908. Studies on the evolution of the angiosperms the relationship of the angiosperms to the Gnetales. Annals of Botany 22:489-515. Bino, R. J., N. Devente & A. D. Meeuse. 1984. Entomophily in the dioecious gymnosperm Ephedra aphylla Forsk.(= E. alte CA Mey.), with some notes on E. campylopoda CA Mey. II. Pollination droplets, nectaries, and nectarial secretion in Ephedra. Proceedings of the Koninklijke Nederlandse Academie van Wetenschappen. Series C: Biological and Medical Sciences 87:15-24. Bolinder, K., A. Humphreys, A. R. Ehrlén J & R. C. Ickert-Bond SM. 2016. From near extinction to diversification by means of a shift in pollination mechanism in the gymnosperm relict Ephedra (Ephedraceae, Gnetales). Botanical Journal of the Linnean Society. DOI: 10.1111/boj.12380 Bolinder, K., K. J. Niklas & C. Rydin. 2015a. Aerodynamics and pollen ultrastructure in Ephedra. American Journal of Botany 102:457-470. Bolinder, K., L. Norbäck Ivarsson, A. M. Humphreys, S. M. Ickert-Bond, F. Han, C. Hoorn & C. Rydin. 2015b. Pollen morphology of Ephedra (Gnetales) and its evolutionary implications. Grana 55:24-51. Bowe, L. M., G. Coat & C. W. dePamphilis. 2000. Phylogeny of seed plants based on all three genomic compartments: extant gymnosperms are monophyletic and Gnetales' closest relatives are conifers. Proceedings of the National Academy of Sciences 97:4092-4097. Buchmann, S. L., M. K. O'Rourke & K. J. Niklas. 1989. Aerodynamics of Ephedra trifurca. III. Selective pollen capture by pollination droplets. Botanical Gazette 150:122-131. Carlquist, S. 1996. Wood, bark, and stem anatomy of Gnetales: a summary. International Journal of Plant Sciences 157:58-76. Carmichael, J. S. & W. E. Friedman. 1996. Double fertilization in Gnetum gnemon (Gnetaceae): its bearing on the evolution of sexual reproduction within the Gnetales and the anthophyte clade. American Journal of Botany 83:767-780. Corlett, R. T. 2001. Pollination in a degraded tropical landscape: a Hong Kong case study. Journal of Tropical Ecology 17:155-161. Coulter, A., B. A. Poulis & P. von Aderkas. 2012. Pollination drops as dynamic apoplastic secretions. Flora 207:482-490. Crane, P. R. 1985. Phylogenetic relationships in seed plants. Cladistics 1:329-348. Crane, P. R. 1996. The fossil history of the gnetales. International Journal of Plant Sciences 157:50-57. Crane, P. R. & S. Lidgard. 1989. Angiosperm diversification and paleolatitudinal gradients in Cretaceous floristic diversity. Science 246:675-678.

36 Hou, C. PhD thesis 2016

Doyle, J. 1945. Developmental lines in pollination mechanisms in the Coniferales. The Scientific proceedings of the Royal Dublin Society 24:43-62. Dutton, J. F. & E. J. Barron. 1997. Miocene to present vegetation changes: a possible piece of the Cenozoic cooling puzzle. Geology 25:39-41. Eames, A. J. 1952. Relationships of the Ephedrales. Phytomorphology 2:79- 100. Endress, P. K. 1996. Structure and function of female and bisexual organ complexes in Gnetales. International Journal of Plant Sciences 157:S113-S125. Forget, P.-M., D. S. Hammond, T. Milleron & R. Thomas. 2002. Seasonality of fruiting and food hoarding by rodents in Neotropical forests: consequences for seed dispersal and seedling recruitment. Pages 241-256 in D. J. Levey, W. R. Silva, and M. Galetti, editors. Seed Dispersal and Frugivory: Ecology Evolution and Conservation. CAB International, Wallingford. Friedman, W. E. 1992. Double fertilization in nonflowering seed plants and its relevance to the origin of flowering plants. Pages 342-348 in S. D. Russell and D. Christian, editors. International review of cytology. Academic Press, INC., California, US. Friis, E. M., P. R. Crane & K. R. Pedersen. 2011. Early flowers and angiosperm evolution. Cambridge University Press, Cambridge, the United Kingdom. Fu, L.-K., Y.-F. Yu & M. G. Gilbert. 1999. Gnetaceae. Pages 102-105 in Z.- Y. Wu and P. H. Raven, editors. Flora of China, Beijing: Science Press; St. Louis: Missouri Botanical Garden Press. Gong, Y. B., M. Yang, J. C. Vamosi, H. M. Yang, W. X. Mu, J. K. Li & T. Wan. 2015. Wind or insect pollination? Ambophily in a subtropical gymnosperm Gnetum parvifolium (Gnetales). Plant Species Biology. DOI: 10.1111/1442-1984.12112 Griffith, W. 1859. Remarks on Gnetum. Transactions of the Linnean Society of London 26:299-312. Guo, S.-X., J.-G. Sha, L.-Z. Bian & Y.-L. Qiu. 2009. Male spike strobiles with Gnetum affinity from the Early Cretaceous in western Liaoning, Northeast China. Journal of Systematics and Evolution 47:93-102. Guo, W., F. Grewe, W. Fan, G. J. Young, V. Knoop, J. D. Palmer & J. P. Mower. 2016. Ginkgo and Welwitschia mitogenomes reveal extreme contrasts in gymnosperm mitochondrial evolution. Molecular Biology and Evolution. DOI: 10.1093/molbev/msw024 Hollander, J. L. & S. B. Vander Wall. 2009. Dispersal syndromes in North American Ephedra. International Journal of Plant Sciences 170:323- 330. Hollander, J. L., S. B. Vander Wall & J. G. Baguley. 2010. Evolution of seed dispersal in North American Ephedra. Evolutionary Ecology 24:333-345.

37 Hou, C. PhD thesis 2016

Hooker, J. D. 1863. On Welwitschia, a new genus of Gnetaceae. Transactions of the Linnean Society of London 24:1-47. Hou, C., A. M. Humphreys, O. Thureborn & C. Rydin. 2015. New insights into the evolutionary history of Gnetum (Gnetales). Taxon 64:239- 253. Hsu, C.-Y., C.-S. Wu, S. Surveswaran & S.-M. Chaw. 2015. The complete plastome sequence of Gnetum ula (Gnetales: Gnetaceae). Mitochondrial DNA. DOI: 10.3109/19401736.2015.1079874 Huang, H., C. Shi, Y. Liu, S.-Y. Mao & L.-Z. Gao. 2014. Thirteen Camellia chloroplast genome sequences determined by high-throughput sequencing: genome structure and phylogenetic relationships. BMC Evolutionary Biology 14:151. Huang, J.-l. & R. A. Price. 2003. Estimation of the age of extant Ephedra using chloroplast rbcL sequence data. Molecular Biology and Evolution 20:435-440. Huang, J., D. E. Giannasi & R. A. Price. 2005. Phylogenetic relationships in Ephedra (Ephedraceae) inferred from chloroplast and nuclear DNA sequences. Molecular Phylogenetics and Evolution 35:48-59. Huang, K. C. 1998. The pharmacology of Chinese herbs. CRC press, Boca Raton. Huang, S.-B., Y.-H. Hu, D. Wu, Q. Tian & H.-Q. Li. 2010. Genetic diversity of Gnetum parvifolium of Fujian by ISSR markers. Guihaia 30:601- 607. Ickert-Bond, S. M., J. Pellicer, A. Souza, J. Metzgar & I. Leitch. 2015. Ephedra - the gymnosperm genus with the largest and most diverse genome sizes driven by a high frequency of recently derived polyploidy taxa and a lack of genome downsizing. Annual Meeting of the Botanical Society of America, Botany 2015, Abstract ID 862, Edmonton, Canada. Ickert-Bond, S. M. & C. Rydin. 2011. Micromorphology of the seed envelope of Ephedra L. (Gnetales) and its relevance for the timing of evolutionary events. International Journal of Plant Sciences 172:36-48. Ickert-Bond, S. M., C. Rydin & S. S. Renner. 2009. A fossil-calibrated relaxed clock for Ephedra indicates an Oligocene age for the divergence of Asian and New World clades and Miocene dispersal into South America. Journal of Systematics and Evolution 47:444- 456. Ickert-Bond, S. M. & M. F. Wojciechowski. 2004. Phylogenetic relationships in Ephedra (Gnetales): evidence from nuclear and chloroplast DNA sequence data. Systematic Botany 29:834-849. Jörgensen, A. & C. Rydin. 2015. Reproductive morphology in the Gnetum cuspidatum group (Gnetales) and its implications for pollination biology in the Gnetales. Plant Ecology and Evolution 148:387-396. Kakiuchi, N., M. Mikage, S. Ickert-Bond, M. Maier-Stolte & H. Freitag. 2011. A molecular phylogenetic study of the Ephedra distachya/E. sinica complex in Eurasia. Willdenowia 41:203-215.

38 Hou, C. PhD thesis 2016

Kato, M. & T. Inoue. 1994. Origin of insect pollination. Nature 368:195. Kato, M., T. Inoue & T. Nagamitsu. 1995. Pollination biology of Gnetum (Gnetaceae) in a lowland mixed dipterocarp forest in Sarawak. American Journal of Botany 82:862-868. Kitani, Y., S. Zhu, J. Batkhuu, C. Sanchir & K. Komatsu. 2011. Genetic Diversity of Ephedra Plants in Mongolia Inferred from Internal Transcribed Spacer Sequence of Nuclear Ribosomal DNA. Biological and Pharmaceutical Bulletin 34:717-726. Kitani, Y., S. Zhu, T. Omote, K. Tanaka, J. Batkhuu, C. Sanchir, H. Fushimi, M. Mikage & K. Komatsu. 2009. Molecular analysis and chemical evaluation of Ephedra plants in Mongolia. Biological and Pharmaceutical Bulletin 32:1235-1243. Kubitzki, K. 1985. Ichthyochory in Gnetum venosum. Anais Da Academia Brasileira De Ciencias 57:513-516. Kubitzki, K. 1990a. Ephedraceae. Pages 379-382 in K. Kubitzki, editor. The families and genera of vascular plants, volume 1. Pteridophytes and Gymnosperms. Springer, Berlin, Heidelberg. Kubitzki, K. 1990b. General Traits of the Gnetales. Pages 378-379 The families and genera of vascular plants, volume 1. Pteridophytes and Gymnosperms. Springer, Berlin, Heidelberg. Kubitzki, K. 1990c. Gnetaceae. Pages 383-386 in K. K.U. and G. P.S., editors. The families and genera of vascular plants, volume 1. Pteridophytes and gymnosperms. Springer, Berlin, Heidelberg. Kubitzki, K. 1990d. Welwitschiaceae. Pages 387-391 The families and genera of vascular plants, volume 1. Pteridophytes and Gymnosperms. Springer, Berlin, Heidelberg. Kurtz, S., J. V. Choudhuri, E. Ohlebusch, C. Schleiermacher, J. Stoye & R. Giegerich. 2001. REPuter: the manifold applications of repeat analysis on a genomic scale. Nucleic Acids Research 29:4633-4642. Leitch, I. J. & A. R. Leitch. 2013. Genome size diversity and evolution in land plants. Pages 314-315 in I. J. Leitch, J. Greilhuber, J. Doležel, and J. F. Wendel, editors. Plant Genome Diversity. Springer, Wien, Austra. Lidgard, S. & P. R. Crane. 1990. Angiosperm diversification and Cretaceous floristic trends; a comparison of palynofloras and leaf macrofloras. Paleobiology 16:77-93. Little, S., N. Prior, C. Pirone & P. von Aderkas. 2014. Pollen-ovule interactions in gymnosperms. Page 97 in K. G. Ramawat, J.-M. Mérillon, and K. R. Shivanna, editors. Reproductive Biology of Plants. CRC Press, Taylor & Francis Group. Liu, A.-L., F. Yang, M. Zhu, D. Zhou, M. Lin, S.-M. Lee, Y.-T. Wang & G.- H. Du. 2010. In vitro anti-influenza viral activities of stilbenoids from the lianas of Gnetum pendulum. Planta Medica 76:1874-1876. Liu, Z.-J. & X. Wang. 2015. An enigmatic Ephedra-like fossil lacking micropylar tube from the Lower Cretaceous Yixian Formation of Liaoning, China. International Journal of Plant Sciences 169:816- 843.

39 Hou, C. PhD thesis 2016

Loera, I., S. M. Ickert‐Bond & V. Sosa. 2015. Ecological consequences of contrasting dispersal syndromes in New World Ephedra: higher rates of niche evolution related to dispersal ability. Ecography 38:001-013. Loera, I., V. Sosa & S. M. Ickert-Bond. 2012. Diversification in North American arid lands: niche conservatism, divergence and expansion of habitat explain speciation in the genus Ephedra. Molecular Phylogenetics and Evolution 65:437-450. Mariac, C., N. Scarcelli, J. Pouzadou, A. Barnaud, C. Billot, A. Faye, A. Kougbeadjo, V. Maillol, G. Martin & F. Sabot. 2014. Cost ‐ effective enrichment hybridization capture of chloroplast genomes at deep multiplexing levels for population genetics and phylogeography studies. Molecular Ecology Resources 14:1103- 1113. Markgraf, F. 1930. Monographie der Gattung Gnetum Series 3. Bulletin du Jardin Botanique de Buitzenzorg 10:407-511. Markgraf, F. 1951. Gnetaceae. Pages 336-347 in C. G. G. Steenis, editor. Flora Malesiana. Djakarta, Noordhoff-Kolff, Batavia Mathews, S. 2009. Phylogenetic relationships among seed plants: persistent questions and the limits of molecular data. American Journal of Botany 96:228-236. McCoy, S. R., J. V. Kuehl, J. L. Boore & L. A. Raubeson. 2008. The complete plastid genome sequence of Welwitschia mirabilis: an unusually compact plastome with accelerated divergence rates. BMC Evolutionary Biology 8:130. McWilliam, J. 1958. The role of the micropyle in the pollination of Pinus. Botanical Gazette 120:109-117. Meeuse, A. D., A. H. De Meijer, O. W. Mohr & S. M. Wellinga. 1990. Entomophily in the dioecious gymnosperm Ephedra aphylla Forsk.(= E. alte CA Mey.), with some notes on Ephedra campylopoda CA Mey. III. Further anthecological studies and relative importance of entomophily. Israel Journal of Botany 39:113- 123. Nepi, M., P. von Aderkas & E. Pacini. 2012. Sugary exudates in plant pollination. Pages 155-185 in J. M. Vivanco and F. Baluška, editors. Secretions and exudates in biological systems. Springer. Nepi, M., P. von Aderkas, R. Wagner, S. Mugnaini, A. Coulter & E. Pacini. 2009. Nectar and pollination drops: how different are they? Annals of Botany 104:205-219. Niklas, K. J. 2015. A biophysical perspective on the pollination biology of Ephedra nevadensis and E. trifurca. The Botanical Review 81:28- 41. Nystedt, B., N. R. Street, A. Wetterbom, A. Zuccolo, Y.-C. Lin, D. G. Scofield, F. Vezzi, N. Delhomme, S. Giacomello & A. Alexeyenko. 2013. The Norway spruce genome sequence and conifer genome evolution. Nature 497:579-584.

40 Hou, C. PhD thesis 2016

O'Leary, S. J., B. A. Poulis & P. von Aderkas. 2007. Identification of two thaumatin-like proteins (TLPs) in the pollination drop of hybrid yew that may play a role in pathogen defense during pollen collection. Tree Physiology 27:1649-1659. O’Leary, S., C. Joseph & P. von Aderkas. 2004. Origin of arabinogalactan proteins in the pollination drop of Taxus × media. Austrian Journal of Forest Science 121:35-46. Owens, J. N., T. Takaso & C. J. Runions. 1998. Pollination in conifers. Trends in Plant Science 3:479-485. Palmer, J. D. 1991. Plastid chromosomes: structure and evolution. Pages 5- 53 in I. K. Vasil and L. Bogorad, editors. The molecular biology of plastids. Academic Press, San Diego. Pearson, H. 1907. The living Welwitschia. Nature 75:536-537. Pearson, H. H. W. 1909. Further observations on Welwitschia. Philosophical Transactions of the Royal Society of London. Series B 200:331-402. Pearson, H. H. W. 1929. Gnetales. Cambridge University Press, Fetter Lane, London. Petit, R. J., J. Duminil, S. Fineschi, A. Hampe, D. Salvini & G. G. Vendramin. 2005. Invited review: comparative organization of chloroplast, mitochondrial and nuclear diversity in plant populations. Molecular Ecology 14:689-701. Porsch, O. 1910. Ephedra campylopoda CA Mey., eine entomophile Gymnosperme. Berichte der Deutschen Botanischen Gesellschaft 28:404-412. Poulis, B. A., S. J. O’Leary, J. D. Haddow & P. von Aderkas. 2005. Identification of proteins present in the Douglas fir ovular secretion: an insight into conifer pollen selection and development. International Journal of Plant Sciences 166:733-739. Powell, W., M. Morgante, R. McDevitt, G. Vendramin & J. Rafalski. 1995. Polymorphic simple sequence repeat regions in chloroplast genomes: applications to the population genetics of pines. Proceedings of the National Academy of Sciences of the United States of America 92:7759-7763. Price, R. A. 1996. Systematics of the Gnetales: a review of morphological and molecular evidence. International Journal of Plant Sciences 157:S40-S49. Prior, N., S. A. Little, C. Pirone, J. E. Gill, D. Smith, J. Han, D. Hardie, S. J. O'Leary, R. E. Wagner & T. Cross. 2013. Application of proteomics to the study of pollination drops. Applications in Plant Sciences 1. DOI: 10.3732/apps.1300008 Provan, J., W. Powell & P. M. Hollingsworth. 2001. Chloroplast microsatellites: new tools for studies in plant ecology and evolution. Trends in Ecology & Evolution 16:142-147. Qin, A. L., M. M. Wang, Y. Z. Cun, F. S. Yang, S. S. Wang, J. H. Ran & X. Q. Wang. 2013. Phylogeographic evidence for a link of species divergence of Ephedra in the Qinghai-Tibetan Plateau and adjacent regions to the Miocene Asian aridification. Plos One 8:e56243.

41 Hou, C. PhD thesis 2016

Ridley, H. N. 1930. Dispersal of plants throughout the world. L. Reeve & Co., Ltd, Ashford, UK. Rodin, R. J. 1953. Distribution of Welwitschia mirabilis. American Journal of Botany 40:280-285. Rydin, C. 2005. The Gnetales: fossils and phylogenies, PhD thesis, Stockholm University, Sweden. Rydin, C. & K. Bolinder. 2015. Moonlight pollination in the gymnosperm Ephedra (Gnetales). Biology letters 11:20140993. Rydin, C. & E. M. Friis. 2010. A new Early Cretaceous relative of Gnetales: Siphonospermum simplex gen. et sp. nov. from the Yixian Formation of Northeast China. BMC Evolutionary Biology 10:183. Rydin, C. & C. Hoorn. 2016. The Gnetales: past and present. Grana 55:1-4. Rydin, C., M. Källersjö & E. M. Friis. 2002. Seed plant relationships and the systematic position of Gnetales based on nuclear and chloroplast DNA: conflicting data, rooting problems, and the monophyly of conifers. International Journal of Plant Sciences 163:197-214. Rydin, C., A. Khodabandeh & P. K. Endress. 2010. The female reproductive unit of Ephedra (Gnetales): comparative morphology and evolutionary perspectives. Botanical Journal of the Linnean Society 163:387-430. Rydin, C. & P. Korall. 2009. Evolutionary relationships in Ephedra (Gnetales), with implications for seed plant phylogeny. International Journal of Plant Sciences 170:1031-1043. Rydin, C., B. Mohr & E. M. Friis. 2003. Cratonia cotyledon gen. et sp. nov.: a unique Cretaceous seedling related to Welwitschia. Proceedings of the National Academy of Sciences of the United States of America 270:S29-S32. Rydin, C., K. R. Pedersen, P. R. Crane & E. M. Friis. 2006a. Former diversity of Ephedra (Gnetales): evidence from Early Cretaceous seeds from Portugal and North America. Annals of Botany 98:123- 140. Rydin, C., K. R. Pedersen & E. M. Friis. 2004. On the evolutionary history of Ephedra: Cretaceous fossils and extant molecules. Proceedings of the National Academy of Sciences of the United States of America 101:16571-16576. Rydin, C., S. Wu & E. Friis. 2006b. Liaoxia Cao et SQ Wu (Gnetales): ephedroids from the Early Cretaceous Yixian Formation in Liaoning, northeastern China. Plant Systematics and Evolution 262:239-265. Sharma, P. & R. Singh. 2015. A new species of Ephedra (Ephedraceae, Ephedrales) from India. Phytotaxa 218:189-192. Stapf, O. 1889. Die Arten der Gattung Ephedra. Denkschrift der Kaiserlichen Akademie der Wissensschaften. Mathematisch- Naturwissenschaftliche Klasse 56:1-112. Strasburger, E. 1872. Die Coniferen und die Gnetaceen: eine morphologsche Studie. pp 155-158, 236-238, 265-273, Jena, Hermann, Germany.

42 Hou, C. PhD thesis 2016

Takaso, T. 1990. " Pollination Drop" time at the Arnold Arboretum. Arnoldia 50:2-7. Tekleva, M. 2015. Pollen morphology and ultrastructure of several Gnetum species: an electron microscopic study. Plant Systematics and Evolution. DOI: 10.1007/s00606-015-1262-6 Tekleva, M. V. & V. A. Krassilov. 2009. Comparative pollen morphology and ultrastructure of modern and fossil gnetophytes. Review of Palaeobotany and Palynology 156:130-138. Thompson, W. 1912. Anatomy and relationship of Gnetales. Annals of Botany 26:1077-1104. Thompson, W. 1918. Independent evolution of vessels in Gnetales and angiosperms. Botanical Gazette 65:83-90. Thureborn, O. & C. Rydin. 2015. Phylogeny of Ephedra revisited. Botany 2015, Abstract 1010, Edmonton, Canada. Wagner, R. E., S. Mugnaini, R. Sniezko, D. Hardie, B. Poulis, M. Nepi, E. Pacini & P. Aderkas. 2007. Proteomic evaluation of gymnosperm pollination drop proteins indicates highly conserved and complex biological functions. Sexual Plant Reproduction 20:181-189. Wang, X. & S. L. Zheng. 2010. Whole fossil plants of Ephedra and their implications on the morphology, ecology and evolution of Ephedraceae (Gnetales). Chinese Science Bulletin 55:1511-1519. Weising, K. & R. C. Gardner. 1999. A set of conserved PCR primers for the analysis of simple sequence repeat polymorphisms in chloroplast genomes of dicotyledonous angiosperms. Genome 42:9-19. Wetschnig, W. & B. Depisch. 1999. Pollination biology of Welwitschia mirabilis Hook. f. (Welwitschiaceae, Gnetopsida). Phyton 39:167- 184. Won, H. & S. S. Renner. 2003. Horizontal gene transfer from flowering plants to Gnetum. Proceedings of the National Academy of Sciences of the United States of America 100:10824-18029. Won, H. & S. S. Renner. 2005a. The chloroplast trnT–trnF region in the seed plant lineage Gnetales. Journal of Molecular Evolution 61:425- 436. Won, H. & S. S. Renner. 2005b. The internal transcribed spacer of nuclear ribosomal DNA in the gymnosperm Gnetum. Molecular Phylogenetics and Evolution 36:581-597. Won, H. & S. S. Renner. 2006. Dating dispersal and radiation in the gymnosperm Gnetum (Gnetales) - Clock calibration when outgroup relationships are uncertain. Systematic Biology 55:610-622. Wu, C.-S., Y.-T. Lai, C.-P. Lin, Y.-N. Wang & S.-M. Chaw. 2009. Evolution of reduced and compact chloroplast genomes (cpDNAs) in gnetophytes: selection toward a lower-cost strategy. Molecular Phylogenetics and Evolution 52:115-124. Wu, H., Z. Ma, M. M. Wang, A.-L. Qin, J.-H. Ran & X.-Q. Wang. 2016. A high frequency of allopolyploid speciation in the gymnospermous genus Ephedra and its possible association with some biological and

43 Hou, C. PhD thesis 2016

ecological features. Molecular Ecology 25:1192-1210. DOI: 10.1111/mec.13538 Yang, Y., L. Lin & D. K. Ferguson. 2015. Parallel evolution of leaf morphology in gnetophytes. Organisms Diversity & Evolution 15:651-662. Yang, Y., L. Lin & Q. Wang. 2013. Chengia laxispicata gen. et sp. nov., a new ephedroid plant from the Early Cretaceous Yixian Formation of western Liaoning, Northeast China: evolutionary, taxonomic, and biogeographic implications. BMC Evolutionary Biology 13:72. Yao, C.-S. & M. Lin. 2005. Bioactive stilbene dimers from Gnetum cleistostachyum. Natural Product Research 19:443-448. Yao, C.-S., M. Lin & L. Wang. 2006. Isolation and biomimetic synthesis of anti-inflammatory stilbenolignans from Gnetum cleistostachyum. Chemical and Pharmaceutical Bulletin 54:1053-1057. Yao, Y.-F., A. A. Bruch, V. Mosbrugger & C.-S. Li. 2011. Quantitative reconstruction of Miocene climate patterns and evolution in Southern China based on plant fossils. Palaeogeography, Palaeoclimatology, Palaeoecology 304:291-307. Yao, Y.-F., Y.-Z. Xi, B.-Y. Geng & C.-S. Li. 2004. The exine ultrastructure of pollen grains in Gnetum (Gnetaceae) from China and its bearing on the relationship with the ANITA Group. Botanical Journal of the Linnean Society 146:415-425. Zachos, J., M. Pagani, L. Sloan, E. Thomas & K. Billups. 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292:686-693. Ziegler, H. 1959. Über die Zusammensetzung des “Bestäubungstropfens” und den Mechanismus seiner Sekretion. Planta 52:587-599.

44 (a) (b)

Ephedra Gnetum (e) Welwitschia

G. gnemon L. 114,405 bp G. parvifolium (Warb.) C.Y.Cheng 114, 950 bp G. luofuense C.Y.Cheng 114,947 bp G. montanum Markgr. 114,922 bp

G. pendulum C.Y.Cheng 114,917 bp (a) (b)

Total number of detected repeats Number of direct repeats 50 20 45 18 40 16 14 35 12 30 10 25 8

20 6 4 15 2 10 0 G. gnemon 102 G. parvifolium 108 G. luofuense 110 G. montanum106 G. pendulum 103 5 30-49bp 7 6 10 8 10 0 G. gnemon 102 G. parvifolium 108 G. luofuense 110 G. montanum106 G. pendulum 103 50-69 3 18 12 13 15 Palindric 16 2 1 6 6 70-99 1 12 15 8 4 Reverse 11 0 0 2 4 100-160 2 5 2 3 3 Forward 13 41 39 32 32

7 8 6 6 5

4 4 3 2 2 1 0 0 G. gnemon 102 G. parvifolium 108 G. luofuense 110 G. montanum106 G. pendulum 103 G. gnemon 102 G. parvifolium 108 G. luofuense 110 G. montanum106 G. pendulum 103 30-49 10 2 0 4 5 30-49 9 0 0 2 4 50-69 3 0 0 2 1 50-69 2 0 0 0 0 70-99 3 0 1 0 0 Table 1. Repeat type and locations detected in the five chloroplast genomes of Gnetum produced in Paper V. Total number of repeat detected: 205. Chloroplast genomes Repeat Length Position A Locus Position B Locus Type (bp) Gnetum gnemon 102 Forward 138 33561 ycf1 33702 ycf1 Forward 120 70431 psbB/rps12 70574 psbB/rps12 Forward 93 5208 trnT/trnK 5298 trnT/trnK Palindromic 87 28754 trnN/ycf1 37407 psaC/trnN Palindromic 86 9491 trnH/trnI 56358 trnH/trnI Palindromic 82 28457 trnN/ycf1 37431 psaC/trnN Forward 63 28476 trnN/ycf1 28754 trnN/ycf1 Forward 55 51966 ycf2 52020 ycf2 Palindromic 54 4474 psbD/trnT 4474 psbD/trnT Reverse 51 4475 psbD/trnT 4475 psbD/trnT Reverse 51 4476 psbD/trnT 4476 psbD/trnT Palindromic 50 4475 psbD/trnT 4475 psbD/trnT Forward 50 4475 psbD/trnT 4477 psbD/trnT Palindromic 50 4477 psbD/trnT 4477 psbD/trnT Forward 49 17493 psbB/rps12 48418 psbB/rps12 Reverse 49 4475 psbD/trnT 4475 psbD/trnT Reverse 49 4478 psbD/trnT 4478 psbD/trnT Palindromic 49 70407 trnL/rps7 70425 trnL/rps7 Palindromic 48 4475 psbD/trnT 4475 psbD/trnT Forward 48 4475 psbD/trnT 4479 psbD/trnT Palindromic 48 4479 psbD/trnT 4479 psbD/trnT Reverse 47 4475 psbD/trnT 4475 psbD/trnT Reverse 47 4480 psbD/trnT 4480 psbD/trnT Palindromic 46 4475 psbD/trnT 4475 psbD/trnT Forward 46 4475 psbD/trnT 4481 psbD/trnT Palindromic 46 4481 psbD/trnT 4481 psbD/trnT Reverse 45 4475 psbD/trnT 4475 psbD/trnT Reverse 45 4482 psbD/trnT 4482 psbD/trnT Palindromic 44 4475 psbD/trnT 4475 psbD/trnT Forward 44 4475 psbD/trnT 4483 psbD/trnT Palindromic 44 4483 psbD/trnT 4483 psbD/trnT Palindromic 44 48434 trnL/rps7 48434 trnL/rps7 Reverse 43 4475 psbD/trnT 4475 psbD/trnT Reverse 43 4484 psbD/trnT 4484 psbD/trnT Forward 43 13310 ycf2 13451 ycf2 Palindromic 43 13310 ycf2 52447 ycf2 Palindromic 43 13451 ycf2 52588 ycf2 Forward 43 52447 ycf2 52588 ycf2 Reverse 39 4490 psbD/trnT 4490 psbD/trnT rpl36/rps1 rpl36/rps1 Forward 39 62671 62709 1 1 G. luofuense 107 (=G. Forward 128 70978 psbB/rps12 70999 psbB/rps12 hainanense) Forward 107 70978 psbB/rps12 71020 psbB/rps12 Forward 92 70447 psbB/rps12 70460 psbB/rps12 Forward 89 34136 ycf1 34250 ycf1 Forward 86 70978 psbB/rps12 71041 psbB/rps12 Forward 86 34112 ycf1 34226 ycf1 Forward 83 28535 trnN/ycf1 28600 trnN/ycf1 Forward 81 70458 psbB/rps12 70471 psbB/rps12 Forward 81 33812 ycf1 33914 ycf1 Forward 79 70447 psbB/rps12 70473 psbB/rps12 Forward 77 33822 ycf1 33924 ycf1 Forward 73 34164 ycf1 34236 ycf1 Forward 70 34136 ycf1 34178 ycf1 Palindromic 69 9382 trnH/trnI 57150 trnH/trnI Forward 69 33366 ycf1 33432 ycf1 Forward 66 33924 ycf1 34344 ycf1 Forward 65 70978 psbB/rps12 71062 psbB/rps12 Forward 64 70462 psbB/rps12 70488 psbB/rps12 Forward 64 70475 psbB/rps12 70488 psbB/rps12 Forward 64 70449 psbB/rps12 70488 psbB/rps12 Forward 62 34061 ycf1 34091 ycf1 Forward 62 34142 ycf1 34298 ycf1 Forward 58 13953 ycf1 13971 ycf1 Forward 58 52587 ycf2 52605 ycf2 Forward 58 33932 ycf2 34352 ycf2 Forward 57 32993 ycf1 33050 ycf1 Forward 57 33842 ycf1 33944 ycf1 Forward 55 70458 psbB/rps12 70497 psbB/rps12 Forward 55 34242 ycf1 34284 ycf1 Forward 55 34256 ycf1 34298 ycf1 Forward 53 70447 psbB/rps12 70499 psbB/rps12 Forward 53 33822 ycf1 34344 ycf1 Forward 53 34184 ycf1 34298 ycf1 Forward 51 70488 psbB/rps12 70501 psbB/rps12 Forward 50 33830 ycf1 34352 ycf1 Forward 49 33941 ycf1 34361 ycf1 Forward 47 34157 ycf1 34313 ycf1 Forward 44 70978 psbB/rps12 71083 psbB/rps12 Forward 40 13953 ycf2 13989 ycf2 Forward 40 52587 ycf2 52623 ycf2 G. montanum 105 Forward 153 34070 ycf1 34184 ycf1 Forward 128 70935 psbB/rps12 70956 psbB/rps12 Forward 107 70935 psbB/rps12 70977 psbB/rps12 Forward 86 70935 psbB/rps12 70998 psbB/rps12 Forward 83 28525 trnN/ycf1 28590 trnN/ycf1 Forward 81 33790 ycf1 33892 ycf1 Forward 77 33800 ycf1 33902 ycf1 Forward 76 34147 ycf1 34261 ycf1 Forward 69 33344 ycf1 33410 ycf1 Forward 68 34033 ycf1 34177 ycf1 Forward 67 33902 ycf1 34322 ycf1 Forward 66 70430 psbB/rps12 70443 psbB/rps12 Forward 65 70935 psbB/rps12 71019 psbB/rps12 Forward 62 34039 ycf1 34069 ycf1 Forward 58 13943 ycf2 13961 ycf2 Forward 58 52562 ycf2 52580 ycf2 Forward 57 33820 ycf1 33922 ycf1 Forward 57 32971 ycf1 33028 ycf1 Forward 55 70441 psbB/rps12 70454 psbB/rps12 Forward 53 33800 ycf1 34322 ycf1 Forward 53 34120 ycf1 34276 ycf1 Forward 53 70430 psbB/rps12 70456 psbB/rps12 Palindromic 50 4516 psbD/trnT 4516 psbD/trnT Forward 50 4516 psbD/trnT 4518 psbD/trnT Forward 50 33919 ycf1 34339 ycf1 Palindromic 48 4516 psbD/trnT 4516 psbD/trnT Forward 47 4521 psbD/trnT 4525 psbD/trnT Forward 47 33820 ycf1 34342 ycf1 Forward 46 34033 ycf1 34291 ycf1 Palindromic 46 4513 psbD/trnT 4525 psbD/trnT Palindromic 46 4516 psbD/trnT 4516 psbD/trnT Forward 44 70935 psbB/rps12 71040 psbB/rps12 Palindromic 44 49008 trnL/rps7 49008 trnL/rps7 Palindromic 44 17529 ycf2 17529 ycf2 Forward 44 17529 ycf2 49008 ycf2 Forward 40 13943 ycf2 13979 ycf2 Forward 40 52562 ycf2 52598 ycf2 Forward 39 70432 psbB/rps12 70471 psbB/rps12 Reverse 37 4521 psbD/trnT 4521 psbD/trnT Reverse 37 4522 psbD/trnT 4522 psbD/trnT G. parvifolium 108 Forward 127 4886 trnT/trnK 5113 trnT/trnK Forward 120 70382 psbB/rps12 70395 psbB/rps12 Forward 111 28934 ycf1 29045 ycf1 Forward 107 70382 psbB/rps12 70408 psbB/rps12 Forward 102 5213 trnT/trnK 5334 trnT/trnK Forward 97 70379 psbB/rps12 70418 psbB/rps12 Forward 94 70395 psbB/rps12 70421 psbB/rps12 Forward 94 70408 psbB/rps12 70421 psbB/rps12 Forward 88 70949 psbB/rps12 70985 psbB/rps12 Forward 84 70379 psbB/rps12 70431 psbB/rps12 Forward 81 70395 psbB/rps12 70434 psbB/rps12 Forward 81 70408 psbB/rps12 70434 psbB/rps12 Forward 80 70949 psbB/rps12 70967 psbB/rps12 Forward 80 34227 ycf1 34257 ycf1 Forward 77 70934 psbB/rps12 70952 psbB/rps12 Forward 72 70965 psbB/rps12 71001 psbB/rps12 Forward 71 70379 psbB/rps12 70444 psbB/rps12 Forward 69 28656 trnN/ycf1 28833 trnN/ycf1 Forward 69 33335 ycf1 33401 ycf1 Forward 68 70395 psbB/rps12 70447 psbB/rps12 Forward 68 70408 psbB/rps12 70447 psbB/rps12 Forward 66 34154 ycf1 34184 ycf1 Forward 65 34036 ycf1 34126 ycf1 Forward 64 70965 psbB/rps12 70983 psbB/rps12 Forward 63 13930 ycf2 13957 ycf2 Forward 63 52546 ycf2 52573 ycf2 Forward 62 70949 psbB/rps12 71003 psbB/rps12 Forward 59 34053 ycf1 34143 ycf1 Forward 58 70379 psbB/rps12 70457 psbB/rps12 Forward 58 34004 ycf1 34034 ycf1 Forward 57 32962 ycf1 33019 ycf2 Forward 56 33903 ycf1 33960 ycf1 Forward 56 34006 ycf1 34126 ycf1 Forward 55 70395 psbB/rps12 70460 psbB/rps12 Forward 55 70408 psbB/rps12 70460 psbB/rps12 Forward 53 34167 ycf1 34197 ycf1 Forward 52 34064 ycf1 34094 ycf1 Forward 51 33960 ycf1 34317 ycf1 Palindromic 50 4523 psbD/trnT 4523 psbD/trnT Forward 50 34012 ycf1 34042 ycf1 Forward 50 34227 ycf1 34287 ycf1 Palindromic 48 4523 psbD/trnT 4523 psbD/trnT Forward 46 4527 psbD/trnT 4529 psbD/trnT G. pendulum 103 Forward 153 34072 ycf1 34186 ycf1 Forward 128 70937 psbB/rps12 70958 psbB/rps12 Forward 107 70937 psbB/rps12 70979 psbB/rps12 Forward 86 70937 psbB/rps12 71000 psbB/rps12 Forward 83 28527 trnN/ycf1 28592 trnN/ycf1 Forward 81 33792 ycf1 33894 ycf1 Forward 77 33802 ycf1 33904 ycf1 Forward 69 33346 ycf1 33412 ycf1 Forward 68 34035 ycf1 34179 ycf1 Forward 67 33904 ycf1 34324 ycf1 Forward 66 70432 psbB/rps12 70445 psbB/rps12 Forward 65 70937 psbB/rps12 71021 psbB/rps12 Forward 62 34041 ycf1 34071 ycf1 Forward 58 13945 ycf2 13963 ycf2 Forward 58 52564 ycf2 52582 ycf2 Forward 57 32973 ycf1 33030 ycf1 Forward 57 33822 ycf1 33924 ycf1 Forward 55 70443 psbB/rps12 70456 psbB/rps12 Forward 53 70432 psbB/rps12 70458 psbB/rps12 Forward 53 33802 ycf1 34324 ycf1 Palindromic 50 4518 psbD/trnT 4518 psbD/trnT Forward 50 4518 psbD/trnT 4520 psbD/trnT Forward 50 33921 ycf1 34341 ycf1 Palindromic 48 4518 psbD/trnT 4518 psbD/trnT Forward 47 4523 psbD/trnT 4527 psbD/trnT Forward 47 33822 ycf1 34344 ycf1 Palindromic 46 4515 psbD/trnT 4527 psbD/trnT Palindromic 46 4518 psbD/trnT 4518 psbD/trnT Forward 46 34035 ycf1 34293 ycf1 Forward 44 70937 psbB/rps12 71042 psbB/rps12 Palindromic 44 17531 trnL/rps7 17531 trnL/rps7 Forward 44 17531 trnL/rps7 49010 trnL/rps7 Palindromic 44 49010 trnL/rps7 49010 trnL/rps7 Forward 42 70443 psbB/rps12 70469 psbB/rps12 Forward 40 13945 ycf2 13981 ycf2 Forward 40 52564 ycf2 52600 ycf2 Forward 39 70434 psbB/rps12 70473 psbB/rps12 Forward 38 70460 psbB/rps12 70473 psbB/rps12 Reverse 37 4523 psbD/trnT 4523 psbD/trnT Reverse 37 4524 psbD/trnT 4524 psbD/trnT Reverse 35 4523 psbD/trnT 4523 psbD/trnT Reverse 35 4526 psbD/trnT 4526 psbD/trnT

Table 2. A list of sequence repeats (SSRs) detected in the five chloroplast genomes of Gnetum produced in Paper V. In total, 283 repeats. Number Number Position Position Loca- Gnetum cp genomes Repeat unit of repeat Region Locus of SSRs A B tion units G. gnemon var. brunonianum 102 A 11 20 1,362 1,372 LSC IGS psbZ/trnS 4,537 4,547 LSC IGS psbD/trnT

4,870 4,880 LSC IGS trnK/trnT

5,112 5,122 LSC IGS trnK/trnT

5,260 5,270 LSC IGS trnK/trnT

5,350 5,360 LSC IGS trnK/trnT

7,901 7,911 LSC IGS psbA/trnK

10,543 10,553 IRb CDS ycf2

33,126 33,136 SSR CDS ycf1

33,309 33,319 SSR CDS ycf1

34,048 34,058 SSR CDS ycf1

35,720 35,730 SSR IGS trnP/ycf1

55,392 55,402 IRa CDS ycf2

58,539 58,549 LSC CDS rps19

62,082 62,092 LSC CDS rps8

67,910 67,920 LSC IGS psbB/psbT

70,502 70,512 LSC IGS psbB/rps12

70,645 70,655 LSC IGS psbB/rps12

92,023 92,033 LSC intron atpF

97,417 97,427 LSC CDS rpoC2

A 12 7 10,128 10,139 IRb CDS ycf2

34,469 34,480 SSR CDS ycf1

55,806 55,817 IRa CDS ycf2

62,204 62,215 LSC IGS infA/rps8

62,452 62,463 LSC IGS infA/rpl36

94,237 94,248 LSC IGS atpl/rps2

107,521 107,532 LSC IGC trnS/ycf3

A 13 1 87,288 87,300 LSC IGS psbL/trnQ

A 14 5 5,129 5,142 LSC IGS trnK/trnT

70,555 70,568 LSC IGS psbB/rps12

75,735 75,748 LSC IGS petL/psbE

107,885 107,898 LSC intron ycf3

109,099 109,112 LSC intron ycf3

A 15 1 5,634 5,648 LSC intron trnK

AT 6 1 85,519 85,530 LSC IGS trnF/trnL

8 1 4,476 4,491 LSC IGS psbD/trnT

9 1 5,045 5,062 LSC IGS trnK/trnT

10 1 72,999 73,018 LSC IGS rpl20/rps12

17 1 4,493 4,527 LSC IGS psbD/trnT

AAAG 4 2 16,822 16,837 IRb IGS trnL/ycf2

49,102 49,117 IRa IGS trnL/ycf2

AAATC 3 1 4,273 4,291 LSC IGS psbD/trnT

AAAAAG 3 1 70,564 70,584 LSC IGS psbB/rps12

AAAAAG 3 2 16,115 16,132 IRb CDS ycf2

49,807 49,824 IRa CDS ycf2

AATTCG 3 3 13,874 13,891 IRb CDS ycf2

52,000 52,017 IRa CDS ycf2

52,054 52,071 IRa CDS ycf2

AAAAAG 3 1 70,564 70,584 LSC IGS psbB/rps12

AAAAAC 3 1 29,498 29,520 SSR CDS ycf1

G. luofuense 110 (=G. hainanense) A 10 21 1,574 1,583 LSC IGS psbC/trnS 4,444 4,453 LSC IGS psbD/trnT

4,574 4,583 LSC IGS psbD/trnT

6,796 6,805 LSC CDS matK

6,920 6,929 LSC CDS matK

29,058 29,067 SSR IGS trnN/ycf1

30,597 30,606 SSR CDS ycf1

31,572 31,581 SSR CDS ycf1

32,737 32,746 SSR CDS ycf1

33,749 33,758 SSR CDS ycf1

35,320 35,329 SSR CDS ycf1

60,624 60,633 LSC intron rpl16

61,861 61,870 LSC IGS rpl14/rpl16

70,436 70,445 LSC IGS psbB/rps12

70,861 70,870 LSC IGS psbB/rps12

87,582 87,591 LSC IGS trnE/trnQ

97,824 97,833 LSC CDS rpoC2

104,426 104,435 LSC CDS rpoB

105,421 105,430 LSC CDS rpoB

108,418 108,427 LSC intron ycf3

108,435 108,444 LSC intron ycf3

A 11 10 5,449 5,459 LSC IGS trnK/trnT

31,158 31,168 SSR CDS ycf1

60,746 60,756 LSC intron rpl16

60,841 60,851 LSC intron rpl16

62,304 62,314 LSC IGS rps8/rps14

62,680 62,690 LSC CDS rps8

66,937 66,947 LSC intron petB

67,396 67,406 LSC intron petB

73,588 73,598 LSC IGS rpl20/rps12

80,502 80,512 LSC IGS psaI/trnR

A 12 5 10,197 10,208 IRb CDS ycf2

35,092 35,103 SSR CDS ycf1

56,396 56,407 IRa CDS ycf2

76,348 76,359 LSC IGS petE/petL

91,977 91,988 LSC IGS atpA/atpF

A 13 3 36,281 36,293 SSR IGS trnP/ycf1

71,224 71,236 LSC IGS psbB/rps12

108,085 108,097 LSC IGS trnS/ycf3

A 14 2 72,855 72,868 LSC IGS psbB/rps12

105,657 105,670 LSC CDS rpoB

A 15 2 5,688 5,702 LSC intron trnK

109,649 109,663 LSC intron ycf3

C 10 2 84,708 84,717 LSC IGS atpE/trnfM

86,470 86,479 LSC intron trnL

C 11 1 84,685 84,695 LSC IGS atpE/trnfM

C 13 2 66,924 66,936 LSC intron petB

86,174 86,186 LSC IGS trnF/trnL

AT 6 1 86,145 86,156 LSC IGS trnF/trnL

9 1 4,538 4,555 LSC IGS psbD/trnT

16 1 5,070 5,101 LSC IGS trnK/trnT

AAG 5 2 16,149 16,164 IRb CDS ycf2

50,453 50,468 IRa CDS ycf2

AAAAAG 3 3 4,287 4,308 LSC IGS psbD/trnT

16,200 16,217 IRb CDS ycf2

50,400 50,417 IRa CDS ycf2

G. parvifolium 108 A 10 27 4,451 4,460 LSC IGS psbD/trnT 6,789 6,798 LSC CDS matK

6,913 6,922 LSC CDS matK

17,508 17,517 IRb CDS rps7/trnL

30,545 30,554 SSC CDS ycf1

31,544 31,553 SSC CDS ycf1

32,709 32,718 SSC CDS ycf1

33,718 33,727 SSC CDS ycf1

34,628 34,637 SSC CDS ycf1

35,277 35,286 SSC CDS ycf1

36,257 36,266 SSC IGS trnP/ycf1

49,050 49,059 IRa IGS rps7/trnL

60,580 60,589 LSC intron rpl16

62,254 62,263 LSC IGS rpl14/rps8

62,750 62,759 LSC IGS infA/rps8

62,994 63,003 LSC IGS infA/rpl36

65,727 65,736 LSC intron petD

70,824 70,833 LSC IGS psbB/rps12

76,017 76,026 LSC IGS petG/petL

84,668 84,677 LSC IGS atpE/trnfM

86,450 86,459 LSC intron trnL

92,615 92,624 LSC intron atpF

97,806 97,815 LSC CDS rpoC2

104,408 104,417 LSC CDS rpoB

105,406 105,415 LSC CDS rpoB

107,234 107,243 LSC IGS psbM/rps4

108,051 108,060 LSC IGS trnS/ycf3

A 11 9 4,248 4,258 LSC IGS psbD/trnT

4,586 4,596 LSC IGS psbD/trnT

62,630 62,640 LSC CDS rps8

66,875 66,885 LSC intron petB

71,171 71,181 LSC IGS psbB/rps12

74,704 74,714 LSC IGS psaJ/rpl33

80,483 80,493 LSC IGS psal/trnR

87,554 87,564 LSC IGS trnE/trnQ

109,622 109,632 LSC intron ycf3

A 12 4 10,190 10,201 IRb CDS ycf2

35,049 35,060 SSC CDS ycf1

56,353 56,364 IRa CDS ycf2

76,330 76,341 LSC IGS petL/psbE

A 13 2 72,822 72,834 LSC IGS psbB/rps12

91,957 91,969 LSC IGS atpA/atpF

A 14 1 105,642 105,655 LSC CDS rpoB

A 16 1 1,363 1,378 LSC IGS psbZ/trnS

A 18 1 5,678 5,695 LSC intron trnK

A 19 1 71,145 71,163 LSC IGS psbB/rps12

A 26 1 72,663 72,688 LSC IGS psbB/rps12

C 11 1 66,875 66,885 LSC intron petB

C 12 1 86,143 86,154 LSC IGS trnF/trnL

AT 13 1 5,074 5,100 LSC IGS trnK/trnT

20 1 4,528 4,567 LSC IGS psbD/trnT

AAG 5 1 50,422 50,437 IRa CDS ycf2

AGA 5 1 16,130 16,145 IRb CDS ycf2

ATA 5 1 14,017 14,033 IRb CDS ycf2

ATT 5 1 52,534 52,550 IRa CDS ycf2

AAAAAG 3 2 16,181 16,198 IRb CDS ycf2

50,369 50,386 IRa CDS ycf2

AAAAGA 3 1 4,286 4,307 LSC IGS psbD/trnT

G. montanum 105 A 10 22 1,574 1,583 LSC IGS psbC/trnS 4,444 4,453 LSC IGS psbD/trnT

4,559 4,568 LSC IGS psbD/trnT

5,136 5,145 LSC IGS trnK/trnT

6,790 6,799 LSC CDS matK

6,914 6,923 LSC CDS matK

30,575 30,584 SSR CDS ycf1

31,550 31,559 SSR CDS ycf1

32,715 32,724 SSR CDS ycf1

33,727 33,736 SSR CDS ycf1

35,298 35,307 SSR CDS ycf1

36,259 36,268 SSR IGS trnP/ycf1

60,818 60,827 LSC intron rpl16

66,919 66,928 LSC CDS petB

67,313 67,322 LSC CDS petB

67,378 67,387 LSC CDS petB

82,488 82,497 LSC IGS atpB/rbcL

87,548 87,557 LSC IGS trnE/trnQ

97,790 97,799 LSC CDS rpoC2

104,392 104,401 LSC CDS rpoB

105,387 105,396 LSC CDS rpoB

107,235 107,244 LSC IGS psbM/rps4

A 11 10 4,580 4,590 LSC IGS psbD/trnT

29,035 29,045 SSR IGS trnN/ycf1

60,599 60,609 LSC intron rpl16

61,837 61,847 LSC IGS rpl14/rpl16

62,655 62,665 LSC CDS rps8

73,545 73,555 LSC IGS rpl20/rps12

73,968 73,978 LSC IGS rpl20/rps18

76,309 76,319 LSC IGS petE/petL

84,662 84,672 LSC IGS atpE/trnfM

108,385 108,395 LSC intron ycf3

A 12 8 1,370 1,381 LSC IGS psbZ/trnS

10,191 10,202 IRb CDS ycf2

31,135 31,146 SSR CDS ycf1

35,070 35,081 SSR CDS ycf1

56,371 56,382 IRa CDS ycf2

60,722 60,733 LSC intron rpl16

91,943 91,954 LSC IGS atpA/atpF

108,403 108,414 LSC intron ycf3

A 13 6 70,416 70,428 LSC IGS psbB/rps12

71,181 71,193 LSC IGS psbB/rps12

80,462 80,474 LSC IGS psal/trnR

86,139 86,151 LSC IGS trnF/trnL

108,052 108,064 LSC IGS trnS/ycf3

109,619 109,631 LSC intron ycf3

A 14 2 72,812 72,825 LSC IGS psbB/rps12

105,623 105,636 LSC CDS rpoB

A 15 1 5,682 5,696 LSC intron trnK

C 10 1 84,673 84,682 LSC IGS atpE/trnfM

C 11 2 84,648 84,658 LSC IGS atpE/trnfM

86,435 86,445 LSC intron trnL

C 13 1 66,906 66,918 LSC CDS petB

AT 6 1 86,110 86,121 LSC IGS trnF/trnL

10 1 5,077 5,096 LSC IGS trnK/trnT

19 1 4,522 4,559 LSC IGS psbD/trnT

AAG 5 2 16,139 16,154 IRb CDS ycf2

50,428 50,443 IRa CDS ycf2

AAAAAG 3 3 4,287 4,308 LSC IGS psbD/trnT

16,190 16,207 IRb CDS ycf2

50,375 50,392 IRa CDS ycf2

G. pendulum 103 A 10 22 1,575 1,584 LSC IGS psbC/trnS 4,561 4,570 LSC IGS psbD/trnT

5,138 5,147 LSC IGS trnK/trnT

6,792 6,801 LSC CDS matK

30,577 30,586 SSC CDS ycf1

31,552 31,561 SSC CDS ycf1

32,717 32,726 SSC CDS ycf1

33,729 33,738 SSC CDS ycf1

35,300 35,309 SSC CDS ycf1

36,261 36,270 SSC IGS trnP/ycf1

60,820 60,829 LSC intron rpl16

62,283 62,292 LSC IGS rpl14/rps8

66,922 66,931 LSC intron petB

67,316 67,325 LSC intron petB

67,381 67,390 LSC intron petB

82,490 82,499 LSC IGS atpB/rbcL

84,664 84,673 LSC IGS atpE/trnfM

87,546 87,555 LSC IGS trnE/trnQ

97,788 97,797 LSC CDS rpoC2

104,390 104,399 LSC CDS rpoB

105,385 105,394 LSC CDS rpoB

108,381 108,390 LSC intron ycf3

A 11 9 4,445 4,455 LSC IGS psbD/trnT

4,582 4,592 LSC IGS psbD/trnT

29,037 29,047 SSC IGS trnN/ycf1

60,601 60,611 LSC intron rpl16

61,839 61,849 LSC IGS rpl14/rpl16

62,658 62,668 LSC CDS rps8

73,547 73,557 LSC IGS rpl20/rps12

73,970 73,980 LSC IGS rpl20/rps18

76,311 76,321 LSC IGS petE/petL

A 12 9 10,193 10,204 IRb CDS ycf2

31,137 31,148 SSC CDS ycf1

35,072 35,083 SSC CDS ycf1

56,373 56,384 IRa CDS ycf2

60,724 60,735 LSC intron rpl16

70,419 70,430 LSC IGS psbB/rps12

91,941 91,952 LSC IGS atpA/atpF

108,049 108,060 LSC IGS trnS/ycf3

108,398 108,409 LSC intron ycf3

A 13 4 1,370 1,382 LSC IGS psbZ/trnS

71,183 71,195 LSC IGS psbB/rps12

80,464 80,476 LSC IGS psaL/trnR

109,614 109,626 LSC intron ycf3

A 14 2 72,814 72,827 LSC IGS psbB/rps12

105,621 105,634 LSC CDS rpoB

A 15 1 5,684 5,698 LSC intron trnK

C 11 1 84,650 84,660 LSC IGS atpE/trnfM

C 13 2 66,909 66,921 LSC intron petB

86,139 86,151 LSC IGS trnF/trnL

AT 6 1 86,110 86,121 LSC IGS trnF/trnL

10 1 5,079 5,098 LSC IGS trnK/trnT

19 1 4,524 4,561 LSC IGS psbD/trnT

AAG 5 2 16,141 16,156 IRb CDS ycf2

50,430 50,445 IRa CDS ycf2

AAAAAG 3 3 4,288 4,309 LSC IGS psbD/trnT

16,192 16,209 IRb CDS ycf2

50,377 50,394 IRa CDS ycf2