A novel case of autogamy and cleistogamy in Dendrobium wangliangii: A rare orchid distributed in the dry‐hot valley

Abstract Dendrobium wangliangii is an epiphytic orchid distributed in the Jinshajiang dry‐hot valley in Luquan County, Yunnan Province, China. Most Dendrobium spp. typically have a low fruit set, but this orchid shows a higher fruit set under natural conditions despite the lack of effective pollinators. The pollination biology of the critically endangered D. wangliangii was investigated in this study. A fruit set rate of 33.33 ± 4.71% was observed after bagging treatment in 2017 and a high fruit set rate (65.72 ± 4.44% in 2011; 50.79 ± 5.44% in 2017) was observed under natural conditions, indicating that D. wangliangii is characterized by spontaneous self‐pollination. The anther cap blocked the growing pollinium; thus, the pollinium slid down and reached the stigmatic cavity, leading to autogamous self‐pollination. Specifically, 51.50% of 162 unopened flowers (total 257 flowers) of this Dendrobium species under extreme water‐deficit conditions developed into fruits, suggesting the presence of cleistogamy in D. wangliangii. Here, cleistogamy may represent the primary mode of pollination for this orchid. Spontaneous self‐pollination and specific cleistogamous autogamy could represent major adaptions to the drought and pollinator‐scarce habitat in the Jinshajiang dry‐hot valley.


| INTRODUC TI ON
The floral diversity and pollination biology of Orchidaceae have long intrigued evolutionary biologists (Cozzolino & Widmer, 2005).
Orchids are well known for their highly evolved flowers that promote cross-pollination. Most orchid species rely on pollinators for sexual reproduction (Jacquemyn, Micheneau, Roberts, & Psiller, 2005;Ren, Wang, & Luo, 2012). However, not all orchids possess adaptations that guarantee cross-pollination, and some have become modified to ensure self-pollination. Autogamy is the preferred mode of pollination and is a rather common phenomenon within the orchid family (Kowalkowska & Margońska, 2012). Between 5% and 20% of orchid species have evolved autogamy (Gale, 2007;Pang et al., 2012;Tremblay, Ackerman, Zimmerman, & Calvo, 2005). This mode of autogamy often occurs under windless conditions and in drought and insect-scarce habitats to ensure reproductive success (Liu et al., 2006).
Cleistogamy is a type of special autogamy. It refers to a pollination system in which unopened flowers are capable of autogamy resulting in fruit productions. This trait is an adaptive mechanism for retaining reproduction under poor growth conditions, such as drought, extreme airflow, and sunlight (Faisal et al., 2018;Liu et al., 2006;Lord, 1981;Zou, Zhou, Angessa, Zhang, & Li, 2018). In a majority of cleistogamous species, closed (cleistogamous, CL) and open flowers (chasmogamous, CH) usually occur on the same individuals during a single flowering season. These flowers comprise multiple strategies and perform the same function simultaneously (Schoen & Lloyd, 1984). Both CL and CH flowers produce fruit and seed depending on environmental conditions, including parameters such as light intensity, photoperiod, water availability and nutrient availability (Morinaga et al., 2008).
Generally, Dendrobium spp. flower much often but exhibit a low fruit set rate; the fruit set rate of natural populations of Dendrobium is approximately 10% to 17% (Zhang, Xia, Zhu, & Zhang, 2000). The low rate of fruit set is mainly attributed to insect-scarce habitats.
Furthermore, self-incompatibility is found in approximately 50% of Dendrobium species. Most Dendrobium species undergo cross-pollination (Brodmann et al., 2009;Niu et al., 2017). The restriction of low numbers of pollinating insects is the main factor leading to low fruit set rates in the field. Dendrobium wangliangii, a recently described species of Dendrobium (Hu, Long, & Jin, 2008), endemic to the Jinshajiang dry-hot valley in Luquan County, Yunnan Province, and found at an altitude of 2,200 m, is an epiphytic orchid with strong drought tolerance (Zhao et al., 2013(Zhao et al., , 2018. Studies on this plant in the wild observed that the rate of fruit set of D. wangliangii was rather high under natural conditions. A common trend of pollinator scarcity is noted under drought conditions (Liu et al., 2006), that may exert selective pressure via pollinator limitations, thus promoting the evolution of autogamy as a reproductive assurance mechanism. Therefore, the reproductive biology of D. wangliangii represents an interesting problem. In the present study, D. wangliangii was used to investigate the following questions based on preliminary observations: (a) What type of mating system does D. wangliangii possess? and (b) If the species exhibits self-pollination, what type of pollination strategy is employed?

| Study sites and species
Studies were conducted in one wild population in northern Yunnan province, China. The exact location of D. wangliangii is not provided here due to the concern of illegal collection of this endangered orchid. The study area has a subtropical monsoon climate with a mean 947.7 mm of rainfall and average temperature of 15.6°C per year.
The forest vegetation at this location is mainly composed of Quercus species and D. wangliangii epiphytes on these deciduous Quercus.
To compare the mating system of D. wangliangii with other species, 23 Dendrobium spp. (Table 3) were selected and collected from Yunnan, Southeast Asia, and other surrounding regions. These species are cultivated in the nursery garden in Kunming, Yunnan, China.
The study area has a monsoonal climate in the northern subtropical low-latitude plateau with a mean 1,031 mm of rainfall and an average temperature of 13.2°C per year.

| Flower visitors and their visiting behavior
To record flower visitors and their visiting behavior, the study site was observed during the flowering periods of D. wangliangii for a total of 120 hr. Thirty blooming flowers were randomly marked and subsequently observed. Daily periods of observation were from 6:00 to 18:00. Visiting behavior of insects was recorded and photographed, and the following factors were noted: visitor species, visitor time, and visitor behavior, especially if the visitors take pollinia away. The fruit set of the experimental series and that of untreated flowers was recorded in July.

| Mating system test
The study of the mating system of 23 Dendrobium spp. was performed at the nursery from May to September 2017. Experimental treatments included hand self-pollination, and hand cross-pollination, controls, the same to treatments 1, 2, and 5 as described above for D. wanglianglii. Ten flowers on 10 inflorescences from five plants per species were randomly selected each time on the morning on their third day of anthesis, with three replicates. Pollination success was estimated with the fruit set after the following 90 days. The fruit set rate was calculated as a number of capsules per number of treated flowers.

| Evaluating fruit set under natural conditions
A total of 257 buds of D. wangliangii were randomly selected and marked. Within the next two weeks, the open state of the marked flowers was recorded. If the marked flowers remained closed throughout the flowering period, they were marked as unopened; on the contrary, they were marked as opened. After two months, the fruit set rate of the flowers in different states was calculated (as a number of capsules that produced fruit per number of marked flowers).

| Statistical analyses
For mating system assays, experiments were conducted by different treatments. Fruit set rate of each treatment was used for statistical analyses. Comparison between the different treatments was performed via one-way ANOVA and LSD with SPSS (ver. 19.0).

| Floral morphology and phenology
Dendrobium wangliangii flowers are pink (red-lilac) with two yellow (greenish yellow) patches on the labellum (Figure 1a). The flower was borne at the upper nodes of an older leafless stem occasionally with two or more flowers on a single inflorescence. The mean number of flowers on a single inflorescence was 1.04 ± 0.05 (n = 520).
Inflorescences were mostly borne on the first or second stem node (Figure 1b), and a few inflorescences were borne on the third node.
According to plant age and number of branches, the number of flowers per plant ranged from 0 to 22. The number of fruits per plant is 2.04 ± 0.12 (n = 254). The most common number of fruits observed was 1 to 3 (Figure 1d), and up to 17 fruit sets were noted per plant.
In this study, the length and width of petals were 30.83 ± 0.89 mm and 23.97 ± 0.41 mm, respectively. The diameter of the labellum was 6.99 ± 0.08 mm (Table S1). Other morphological characteristics were detailed in Table S1.

| Flower visitors and their visiting behavior
Two types of visitors were found, that is, fly and the cockchafer.
However, the insects did not enter the flower passage of D. wangliangii but only crawled outside the flowers, leaving after a few seconds. During the observation period, the pollens removal by any insect and pollination were never observed during either the day or night. In the natural habitat, effective pollinators were scarce, and no insect was observed pollinating D. wangliangii.

| Mating system
The fruit set rate was 65.72 ± 4.44% under artificial self-pollination in 2011, which was the highest in all treatments. The difference between different treatments was not significant in 2011 (p > .05). In 2017, the difference was not significant between artificial self-pollination and natural pollination (p > .05). There was no difference between hand cross-pollination and agamospermy (p > .05). The difference between other treatments was statistically significant (p < .05). The rates of fruit set under hand and spontaneous self-pollination were significantly higher than that under hand cross-pollination in 2017 (p < .05). Fruits were not formed under agamospermy in D. wangliangii (Table 1).
Both opened and unopened flowers of D. wangliangii developed into fruits (Table 2). The rate of fruit set displayed no significant difference between opened and unopened flowers (p > .05). The unopened flowers produced fruits, indicating that D. wangliangii exhibited cleistogamous reproduction.
In nursery experiments, fruit set in 14 Dendrobium spp. was formed under both self-and cross-pollination, while fruit sets in nine species did not form under self-pollination (Table 3). No fruits were formed under the control treatment in 23 Dendrobium species.
The fruit set rate under cross-pollination was significantly higher than that under self-pollination in 23 Dendrobium spp. (p < .05), revealing that the mating system of Dendrobium was dominated by cross-pollination (Table 4).

| D ISCUSS I ON
The present investigation provides evidence for autonomous selfpollination in D. wangliangii. Self-compatibility is thought to be relatively common in orchids, while only 10% of orchid species exhibit self-incompatibility (SI). Self-compatibility has few been reported in a typical genus with SI, Dendrobium (Catling, 1990;Liu et al., 2006;Niu et al., 2017).
Dendrobium wangliangii was self-compatible but also cross-compatible for fruit set, and the fruit set rate under hand self-pollination was higher than that under hand cross-pollination. Inversely, the fruit set rate under cross-pollination was significantly higher than that under self-pollination in 23 tested Dendrobium spp., revealing that Dendrobium mating system mainly involved cross-pollination. wangliangii were self-compatibility but also cross-compatible (Li et al., 2009;Lian & Li, 2003).
Furthermore, D. wangliangii evolves a trait of autonomous selfing.
Therefore, it is tempting to speculate that D. wangliangii is a special case of evolution to autogamy. Darwin (1890) realized that autogamy could represent an adaptation to reproduction if pollinator service was lost or extremely unpredictable, thus providing reproductive assurance. Both abiotic and biotic factors can influence mating-system evolutionary transitions from outcrossing to selfing, such as pollinator absence or low pollinator abundance (Kalisz & Vogler, 2003). Fishman and Willis (2008) reported that pollen limitation intensified selection on attractive floral traits with increasing pollen limitation, but severe pollen limitation may favor traits associated with autonomous self-fertilization. Here, autogamous self-pollination of D. wangliangii may be an adaptive evolution to the orchid's dry and insect-limited habitat.

No effective pollinator was observed in our investigation. Most
Orchidaceae species exhibit particular types of floral morphology, color patterns, fragrances, or rewards for cross-pollination; thus, the tendency of automatic self-pollination is reduced (Kowalkowska & Margońska, 2012). However, automatic self-pollination within Orchidaceae is a rather common mode in windless, drought conditions, and insect-limited habitats (Liu et al., 2006). Numerous studies have documented the tendency of plants to autonomous self-pollination when grown in dry and insect-limited conditions (Dole, 1990;Elle & Carney, 2003;Elle & Hare, 2002;Herrera et al., 2001;Holsinger, 2000;Liu et al., 2006). Flowers can attract pollinators based upon attraction and reward (Monty, Saad, & Mahy, 2006;Pang et al., 2012). However, in the study, D. wangliangii lacked nectar and odors. The loss of attraction and reward for D. wangliangii TA B L E may lead to a deficiency of effective pollinators under natural conditions, which was coincident with previous findings demonstrating a common trend of pollinator scarcity in drought conditions and high mountains (Blionis & Vokou, 2008;Li, Zheng, Dafni, & Luo, 2010;Liu et al., 2006;Totland, 2001). Pollinator scarcity may drive evolutionary shift from cross-pollination to self-pollination (Xiong, Fang, & Huang, 2013). Most of the mycoheterotrophic species (especially nectarless species) seem to have abandoned insect pollinators in favor of self-pollination (Suetsugu, 2016). In D. wangliangii, autonomous self-pollination could represent an evolutionary response to ensure reproductive success to compensate for pollinator limitation.
Commonly, selfing is considered to decrease genetic variation and increased genetic drift, high levels of inbreeding depression (Culley & Klooster, 2007). However, when pollinators are rare or absent, cleistogamous selfing can exhibit fitness advantages distinct from out-crossing, owing to the increase of seed production, less energetical cost, and inherent automatic selfing advantage (Culley & Klooster, 2007;Suetsugu, 2014Suetsugu, , 2016Waller, 1984). Outcrossing in the orchid may gradually lose its position under pollinators scarcity and extreme water-deficit conditions. Finally, D. wangliangii may evolve a specific mating system for reproductive assurance, cleistogamy. The breeding system evolved at least six to eight times within monocots (Culley & Klooster, 2007;Hilu et al., 2003;Soltis et al., 2000). Consistent cleistogamous selfing can eliminate deleterious recessive alleles within populations; over time, this could lead to a decrease in the level of inbreeding depression (Clay & Antonovics, 1985;Culley & Klooster, 2007). Here, cleistogamy in D. wangliangii could ensure population purity and germplasm stability. The evolutionary pathway to cleistogamy remains relatively understudied, so cleistogamous species can be used to test evolutionary theories and origins of cleistogamous taxa. TA B L E 4 Variance analysis of selfpollination rate and cross-pollination rate F I G U R E 2 The process of automatic pollination of Dendrobium wangliangii. (a) The anther cap was covered with pollen, and the pollen did not adhere to the stigma, exhibiting a pale yellow color; (b) the anther cap blocked the growing pollinium and the pollinium slid down and reached the stigmatic cavity; (c) after the pollen grains absorbed water on the stigma, their volume expanded; (d) the pollen grains were gradually integrated into the stigma and the color faded

(a) (b) (c) (d)
Dendrobium wangliangii thus exemplifies a unique mode of autogamy, cleistogamy, in the genus Dendrobium. This can provide a novel breeding mechanism in Dendrobium. Future research will explore the importance of climate and development time for mating system evolution, the extent of dependency on autonomous selfpollination for reproductive success and the molecular mechanism of cleistogamy in this species.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
Shicheng Shao, Yuan Su, and Xueli Hu conducted the field and greenhouse research. Qiuxia Wang performed data analysis and wrote the manuscript. Dake Zhao and Yong Shen designed the research. All authors read and approved the final manuscript. Qiuxia Wang and Shicheng Shao contributed equally to this work.

DATA AVA I L A B I L I T Y S TAT E M E N T
The mating systems, fruit set rates, and morphological characteristics data used in the manuscript can be accessed at the Dryad Digital Data Repository: Dryad https ://doi.org/10.5061/dryad.11n8t6n.