**1. Introduction**

Over the past 200 years, human activities have altered more than 80% of the terrestrial world, leading to increasing habitat reduction and fragmentation of almost all ecosystems [1–3]. Theoretically, habitat fragmentation has severe negative impacts on community structures, population distributions and abundances, patterns of genetic variation, gene flow, etc. [4,5]. These changes, especially genetic alterations, can reduce the individual fitness and evolutionary potential of populations, further impeding the long-term persistence of species and increasing the risk of local extinction [6,7].

Detecting the genetic effects of habitat fragmentation on natural populations is challenging in practice, as there are many confounding factors involved [2]. For instance,

**Citation:** Yuan, N.; Wei, S.; Comes, H.P.; Luo, S.; Lu, R.; Qiu, Y. A Comparative Study of Genetic Responses to Short- and Long-Term Habitat Fragmentation in a Distylous Herb *Hedyotis chyrsotricha* (Rubiaceae). *Plants* **2022**, *11*, 1800. https://doi.org/10.3390/ plants11141800

Academic Editors: Aiping Song and Yu Chen

Received: 23 May 2022 Accepted: 5 July 2022 Published: 7 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

life-history traits such as life span, pollination, and seed dispersal, are essential in determining the magnitude of the plant's response to habitat fragmentation. Short-lived herbs experience more generations and may show a faster decrease in genetic diversity than longer-lived trees or shrubs over the same timescale of fragmentation [8]. In addition, plants with abiotic-mediated pollination and seed dispersal are expected to be less susceptible to habitat fragmentation than those with animal-mediated pollination and seed dispersal [9,10]. For example, Wang et al. (2011) [11] observed no significant difference in genetic diversity between pre- and post-fragmentation cohorts of a wind-pollinated tree *Castanopsis sclerophylla* (Lindl. & Paxton) Schottky in a recently fragmented island system. Due to the disappearance of a genetic barrier, the genetic structure of post-fragmentation cohorts was significantly weakened by increased wind speeds and easier pollen movement over water. In contrast, our previous study on the distylous herb *Hedyotis chrysotricha* (Palib.) Merr in the Thousand-Island Lake (TIL) region revealed that 53 years of fragmentation has led to the loss of c. 17.4% of the initial genetic diversity in the island populations [12]. *Hedyotis chrysotricha* is insect-pollinated, with seeds that are dispersed by wind or (and) water [13,14]. Although patterns of gene flow have not been greatly modified among the *H. chrysotricha* populations, it is uncertain whether extensive seed flow can counteract the negative effects of genetic drift to reach a new dynamic equilibrium and assist the long-term persistence of remnant populations.

The various island systems in southeastern China provide natural laboratories for testing the genetic responses of plants to habitat fragmentation. For example, a longestablished hotspot for testing the predictions of population genetics theory on recent anthropogenic habitat fragmentation is the TIL region in southeast China (Chun'an County, Hangzhou City, Zhejiang Province, China). This artificial lake (c. 573 km2), with about 1078 islands of different sizes (0.25–1320 ha) and shapes (c. 83 km2 in total), was formed in 1959 after the construction of the Xinanjiang River hydroelectric power station [15]. In contrast, the Zhoushan Archipelago (ZA), located in the East China Sea off the northern coast of Zhejiang Province, is an ideal candidate for examining the genetic responses to long-term habitat fragmentation. It was originally an extended part of the nearby continent and was formed by rising sea levels during the early Holocene, c. 7000–9000 years ago. It consists of 1339 islands and 3306 reefs, covering about 1440 km2 in total [16,17]. The two island systems differ greatly in spatial scale and temporal origin, making them ideal systems for comparative studies detecting both short-term and long-term responses of species to habitat fragmentation. For example, Yuan et al. (2015) [18] compared the patterns of nuclear microsatellite (SSR) variation and the population structure of an evergreen shrub, *Loropetalum chinense* (R. Br.) Oliver in these two island systems. The results showed that this species can mitigate the negative effects of millennia-old natural habit fragmentation via water-facilitated seed dispersal. Similar results were also observed in a comparative genetic study of the perennial vine *Actinidia chinensis* Planch. [19]. For long-lived species, the occurrence of gene flow among populations has great potential to mitigate the genetic erosion caused by both recent and historical habitat fragmentation [20]. However, similar comparative genetic studies to determine whether the same is true for short-lived, heterostylous plants, e.g., *H. chrysotricha*, are still lacking.

*Hedyotis chrysotricha* is a short-lived, subtropical forest understory herb of southeast China, with a wide distribution in both the TIL and ZA regions [21,22]. This insectpollinated species has two floral morphs, i.e., long-styled vs. short-styled, which differ reciprocally in the placement of stigmas and anthers to enhance outcrossing. In an ideal population, an equal morph ratio is expected for the availability of compatible pollen and reproductive success [23]. However, skewed morph ratios have been observed in small and fragmented populations of many heterostylous species as a result of demographic stochasticity [24,25]. Morph types differ not only in herkogamy, but also in self- and intramorph compatibility levels, etc. This may further affect levels of inbreeding and population genetic diversity, and the ability to cope with the effects of habitat fragmentation [26–28]. For instance, van Rossum and Triest (2008) [29] found that pin and thrum individuals of distylous

*Primula veris* differed in fine-scale spatial genetic structure patterns at a small scale, which may be explained by partial self-compatibility of the pin morph combined with restricted seed dispersal and pollinator behavior. In our previous genetic study [13], we surveyed populations of *H. chrysotricha* from the recently fragmented TIL region only, using nuclear SSR markers. Although SSRs are highly polymorphic markers, a limited number of markers might result in low power for addressing landscape genetic questions at large spatial and temporal scales [30–32]. With the decreasing cost of next-generation sequencing, restriction site-associated DNA sequencing (RAD-Seq), which enables the fast identification of thousands of single nucleotide polymorphisms (SNPs) in non-model organisms without any prior information, has become a cost-effective approach in phylogeographic and population genetic studies [33]. In this study, we employed genome-wide SNPs derived from RAD-Seq data to examine and compare short-term vs. long-term effects of habitat fragmentation on *H. chrysotricha* in the TIL vs. ZA regions. Specifically, we aimed to: (1) investigate the effects of short-term vs. long-term habitat fragmentation on the species' population genetic variation (diversity, inbreeding) patterns; (2) examine the genetic differentiation and genetic structure of *H. chrysotricha* populations at small and large spatial scales; and (3) evaluate patterns of gene flow and demographic changes of *H. chrysotricha* populations in the two island systems. This comparative study will increase our understanding of the genetic responses of short-lived plants with strong seed dispersal capabilities to fragmentation and provide an insight into the management and conservation of subtropical forest dwellers in fragmented habitats at both small and large spatial–temporal scales.
