Taxonomic relevance of leaf surface micromorphology in Korean Clematis L. (Ranunculaceae)

Introduction

Clematis L. ranks among the largest genera in the Ranunculaceae family, comprising around 280–350 species (Tamura, 1987, 1995; Wang & Li, 2005). Most of the Clematis taxa are woody or herbaceous vines, but a few are shrubs, subshrubs, or erect perennial herbs. This cosmopolitan genus has a broad global distribution, with significant variations in diversity in temperate and subtropical regions of the Northern Hemisphere, particularly in eastern Asia. These significant morphological variations among Clematis species have attracted research attention since the early 19th century, resulting in numerous infrageneric revisions (Prantl, 1888; Tamura, 1987, 1995; Johnson, 1997; Grey-Wilson, 2000; Johnson, 2001; Wang & Li, 2005). Tamura (1987) divided this genus into four subgenera and 16 sections, some of which were subdivided into subsections and series. Grey-Wilson (2000) conducted a grouping of 297 Clematis species into nine subgenera, 16 sections, and 26 subsections. Conversely, Johnson (2001) identified 325 species, categorizing them into 18 sections and 36 subsections. More recently, Wang & Li (2005) proposed a new classification system based on analyses of morphological and palynological traits of 345 Clematis species. They established four subgenera, two of which were similar to Tamura’s (1987) subgenera, and introduced two new subgeneric names within the genus. They further divided the subgenera into 15 sections and numerous subsections and series.

Numerous systematic studies have been conducted on Clematis, focusing on palynology, cytology, and the anatomy of various plant parts (Tobe, 1974, 1979, 1980a, 1980b, 1980c, 1980d; Essig, 1991; Zhang & He, 1991; Yano, 1993; Yang & Moore, 1999; Shi & Li, 2003; Xie & Li, 2012; Ghimire et al., 2020; Park, Son & Ghimire, 2021). Unfortunately, none of these morphological characteristics seem to provide enough details to resolve the infrageneric uncertainty of this large genus. Because different taxonomists interpret morphological and anatomical features differently, it is very difficult to accurately classify this genus, especially at the sectional level. In addition, the existing classification systems for this genus vary due to the complex morphological diversity within the genus and the different characteristics emphasized in each system (Tamura, 1987; Wang & Li, 2005; Johnson, 1997, 2001; Prantl, 1888). In recent years, multiple molecular phylogenetic analyses (Miikeda et al., 1999, 2006; Slomba, Garey & Essig, 2004; Xie, Wen & Li, 2011; Jiang et al., 2017; Lehtonen, Christenhusz & Falck, 2016) have predominantly supported the monophyly of Clematis, recommending the segregation of Archiclematis and Naravelia into a distinct section within the genus (Miikeda et al., 2006; Xie, Wen & Li, 2011; Jiang et al., 2017). However, the phylogenetic hypotheses based on morphological data are largely inconsistent with those derived from molecular data.

China stands out as the center of diversity of the genus Clematis, harboring 147 species, 93 of which are endemic to the country (Grey-Wilson, 2000; Johnson, 2001; Wang & Bartholomew, 2001). In a synoptic overview of Korean flora, Nakai (1952) documented 21 species and 14 varieties of Clematis, but later, Lee (1967) revised the count to 16 species, 11 varieties, and five formae within the genus. Lee’s (2006), New Flora of Korea introduced 18 taxa, including C. taeguensis Y. Lee, initially described by Lee (1982). The Checklist of Vascular Plants in Korea by the Korea National Arboretum (2017) listed 17 species and five varieties of Clematis, while Kim (2017) detailed 13 species and seven varieties in The Flora of Korea. After a thorough review of Lee (2006), Chang, Kim & Chnag (2011), the Korea National Arboretum (2017), and Kim (2017), this study encompasses 16 species and three varieties.

Plant micromorphological features, including vegetative organs, can be valuable tools for taxonomy and species identification (Ullah et al., 2018; Ghimire et al., 2020; Park, Son & Ghimire, 2021; Zhao et al., 2022; Suwanphakdee, Sutthinon & Hodkinson, 2024). Despite receiving less attention compared to other reproductive characteristics (such as flowers, fruits, and seeds) in terms of taxonomic information, leaf micromorphology has been shown to offer valuable diagnostic and taxonomic traits (Stace, 1965; Wilkinson, 1979; Heywood et al., 2007). Studies have demonstrated that epidermal cells, stomatal complex (Carpenter, 2005; Garcia Alvarez et al., 2009; Grace et al., 2009; Yang et al., 2012; Lee & Oh, 2019; Nisa et al., 2019), trichomes (Webster, Del-Arco-Aguilar & Smith, 1996; Hu, Balangcod & Xiang, 2012; Eiji & Salmaki, 2016; Mannethody & Purayidathkandy, 2018; Ashfaq et al., 2019), and epicuticular waxes (Barthlott et al., 1998; Wissemann, 2000; Tomaszewski & Zielinski, 2014; Tomaszewski, Byalt & Gawlak, 2019) on the leaf surface are valuable for species identification. This approach has been used in the genus Clematis, where the leaf epidermis is used as a distinguishing criterion for identifying the subsections in section Meclatis (Spach) Tamura and section Fruticella Tamura (Decamps, 1974; Grey-Wilson, 2000; Shi & Li, 2003). Additionally, this approach supports the notion of several distinct evolutionary processes within the genus Clematis (Shi & Li, 2003). Due to the considerable diversity in morphological characteristics among Clematis taxa, the classification system requires additional evidence to establish a more accurate and comprehensive systematic framework. Despite the potential taxonomical significance of leaf micromorphological characteristics in Clematis taxa, no such studies have been conducted in Korean Clematis. We hypothesize that incorporating leaf epidermal characteristics, including qualitative and quantitative stomatal features, will greatly enhance the systematic classification of Clematis species distributed in the Korean Peninsula. This study sought to: (1) evaluate the variation in the adaxial and abaxial leaf surface in Clematis including cell sculpture, stomata structure, stomata density, leaf margin, and trichomes; (2) determine the infrageneric and infraspecific variation of these characteristics; and (3) analyze the taxonomic importance of leaf epidermal characteristics by establishing species groups based on these traits.

Materials and Methods

Results

There was significant variation in the leaf surface morphology and stomata among the investigated Clematis species. Observations of leaf surface micromorphology and stomatal characteristics of 19 Clematis taxa are presented in Tables 2 & 3 and Figs. 1–10, and comparative descriptions of selected features are presented below.

Discussion

This study presented new data on leaf micromorphological features in Clematis taxa distributed in Korea and attempted to assess their systematic significance in the genus. While leaf micromorphological characteristics could aid in establishing systematic relationships among species, there has been a lack of extensive research conducted within this genus. A few fragmented studies on leaf surface morphology of Clematis have been published providing evidence for the taxonomic discrimination of several taxa (Decamps, 1974; Grey-Wilson, 2000; Shi & Li, 2003). However, numerous environmental factors can significantly influence a plant’s ability to thrive in a specific habitat, prompting potential alterations in leaf morphological characteristics as an adaptation to distinct environmental conditions. As leaves are a part of the plant exposed to the environment, some leaf features, such as trichome density and cuticular striation, are considered the result of plant adaptation to a habitat and are affected by environmental factors (Stace, 1965; Parkhurst, 1978; Takahashi & Miyajima, 2008; Zhu, Kang & Liu, 2011). However, the type and size of stomata, the presence and abundance of trichomes, and the shape of epidermal cells are usually consistent at the species level, and are often used as key characteristics in species identification (Okpon, 1969; Stace, 1965; Shi & Li, 2003; Ayodele & Olowokudejo, 2006).

All the species included in this study were hypostomatic, although amphistomatic species have been recorded in previous studies for this genus (Shi & Li, 2003). Shi & Li (2003) observed 10 Clematis taxa with amphistomatic leaves (C. fruticosa, C. tomentella, C. tanguticai, C.akebioides, C. glauca, C. intricate, C. orientalis, C. cadmina, C. gentianoides, and C. lingusticifolia) in their 77 sampling taxa. Hypostomaty is regarded as the primitive trait in land plants, with amphistomaty being considered the derived state for stomatal distribution. As a result, vascular cryptogams are predominantly characterized by hypostomaty (Mott & O’leary, 1982). Furthermore, hypostomaty has been observed in the tallest trees, which experience elevated levels of overhead irradiance across the entire crown (Rozendaal, Hurtado & Pooter, 2006). Trees generally exhibit hypostomatic characteristics, while non-woody plants tend to possess hypostomatic or amphistomatic leaves (Peat & Fitter, 1994). Stomata on both leaf surfaces have been observed in various unrelated groups in the Ranunculaceae family, implying that this trait did not originate from a single common ancestor but instead evolved, independently, multiple times within the family (Grey-Wilson, 2000; Shi & Li, 2003). Furthermore, the distribution of stomata on the leaf might be linked to habitat-related factors, given that amphistomatic species are typically associated with xerophytic environments (Parkhurst, 1978).

Unlike the position of stomata, stomata size seems to have taxonomic scope in the family as more primitive taxa bear larger stomata than advanced ones in tribe Anemoneae of the family Ranunculaceae and in section Cheriopsis of the genus Clematis (Wang, 1998; Fu & Hong, 2001; Wu & Ravan, 2001; Shi & Li, 2003). None of the species used in the present study belonged to the sect. Cheriopsis, and although there were no significant differences in stomatal size between the observed species, the stomatal area varied from 855.5 ± 102.61 µm² to 1697.1 ± 210.71 µm², with the smallest stomata observed in C. apiifolia of sect. Clematis and the largest in C. calcicola of sect. Atragene. Of greater significance, the second largest stomata were measured in the section identical to C. apiifolia, namely sect. Clematis. Although the specimens sampled in this study represented a limited range of taxa, the results disagreed with the concept of primitive taxa bearing larger stomata proposed by previous studies regarding the evolutionary patterns of stomatal size in the genus (Wang, 1998; Fu & Hong, 2001; Wu & Ravan, 2001; Shi & Li, 2003). Alternatively, plants grown in dry habitats can have smaller stomatal sizes to allow greater water-use efficiency as they open and close more rapidly, whereas those grown in moist habitats have larger stomata (Hodgson et al., 2010).

There was significant variation in stomatal density among the species, as the highest stomatal density, found in C. serratifolia (88 ± 6.92), was four times higher than the lowest density, found in C. urticifolia, (21.33 ± 2.55). Stomatal density in the leaves has been shown to be highly influenced by environmental factors such as CO₂ levels (Royer, 2001) and light intensities (Rossatto & Kolb, 2010). Higher stomatal density is associated with the plant’s habitat, and high light intensities are crucial for the exchange of gases during photosynthesis and the release of water vapor during transpiration (Galmes et al., 2007; Camargo & Marenco, 2011). Therefore, plants that are distributed in sunny areas usually have higher stomatal density than the species that are found in shady areas (Song, Yang & Choi, 2020). Among the observed species in the current study, C. serratifolia usually grows in dry forests and slopes while C. urticifloia grows on the edge of bushes under deciduous trees. Although the largest stomatal area did not correspond with higher density in this study, there was a negative correlation between stomatal size and stomatal density (P < 0.0001). This inverse relationship between stomatal size and density has also been observed in other groups of taxa (Hetherington & Woodward, 2003; Nejad & van Meeteren, 2005; Pearce et al., 2006; Camargo & Marenco, 2011; Song & Hong, 2017; Kong & Hong, 2019).

The taxonomic significance of foliar trichomes has been studied for many groups of taxa (Metcalfe & Chalk, 1950; Abu-Assab & Cantino, 1987; Cantino, 1990; Navarro & Oualidi, 2000; Khokhar, 2009; Khokhar, Rajput & Tahir, 2012), although environmental conditions and the adaptation mechanisms of the plants often affect the degree of hairiness (Cantino, 1990; Ayodele & Olowokudejo, 2006). In a particular group of the taxa, the structure and density of trichomes can be diverse, and sometimes too variable for taxonomic use (Guerin, 2005). However, this variation typically falls within definable limits, and the presence, density, and dominance of particular trichome types serve as valuable indicators for differentiation between specific taxa, particularly when combined with morphological characteristics (Steyn & VanWyk, 2021). In the present study, considerable variation in the density of trichomes (both eglandular and glandular) was observed. In the majority of taxa, the density of eglandular trichomes was greater than that of glandular trichomes. Conversely, in C. seratifolia and C. fusca var. fusca, the density of glandular trichomes was greater than that of eglandular trichomes. Glandular trichomes secrete secondary metabolites, such as terpenes, phenolics, alkaloids, lipophilic, and other substances that the plants use as self-defense against herbivores and pathogens (Levin, 1973; Werker, 2000; Tissier, 2012; Ma et al., 2016). However, there is significant diversity among species, within species, and even among individual plants in the shapes and compounds they secrete (Tissier, 2012). It is thought that glandular trichomes have evolved from eglandular trichomes through the differentiation of apical cells into secretory cells (Uphof, 1962; Fahn & Shimony, 1997). While there is no specific evidence for this, especially for the genus Clematis within the Ranunculaceae family, the glandular trichomes seem to represent an apomorphy resulting from the enlargement of eglandular trichomes (Aleykutty & Inamdar, 1980; Hoot, 1991). However, this generalization may not hold for all taxonomic groups. For instance, in the genus Vitis (Vitaceae), glandular trichomes appear to be plesiomorphic but are lost in the New World Vitis subgenus Vitis (Ickert-Bond et al., 2018).

Noticeable differences were observed between the study taxa in foliar epidermal features such as epidermal cell shape (ECS), epidermal cell boundary (ECB), number of epidermal cells connected with subsidiary cells (ECSC), abundance of trichomes in the margin (ATM), marginal cell types (CM), and cells of the midvein (CMV). Micromorphological variances such as these have been effective tools in proposing new alignments among species and within species (Hoot, 1991; Grace et al., 2009; Ickert-Bond et al., 2018; Hanafy et al., 2019; Song, Yang & Choi, 2020). The epidermal cell structure on the abaxial and adaxial surfaces of the Clematis displayed some consistency, with the majority of the taxa having irregular epidermal cell structure, followed by elongated and then irregular and polygonal-shaped cells. The epidermal cell boundaries on both surfaces showed similar patterns as well, as the majority of species had a combination of omega (Ω) shaped and U-shaped striation, followed by a combination of straight and undulate striation (especially on the abaxial surface). This variation in the epidermal shape and cell boundaries develops various interlocking lobes and indentations, which often resemble puzzle pieces in leaf epidermal cells. This complex geometry offers biomechanical advantages; interlocking increases resistance to tension and maintains the structural integrity of the leaf surface (Malinowski, 2013). The morphology of pavement cells may serve as a substitution for the mechanical properties of the leaf epidermis. Although several parameters have been used to describe these pavement cells, including aspect ratio, number of lobes and indentations, circularity, solidity, and convexity (Kuan, Yang & Ho, 2022), the evidence for taxonomic relevance of such parameters is still unexplored.

All of the observed species had deep anticlinal walls on the abaxial surface with either wavy or interlocked outlines. However, on the adaxial surface, most taxa had shallow anticlinal walls. Patterns of the periclinal wall did not differ between most species, with the periclinal wall on the abaxial surface having raised or convex dominance and the periclinal wall on the adaxial surface having flat to slightly raised dominance. Clematis taeguensis was the only species with exclusively papillose and flat periclinal walls on the abaxial and adaxial surfaces, respectively. This Korean endemic plant is restricted to certain areas and distinguishable from its closely related taxa by lacking shoot remnants in the winter season, having a hairy abaxial surface and margin of the sepal, and having yellow to brown hairs on the style (Ko & Park, 2018; Park et al., 2021). The combination of the convex and papillose periclinal wall was also found in C. hexapetala. However, C. hexapetala is more distinct from other taxa because it has three to five subsidiary cells (C. trichotoma has three to six subsidiary cells) and solely papillose leaf marginal cells. A similar result was found by Park, Son & Ghimire (2021) who found that among the 19 species they investigated, C. hexapetala was the only species with glabrous petioles (although sparsely distributed trichomes could be observed at the base of the leaflets). In the present study, the observed species showed remarkable similarity in the structure of the stomatal complex even though there was a slight variation in the number of subsidiary cells, with minor differences in the shape, size, and surface features of subsidiary cells in the same stomatal type. These results indicate these characteristics may not have specific relevance to the taxonomical study of the genus Clematis.

The cluster analysis based on 13 leaf micromorphological features generated at least four clusters. Of these, Clematis calcicoa solely represented a cluster, which was separate from the rest of the species (Fig. 11). One of the crucial leaf micromorphological features of this species is the rare occurrence of glandular trichomes, although C. mandshurica and C. hexapetal also share a similar feature. Most of the infra-generic classifications of the genus showed that C. calcicola is closely allied with C. koreana and C. ochotensis (Lehtonen, Christenhusz & Falck, 2016; Tamura, 1987; Johnson, 1997; Wang & Li, 2005). Ghimire et al. (2020) also demonstrated that these three species grouped within the same cluster and share some achene features, like permanent style length, reticulate primary surface sculptures, identical anticlinal walls, and endocarp cell structures. However, this was not the case in the present study as C. koreana and C. ochotensis grouped with C. patens, C. fusca, and C. terniflora complex. This indicates that these leaf features might not be useful for infra-generic discrimination in the Clematis. In another report, based on petiole anatomy, Park, Son & Ghimire (2021) found that C. calcicola, C. koreana, and C. ochotensis, which belong to section Astragene, are clustered with C. fusca complex (section Virona) and C. brachurya (section Pterocarpa). Nevertheless, according to infrageneric classifications of the genus, C. brachurya is closely allied with C. taeguensis, C. hexapetala, C. mandshurica, and C. terniflora, which belong to section Flammula (Tamura, 1995; Johnson, 1997; Wang & Li, 2005; Xie, Wen & Li, 2011). Notably, in this study, C. brachurya was observed to be grouped within a larger cluster that included C. taeguensis (sect. Flammula–Tamura, 1995, Johnson, 1997; Clematis–Wang & Li, 2005), C. serratifolia (sect. Meclatis), C. apifolia, C. bravicaudata, C. trichotoma (sect. Clematis), C. pseudotubulosa, and C. urticifolia (sect. Tubulosae). Such uncertainty is a common and typical interpretation for this large genus, a trend observed in previous morphological and molecular studies (Grey-Wilson, 2000; Wang & Li, 2005; Miikeda et al., 2006; Xie, Wen & Li, 2011; Xie & Li, 2012; Lehtonen, Christenhusz & Falck, 2016; Ghimire et al., 2020; Park, Son & Ghimire, 2021).

The principal component analysis (PCA) exhibited a similar result as that of the UPGMA cluster analysis. According to the PCA, the taxa from section Virona are closely associated with C. terniflora (sect. Flammula) and C. pseudotubulosa (sect. Tubulosae) in axis 2 (Fig. 12). Nevertheless, C. mandshurica and C. hexapetala, which belong to the sect. Flammula, remained close to the negative side of axis 2. Previous research has shown that the infrageneric classification of Clematis lacks robust support from both molecular analyses and morphological data. Aligning with the findings of previous studies, the results of this analysis confirm that attempting infrageneric classification of this genus exclusively based on morphological characteristics poses significant challenges. In this analysis, out of the three taxa from section Tubulosae, C. pseudotubulosa and C. tubulosa were grouped with C. apiifolia (section Clematis), whereas C. urticifolia remained distantly on the positive side of axis 1. According to Lehtonen, Christenhusz & Falck (2016), these three species, C. pseudotubulosa, C. tubulosa, and C. apiifolia, were found to be grouped within clade C, alongside C. bravicaudata, C. trichotoma, and C. urticifolia. Remarkably, the present study revealed that these six taxa, along with C. brachurya, C. serratifolia, and C. taeguensis, collectively formed a large cluster. However, the grouping between these taxa seems arbitrary here because—except for abundant glandular trichomes, epidermal cell boundaries on the adaxial surface, absence of wax in the leaf surface, and comparable stomatal width, which are frequently evolved in other taxa as well—there is no single significant leaf micromorphological feature common among these taxa. In addition, these nine taxa belong to at least four different sections and/or clades in previous infrageneric classifications (Tamura, 1995; Johnson, 1997; Wang & Li, 2005; Lehtonen, Christenhusz & Falck, 2016).

Conclusion

This study showed that the type and abundance of leaf trichomes, stomata size and density, epidermal cell types and structure, secondary and tertiary walls of epidermal cells, cell structure in the midvein, and presence or absence of wax in the leaf were useful characteristics for species discrimination in Korean Clematis. The results of this study also revealed resemblances in some key leaf micromorphological features among taxa belonging to the sections Tubulosae, Clematis, and Virona. Nevertheless, it is important to note that the number of taxa investigated within these groups was relatively limited, rendering any interpretation based on this data somewhat arbitrary. The main conclusion of this study is that relying solely on leaf micromorphology is not enough to clarify the challenging infrageneric relationships within the genus Clematis. However, this study offers novel and intriguing perspectives on these data, serving as a potential source of descriptive and/or diagnostic features for specific taxa within the genus. Additionally, it makes a number of exciting recommendations for future research that could significantly advance our knowledge of the ecology, evolution, and taxonomy of Clematis. In order to address the issues of infrageneric differentiation in this diverse and complex genus, future studies should integrate multiple lines of data and expand the scope of investigations.

Supplemental Information