﻿Micromorphological leaf epidermal traits as potential taxonomic markers for infrageneric classification of Oxytropis (Fabaceae)

﻿Abstract The characteristics of the leaf epidermis have proven to be useful criteria to support taxonomic studies within Fabaceae. However, there are few systematic studies on the taxonomic significance of leaf epidermis of Oxytropis DC. Here, we used light and scanning electron microscopy to investigate leaf epidermal characteristics of 18 species of genus Oxytropis from the Northeastern Margin of Qinghai-Tibet Plateau. Our examination showed two main types of leaf epidermal cells: polygonal and irregular, and four different patterns of anticlinal walls: straight-arched, sinuolate, undulate, and sinuate. All species studied possess anomocytic stomata. Two trichome shapes were identified: strip-like trichomes, that were present only in O.ciliata, and cylindrical trichomes, present in all other species. Epidermal cell shape and anticlinal wall pattern were constant within species and are useful for species delimitation within genus Oxytropis, when combined with other macroscopic traits. The shape of trichomes can be useful for distinguishing O.ciliata from the other investigated species. Stomatal type was the same within the genus and may be used to elaborate the phylogenetic relationships between genera in combination with data on stomata from other genera. Cluster analysis results were largely consistent with the classification of species and sections based on macro morphological data, indicating that foliar epidermis characteristics of Oxytropis can be used as markers for taxonomic identification at the infrageneric classification level. Lastly, our results support the delineation of the section Leucopodia as an independent section but do not support the merging of section Gobicola into section Baicalia.


Introduction
Genus Oxytropis DC. is one of the largest groups within Fabaceae, with approximately 330 species occurring in the cold mountainous regions of Europe, Asia, and North America, and also concentrated in Central Asia (Polhill 1981;Zhu et al. 2010). The genus was established by De Candolle (1802), who distinguished it from Astragalus based on differences in keel-petals and legumes. In China, Oxytropis species are mainly distributed in the north and northwest regions (Zhang 1998;Zhu and Ohashi 2000). However, there is some debate regarding the delimitation and identification of species within this genus. In China, the genus was first reported by Peter-Stibal (1937), who recorded two subgenera, 11 sections, and 27 species. Wang and Tang (1955) recorded 27 species of Oxytropis and 1 variety in China. In contrast, in Flora Reipublica Popularis Sinicae (FRPS), Zhang (1998) divided Oxytropis into six subgenera, 22 sections, and 146 species. While Zhu and Ohashi (2000) recognized 125 species and 4 varieties, a recent study (Zhu et al. 2010) on the Flora of China (FOC) reported that the genus Oxytropis consisted of three subgenera and 20 sections containing 133 species. Therefore, different species delimitations in China have been proposed by taxonomists, whereby the infrageneric delimitation of Oxytropis remains controversial.
Leaf epidermal features are also valuable for classification at the species level in Fabaceae. For example, Silva et al. (2018) found that leaflet anatomy is taxonomically useful at both genus and species levels in the Dipterygeae clade. Similarly, Rashid et al. (2019) concluded that the combination of leaf epidermal characteristics and other traits has potential for taxonomic resolution at the species level in the tribe Trifolieae. Additionally, in Crotalaria, leaf characteristics, such as texture, venation pattern, and epidermis, have shown potential for aiding the circumscription of some species (Devecchi et al. 2014).
Previous studies have investigated different aspects of Oxytropis, including cytology (Ledingham 1957(Ledingham , 1960Ledingham and Rever 1963;Ranjbar et al. 2010;Liu et al. 2011;Martin et al. 2015), molecular phylogeny (Jorgensen et al. 2003;Archambault and Strömvik 2012;Dizkirici et al. 2016), pollen (Zhu and Ohashi 2000;Wang 2005;Ceter et al. 2013), and seed micromorphology (Solum and Lockerman 1991;Bojňanský and Fargašová 2007;Meyers et al. 2013;Erkul et al. 2015). On the other hand, relatively few studies on leaf epidermal anatomy have been reported in Oxytropis species (Karaman et al. 2009;Lu 2011), which described leaf epidermal traits by light or scanning electron microscopy but lacked systematic analysis. In addition, the Northeastern Margin of the Qinghai-Tibet Plateau is located in the transition zone between the Qinghai-Tibet Plateau and the Loess Plateau, the two main distribution areas of Oxytropis in China (Fig. 1). This region includes the northeastern part of the Qinghai-Tibet Plateau and the western part of the Loess Plateau (Tian et al. 2021) (Fig. 1). It is one of the regions with rich diversity of Oxytropis (Zhang 1998;Zhu et al. 2010). However, little research has been conducted on the epidermal traits of Oxytropis in this region. Thus, we provide the first systematic comparison and microscopic investigation of 18 species of Oxytropis from this region using light and scanning electron microscopy to elucidate the taxonomic significance of leaf micromorphology and test the recent taxonomic treatment (Zhu et al. 2010).  Table 1.

Materials and methods
All leaf samples were obtained from specimens deposited at the herbarium of the Northwest Normal University. The materials investigated are listed in Table 1, and the infrageneric classification by Zhu et al. (2010) was adopted (Fig. 1). Those used for analysis by light microscopy were soaked in water at 37 °C for 12-18 h; removed and placed under a dissecting microscope to separate the epidermal tissue from the leaf body, followed by maceration in 1% safranine solution. To check the consistency of the epidermal structure under a light microscope (DM6 B Leica, Leica Microsistemas S.L.U., Barcelona, Spain), at least 20 slides were prepared from different parts of a single leaf, and from different leaves of each species. The number and size of the stomata on each slide were counted. Materials for observation by scanning electron microscopy (SEM) were mounted directly on the stubs without treatment. After gold sputtering, the specimens were examined and imaged using a field emission (FE-SEM) Zeiss Ultra Plus instrument (Zeiss, Germany). Quantitative and qualitative traits were selected when performing cluster analysis.
Leaf epidermal terminology was based on the classification proposed by Dilcher (1974) and Wilkinson (1980). Stomatal index (SI) was calculated using the following equation: where, SI is the stomatal index, S is the number of stomata per unit area, and E is the number of epidermal cells per unit area. Stomatal density (SD) was expressed as the number of stomata per unit leaf area.

Data analysis
Statistical data was processed by the Origin 2021 software (OriginLab Corporation 2021). The raw data matrix includes quantitative and qualitative characters, and qualitative traits were coded using a presence/absence (0/1) matrix ( Table 2). The Euclidean distance is one of the most commonly used distance measurement methods in hierarchical clustering, which can reflect the absolute difference of individual numerical characteristics and is suitable for the analysis that needs to reflect the difference from the numerical size of the dimension (Raymond and Sylvia 1993;Farhana and Safwana 2018). The Ward error sum of squares method applies the idea of ANOVA to classification, and the obtained clustering information is more abundant and rarely affected by abnormal data (Ward 1963;Szekely and Rizzo 2005). Therefore, the Ward's method was used for cluster analysis using the squared Euclidean distance to interpret the morpho-anatomical similarity among species in this study.

Epidermal cell characters
Epidermal cell traits varied within a wide range. The shape varied from polygonal to irregular with straight arched, sinuolate, undulate, and sinuate wall patterns (    (Figs 2-4).
In addition, SEM analysis showed that, based on the shape, the waxy layer on epidermal cells could be separated into three groups (Table 4)

Stomatal characters on the epidermis
With respect to stomata, all species of Oxytropis studied here were anomocytic, and stomatal index (SI) and stomatal density (SD) of the adaxial epidermis were greater than those of the abaxial epidermis (Table 3). In most of the examined species,  remarkable variation was observed in stomatal size and number. Specifically, SD was lowest in O. coerulea (Table 3; Fig. 3) and largest in O. melanocalyx (Table 3; Fig. 4)  observation, the inner margin of the outer stomatal rim was either undulate or smooth (Table 4). Five species, including O. ciliata, O. muricata, O. ochrantha, O. glabra, and O. latibracteata, showed smooth and undulating inner margins of the outer stomatal ledge (Figs 5,7,9), while O. bicolor showed only a smooth inner margin of the outer stomatal ledge (Figs 5, 6). In contrast, the remaining species  had an undulate inner margin of the outer stomatal ledge. Ornamentation of the outer stomatal ledge was smooth or granular in most species under this study. O. coerulea was a notable exception with a banded sediment ornamentation of the outer stomatal ledge (Fig. 7).

Trichome characters on the epidermis
Most of the species observed showed trichomes, except for O. squammulosa (Table  4; Fig. 9). Two trichome shapes were identified in this genus. Strip-like trichomes, that were present only in O. ciliata (Table 4; Fig. 5), and cylindrical trichomes, that  were present in all other species (Table 4). Trichrome ornamentation of O. ochrantha, O. ciliata, and O. myriophylla was striate (Table 4; Figs 5, 6), while the remaining species were striate and granular (Table 4).

Cluster analysis
Cluster analysis reflects the similarity among species based on anatomical characteristics and delimitation of these groups. The phenograms of the quantitative and qualitative data provided four principal clusters (Fig. 10)

Discussion
Leaf characteristics, such as epidermal micro-and macro-hairs, and stomata, are important for the classification of many genera (Dickison 2000;Yang and Lin 2005;Kadiri and Muellner-Riehl 2021). Previous studies have shown that the anatomical features of the leaf epidermis, such as the shape and anticlinal walls of epidermal cells, are taxonomically significant and can therefore be used for the classification of taxa at the genus or even at the species level (Barthlott et al. 1998;Wissemann 2000;Tomaszewski and Zieliński 2014;Tomaszewski et al. 2019). In this study, there were two main types of leaf epidermal cells: polygonal and irregular; and four different types of pattern of anticlinal walls: straight-arched, sinuolate, undulate, and sinuate. It has been proposed that the pattern of the anticlinal wall may be influenced by habitat; specifically, species in dry environments tend to have a straight arched anticlinal wall, whereas those in humid areas tend to have undulating to sinuous anticlinal walls (Stace 1965;Gifford and Foster 1989 (Zou et al. 2008;Ren et al. 2007). Therefore, epidermal cell shape and anticlinal wall patterns need to be considered in combination with other macro-morphological features classifying the species within the genus Oxytropis. Studies on stomata can have great taxonomic significance for the delimitation of different levels of taxa (Kothari and Shah 1975). Carpenter and Smith (1975) showed that variability in stomatal frequency is taxonomically important at the genus level, whereas Carlquist (1961) emphasized the contribution of stomatal size variation to delimiting species within a genus. In Oxytropis, a wide range of variability was observed for stomatal quantitative parameters, such as stomatal density, size, and index. Our results indicated that the quantitative stomatal traits have limited taxonomic value, as they are strongly influenced by environmental factors, such as CO 2 levels and light intensity (Metcalfe and Chalk 1950;Royer 2001;Rossatto and Kolb 2010). However, stomatal distribution and types are considered an important taxonomic criterion for taxonomic value, especially at higher taxa (Metcalfe and Chalk 1950;Patil and Patil 1987). Thus, for example, we found that the anomocytic stomata type is a common feature in Oxytropis that may be used to elaborate the phylogenetic relationships among genera, in combination with stomatal data from other genera. These findings support the concept that genus Oxytropis is a monophyletic group (Zhu and Ohashi 2000).
Further, trichomes and their characteristics provide important information for plant identification. The type of indumentum and its presence or absence may serve as diagnostic features for species or genus recognition, as has been recognized in some groups such as Asteraceae (Adedeji and Jewoola 2008;Krak and Mráz 2008), Brassicaceae (Beilstein et al. 2006), Fabaceae (Chukwuma et al. 2014), and Lamiaceae (Eiji and Salmaki 2016). In addition, large plant taxa often share a common pattern of trichome structure. For example, chandelier-shaped trichomes with branches of whorls are characteristic of Platanaceae ; peltate or scale-like hairs are typical of Eleagnaceae (Mishra 2009), and three-celled uniseriate hairs are common in Proteaceae (Johnson and Briggs 1975). In the genus Oxytropis, the trichome type of the investigated species was simple hair. This is consistent with the results of previous studies on Oxytropis (Karaman et al. 2009;Lu 2011). Furthermore, we found that O. ciliata, belonging to Section Xerobia (Zhu et al. 2010), has strip-like trichomes (margin ciliates) that distinguish this species from other species in this study. The trichrome ornamentation of most Oxytropis species was consistent, indicating that trichrome ornamentation appears to be of a low taxonomic value for distinguishing sections and species. However, owing to sample size limitations, the systematic significance of Oxytropis trichomes needs to be based on a more comprehensive sampling.
In this study, six species, including O. glabra, O. kansuensis, O. melanocalyx, O. taochensis, O. ochrocephala, and O. xinglongshanica, all belonging to section Mesogaea, clustered together. Our results of cluster analysis are largely consistent with that of the classification of species and sections based on macro morphological data (Zhu et al. 2010), indicating that leaf epidermal micro characteristics might be valuable in understanding systematics of genera at the section level. Bunge (1874) established the section Gobicola in 1874, which contained only O. racemosa. This treatment was recognized by the FRPS, but section Gobicola was merged into section Baicalia in Flora IntraMongolica and FOC (Fu 1989;Zhang 1998;Zhu et al. 2010). However, our results do not support the interpretation of Flora IntraMongolica and FOC. In this study, O. racemosa and some species of the section Mesogaea, such as O. glabra and O. kansuensis, clustered together into one clade, indicating that the systematic position of O. racemosa needs to be reconsidered (Fig. 10). In addition, based on the results of quantitative taxonomy, Wang (2005) advocated that section Leucopodia, which only contains O. squammulosa, should be merged with section Xerobia. Our results clearly do not support this treatment, because O. squammulosa did not cluster together with O. ciliata in section Xerobia (Fig. 10). Moreover, different populations of O. ochrocephala clustered together into one group, which demonstrates that leaf epidermal traits are useful for the identification of taxa at the species level. Therefore, foliar epidermis traits of Oxytropis can be used as taxonomic markers for identification at the infrageneric classification level.
There is no comprehensive phylogenetic study on the genus Oxytropis. Furthermore, although several studies have applied DNA barcodes such as ITS, trnL-F, and psbA-trnH to explore the molecular phylogeny of Oxytropis in Northwestern China, the low genetic divergence of the above barcodes among the species makes it difficult to distinguish species within the genus as well as to resolve phylogenetic relationships between sections (Li et al. 2011;Gao et al. 2013;Lu et al. 2014). Therefore, the reliability of epidermis characters in terms of phylogeny cannot be affirmed. More detailed molecular phylogenetic studies with a broader taxon sampling are required to find correlations between epidermis characteristics and classification of the genus.

Conclusions
Our results suggest that leaf epidermis can be used as potential taxonomic markers for infrageneric classification of Oxytropis. The shape of epidermal cells and the pattern of the anticlinal wall were constant within species, and are useful for species delimitation in the genus Oxytropis when combined with other macroscopic traits. Trichome shapes can be useful characteristics to distinguish O. ciliata from other investigated species. Although quantitative stomatal characteristics were not effective diagnostic characteristics because of the considerable variation within the same taxa, it nevertheless plays an important role in cluster analysis. Results of cluster analysis are largely consistent with the classification of species and sections based on macro morphological data, indicating that foliar epidermis characteristics of Oxytropis can be used as taxonomic identification markers infrageneric classification level. Lastly, our results support the delineation of the sect. Leucopodia as an independent section, while not supporting the treatment of merging the sect. Gobicola into the sect. Baicalia.