**Documenting Greek Indigenous Germplasm of Cornelian Cherry (***Cornus mas* **L.) for Sustainable Utilization: Molecular Authentication, Asexual Propagation, and Phytochemical Evaluation**

**Eleftherios Karapatzak 1, Nikos Krigas 1,\*, Ioannis Ganopoulos 1, Katerina Papanastasi 1, Dimitris Kyrkas 2, Paraskevi Yfanti 2, Nikos Nikisianis 3, Antonis Karydas 1, Ioannis Manthos 1, Ioanna S. Kosma 4, Anastasia V. Badeka 4, Dimitrios Fotakis 5, Eleni Maloupa <sup>1</sup> and Giorgos Patakioutas 2,\***


**Abstract:** Wild-growing Cornelian cherries (*Cornus mas* L., Cornaceae) are well-known native fruits in Greece since ancient times that are still consumed locally nowadays. Modern research has highlighted the value of Cornelian cherries as functional food with exceptional health benefits on account of the fruits' biochemical profile. However, apart from local consumption directly from wild growing individuals, Greek native *C. mas* populations have not yet been investigated or sustainably utilized. A multifaceted evaluation was conducted herein including authorized collection-documentation, taxonomic identification, and molecular authentication (DNA barcoding), asexual propagation via cuttings and phytochemical evaluation (multiple antioxidant profiling) of neglected and underutilized Greek native *C. mas* germplasm sources. Successive botanical expeditions resulted in the collection of 18 samples of genotypes from distant *C. mas* populations across different natural habitats in Greece, most of which were DNA fingerprinted for the first time. Asexual propagation trials revealed high variability in rooting frequencies among Greek genotypes with low (<25%), average (25–50%), and adequate propagation potential (>50%) using external indole-3-butyric acid (IBA) hormone application on soft- or hard-wood cuttings. The comparative phytochemical evaluation of the studied Greek genotypes showed significant potential in terms of antioxidant activity (>80% radical scavenging activity in 13 genotypes), but with variable phenolic content (47.58–355.46 mg GAE/100 g), flavonoid content (0.15–0.86 mg CE/100 g), and vitamin C content (1–59 mg AAE/100 g). The collected material is currently maintained under ex situ conservation for long-term monitoring coupled with ongoing pilot cultivation trials. The pivotal data create for the first time a framework for the sustainable utilization of Greek native C. mas germplasm as a superfood with significant agronomic potential.

**Keywords:** neglected and underutilized plants; phytogenetic resources; DNA barcoding; forest berries; protocols; nutraceutical potential; genotype selection; multifaceted evaluation

**Citation:** Karapatzak, E.; Krigas, N.; Ganopoulos, I.; Papanastasi, K.; Kyrkas, D.; Yfanti, P.; Nikisianis, N.; Karydas, A.; Manthos, I.; Kosma, I.S.; et al. Documenting Greek Indigenous Germplasm of Cornelian Cherry (*Cornus mas* L.) for Sustainable Utilization: Molecular Authentication, Asexual Propagation, and Phytochemical Evaluation. *Plants* **2022**, *11*, 1345. https://doi.org/ 10.3390/plants11101345

Academic Editors: Sofia Rhizopoulou, Maria Karatassiou and Efi Levizou

Received: 13 April 2022 Accepted: 12 May 2022 Published: 19 May 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/).

## **1. Introduction**

Apart from widely used crop varieties, phytogenetic resources also include wildgrowing plant species which are often neglected or underutilized, especially in biodiversityrich areas [1]. Such germplasm resources involve a wide variety of plant species, including small fruit trees of potentially high alimentary value, relating to nutritional health benefits, but also pharmaceutical value [2–6]. Local native fruit species can be an important source of natural food products rich in antioxidants such as polyphenols with known importance for human health [7–9]. The natural heritage of Greece and its rich biodiversity at all levels offer new possibilities for evaluating and utilizing valuable wild germplasm such as small stone fruits and berries with high nutritional and medicinal (in one word nutraceutical) value [6,10]. Prior to domestication and sustainable exploitation of wild germplasm resources of interest, multifaceted research approaches are required in terms of taxonomic validation and molecular authentication of wild plant material, agronomical assessment propagation-wise, and comprehensive phytochemical evaluation aiming at the selection of promising genotypes with high potential for further applied research and future breeding. To this end, Greek native *Rosa canina* and *Sambucus nigra* germplasms constitute recently developed example cases [6,10].

Cornelian cherry (*Cornus mas* L., Cornaceae) is a rather common and wild-growing deciduous shrub or occasionally a small tree 2–4(–6) m tall in scrub and woodlands of central and southeastern Europe and western Asia, often occurring in ravines or near streams (areas with sufficient rainfall) from low (100–200 m) to intermediate and higher altitudes (800 m, occasionally up to 1700 m) [11]. In Greece, Cornelian cherry is a traditionally wellknown native fruit tree that has been consumed locally for centuries and it has been known since ancient times, i.e., the Trojan horse in Homeric times was believed to be a *C. mas* woodcraft [12]. The trees bloom by the end of winter and in early spring and set fruits (ellipsoid to broadly cylindrical drupes 12–15 mm long, becoming red and cherry-like) in late summer through fall. Cornelian cherry fruits (both wild-growing and cultivated) and their food derivatives (e.g., jams, jellies, juice) contain antioxidants, phenolic compounds, iron, potassium, vitamin C, flavonoids, and many other substances of high nutritional value [13–17]. In addition, *C. mas* fruits to date have validated pharmaceutical value in terms of free radical scavenging potential among others, resulting in several health benefits including antimicrobial, antidiabetic, anti-inflammatory, anticancer, and cardioprotective activity; these assets render *C. mas* fruits as functional food products and as 'superfood' with intensifying relevant research over recent years [18–20]. Apart from the flesh of the fruits, the endocarp of *C. mas* drupes has also been demonstrated to have nutraceutical potential as a source of bioactive compounds with strong antioxidant activity [21]. Sustainable utilization efforts of *C. mas* germplasm have been attempted in eastern European countries with promising results [22–24]. In more recent studies in the same area, the phytochemical evaluation of *C. mas* domesticated germplasm (cultivars and/or hybrids) has been conducted [22]. Furthermore, *C. mas* germplasm has been characterized in terms of the nutritional value of the fruits across a wider spectrum of its distribution range [25–28] and the results showed the nutritional superiority of wild-growing germplasm [29]. However, in these studies no assessment was included regarding the native *C. mas* wild-growing genotypes occurring in Greece.

To support classical taxonomic identification and to enable the establishment of a distinct genetic identity, molecular authentication can be performed through novel DNA barcoding and bioinformatic analysis methods in selected plant materials collected from natural habitats [30]. DNA barcoding as a tool for molecular authentication of plant germplasm has been greatly developed and to date is being widely used for many medicinal plant species and phytogenetic resources [6,10,31–35]. There are also reports of DNA barcodes being successfully used in identification of ancient DNA (aDNA) of *C. mas* from fossils [36]. However, such molecular tools are still underused in authenticating native *C. mas* germplasm of different regions.

To date, wild-type populations of Cornelian cherry (*C. mas*) from several parts of the world have been successfully propagated via cuttings in the past with the use of external hormone application on hardwood and soft wood cuttings taken from mature individuals growing in the wild [37,38]. A popular hormonal substance that has been used in the past is indole-3-butyric acid (IBA) [39,40], frequently achieving high rooting rates.

In this framework, the study herein was focused on the collection and multifaceted documentation of *C. mas* plant material from geographically separated Greek populations (different genotypes) with the aim to provide: (1) molecular authentication of the population samples via DNA barcoding; (2) asexual propagation protocols for the collected genotypes via propagation trials using cuttings (dormant hardwood twigs, fresh softwood plant parts) facilitating the ex situ conservation and evaluation of the collected germplasm; and (3) comparative phytochemical evaluation of their fruits in terms of nutraceutical properties (total phenolic content, antioxidant activity, total flavonoids, and vitamin C content) to assess their potential as germplasm sources for artificial selection of genotypes and future breeding efforts. The overall work provided first-time insight into Greek native *C. mas* (Cornelian cherry) germplasm and these efforts are aimed at paving the way for its sustainable utilization as a superfood with significant agronomic potential.

## **2. Results**

## *2.1. Molecular Authentication of Greek Native Cornus mas Genotypes*

BLAST1 and distance-based methods were selected to validate the authentication efficiency of ITS2 for the selected Greek native *Cornus mas* genotypes. The ITS2 barcode with BLAST1 showed a high competence (97% and 100%) in sample identification at species and genus levels, respectively. The ITS2 barcode with the DISTANCE method showed almost similar identification efficiency (98%) at the species level. Thus, barcode ITS2 using the NJ tree method was proved able to distinguish Greek native *C. mas* genotypes from other genotypes of different origin, and similarly, from other *Cornus* spp. not present in Greece (Figure 1A). The neighbor-joining (NJ) phylogenetic tree obtained from DNA barcoding application with the ITS2 region discerned two main groups among the samples of *Cornus* spp. used herein (*n* = 21), and clearly classified the Greek native *C. mas* genotypes of this study as distinct sub-group with respect to other samples of *C. mas* of different origins that were sourced from databases (Figure 1A). The nucleotide differences in Greek genotypes are depicted in Figure 1B. The separation of the genotypes in Figure 1A presents similarity to an extent with the geographical separation of the respective habitats since genotypes GR-1- BBGK-19,195, GR-1-BBGK-19,197, GR-1-BBGK-19,198, and GR-1-BBGK-19,190 coupled with genotype GR-1-BBGK-19,196 originate from areas of the same prefecture, whereas genotype GR-1-BBGK-19,72 is both genetically and geographically further apart (Figure 1A). The same holds true for genotype GR-1-BBGK-19,502 which is geographically separated coming from the lowest altitude. The bootstrap values seem to further validate this classification scheme. Despite the fact that evolutionary relationships may also be analyzed through the neighbor-joining tree, the key function applied herein was to repeatedly evaluate bootstrap values as a measure of informed distinction of the Greek native germplasm (see distinct clades). The results of this study outlined the ITS2 gene as an efficient tool (100%) regarding the distinction of the species in concern (*C. mas*) among other *Cornus* spp. (see monophyletic clusters).

**Figure 1.** (**A**) Phylogenetic tree on the basis of ITS2 regions regarding the Greek native Cornus mas genotypes in contrast with other C. mas and Cornus spp. genotypes of different origin retrieved from NCBI. (**B**) Overview of the genotypes analyzed in this study with multiple sequence alignment of their ITS2 barcode region. Results from neighbor-joining (NJ) bootstrap analyses with 1000 replicates were used to assess the strength of the nodes. The node numbers indicated the bootstrap value of NJ. The distinct genotypes of this study are highlighted with blue.

## *2.2. Phytochemical Evaluation of the Greek Native Cornelian Cherries*

The phytochemical screening of the fruits of Greek native *C. mas* samples revealed a significant level of variability between genotypes in terms of total phenolic content (TPC), vitamin C content, antioxidant activity (AA), and total flavonoids (TF) (Table 1, *p* < 0.05). No significant correlation with altitude was observed for any of the phytochemical parameters measured. Similarly, no significant correlation of the maturity index of the fruits with any of the phytochemical parameters measured was observed (Table 2, *p* < 0.05, *p* < 0.01). As expected, TPC was significantly positively correlated with TF (*p* < 0.01), while total dissolved solids (TDS) were significantly positively correlated with vitamin C content (Table 2, *p* < 0.05).


**Table 1.** Total phenolic content (TPC), antioxidant activity (AA), total flavonoids (TF), and vitamin C content assessed in fruit samples of wild-growing Greek genotypes of *Cornus mas*.

Values represent mean values ± standard deviation (S.D.) of samples analyzed in triplicate (*n* = 3); values within the same column that do not share the same letter are significantly different (Tukey post-hoc test, *p* < 0.05). For genotypes GR-1-BBGK-19,190, GR-1-BBGK-19,633, and GR-1-BBGK-19,638, capital letters A and B denote two consecutive years that fruits were measured.

**Table 2.** Pearson's correlation coefficients for the Greek *C. mas* genotypes (*n* = 14) between altitudes (m), four fruit nutraceutical properties assessed namely total phenolic content (TPC, mgGAE/100 g), antioxidant activity (AA, %RSA), total flavonoids (TF, mgCE/100 g), and vitamin C content (Vit C, mgAAE/100 g) as well as two complementary fruit phytochemical properties measured, i.e., maturity index (MI) expressed as the ratio of sugar content (◦Brix) to malic acid content (gMA/100 g) and total dissolved solids (TDS, mg/L).


Respective *p*-values are shown in parentheses. \*\* Correlation is significant at the 0.01 level (2-tailed). \* Correlation is significant at the 0.05 level (2-tailed).

Antioxidant activity (AA) among all assessed genotypes ranged from 55.46 %RSA (radical scavenging activity) in GR-1-BBGK-19,638A to 95.94 %RSA in GR-1-BBGK-19,844 with most genotypes, however, presenting AA above 80% (Table 1, *p* < 0.05). TPC ranged from 29.93 mg GAE/100 g in GR-1-BBGK-19,638A to 355.46 mg GAE/100 g in GR-1-BBGK-19,669. Vitamin C content ranged from 0.95 mg AAE/100 g in GR-1-BBGK-19,847 to 58.97 mg AAE/100 g in GR-1-BBGK-19,638B. Similarly, TF ranged from 0.11 mg CE/100 g in GR-1-BBGK-19,847 to 0.86 mg CE/100 g in GR-1-BBGK-19,669. Most of the genotypes showed very high AA (>80 %RSA) except genotypes GR-1-BBGK-19,638A, GR-1-BBGK-19,753, GR-1-BBGK-19,847, and GR-1-BBGK-19,848 (*p* < 0.05). The same four genotypes also presented the lowest TPC, as expected (Table 1, *p* < 0.05). Very high values of TPC were found in three cases, namely 304.73, 337.14, and 355.46 mg GAE/100 g in genotypes GR-1-BBGK-19,190A, GR-1-BBGK-19,638B, and GR-1-BBGK-19,669, respectively. Concerning vitamin C content, a messier picture emerged with four genotypes presenting very low

values of <1.5 mg AAE/100 g including GR-1-BBGK-19,753 and GR-1-BBGK-19,847, similar to AA and TPC results but also genotypes GR-1-BBGK-19,72 and GR-1-BBGK-19,590 which performed better in TPC and AA. All other genotypes showed vitamin C content values from 10 up to 30 times higher than 1.5 mg AAE/100 g (Table 1, *p* < 0.05). Concerning TF, genotypes GR-1-BBGK-19,847 and GR-1-BBGK-19,848 along with GR-1-BBGK-19,195 were among the poorest with 0.11, 0.17, and 0.15 mg CE/100 g, respectively (Table 1, *p* < 0.05).

In addition to the above results, significant year-to-year variations were observed in genotypes that were measured for two consecutive years, especially in TPC. Particularly, genotype GR-1-BBGK-19,190 showed a drop in TPC from 304.73 mg GAE/100 g in 2019 to 49.29 mg GAE/100 g in 2020 coupled with a drop in TF from 0.73 to 0.22 mg CE/100 g (Table 1, *p* < 0.05). On the contrary, genotypes GR-1-BBGK-19,633 and GR-1-BBGK-19,638 showed an increase in TPC between 2019 and 2020 from 52.31 to 195.2 mg GAE/100 g and from 29.93 to 337.14 mg GAE/100 g, respectively, with the latter showing also a significant increase in AA (Table 1, *p* < 0.05).

## *2.3. Propagation of Greek Native Cornus mas Genotypes with Cuttings*

A variation in rooting capacity was observed between different genotypes of *C. mas*. Rooting frequencies varied from 1.19% in genotype GR-1-BBGK-19,641 with 6000 ppm IBA/softwood cuttings in summer to 69.93% in genotype GR-1-BBGK-19,638 with 10,000 ppm IBA/softwood cuttings in late summer which was the highest rooting capacity observed (*p* < 0.05). The next best rooting capacity (58.33%) was observed in genotype GR-1-BBGK-19,753 with 4000 ppm IBA/softwood cuttings in early autumn (Table 3, *p* < 0.05). Hardwood, dormant cuttings during the winter presented generally low rooting rates which took months to be reached since cuttings remained under mist throughout winter and started to root in the following spring. The highest rooting frequency observed within this group was 20.93% (*p* < 0.05) in genotype GR-1-BBGK-19,198. For the rest of the genotypes that were tested via softwood non-dormant cuttings during the summer through early autumn, rooting was still diverse, but more cases of higher rates were found compared to genotypes tested in winter, with three and two genotypes presenting >30% and 50% rooting, respectively (Table 3, *p* < 0.05, Figure 2).

**Table 3.** Overview of the propagation results achieved in terms of highest rooting frequencies after experimental setups in different seasons (winter 2018, spring–late summer 2019, and early–late autumn 2019) with different types of initial material (softwood or hardwood cuttings) collected directly from wild-growing Greek native populations of *Cornus mas* (18 genotypes).


The symbols † and ‡ denote the highest rooting frequencies for hardwood and softwood cuttings, respectively, following pairwise comparisons of the observed rooting frequencies via Pearson X2 tests. All cases were tested against a control treatment with no hormone application and no rooting.

**Figure 2.** Softwood propagation material collected from representative wild-growing Greek native *C. mas* genotypes (**A1**: GR-1-BBGK-19,72 and **A2**: GR-1-BBGK-19,632) during botanical expeditions. Cutting preparation for genotype GR-1-BBGK-19,502 (**B1**) and GR-1-BBGK-19,72 (**B2**). Representative rooted cutting of GR-1-BBGK-19,72 (**C1**) and cutting that failed to root of GR-1-BBGK-19,72 (**C2**). Transplanted well-rooted plants under ex situ conservation of genotype GR-1-BBGK-19,72 (**D1**) and GR-1-BBGK-19,753 (**D2**). Bars in photos B1, B2, C1, and C2 represent 10 cm.

## *2.4. Multifaceted Evaluation of Greek Native Cornus mas Genotypes*

An overview of the sustainable exploitation potential of the native *C. mas* genotypes assessed in this study is summarized in Table 4 in a multifaceted way based on the obtained molecular authentication results, the propagation success achieved, and the phytochemical profiles assessed. From the effectively authenticated genotypes, GR-1-BBGK-19,72 showed very high antioxidant activity and average propagation potential (Table 4, Figure 2) and could be prioritized for further research. However, genotypes GR-1-BBGK-19,190 and GR-1-BBGK-19,195 also showed high antioxidant activity but it is unclear from the current data whether the low propagation success observed was due to genotype effect or due to winter hardwood cuttings since these genotypes were only tested during the winter in contrast with GR-1-BBGK-19,72 (Table 1, Table 3, Table 5). As far as the rest of the genotypes are concerned, genotype GR-1-BBGK-19,638 stands out with very promising phytochemical potential and high propagation success and, as such, this genotype merits further investigation. Another genotype that showed a promising potential is GR-1-BBGK-19,753 with high propagation success and strong antioxidant activity (Table 3, Figure 2).

**Table 4.** Multifaceted evaluation of Greek native *Cornus mas* genotypes based on molecular authentication effectiveness (Figure 1), fruit phytochemical potential expressed as antioxidant activity (AA, %RSA: low ≤50, average 51–70, high 71–90, very high >90), total phenolic content (TPC, mg GAE/100 g: low ≤50, average 51–100, high 101–200, very high >200) and vitamin C content (mg AAE/100 g: low ≤30, average 31–50, high >50) (Table 1), and asexual propagation potential expressed as hormone-induced rooting under the most successful application (very low < 10%, low 11–30%, average 31–55%, or high >55%), (Table 3).


**Table 5.** IPEN accession number, location, altitude, and sampling details of the genotypes (population samples) of Greek native *Cornus mas* germplasm collected from various areas and habitats of Northern Greece in 2018–2019.



**Table 5.** *Cont.*

RFS: Ripe cornelian cherry fruits sample for chemical analysis; HWSC: hardwood stem cuttings; SWSC: softwood stem cuttings for propagation; LS: leaf samples for DNA analysis

## **3. Discussion**

## *3.1. Molecular Authentication of Greek Native Genotypes of Cornus mas*

In general, DNA barcoding may offer support to classical morphological identification of samples and insight into phylogenetic relationships of closely related species; thus it is an efficient method for the discernment of various genotypes independently from the stage of plant development [6,10]. Herein, we provide the first-ever report regarding the molecular authentication of Greek native germplasm of *Cornus mas*, with the NJ (Neighbor-Joining) tree classification obtained from the ITS2 barcoding discerning clearly the DNA fingerprints of the Greek genotypes. Undoubtedly, more genotypes from different habitats in different regions of Greece should be evaluated to further confirm this distinctiveness. Regardless, as reported herein, the ITS2 gene can be an effective marker for the identification of *Cornus* spp. and of different genotypes thereof; thus, offering to deciphering their evolutionary relationships.

## *3.2. Nutraceutical Potential of Greek Native Genotypes of Cornus mas*

The current work presents for the first time comprehensive phytochemical data for an extended range of Greek native wild-growing *Cornus mas* genotypes assessing their potential as a functional food or superfood for ex situ conservation and future breeding efforts. The parameters assessed herein revolve around human oxidative stress and particularly plant secondary metabolites with known free radical scavenging activity such as phenolic compounds and flavonoids coupled with vitamin C content [41]. The results showed that at least five Greek genotypes with significant nutraceutical potential can be distinguished, namely GR-1-BBGK-19,72, GR-1-BBGK-19,190, GR-1-BBGK-19,195, GR-1-BBGK-19,638, and GR-1-BBGK-19,753 (Table 4). Similarly to the current results, a wide range of TPC of the fruits was detected in studies dealing with wild-growing *C. mas* germplasm in Romania, coupled with significant variability among genotypes in other fruit phytochemical traits [42,43]. Some similarities can also be found between *C. mas* wild-growing genotypes of Serbia with those studied herein in terms of vitamin C content that also show higher sugar content due to higher fruit maturity index at the time of collection [15].

Although variable among genotypes, the phytochemical properties of the Greek germplasm are generally similar with both wild-growing and cultivated *C. mas* genotypes reported from countries at a similar latitudinal range as Greece such as Turkey [13]. However, studies on Central European genotypes of *C. mas* cultivated in higher latitudes have shown higher antioxidant activity (AA) of the fruits that ranged from 3.30 to 9.54 g AAE/1000 g, which implies an interaction of genotype with climatic conditions affecting fruit phytochemical profile [14]. Evidence regarding the effects of the climatic conditions in interaction with genotype on fruit phytochemical profile can also be seen through the significant correlation of total dissolved solids (TDS) with vitamin C content observed among genotypes herein, since TDS is physiologically connected to climatic factors [14]. In a more recent study concerning an Italian cultivated genotype of *C. mas*, total phenolic content (TPC) was found at 196.68 mgGAE/100 g [44], which is similar to some genotypes assessed herein and even lower than three of the studied genotypes, while vitamin C content was at similar ranges. Recently, the phytochemical profile of fruits was comparatively evaluated between cultivars of *C. mas*, genotypes of the Chinese endemic *Cornus officinalis* Siebold and Zucc., and hybrids thereof (*C. mas* X *C. officinalis*). *C. mas* cultivars in that study have shown a similar TPC range with the studied Greek genotypes apart from one cultivar which was higher, yet the TPC and AA of the hybrids was generally higher, indicating the significance of conservation of *C. mas* germplasm stocks for breeding programs [22]. The nutraceutical potential of native genotypes has been similarly assessed on other Greek germplasm resources such as, for example, *Rosa canina* [10] with slightly higher potential of AA and TPC detected and comparable to some extent with that of *C. mas*, but with very high content of vitamin C for *R. canina* [10].

The climatic conditions (most prominently temperature along with sunshine) during the last stages of fleshy fruit ripening in cornelian cherries may affect acid reduction and sugar increase, which in turn affects biosynthesis of secondary metabolites [45,46]. As such, the climatic conditions of the natural habitats of wild populations on one hand and the stage of fruit collection and fruits' ripening status on the other hand are considered as significant factors in terms of comparative phytochemical evaluations. However, in this study no significant correlations were observed between the fruit phytochemical properties measured and the altitude from which they were collected in the wild or their maturity index determined as the ratio of sugar content to malic acid content. This observation indicates that the variability of phytochemical properties observed among Greek native genotypes of *C. mas* may be attributed to genetic factors rather than timing of fruit harvest or altitudinal differences between collection sites. Nevertheless, further ecological profiling work in terms of correlation studies with more environmental/topoclimatic parameters and more fruit phenological traits on Greek *C. mas* wild populations is suggested.

## *3.3. Propagation Potential of Greek Native Genotypes of Cornus mas*

A further component of the multifaceted evaluation of wild Greek *C. mas* genotypes investigated herein is the asexual propagation via cuttings which was conducted for the first time in Greek germplasm. Similarly to fruit phytochemical data, the genotypes of *C. mas* tested showed variable rooting capacity with a number of them being below the commercially accepted threshold [47]. In all cases, however, external hormone application was necessary to achieve even low rooting frequencies which are in agreement with similar studies [38,39,48]. Hardwood dormant cuttings tested in the current study remained under mist for the duration of winter and started to root the following spring giving low rooting rates (<25%) under 10,000 ppm IBA treatment. This observation concurs with similar studies on cutting propagation of *C. mas* genotypes in other countries where softwood cuttings surpassed hardwood cuttings under the 10,000 ppm IBA treatment [38]. The poor results on hardwood cuttings indicate that it is not economically efficient to produce new plants via winter dormant cuttings [37,40], a fact that also applies for the Greek germplasm. Softwood cutting material with a small degree of lignification has known advantages in rooting quality with external hormone utilization [49], producing higher rooting rates in *Cornus alba* [50]. Even in other genera, cutting type affected by the status of mother's plant growth has been linked to rooting capacity and quality [51–53]. Consequently, since only two of the selected genotypes evaluated herein showed appreciable propagation potential (>50% rooting in genotypes GR-1-BBGK-19,638 and GR-1-BBGK-19,753) with softwood cuttings (Table 3), further research is required to explore pathways on improving rooting capacity but also to investigate the relationships between cutting type and rooting capacity and quality in more detail. Additionally, further research is proposed on the correlation of rooting factors with early plant establishment and survival in *C. mas* propagation since such a correlation has been shown to be significant in other Greek native germplasm resources such as *Sambucus nigra* [6].

Finally, asexual propagation through grafting has been attempted on European *C. mas* germplasm with positive results [54], showing that successful grafting in *C. mas* can enhance growth and development of the scion [55]. For the Greek native *C. mas* genotypes studied herein, the method of grafting a genotype with high phytochemical potential but low rooting capacity onto another genotype with improved rooting capacity may provide a good option in dealing with difficult-to-root genotypes but with valuable fruit traits. However, further research is required to systematically test this strategy.

## **4. Materials and Methods**

## *4.1. Collection and Documentation of Plant Material*

A plethora of botanical expeditions took place covering a wide range of geographically distant and altitudinally different natural habitats of *C. mas* across Northern Greece reached at consecutive stages across two years (Table 5). Targeted and variable types of plant material were collected (Table 5): (i) leaves from 20 individuals from each genotype for DNA analysis; (ii) hardwood dormant twigs for propagation from six genotypes (winter 2018); (iii) softwood stem cuttings for propagation from 14 genotypes (14 in spring–late summer and 5 in autumn of 2019); (iv) ripe fruits from nine genotypes for phytochemical evaluation (autumn 2019). The work resulted in the documentation of 18 *C. mas* genotypes (population samples) in total from healthy wild-growing individuals (Figure 3). After each expedition, the collected material was promptly transferred to the laboratory where it was taxonomically identified [56] and was assigned with a unique IPEN (International Plant Exchange Network) accession number given by the Balkan Botanic Garden of Kroussia of the Institute of Plant Breeding and Genetic Resources (IPB&GR) of the Hellenic Agricultural Organization-Demeter (ELGO-Demeter). Plant material collection in all cases was conducted under a special permit to the IPB&GR, ELGO-Demeter (Permit 82336/879 of 18/5/2019 and 26895/1527 of 21/4/2021) issued by the Greek Ministry of Environment and Energy. The overall work was conducted under the auspices of the research project "Highlighting local traditional varieties and wild native forest fruit trees and shrubs" (acronym: EcoVariety, T1EΔK-05434).

## *4.2. DNA Isolation*

A standardized commercial DNA extraction kit (Nucleospin Plant II, Macherey-Nagel) was used for DNA extraction following the manufacturer's instructions using approximately 30 mg of dried leaf sample from each *C. mas* genotype that was previously ground in liquid nitrogen.

(**A**)

(**B**) (**C**)

**Figure 3.** (**A**) Geographical distribution of the *Cornus mas* Greek native genotypes sampled for taxonomic identification, DNA barcoding, phytochemical assessment, and asexual propagation trials. (**B**) Typical habit of wild-growing individuals; morphology of flowers (**C**) leaves and fruits (**D**) in early (**E**) and full ripening (**F**).

(**D**) (**E**) (**F**)

## *4.3. Polymerase Chain Reaction (Pcr) Amplification and Sequence Analysis*

PCR amplification was conducted according to [57] using one primer set of the nuclear ITS2 barcode region [58]. The resulting PCR products for each genotype were consecutively sequenced using an automated ABI 3730 sequencer (PE Applied Biosystems), in two directions of each fragment with a Big Dye terminator v3.1 Cycle sequencing kit (PE Applied Biosystems, Foster City, CA, USA). The sequences were aligned using the CLUSTAL W program.

## *4.4. Molecular Data Analysis*

Following the results of the sequence analysis, the molecular authentication of the *Cornus mas* genotypes studied was conducted using three methods: (i) evaluation against the nucleotide database at NCBI using the Basic Local Alignment Search Tool (BLAST); (ii) using maximum-likelihood models for the genetic divergence method; and (iii) tree topology analysis using the Neighbor-Joining (NJ) method based on different loci in MEGAX with the K2P distance model and 1000 bootstrap replications. The generated ITS2 DNA barcoding sequences (without primers used for PCR amplification) were deposited to the NCBI-Genbank (https://www.ncbi.nlm.nih.gov/BankIt/, accessed 5 May 2022) under the accession numbers MZ35480-MZ35488.

## *4.5. Phylogenetic Relationships*

The phylogenetic relationship of *Cornus* spp. was inferred using the Neighbor-Joining (NJ) method [59]. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches [60]. The evolutionary distances were estimated using the Maximum Composite Likelihood (MCL) method [61]. All ambiguous positions were removed for each sequence pair (pairwise deletion option). The phylogenetic dendrogram was constructed using the MEGA X software [62].

## *4.6. Propagation Trials*

Depending on the propagation material obtained from each expedition, a variety of targeted experiments were conducted using different external hormone application treatments of indole-3-butyric acid (IBA) [50,63] and different cutting types taken at different seasons from varying stages of mother plant growth (Table 5). Cuttings were set for rooting in propagation trays under mist where relative humidity (RH) was maintained at >85% within a greenhouse at ambient temperature. The substrate used was peat (Klasmann, KTS 1):perlite at 1:3 *v/v*. Cuttings were attended weekly to assess their rooting capacity. The produced mother plants were kept ex situ at the laboratory's nursery in the grounds of IPB and GR under ambient conditions for plant adaptation. The plants were watered regularly and were grown in 3 L pots using a mixture of peat (Klasmann, KTS 2) and perlite (3:1, *v/v*).

## *4.7. Phytochemical Analysis of Cornus mas Fruits*

A modified method described in [64] was used for preparation of the extracts. An appropriate volume of MeOH/H2O (60:40) was mixed with 2–5 g of homogenized sample and was centrifuged at 4000rpm at 4◦C. The collected supernatant was made up to 20 mL volume and was used for the determination of total phenolic content (TPC), total flavonoids (TF), and antioxidant activity (AA). For the determination of TPC, phenolic extract 0.20 mL along with 2.3 mL of H2O and 0.25 mL of Folin–Ciocalteu reagent were added in a volumetric flask [64] in which, after 3 min, 0.50 mL of 20% Na2CO3 was added, and the volume was made up to 5 mL. The above solution was stored in a dark place for 2 h and the absorbance was measured at 725 nm against blank solution. The calculation of TPC was conducted using a standard curve of gallic acid at various concentrations giving the results expressed as gallic acid equivalents (GAE)/100 g of sample. All analyses were carried out in triplicate.

A modified method for the determination of TF was used according to [65]. In a test tube, 5 μL of the above extract were added along with 3270 μL of H2O and 75 μL of 5% NaNO2, stirred, and stored in the dark for 5 min which was then added with 150 μL of 10% AlCl3-6H2O, mixed, and stored in the dark for a further 6 min. Consequently, it was added with 500 μL of 1 M NaOH and the absorbance was measured at 510 nm against H2O as blank. The calculation of TF was conducted using a standard curve of catechin at various concentrations giving results expressed as catechin equivalents (CE)/100 g of sample. All analyses were carried out in triplicate.

A modified method for the determination of AA was employed described in [66]. In a 5 mL plastic cuvette, 0.1 mL of phenolic extract was added along with 2.9 mL 0.10 mM (2,2-diphenyl-1-picrylhydrazyl (DPPH)) in MeOH, stored in the dark for 15 min and then the absorbance was measured at 517 nm against MeOH as blank. The control sample was prepared using 0.10 mM DPPH in MeOH. The percentage of radical scavenging activity (%RSA) was carried out in triplicate and was calculated using the following equation:

$$\% \text{RSA} = \frac{A\_{\mathcal{o}} - A\_{\mathcal{s}}}{A\_{\mathcal{o}}} \ast 100$$

where

*Ao* = absorbance of control sample.

*As* = absorbance of the sample after 15 min of incubation.

A modified method for the determination of vitamin C presented in [67] was used. A mixture of homogenized sample (2–5 g) and 5 mL of 4.5% *w/v* metaphosphoric acid (MPA) solution was stirred and centrifuged at 8000 rpm at 4◦C for 20 min. The supernatant (1 mL) was taken and diluted up to 10 mL with 4.5% MPA solution and it was filtered through 0.45 μm polyethersulfone filters. The vial was covered with aluminum foil to prevent oxidation of ascorbic acid and stored at 4 ◦C until HPLC-DAD analysis. HPLC-DAD analysis conditions: Column (Agilent Eclipse XDB-C18) 4.6 mm × 150 mm, 5μm, elution solvent: aqueous 0.005 M H2SO4 solution at a flow rate of 0.5 mL/min (isocratic) and wavelength 245 nm. A standard curve of ascorbic acid at various concentrations was used for vitamin C content calculation which gave results expressed as ascorbic acid equivalents (AAE)/100 g of sample.

Sugar content of fruit samples was measured via Brix analysis (oBx), and the maturity index (MI) of fruits was calculated as the ratio of sugar content in ◦Bx to malic acid (MA) content expressed as grams of MA per 100 g.

All analyses were carried out in triplicate.

## *4.8. Experimental Design and Statistical Analysis of Propagation and Phytochemical Data*

The propagation trials followed a complete randomized design. The number of replicate cuttings per treatment varied according to the volume of material obtained from each expedition (Table 5). Targeted IBA hormone application treatments were applied ranging from 2000 to 10,000 ppm depending on the season and whether the cuttings were softwood or hardwood, respectively (Table 5). Each trial included a control with an equal number of replicate cuttings. The rooting frequencies obtained were compared through pairwise Pearson Chi-Square tests.

For the phytochemical data that were measured in triplicate, a mean coupled with its standard deviation (±S.D.) was calculated in each case and means were compared using Tukey's HSD post hoc test. In addition, the phytochemical data were subjected to Pearson's correlation coefficient analysis to determine the correlations between the different fruit phytochemical traits measured along with the differences in altitude between the natural habitats of *C. mas* that collection occurred in. All analyses were conducted using the IBM-SPSS 23.0 software.

## **5. Conclusions**

In conclusion, the multifaceted evaluation that was conducted provides a first depiction of the sustainable exploitation potential of Greek native *Cornus mas* genotypes. All genotypes collected from the wild are under ex situ conservation at the IPB and GR for future monitoring, since juvenile–mature correlations are known in perennial crops from the literature [68,69]; such aspects should be taken into account for future long-term breeding programs. The transition from the juvenile to the mature (producing) stage followed by several phenological and developmental changes was found to vary among genotypes mainly in forest species but also in cultivated deciduous tree species with stone fruits such as apples [69]. Consequently, further work on molecular authentication, ecological profiling, and phytochemical profiling is proposed coupled with further research on propagation. In addition, the long-term monitoring of the current Greek genotypes is needed, and in fact, pilot cultivation trials of the genotypes prioritized herein have been established (ongoing process). The data obtained in the current study along with the conservation of the plant material collected can serve as a basis for future breeding efforts, creating for the first time a framework for the sustainable utilization of Greek native *C. mas* germplasm as a superfood with significant agronomic potential.

**Author Contributions:** Conceptualization, N.K., E.K. and I.G.; methodology, E.M., E.K., I.G., A.K., K.P., I.M., D.K., P.Y., N.N., I.S.K., A.V.B., G.P., I.G. and N.K.; software, I.G., and D.F.; validation, E.K., I.G., A.K., K.P., I.M., D.K., P.Y., N.N., I.S.K., A.V.B., G.P. and D.F.; formal analysis, E.K., I.G., I.S.K. and A.V.B.; investigation, E.K., I.G., A.K., K.P., I.M., D.K., P.Y., N.N., I.S.K., A.V.B. and G.P.; resources, E.M., I.G. and A.V.B.; data curation, N.K., E.K., K.P., I.G., N.N., I.S.K. and A.V.B.; writing—original draft preparation, E.K., I.G., D.F., A.V.B. and N.K.; writing—review and editing, E.M., E.K., I.G., A.K., K.P., D.K., P.Y., N.N., I.S.K., A.V.B., I.M., G.P., D.F. and N.K.; visualization, N.K., E.K., I.G. and D.F.; supervision, E.M., A.V.B., G.P. and N.K.; project administration, E.M. and G.P.; funding acquisition, E.M., N.N. and G.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was co-financed by the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH— CREATE—INNOVATE (project code: T1EDK-05434), entitled "Highlighting of local traditional and native wild fruit trees and shrubs".

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** All data supporting the results of this study are included in the manuscript and datasets are available upon request.

**Acknowledgments:** The authors would like to thank the staff of the Institute of Plant Breeding and Genetic Resources, Hellenic Agricultural Organization-Demeter for administrative and technical support.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **Abbreviations**

All abbreviations are explained at their first mention in the manuscript.

## **References**


## *Article* **Hydraulic Response of Deciduous and Evergreen Broadleaved Shrubs, Grown on Olympus Mountain in Greece, to Vapour Pressure Deficit**

**Maria Karatassiou 1,\*, Panagiota Karaiskou 1, Eleni Verykouki <sup>2</sup> and Sophia Rhizopoulou <sup>3</sup>**


**Abstract:** In this study, leaf hydraulic functionality of co-occurring evergreen and deciduous shrubs, grown on Olympus Mountain, has been compared. Four evergreen species (*Arbutus andrachne*, *Arbutus unedo*, *Quercus ilex* and *Quercus coccifera*) and four deciduous species (*Carpinus betulus*, *Cercis siliquastrum*, *Coronilla emeroides* and *Pistacia terebinthus*) were selected for this study. Predawn and midday leaf water potential, transpiration, stomatal conductance, leaf temperature and leaf hydraulic conductance were estimated during the summer period. The results demonstrate different hydraulic tactics between the deciduous and evergreen shrubs. Higher hydraulic conductance and lower stomatal conductance were obtained in deciduous plants compared to the evergreens. Additionally, positive correlations were detected between water potential and transpiration in the deciduous shrubs. The seasonal leaf hydraulic conductance declined in both deciduous and evergreens under conditions of elevated vapor pressure deficit during the summer; however, at midday, leaf water potential reached comparable low values, but the deciduous shrubs exhibited higher hydraulic conductance compared to the evergreens. It seems likely that hydraulic traits of the coexisting evergreen and deciduous plants indicate water spending and saving tactics, respectively; this may also represent a limit to drought tolerance of these species grown in a natural environment, which is expected to be affected by global warming.

**Keywords:** water potential; stomatal conductance; transpiration; leaf hydraulic conductance; drought

## **1. Introduction**

The climate in the Mediterranean region is characterized by a prolonged summer drought period, which according to the majority of climate change scenarios, is expected to become more severe in the future [1–3]. Drought is the main environmental stress directly linked to plant survival, growth, competitiveness, persistence, and productivity in Mediterranean habitats that are under an increasing risk of degradation [1,4–7]. On the other hand, plants have evolved a variety of morphological and physiological tactics [8] to withstand drought stress [9,10]; these include (a) hydraulic strategies via deep, tap root systems to extract water from deep soil layers, enabling the plants to sustain elevated water potential and xylem pressure during drought and (b) morphological and physiological adaptations at the leaf level in order to reduce water loss [11,12]. The importance of the hydraulic system in water consumption has led to the hypothesis of the functional convergence in the regulation of water use among phylogenetically diverse species [13,14]. Nonetheless, few studies have been performed to characterize the differences in leaf hydraulic conductance and hydraulic status among phylogenetically different plant species,

**Citation:** Karatassiou, M.; Karaiskou, P.; Verykouki, E.; Rhizopoulou, S. Hydraulic Response of Deciduous and Evergreen Broadleaved Shrubs, Grown on Olympus Mountain in Greece, to Vapour Pressure Deficit. *Plants* **2022**, *11*, 1013. https:// doi.org/10.3390/plants11081013

Academic Editor: Daniela Trono

Received: 5 March 2022 Accepted: 5 April 2022 Published: 8 April 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/).

such as deciduous and evergreen shrubs, in Mediterranean habitats [15,16]. Leaf is a substantial organ for the transport of water throughout the plant and hence leaf hydraulic conductance is an important parameter to determine plant water status [15,17–20]. To the best of our knowledge, a comparative study among hydraulic characteristics at the leaf level of evergreen and deciduous species co-occurring in the same habitat, with respect to the impact of climatic stimuli on functionality, has not hitherto been published.

In order to comprehensively address differences between evergreen and deciduous plants concerning physiological responses to drought, the hydraulic and stomatal performance should be examined with well-established approaches. The water status for each individual, specific plant depends on the difference between transpiration (E) and water absorption (A), but it is not clear which of these factors is more important for plants' gas exchanges in response to drought stress [12]. If the soil water shortage increases, water stress will increase over time, negatively affecting many physiological and metabolic aspects of the plants [8,21,22]. Plants close the stomatal pores to regulate water loss [12,23] through transpiration when the available soil water decreases and/or an increase in the difference between the saturation (i.e., amount of water vapor that the air can hold, namely the saturation vapor pressure) and actual vapor pressure, i.e., the Vapour Pressure Deficit (VPD), is detected.

VPD has been recognized as an important parameter for plant functionality and survival, which is influenced by the so-called hydraulic failure [24–29]. Oren et al. [30] fount in drought tolerant species a regulation of water potential (Ψ) as VPD increases and recommended different sensitivity of stomatal apparatus to VPD among different functional groups. However, E could either increase or decrease in response to an increased VPD [27,31]; the first response is known as "feedback" response while the second as "feed forward" response. The transpiration rate is influenced by atmospheric conditions and, over a short time-scale, is regulated by the function of stomatal apparatus [32]. It has been argued that a declining stomatal conductance (gs) concomitantly with increasing VPD would rather occur as a feedback response to E and water loss from the leaf, than as a direct response to humidity [33–35].

The function of the stomatal apparatus under global climate change is an essential subject in plant ecophysiological research because it affects plant growth, vegetation distribution and ecosystem function [36]. The stomatal response to VPD varies either across and within species, or within the same species [37–40]. Partial stomatal closure under elevated VPD, especially during midday, will negatively affect CO2 assimilation rate [7,41].

The movement of water in the soil–plant–atmosphere (SPA) continuum is a function of the difference between hydrostatic pressure and hydraulic conductance along this pathway. Water flows through the SPA continuum driven by a gradient in Ψ, which depends on the water flow rate and the hydraulic conductivity of the different pathways [32]. The amount of water that will be absorbed by the roots of the plants and the amount that will move from the roots to the leaves and then to the atmosphere depends mainly on gs, VPD and hydraulic conductance (K). The differences between plant species in Ψ, E and gs, determine to a greater or lesser extent the plants' response to various water stresses [12,42–45]. However, the effects of rising VPD on vegetation and hydraulic dynamics remain poorly studied. It has been reported that significant higher leaf and stem hydraulic conductance [25,46,47] under increased irradiance are due to the regulation of aquaporins [46–48]. A midday decline of gs has been obtained in many plant species and has been related with variation in midday stem water status [49,50]; this aspect supports the idea that stomatal response to VPD is substantially related to the hydraulic characteristics at both the whole plant scale and the leaf level [27,51]. There is a fundamental role of hydraulics on stomatal sensitivity to VPD [27,41], while inverse and non-linear relationship between conductance and VPD have been observed [36]. The opening and closing of the stomata seem to be controlled by complex mechanisms, which include chemical and hydraulic signals from roots, to shoots and leaves [52]. In this frame, hydraulic traits could be an important factor buffering the negative impact of drought on plant function [53].

A lot of research has been devoted to understanding how plants' hydraulic systems have evolved to accommodate survival under different environments. However, concerning old-grown, trees and shrubs the relationship between K and the response of gs to VPD has not explicitly evaluated, in situ. Despite the above-mentioned trends, species grown under the same environmental conditions may exhibit entirely different hydraulic properties [54,55]. This interspecific variation sometimes may be ascribed to different functional types, such as deciduous and evergreen [56]. Nevertheless, variation of hydraulic traits cannot only be explained by categorization of species in functional types [56,57]. Forest and shrub communities' response to climate change are most closely related to microclimatic change and not to macroclimatic change [58,59].

In considering that the water balance (W) is expressed by the relationship W=A − E, it is very interesting to study whether: (1) variation of hydraulic conductance in Mediterranean shrubs during the dry summer period is reflected in the leaf phenology (i.e., evergreen vs. deciduous), and (2) changes of VPD and consequently of microclimatic conditions influence the physiological mechanism and the performance of evergreen vs. deciduous shrubs.

The main objective of this study was to compare the ability of co-existing deciduous and evergreen broadleaved shrubs grown on the Olympus Mountain to maintain their water status and control their stomatal conductance throughout the dry season, as well as to evaluate the response of deciduous and evergreen shrubs to enhanced vapor pressure deficit. It is likely that hydraulic responses of co-occurring deciduous and evergreen shrubs, which are to some extent linked to water exploitation and effectiveness of plants' life forms during a period of soil drying, have not been reported.

## **2. Results**

## *2.1. Climatic Conditions*

The seasonal pattern of VPD during the experimental period is given in Figure 1. Overall, predawn vapour pressure deficit (VPDp) was significantly lower (*p* < 0.05) than midday vapour pressure deficit (VPDm), except from the first measurement (mid-May) (*p* ≥ 0.05). Predawn VPDp ranged from 0.769 ± 0.019 kPa to 1.731 ± 0.054 kPa, while VPDm ranged from 0.974 ± 0.044 kPa to 4.043 ± 0.106 kPa. The relative humidity (RH) followed a reverse course relatively to VPD during the experimental period (Figure 1). The Photosynthetic Photon Flux Density (PPFD) was maintained at a relatively high level (i.e., >1500 μmol m−<sup>2</sup> s−1) (Figure 2) from the end of May to the end of August; only during the first measurement (end of spring) PPFD was relatively low, approximately 800 μmol m−<sup>2</sup> s<sup>−</sup>1.

Data analysis revealed that only the date was a significant predictor for leaf temperature (Tl) (*p* < 0.001), whereas the shrub life form (deciduous, evergreen) and the interaction between date and shrub life form were not significant (*p* ≥ 0.05). Leaf temperature (Tl) during the study period almost coincided with the midday ambient air temperature (Tm), ranging from 19.86 ◦C to 33.04 ◦C and 19.29 ◦C to 33.61 ◦C, respectively (Figure 2). Predawn air temperature (Tp) ranged from 17.80 ◦C to 23.21 ◦C. The differences between Tp and Tm were statistically significant (*p* < 0.05) throughout the experimental period, except from the first measurement (mid-May).

**Figure 1.** Seasonal predawn (VPDp) and midday (VPDm) vapour pressure deficit and relative humidity (RH) in the study site, during the experimental period, from 27 May to 28 August. The values are means ± SE (*n* = 24).

**Figure 2.** Seasonal Photosynthetic Photon Flux Density (PPFD), leaf temperature (Tl), morning (Tp) and midday (Tm) ambient air temperature during the experimental period from 27 May to 28 August. The values are means ± SE (nTp,Tm = 24, nTl = 48).

## *2.2. Physiological Parameters*

The principal component analysis (PCA, Figure 3) of the physiological parameters showed that the eight species are classified into the groups they belong to, deciduous and evergreen, and characterize the first principal component (Axis x), i.e., the species *Coronilla emeroides* Boiss. and Spruner, *Carpinus betulus* L., *Pistasia terebinthus* L. and *Cercis siliquastrum* L. are deciduous, while the species *Arbutus unedo* L., *Arbutus adrachnae* L., *Quercus coccifera* L. and *Quercus ilex* L. are evergreens. Additionally, variables such as predawn leaf water potential (Ψp), gs and leaf hydraulic conductance (KLeaf) are negatively correlated with VPDp, VPDm, the Julian date of the measurements and the second principal component (Axis y). Midday leaf water potential (Ψm) characterizes both components, while E does not characterize any of the two components. The deciduous species presented higher values (less negative) of Ψp, Ψm, gs and KLeaf relatively to evergreen species (Table 1).

**Figure 3.** Principal component analysis (PCA) loading plot representing the variables and the species characterizing the two components. Black dots indicated predawn (Ψp) and midday (Ψm) leaf water potential, midday transpiration rate (E), midday stomatal conductance (gs), midday leaf hydraulic conductance (KLeaf) and midday vapour pressure deficit (VPDm); brown squares indicated the deciduous species: *Coronilla emeroides* (*C.e*.), *Carpinus betulus* (*C.b*.), *Pistasia terebinthus* (*C.t.*). *Cercis siliquastrum* (*C.s.*), while green triangles indicated the evergreen species: *Arbutus unedo* (*A.u*.), *Arbutus adrachnae* (*A.a.*), *Quercus coccifera* (*Q.c*.) and *Quercus ilex* (*Q.i.*).


**Table 1.** Mean values of predawn (Ψp) and midday (Ψm) leaf water potential, midday transpiration rate (E), midday stomatal conductance (gs), midday leaf hydraulic conductance (KLeaf) and midday vapour pressure deficit (VPDm) in deciduous and evergreen shrubs during the study period.

\* Significant for *p* < 0.05, \*\* significant for *p* < 0.001, NS—no significance. Mean values were compared with independent samples Student's *t*-test.

## *2.3. Leaf Water Potential*

On one hand, data analysis revealed that the date was a significant predictor for Ψ<sup>p</sup> and Ψ<sup>m</sup> (*p* = 0.021 and *p* < 0.001, respectively). On the other hand, the shrub life form (deciduous, evergreen) significantly affected the Ψ<sup>m</sup> (*p* < 0.001). The interaction between date and shrub life form was not significant concerning the two variables (*p* ≥ 0.05) (Figure 4a). The estimated Ψ<sup>m</sup> values were significantly lower in the considered evergreen shrubs in comparison to the deciduous shrubs during the experimental period (*p* < 0.001), (Figure 4b). Ψ<sup>p</sup> ranged in deciduous shrubs from –0.87 MPa to −1.21 Mpa and in the evergreens from −0.92 Mpa to −1.14 Mpa, while Ψ<sup>m</sup> from −1.4 Mpa to −2.59 Mpa and from −1.88 Mpa to −2.86 Mpa, respectively. The average values of the studied parameters are presented in Table 1.

**Figure 4.** Seasonal (**a**) predawn (Ψp) and (**b**) midday (Ψm) leaf water potential of deciduous and evergreen shrubs during the experimental period from 27 May to 28 August. The values are means ± SE (*n* = 24).

## *2.4. Transpiration Rate, Stomatal and Leaf Hydraulic Conductance*

Data analysis revealed that the date was a significant (*p* < 0.001) predictor for E, gs and KLeaf. The shrub life form (deciduous, evergreen) significantly affected variable KLeaf (*p* < 0.001). The interaction between date and shrub life form was significant only for gs (*p* < 0.05). The seasonal pattern of E (at solar noon) of evergreen and deciduous shrubs did not fluctuate substantially during the experimental period and significant difference was not observed in E between deciduous and evergreen shrubs (*p* > 0.05). The average values of E were found 16.12 ± 0.89 and 15.87 ± 0.68 mmol m−<sup>2</sup> <sup>s</sup>−<sup>1</sup> for deciduous and evergreen shrubs, respectively (Table 1). In particular during the dry season, E ranged from 11.71 to 20.29 mmol m−<sup>2</sup> s−<sup>1</sup> in the deciduous and from 13.04 to 17.02 mmol m−<sup>2</sup> s−<sup>1</sup> in the evergreen shrubs (Figure 5).

**Figure 5.** Seasonal transpiration rate (E) of deciduous and evergreen shrubs during the experimental period, from 27 May to 28 August. The values are means ± SE (*n* = 24).

Seasonal pattern of gs was different (*p* < 0.05) between the two groups of plants. The deciduous shrubs presented significantly lower values (*p* < 0.05) of gs in relation to the evergreens during dry period (from July until mid-August). The gs during dry period ranged from 416.5± 22.7 to 585.25 ± 42.5 mmol m−<sup>2</sup> <sup>s</sup>−<sup>1</sup> in deciduous and from 413.5 ± 27.4 to 644.75 ± 43.5 mmol m−<sup>2</sup> <sup>s</sup>−<sup>1</sup> in evergreen shrubs (Figure 6).

**Figure 6.** Seasonal stomatal conductance (gs) of deciduous and evergreen shrubs during the experimental period, from 27 May to 28 August. The values are means ± SE (*n* = 24).

The KLeaf was significantly higher in deciduous compared to evergreen shrubs (*p* < 0.001), (Figure 7), especially during the dry season. The mean KLeaf was significantly higher in deciduous 17.18 ± 1.45 mmol Mpa−<sup>1</sup> <sup>m</sup>−<sup>2</sup> <sup>s</sup>−<sup>1</sup> compared to evergreens 13.15 ± 1.57 mmol Mpa−<sup>1</sup> m−<sup>2</sup> s−<sup>1</sup> (*p* < 0.001), (Figure 7, Table 1). However, the seasonal changes of KLeaf in both groups of plants revealed a decrease in the KLeaf when VPDm increased during the dry period (Figure 7).

**Figure 7.** Seasonal midday vapour pressure deficit (VPDm) and leaf hydraulic conductance (KLeaf) of the considered deciduous and evergreen shrubs during the experimental period, from 27 May to 28 August. Values are means ± SE (nKleaf = 24, nVPDm = 48).

The relationship between KLeaf and Ψ<sup>m</sup> is presented in Figure 8. It is likely that as the growing season proceeded KLeaf decreased and exhibited the lowest values concomitantly with the lowest Ψ<sup>m</sup> values. It is worth mentioning that for the same low values of Ψm, the deciduous shrubs exhibited higher KLeaf in relations to the evergreens.

**Figure 8.** The relationship between midday leaf water potential (Ψm) and hydraulic conductance (KLeaf) of deciduous and evergreen shrubs during the experimental period, from 27 May to 28 August. Values are means ± SE (*n* = 24). Polynomial regression analysis was used to assess the relationship between KLeaf and Ψm.

In Table 2, the Pearson correlation between physiological parameters and VPDm is presented. In deciduous species a significant positive correlation was detected between VPDm and E and a negative between Ψ<sup>m</sup> and gs, while in evergreen shrubs VPDm was negatively correlated with Ψ<sup>m</sup> and KLeaf.


**Table 2.** Pearson correlation between predawn (Ψp) and midday leaf water potential (Ψm), transpiration rate (E), stomatal conductance (gs), leaf hydraulic conductance (KLeaf) and midday Vapour Pressure Deficit (VPDm) for deciduous (left) and evergreen (right) shrubs.

\* Significant for *p* < 0.05, \*\* significant for *p* < 0.001.

## **3. Discussion**

The results of this study demonstrate that the water status of co-occurring evergreen and deciduous broadleaved shrubs in semi-arid Mediterranean conditions under rising VPDm is different. Additionally, it seems likely that deciduous shrubs control more efficiently their water status during the dry season, by exhibiting lower stomatal conductance and higher hydraulic conductance than the evergreens. In other words, the deciduous plants possess a more water spending behaviour than the evergreens.

The increase of air temperature accompanied by decreasing RH and increasing heat load due to incident radiation on the leaf, in combination with air speed, may elevate VPD in the atmosphere [60]. It has been reported that RH concerning future climatic scenarios will remain either constant at the global scale [61], or a negative [62] and/or positive trend between VPD and RH at a regional scale [27] will be observed. In our research, the values of PPFD were generally maintained at high levels, as it was expected concerning the characteristics of the Mediterranean climate in the experimental site. The same values in air and leaf temperature imply that broadleaved shrubs possess such leaf tissues, where the transpiration flow is not sufficient to reduce leaf temperature in relation to the ambient temperature. The increase of Tm and VPDm under drought conditions (after June) may be linked to the physiological performance and survival of deciduous and evergreen shrubs via either reducing gs and gas exchanges (feedforward mechanism), or increasing plant water loss (feedback mechanism) [63].

The seasonal patterns of Ψ<sup>p</sup> and Ψ<sup>m</sup> suggest that the co-occurring deciduous and evergreen shrubs on Olympus Mountain were able to regulate water loss and ensuring adequate hydration of leaf tissues overnight, even during the dry season, therefore Ψ<sup>p</sup> was retained to approximately −1.0 MPa; such scale of the Ψ<sup>p</sup> values is sufficient for the Mediterranean region (*Quercetalia ilicis* zone) where the studied species are growing [64]. Ψ<sup>p</sup> is an essential index to study the response of plants to drought, because its value can be considered equal to soil water potential, since plants and soil reach a hydration equilibrium during the night [65–68].

The higher (less negative) Ψ<sup>m</sup> in deciduous shrubs indicates higher values of relative water content, E and gs in comparison to the evergreen shrubs. The Ψ<sup>m</sup> seems to be the main factors for stomatal regulation because is directly linked to the turgor of guard cells [49,69,70]. After sunrise, the leaf water potential in Mediterranean plants decreases steadily, reaching its lower value at noon, while it begins to recover again during the afternoon [71–73]. This reflects a decline of stored water, and therefore water shortage [60]. The plants are the best indicators of soil water availability, which affect their water status [74,75]. Plants with both elevated (less negative) Ψ<sup>m</sup> and water transport capacity can better control their leaf water status during midday, and hence they experience a smaller decline of midday gs [49].

It is likely that the relatively higher E in deciduous species during midday is due to their elevated water status, i.e., higher Ψ<sup>m</sup> in leaves (Figures 4a and 5); a slight variation of E values was detected between deciduous and evergreen shrubs, while the lowest values of E occurred in the two groups of shrubs during the dry season (June–August), (Figure 5). Apparently, low E values during the dry season remained rather stable and above zero values, which indicates influx of carbon dioxide [49,76].

The deciduous shrubs exhibited higher Ψ<sup>m</sup> and lower midday gs than the evergreens; this may indicate that midday stomatal conductance is more related to stem rather than leaf water status, which is in accordance with earlier results [49]. In fact, gs was negatively correlated with VPDm (Table 2), suggesting a response to environmental conditions, which is in agreement with the published results by Auge et al. [77] from other deciduous species. This trait may help deciduous species to regulate their stomatal apparatus in order to maintained Ψ<sup>m</sup> in a range of values that will support their water status and avoid xylem embolism under drought conditions [49,78]. Our results show that there was not any synchronization between gs and Ψ<sup>m</sup> in both shrub groups, which may also be due to the fact the evergreens are hypostomatic plants [79–81]. However, it is not clear which microclimate parameter, i.e., temperature and/or relative humidity, was directly linked to stomatal behaviour. It has been proposed that temperature had a greater impact on stomatal conductance and assimilation rate in the amphistomatic leaves of *Eucalyptus tetrodonta* independently of VPD [82,83]. Addington et al. [35] reported that the stomatal apparatus controls Ψ in a way that the tension on the water column created by decreasing Ψ did not cause extreme xylem cavitation. Several studies suggested that high VPD reduces gs, consequently affecting assimilation rate and growth [84,85]. On the contrary, it has been argued that the impact of high VPD on gs may not possess any impact on assimilation rate and growth [86]. Nevertheless, the impact of VPD on gs and/or assimilation rate varies amongst plant species [82]. In addition, the stomatal anatomy and structure affect water loss and carbon assimilation, demonstrating the evolution and adaptations of the plants to environmental conditions [84,87].

Plants in order to grow and survive under water limited conditions evolved mechanisms to control stomatal aperture and xylem water capacity [88]. The differences in wateruse strategies might be partially due to various hydraulic properties between the considered groups. The hydrodynamic status of leaf tissue expressed via leaf water potential [89] is determined by two mail functions taking place in the SPA continuum, i.e., transpiration rate and hydraulic conductance. Therefore, the favourable water status in deciduous shrubs could be attributed to the higher values of transpiration rate and/or at the highest values of the hydraulic conductance. The deciduous shrubs exhibited higher values of KLeaf when compared to evergreen shrubs, especially during the dry summer period. This advantage of deciduous shrubs could be attributed to their water status and anatomical features. The positive correlation between KLeaf and Ψm, and E in deciduous shrubs (Table 2) and the negative correlation between KLeaf and VPDm suggest that the water transport from root to the leaf plays a role to the fluctuations of leaf water status. The seasonal KLeaf declined in both groups of plants when VPD increased during the summer dry period (Figure 7) especially at the end of July, when the highest value of VPD (3.99 KPa) was recorded. Probably, at a given transpiration rate in deciduous shrubs the leaf water status is maintained due to high hydraulic conductance [17,90,91]. Manzoni et al. [88] reported that, in some ecosystems, deciduous species have been found to be more hydraulically efficient than evergreen species. Choat et al. [92] suggested that deciduous species are more hydraulically efficient, but also more vulnerable to drought-induced embolism, than co-existing evergreens in a rainforest. It is well known that embolism occurs in plants under drought conditions when Ψ reaches very low values; however, the plants have the capacity to repair damage when the environmental conditions are favourable [89].

Our data are somehow in agreement with BIumler [93], who argued that although the Mediterranean climate is associated with evergreen species, some coexisting deciduous species exhibit some advantages in response to drought [92,94]. It is noteworthy that in our work deciduous and evergreen species are presented as two distinct life-history groups [91], although they probably form a continuum of variation in leaf life-history span [95].

## **4. Materials and Methods**

*4.1. Study Area and Climate*

The study was conducted on Olympus Mountain (40◦06 54 N, 22◦28 42 E), which is a great, long-lived natural laboratory [96–98] in 2009, at an altitude of 554 m a.s.l., in an area located 5 Km from the town of Litochoro, 95 km south-east of Thessaloniki, in Greece. The climate of the study area is characterized as Cfa in the Köppen-Geiger system (http://www.en.climate-data.org, 7 January 2022) and as Mediterranean with cold and wet winters, dry and warm summer according to the bioclimatogram of Emberger. The mean annual rainfall in the study area ranges from 800 to 1000 mm, while substantially elevated precipitation is recorded during winter. The minimum air temperature ranges from 13 to 6 ◦C (during summer and winter, respectively). The maximum air temperature ranges from 4.9 to 6.7 ◦C during winter, and from 20 to 26 ◦C during summer. The warmest month is July and the coldest is December. The mean annual relative humidity (RH) ranges from 75 to 80%. Average RH values during the most humid month (December) is 85–90% and during the driest month (July) 30–50%. The monthly changes of temperature and precipitation (ombrothermic diagram) during the year of the measurements, from the nearest meteorological station of Dion (40◦12 00 N, 22◦30 00 E), are presented in Figure 9.

**Figure 9.** Ombrothermic diagram in the study area throughout a year.

The microclimatic conditions (temperature, humidity), in the study area during the experimental period was measured with the Hobo H8 Pro (Hobo H8 Pro Series 1997–2003, Onset Computer Coorporation, Bourne, MA, USA). Furthermore, on specific Julian dates and hours when plant physiological parameters were investigated, ambient air predawn and midday temperature and RH were measured using a Novasima MS1 microclimatic sensor (Novatron Scientific Ltd., Horsham, UK). Predawn (VPDp) and midday (VPDm) vapour pressure deficit were calculated according to Abtew and Melesse (2013). The Photosynthetic Photon Flux Density was measured with the steady-state diffusion porometer LI-1600 (LI- COR, Inc., Lincoln, NE, USA). The presented values of VPDp, VPDm, RH, PPFD, Tp and Tm are means of twenty-four measurements.

## *4.2. Plant Material*

The study area is part of the Mediterranean zone of the evergreen broadleaved plants *Quercus ilex* L. and *Arbutus andrachne* L. Deciduous and evergreen shrubs and small trees such as *Acer monspesulanum* L., *Carpinus betulus* L., *Cercis siliquastrum* L., *Cotinus coggygria* Scop., *Fraxinus ornus* L., *Pistacia terebinthus* L., *Coronilla emeroides* Boiss. and

Spruner, *Quercus coccifera* L., *Arbutus unedo* L., *Phillyrea media* L., and *Juniperus oxycedrus* L. are widespread in the study area.

Four evergreen shrubs *Arbutus andrachne* (commonly known as Grecian strawberry tree), *Arbutus unedo* (strawberry tree), *Quercus ilex* (holm oak) and *Quercus coccifera* (kermes oak) and four deciduous shrubs *Carpinus betulus* (common hornbeam), *Cercis siliquastrum* (Juda's tree), *Coronilla emeroides* (scorpion senna), and *Pistacia terebinthus* (terebinth) were selected for this study. The shrubs were randomly selected for sampling. More specifically, the area of interest was equal to 20 ha. We divided this area into 20 tiles, 1 ha each, and then we randomly selected 6 tiles where each one of the eight species was randomly chosen. Concerning dendrometric parameters, the considered species possessed the same height (3–4 m) and diameter (1.5–2.0 m).

## *4.3. Physiological Parameters*

Leaf water potential (Ψ), transpiration rate (E), stomatal conductance (gs) and leaf temperature (Tl) were measured on the newest fully developed mature leaves of six different individuals from sun-exposed terminal branches. Using the pressure-bomb technique (PMS, Albany, OR, USA), leaf water potential was measured twice during the day, i.e., at predawn (Ψp) and midday (Ψm). Transpiration rate, gs and Tl were measured using steady state porometer (Li1600, LI-COR Lincoln, NE, USA). Seasonal measurements of Ψ<sup>p</sup> were obtained before sunrise, while the measurements of Ψm, E, and gs were obtained on clear sunny days at around solar noon (12:00–14:30 h), approximately 10–15 days intervals. The presented values, for each of the parameters, are means of six replications per studied shrub.

Leaf hydraulic conductance (KLeaf) was calculating according to Ohm's law following the formula:

$$\mathcal{K}\_{\text{Lauf}} = \frac{\mathcal{E}}{(\Psi\_{\text{soil}} - \Psi\_{\text{Lauf}})} = \frac{\mathcal{E}}{(\Psi\_{\text{P}} - \Psi\_{\text{m}})} \tag{1}$$

It has been assumed that Ψsoil is in equilibrium with Ψ<sup>p</sup> and the lowest diurnal Ψleaf is equal to Ψ<sup>m</sup> [66,67,99]. However, sometimes the first assumption may lead to overestimation of Kplant [100]. Additionally, values of KLeaf reflect the capacity of evergreen and deciduous plants, grown under ambient conditions, for water exploitation during a period of soil drying.

## *4.4. Statistical Analysis*

The Kolmogorov–Smirnov test was employed to evaluate data normality. To explore whether the ecophysiological response of deciduous and evergreen shrubs vary during the dry season, a two-way analysis of variance (ANOVA) was performed on the studied parameters (TL, Ψp, Ψm, E, gs, Kleaf) [81]. Student's *t*-test for independent samples was used to examine differences in TL, E, gs, Ψm, Ψ<sup>p</sup> and KLeaf between deciduous and evergreen shrubs. The climatic parameters VPDp, VPDm, Tp, and Tm at each sampling date were compared using also the Student's *t*-test for independent samples. Polynomial regression analysis was used to determine the relationship between KLeaf and Ψm, between deciduous and evergreen shrubs. In addition, a Principal Component Analysis (PCA) with varimax rotation was used to assess the relationships among the measurements of interest between the eight shrubs to see whether they could be classified according to their ecophysiological response in the life form in which belong (deciduous, evergreen). Pearson correlation was used to explore links among VPDm, E, gs, Ψp, Ψ<sup>m</sup> and KLeaf. P-values lower than 0.05 were considered statistically significant. All statistical analyses were performed using the SPSS statistical package v. 27.0 (IBM Corp. in Armonk, NY, USA).

## **5. Conclusions**

The results of this research work demonstrate different hydraulic strategies between cooccurring deciduous and evergreen shrubs grown on Olympus Mountain. The deciduous life-form presented a strategy of higher hydraulic and lower stomatal conductance, while the evergreen exhibited lower hydraulic and higher stomatal conductance. Although, seasonal leaf hydraulic conductance declined in both groups of plants when vapor pressure deficit increased during the summer dry period, evergreen shrubs sustain a water transport to their foliage at a rate sufficient to prevent severe damage due to desiccation.

**Author Contributions:** Conceptualization: M.K.; methodology: M.K., E.V. and S.R.; software: M.K., P.K. and E.V.; validation: M.K.; formal analysis: M.K., E.V. and S.R.; investigation: M.K. and P.K.; resources: M.K.; writing—original draft preparation: M.K. and S.R.; writing—review and editing: M.K., E.V. and S.R.; visualization: M.K.; supervision: M.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable for studies not involving humans or animals.

**Informed Consent Statement:** Not applicable for studies not involving humans or animals.

**Data Availability Statement:** The data presented in this study are available in figures and tables provided in the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


## *Article* **Impact of Grazing on Diversity of Semi-Arid Rangelands in Crete Island in the Context of Climatic Change**

**Maria Karatassiou 1,\*, Zoi M. Parissi 2, Sampson Panajiotidis <sup>3</sup> and Afroditi Stergiou <sup>1</sup>**


**Abstract:** The rangelands of Crete island (Greece) are typical Mediterranean habitats under high risk of degradation due to long-term grazing and harsh climatic conditions. We explored the effect of abiotic (climatic conditions, altitude) and biotic factors (long-term grazing by small ruminants) on the floristic composition and diversity of selected lowland (Pyrathi, Faistos) and highland (Vroulidia, Nida) rangelands. In each rangeland, the ground cover was measured, and the floristic composition was calculated in terms of five functional groups: grasses, legumes, forbs, phrygana, and shrubs. The aridity index, species turnover, species richness, Shannon entropy, and Gini–Simpson index (with the latter two converted to the effective number of species) were calculated. Our results reveal that highlands are characterized by the highest aridity index (wetter conditions). Lowland rangelands, compared to highland, exhibited a higher percentage contribution of grasses, legumes, and forbs, while species turnover decreased along the altitudinal gradient. The Shannon entropy index was correlated (a) positively with Gini–Simpson and mean annual temperature and (b) negatively with mean annual precipitation, aridity index, and altitude. Moreover, the Gini–Simpson index correlated positively with mean annual temperature and negatively with altitude. Our results could help to understand the effects of grazing on rangeland dynamics and sustainability in semi-arid regions in the context of climatic change.

**Keywords:** aridity index; effective number of species; Shannon entropy; richness; Gini–Simpson

## **1. Introduction**

Arid and semi-arid rangelands occupy approximately 40% of the Earth's land surface and influence the livelihood and well-being of one-fifth of the world's human population [1,2]. More than one billion people rely on rangelands for their living, primarily through extensive livestock production, and roughly two billion acquire animal protein, water, or other resources from these biomes [3,4]. Rangelands comprise many habitats and host economically important species offering support to approximately 50% of the world's livestock, providing forage production for both domestic and wildlife populations [5–7].

Despite their high importance, most of the non-marketed services of these rangelands and their economic value have often been neglected [8,9]. Moreover, they have faced increased risks resulting from overutilization and degradation [10–12]. The estimated extent of rangeland degradation varies extensively, from as little as 10–20% to as much as 70–80% [3]. Desertification is a cumulative threat that includes both climatic and land-use drivers that interact in space and time [13].

It is well demonstrated that rangelands are maintained by grazing. However, they can be severely affected by the high intensity of the latter, climate change, soil quality, nutrient depletion, fire, habitat fragmentation, as well as human activities [4,7,14]. In

**Citation:** Karatassiou, M.; Parissi, Z.M.; Panajiotidis, S.; Stergiou, A. Impact of Grazing on Diversity of Semi-Arid Rangelands in Crete Island in the Context of Climatic Change. *Plants* **2022**, *11*, 982. https://doi.org/10.3390/ plants11070982

Academic Editor: Giuseppe Fenu

Received: 29 December 2021 Accepted: 1 April 2022 Published: 4 April 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/).

most rangelands, precipitation [15] and grazing [16,17] are the most important factors determining species diversity and ecosystem function [18,19]. On the other hand, altitude, which greatly affects the abiotic environment by modifying climatic variables and the topography [20,21], is an indirect gradient that is correlated with resources and regulators of plant growth [22,23] and species composition [24].

Although it is known that the richness of vascular plant species decreases with an increase in altitude [25–27], the patterns of response are rather fickle [28]. Changes in plant species richness along altitudinal transects are of great importance in the study of global climate change [29,30]. The spatial change in species composition involves the study of beta diversity and species turnover. Because of the greater diversity of habitat conditions, mountains have higher levels of species turnover than lowland areas [31], and under a climate change scenario, mountains are considered significant for the maintenance of biodiversity [32–34]. The relationship between climatic conditions and species turnover is described by the relationship between climatic factors and regional species richness [35].

To assess the impact of grazing on vegetation, the effect of precipitation on species diversity should be thoroughly studied and considered [36]. Both low ground cover and plant diversity increase the vulnerability of rangelands to climate change [37,38]. Overgrazing, which is prescribed as a decrease in productivity [39] and loss of biodiversity [40,41], is considered one of the main causes of land degradation in arid and semi-arid regions worldwide [42]. Heavy grazing directly changes the floristic composition of plant communities selectively, changing the structure and composition of communities at the expense of palatable species [43,44], and may also indirectly modify the outcome of competitive interaction by changing light availability [45]. The impact of grazing intensity on plant diversity varies along the precipitation gradient [46,47].

There is a list of methods employed to study diversity, which is a multi-dimensional phenomenon [48]. The simplest measure of diversity is to calculate the number of species (richness) in an area, which, however, does not take into consideration species abundances and is sensitive to sample size. Other approaches consider species abundance (Shannon index) or give weight to dominant species (e.g., Gini–Simpson). The Shannon and Gini–Simpson measures of diversity are themselves mere indices and not "true" diversities [49–53]. The true diversity of an investigated community is simply the community of equally common species (effective number of species, ENS) required to give the same value of an index calculated for the community in question [52,54,55]. In recent years, the use of ENS has been established in ecological studies. After the conversion of classical indices (Shannon and Simpson) to ENS, diversity is always measured in the number of species, providing more interpretable and comparable assessments of diversity [54,56,57].

Thirty-five percent of the Greek land, and more specifically 37–50% of the land in Crete, is characterized as critically susceptible to desertification due to the combination of a warming climate with low precipitation and intensified human activities [58,59]. To the best of our knowledge, the effect of grazing and climatic conditions on grassland biodiversity has not yet been studied in Crete, a vulnerable Mediterranean region.

The current study aimed to investigate the effect of abiotic (climatic conditions, altitude) and biotic factors (long-term grazing) on the floristic composition and diversity of lowland and highland rangelands on the island of Crete, Greece, which are typical Mediterranean habitats at high risk of degradation. We aimed to answer the following questions:


## **2. Results and Discussion**

The current study indicates that the existing high grazing pressure, in combination with climatic conditions, could result in rangeland degradation on Crete island.

Diverse climatic conditions prevail among the four studied rangelands (Figure 1). At Faistos and Pyrathi, the mean annual temperature was 19.17 ± 1.24 and 17.38 ± 1.18 ◦C, and the mean monthly precipitation was 45.77 ± 11.05 and 60.67 ± 15.1 mm, respectively. At Vroulidia and Nida, an inverse trend was observed as the mean annual temperature was 12.67 ± 1.98 ◦C with an average monthly precipitation of 107.05 ± 23.93 mm. Pyrathi had higher rates of precipitation and mean monthly temperature compared to Faistos (Figure 1b,c). The climatic data indicated a shorter drought period in Vroulidia and Nida, which implied that plant species faced a water deficit for a shorter period in these areas.

**Figure 1.** Monthly means of air temperature (◦C) and precipitation (mm) at (**a**) Vroulidia—Nida, (**b**) Pyrathi, and (**c**) Faistos rangelands during the experimental period.

The aridity index (IdM) classifies the type of climate in relation to water availability, and it is a crucial environmental factor affecting the growth of natural vegetation. In the present study, IdM was negatively correlated with mean air temperature and positively with altitude and precipitation (Table 1). The Nida rangeland scored the highest aridity index, followed by Vroulidia, while Faistos had the lowest one (Figure 2). The higher values of IdM in the highlands indicated higher humidity [60] and better climatic conditions for plant growth and development. Mallen-Cooper and coauthors [61] found similar results in eastern Australia, which support that aridity decreases when the height of precipitation and absorptivity of water increase. The higher IdM correlated with the higher available water resources over time and, consequentlym lower vulnerability to desertification.

**Table 1.** Pearson correlation between richness (R), effective Shannon entropy (SE) and Gini–Simpson (GS), Martonne aridity index (IdM), mean annual temperature (T), mean precipitation (P), and altitude for the four studied rangelands.


\* Significant for *p* < 0.05, \*\* Significant for *p* < 0.001.

**Figure 2.** Martonne's aridity index (IdM) for the studied rangelands for the period 1979–2013.

The aridity level interacts with plant traits related to stress resistance to determine the floristic composition and vegetation responses to domestic animal grazing [62–66]. The impact of grazing on species richness and composition under high-aridity conditions could be either high [67] or low [65].

Generalized Linear Model analysis showed significant differences (*p* < 0.001) in forage production in both fenced plots and grazed sites and for FUP (Table 2) among the studied rangelands. Additionally, there was a significant interaction between rangeland and year (*p* < 0.001), but only for forage production in grazed sites. The forage production in fenced plots ranged from 189.2 ± 24.2 to 116.5 ± 7.3 g m−2, while in grazed from 128.9 ± 4.8 to 15.8 ± 1.4 g m−2. The lowland rangelands had higher forage production in relation to the highland. This is in agreement with Bhandari and Zhang [68], who demonstrated that altitude is negatively related to aboveground biomass. Concerning the year, it was a significant predictor (*p* < 0.001) only for FUP. The FUP was 84.4–87.1% at Nida, 74.1–76.1% at Vroulidia, 74–75% at Pyrathi, and 15–17% at Faistos for 2014 and 2015, respectively (Table 2). The highest value of FUP was presented in Nida and the lowest in Faistos. In many cases, the high FUP is related to low vegetation percentage cover, a result of overgrazing [69].

**Table 2.** Forage production (g m<sup>−</sup>2) in fenced plots and grazed sites and forage utilization percentage (FUP %) in the four studied rangelands in the study years. Values represent means ± SE (n = 9). Different letters in the same column indicated significant differences (*p* < 0.05).


\* Significant for *p* < 0.05, \*\* Significant for *p* < 0.001, ns not significant.

It is known that the cover of vegetation is a health indicator of the rangelands. Our data analysis revealed significant differences (*p* < 0.05) in vegetation cover among the studied rangelands. The vegetation cover recorded in Nida and Vroulidia scored the lowest values, 52% to 66% and 64% to 70%, for 2014 and 2015, respectively, in comparison to the lowland sites of Pyrathi (78–90%) and Faistos (92–99%). The low vegetation cover and the high FUP at Nida and Vroulidia are likely the results of overgrazing, as the number of transhumant small ruminants, and, consequently, the grazing pressure is immoderate in the Psiloritis mountain [70]. Papanastasis and coauthors [71] point out that the Psiloritis mountain is overgrazed as the stocking rate is four times higher than the grazing capacity. Ojima and coauthors [72] found that overgrazing results in the loss of vegetation cover and increased erosion. The low vegetation cover provides low protection from soil erosion and a high risk for degradation [13] and reduced soil porosity [46,73]. This reduced vegetation cover and the lower plant diversity probably increase the susceptibility of rangelands to the effects of climate change as well [37,38,46].

Data analysis revealed that in all functional groups, there were no significant differences between years and no significant interaction between rangeland and year (*p* ≥ 0.05). On the contrary, there was a significant interaction between rangeland and functional groups (*p* < 0.001) (Table 3, Figure 3). Overall, lowlands, compared to highlands, present a significantly (*p* < 0.05) higher percentage contribution of grasses, legumes, and forbs (Figure 3). On the other hand, shrubs had a significantly (*p* < 0.05) higher percentage in the highlands compared to the lowlands, while phrygana [74,75] had similar participation in both lowlands and highlands. Concerning the contribution of functional groups separately in highland and lowland, shrubs were significantly (*p* < 0.05) higher at Nida compared to Vroulidia, while the opposite trend was detected for forbs in highlands for both experimental years (Figure 3). As others point out as well, the Cretan landscape, especially in high elevations, is a mixture of woodland and open vegetation, where many woody species are found in various forms (from trees to small or dwarf shrubs) [76–78]. Many of these shrubby taxa are shaped by grazing and are adapted to this pressure, which includes prescribed fires. Moreover, woody plants are able to 'colonize' rocky places where soil can be scarce. Moreover, Papanastasis and coauthors [77] found that woody species on the Philoritis mountain cover 30% of the soil. Regarding the lowlands, there is significantly higher participation of grasses (*p* < 0.05) in the Faistos rangeland compared to Pyrathi (Figure 3). On the contrary, the participation of phrygana was higher at Pyrathi compared to Faistos.

**Figure 3.** Floristic composition per functional groups (grasses, legumes, forbs, phrygana, shrubs) (%) at the four studied rangelands for the experimental period. Values represent means ± SE (n = 6). Different letters in columns indicate significant differences for the same parameter (*p* < 0.001).


**Table 3.** General linear model analysis for the effects of rangeland and year on participation of functional plants groups.

\* Significant for *p* < 0.001, ns not significant.

On Crete island, as elsewhere in Greece, farmers traditionally improve grassland productivity [79] and quality through fire management of vegetation, mainly phrygana, which enables the modification of the floristic composition. The fires decrease the percentage of shrubs and phrygana and drive the ecosystem to a previous successional stage (secondary succession), where the percentage of grasses and legumes is higher, leading to higher herbage biomass production in terms of quantity and quality [80,81]. The floristic composition of the studied rangelands is strongly linked to habitat characteristics (abiotic factors: altitude, climatic conditions) and primary consumers (biotic factors) [82].

A drop in species turnover is observed between pairs of lowland to highland rangelands (Figure 4). Species turnover presented the highest value at an intermediate altitude from 355 to 1100 m a.s.l.; and decreased at higher altitudes, from 1100 to 1530 m a.s.l., due to the range of ecological adaptation and growth of plants at different altitudes. The same results were found by Mena and Vázquez-Domínguez [83] when the species turnover in mammals was studied, more specifically, small rodents, along an altitudinal gradient. Our results support the hypothesis that species turnover decreases with altitude only at the higher altitudinal zone. The lower species turnover in highlands could be attributed to the presence of sparse vegetation in mountainous areas generally [79], and probably, species turnover correlates with different rangeland management [84].

**Figure 4.** Species turnover in the studied rangelands Faistos—Pyrathi (F-P), Pyrathi—Vroulidia (P-V) and Vroulidia—Nida (V-N) for the experimental period.

Diversity in terms of abundance (ENS Shannon entropy) was lower than species richness, while diversity in terms of dominance (ENS Gini–Simpson index) was lower than the Shannon entropy for both years of the study (Figure 5). This result indicated that there is species dominance in all study areas. The greater the dominance in the community, the greater the differences among these three parameters [51,52]. In both years of the study, the species richness, Shannon entropy, and Gini Simpson were higher in the lowlands of Faistos and Pyrathi compared to Vroulidia, while at Nida, the highest species richness was recorded. It is noteworthy that there are more plant species at Pyrathi than at Faistos, and the same trend for diversity (Figure 5). This was contrary to the theory that species richness and diversity decreased with an increase in altitude. The species richness may not be related to altitude, as it is demonstrated by Zawierucha and coauthors [85]. This unexpected result is probably due to the fire set by shepherds to improve forage production at Pyrathi last year and led to a change in the floristic composition. It has been proven that fire has important effects on diversity and plant community composition [86–90]. Shannon entropy could be used in situations where rare and abundant species or traits are expected to be equally important [91]. However, if dominant species or traits are expected to be more essential, then Gini–Simpson would be more relevant. Both indices were smaller than richness for all rangelands (Figure 5), as they were based more on abundant and dominant species, respectively [82]. Although the environmental conditions favored plant growth in mountainous areas, the grazing pressure significantly decreased species diversity. Nida had the highest species richness in relation to the other studied rangelands, but its species abundance (Shannon entropy) and dominance presented the lowest value. These results could be attributed to high FUP (overgrazing), lower vegetation cover, soil erosion, and unpalatable plant species encroachment that will be exacerbated by climate change [92].

Heavy grazing (high FUP) may result in high-level species replacement [93]. Plants at Pyrathi and Vroulidia are grown under different climatic conditions but under similar FUP, presented different species richness but similar diversity in terms of species abundance and dominance. At Faistos, under light grazing pressure, the rangeland presented similar high ratios of abundant and dominant species to total species recorded (richness). These results could be verified from the ratios of Shannon entropy/richness and Gini Simpson index/richness (Table 4). The highest ratio was recorded at Faistos and the lowest one at Nida. Faistos, with the longer semi-arid period under low FUP, presented a very diverse vegetation pattern, with three-quarters of all species showing the same abundance, while more than half were also dominant (Table 3). As grazing intensity escalates from Pyrathi to Nida, ratios of abundance diversity/richness and dominance diversity/richness decrease; this is more evident in terms of absolute ENS values of abundance and dominance diversity (Figure 5). Vroulidia shows higher ratios than Pyrathi but similar absolute ENS values for dominance diversity and lower for dominance diversity. These rangelands, without these disturbances (climate, grazing), would gradually decline due to the successional process to the next successional stages [94,95]. Animal grazing is a key factor in avoiding the successional processes of vegetation [82].


**Table 4.** Ratios of Shannon entropy/richness (HE/R) and Gini–Simpson (GS/R) index/richness for all of the studied rangelands for the experimental period.

\* Gini–Simpson and Shannon entropy are given as effective number of species (ENS).

**Figure 5.** (**a**) Gini–Simpson vs. richness index and (**b**) Shannon entropy vs. richness index for all studied rangelands for the experimental period, 2014 and 2015 (symbols encircled). Cycle size is proportional to richness. Gini–Simpson and Shannon entropy are given as effective numbers of species (ENS).

According to the Pearson correlation coefficient, the Shannon entropy index was positively correlated with the Gini–Simpson and mean annual temperature and negatively with mean annual precipitation and altitude, while the Gini–Simpson index correlated negatively with altitude (Table 1). According to the results, the species diversity decreased with an increase in altitude and precipitation. This is in agreement and supports the theory that species richness and diversity decrease along the altitude gradient [27,96,97]. Altitude probably has the strongest effects on species richness, abundance, and ground cover [98]. Nevertheless, other studies found that overgrazing affects functional diversity more than climate, and species diversity declines with an increase in grazing intensity in areas with

different climatic conditions [41,46,99]. It is well known that the relationships between diversity indices do not always follow mathematically predicted patterns [100,101].

## **3. Materials and Methods**

## *3.1. Study Area*

The research was conducted during 2014–2015 on the island of Crete, the southernmost part of Greece. The selected experimental sites were two lowland rangelands of Heraklion prefecture: Faistos (F) (24◦51 20 E, 35◦06 30 N) 155 m a.s.l. and Pyrathi (P) (25◦11 21 E, 35◦05 52 N) 355 m a.s.l., and two in the highlands of Psiloritis mountain (Rethymnon prefecture): Vroulidia (V) (24◦47 02 E, 35◦10 58 N) 1100 m a.s.l. and Nida (N) (24◦50 33 E, 35◦12 48 N) 1530 m a.s.l. that have been subjected to grazing (Figure 6).

**Figure 6.** The experimental rangelands on the island of Crete, Greece.

The livestock farming system was introduced on the island about 8000 years ago, and animal husbandry has been used by humans to transform natural ecosystems to produce more grazing material and, therefore, more animal products for their own consumption and survival. Through these processes, the extensive forests of the island were turned into rangelands, while the abandoned fields due to grazing could not be reforested. Uncontrolled and random, both spatially and temporally, grazing is the rule on the island. In lowlands, e.g., Faistos, in recent years, a change in land-use has been observed with farming replacing pastoralism, so there is low grazing intensity in the area. On the other hand, the lowlands of Pyrathi are heavily grazed all year round by sheep and goat flocks. Concerningly, the highlands of Vroulidia are grazed all year by sheep and goats, while Nida, from April to October, by transhumant small ruminant flocks. The highlands are characterized by a long history of small ruminant overgrazing [71].

The climate of the lowland and highland rangelands is characterized as Csa and Csb, respectively, in the Köppen–Geiger system (www.en.climate-data.org, 12 December 2021). The daily climatic data (precipitation, average temperature) for the two lowland rangelands (P, F) (Figure 1b,c) were obtained from the nearest meteorological stations, while for the highlands (N, V) (Figure 1a) from the only one available meteorological station located between them, and are reported as mean monthly data for the period in which the study was conducted.

## *3.2. Field Data*

The vegetation (ground) cover was measured at the end of the growing season according to the line and point method [102]. Three experimental transects (25 m each) [103,104] were established in each rangeland, as the habitats were homogeneous. After that, the floristic composition was calculated and presented in five functional plant groups: (1) grasses, (2) legumes, (3) forbs, (4) phrygana, and (5) shrubs, according to their life form and by distinguishing legumes from forbs based on their nutritional value for small ruminants (Table S1). Moreover, two sampling quadrats of 0.35 x 0.35 m were established in every transection of each rangeland at 8 and 16 m in order to calculate: (a) species richness (equivalent to its own numbers) and (b) species diversity indices (Shannon entropy and Gini–Simpson), which were converted to the effective number of species (ENS). Shannon entropy was calculated following the formula in Equation (1) below

$$\mathbf{H} = -\sum\_{i=1}^{S} p\_i \ln p\_i \tag{1}$$

and was converted to ENS by taking its exponential exp(H) (exponential of Shannon entropy index), where pi is the population frequency of the ith species. The Gini–Simpson index (HGS) was converted by the transformation

$$1/(1-\text{H}\_{\text{GS}}),$$

which is the inverse of the index [51,52,57,105]. These measures easily pass interpretable counts and provide information at three different levels based on how rare and abundant taxa are weighted [53,105–107].

$$1/\left(\sum\_{i=1}^{s} p^{p\_i^2}\right)$$

For every studied rangeland, the aridity index (de Martonne index, IdM) was calculated following the formula in Equation (2) below [60]:

$$\mathbf{I}\_{\rm dM} = \mathbf{P}/(\mathbf{T} + 10) \tag{2}$$

where P is the mean annual precipitation (mm), and T (◦C) is the mean annual air temperature. The values of T and P for every rangeland were downloaded from Climatologies, at high resolution (30 × 30 s), for the Earth's land surface areas (CHELSA, http://chelsa-climate.org/, 2 February 2022), which is a global climate database covering the period 1979 to 2013.

The species turnover was calculated as the gain and loss of species between altitudes following the formula in Equation (3) below [108]:

$$\beta(\mathsf{H}) = (\mathsf{g}(\mathsf{H}) + \mathsf{l}(\mathsf{H})) / (\infty(\mathsf{H}) + \infty(\mathsf{H} \cdot \mathsf{l})) \tag{3}$$

where g(H) and l(H) are the number of species gained and lost, respectively, from altitude H-1 to altitude H, while α(H) and α(H-1) is the species richness at altitude H and H-1, respectively [108].

In order to estimate the forage utilization percentage (FUP) in the spring of 2012 in each of the four rangeland's three plots, 9 m2 were fenced to be protected from grazing. The above-ground herbage production was collected by clipping three 0.5 × 0.5 m quadrats in each fenced plot (i.e., nine quadrats per fenced plot). In the same period into grazed rangelands (sites), the remaining above-ground biomass after grazing was collected by clipping in three similar quadrats in each transect (i.e., nine quadrats per rangeland), in May 2014–2015. Consequently, grazing intensity was expressed by FUP. The difference among herbage yields of fenced (UG) and grazed sites (G) was used to calculate FUP from the formula of Equation (4) below [109]:

$$\text{FUP} = \left[ \left( \text{UG-G} \right) / \text{UG} \right] \times 100 \tag{4}$$

## *3.3. Statistical Analysis*

The Generalized Linear Model (GLM), assuming a normal distribution, was used to assess whether the altitude of each rangeland, functional group, and year were significant predictors of ground cover and floristic composition. Before analysis, the data were converted to ln + 1 to meet assumptions of normality (tested with the Kolmogorov–Smirnov

test) and homogeneity of variances (Levene's test). Estimated marginal means for all the above factors were calculated with pairwise contrasts, and LSD adjustment was applied for the multiple comparisons (α = 0.05). The data in the figures and tables depict values before the transformation. Pearson correlation was used to explore links among R, SE, GS, IdM, T, P, and altitude. All statistical analyses were performed using the SPSS statistical package v. 27.0 (IBM Corp. in Armonk, NY). The Paleontological statistics software package for education and data analysis (Past) was used to calculate the diversity indices.

## **4. Conclusions**

The Mediterranean basin includes a wide range of vegetation, climatic, and edaphic conditions that have been shaped by natural selection under the pressure of a distinct climate and human activities. Our results demonstrate the strong relationship between diversity and temperature and agree with the fact that vegetation diversification is strongly related to the climatic gradient and is more related to temperature than precipitation. Moreover, this research could help to understand how grazing intensity and climatic conditions interactively influence rangelands dynamics in semi-arid regions and monitor the livestock management and decision making in these areas.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/plants11070982/s1, Table S1: Plant species in floristic composition at the four studied rangelands (life form: Grass (G), Legume (L), Forb (F), Phrygana (PH), Shrub (S).

**Author Contributions:** Conceptualization, M.K.; methodology, M.K., Z.M.P. and S.P.; software, M.K., S.P. and A.S.; validation, M.K.; formal analysis, M.K., Z.M.P., S.P. and A.S.; investigation, M.K., Z.M.P. and A.S.; resources, M.K. and Z.M.P.; writing—original draft preparation, M.K., Z.M.P. and S.P.; writing—review and editing, M.K., Z.M.P., S.P. and A.S.; visualization, M.K.; supervision, M.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors gratefully acknowledge the financial support of the European Union through the Action "THALIS" of the Programme "Education and Life-long learning".

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in the figures and tables provided in the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**

