Chara zeylanica J.G.Klein ex Willd. (Charophyceae, Charales, Characeae): First European Record from the Island of Sardinia, Italy

The first record of a species belonging to the genus Chara L. subgenus Chara R.D.Wood section Grovesia R.D.Wood subsect. Willdenowia R.D.Wood from Europe is presented here, thus challenging the interpretation of its distribution pattern as an intertropical group of charophytes. The morphological characters of the specimens, as well as the results of a phylogenetic analysis, clearly identified them as Chara zeylanica J.G.Klein ex Willd. Although the subsection Willdenowia has yet to receive a thorough taxonomic treatment, a discussion of its relationship to other taxa of this subsection is provided despite the lack of a commonly agreed upon taxonomic concept. The ecological conditions of the Sardinian site of C. zeylanica are presented. Moreover, the status of and threats to this taxon, and hypotheses regarding potential pathways through which it reached Europe, are discussed.


Introduction
Charophytes are morphologically complex macrophytic green algae with a worldwide distribution. Because they are close relatives of the earliest land plants [1], they have attracted growing scientific interest in recent decades. However, in addition to becoming a subject of academic interest, charophytes play a major role in bioindication systems due to their species-specific pattern of niche occupation [2,3]. Moreover, Characeae are among the most threatened groups of organisms on earth [4][5][6], and have thus been targeted by nature conservation actions [7][8][9]. As charophytes occur in an astonishingly wide variety of habitats, ranging from ultraoligotrophic freshwater to hypersaline and hypertrophic environments, their presence is often measured in water quality assessments and other related fields [10,11].
For the development of such bioindication systems, having comprehensive and reliable knowledge about the habitat preferences and distribution patterns of the individual species is essential, as is the accurate identification of charophyte species, and the formulation of a sound taxonomic concept. In recent decades, a large number of studies have attempted to fulfil these requirements [12][13][14][15][16][17][18][19]. As a result, our knowledge about the biogeography of charophytes has increased substantially. However, whereas in the past site-specific information about the occurrence of the individual species was provided [20], recent treatments have led to the development of large-scale distribution grid maps and detailed descriptions of the species' preferred habitat conditions [21].
For several species, a strong correspondence between the distribution range and the niche structure was found. For example, the strictly circumpolar distribution of Tolypella normaniana Nordst. can be explained by its temperature preference (cold-stenothermic). Moreover, it has been shown that species such as Chara vulgaris L. or C. braunii C.C.Gmelin occur in a broad range of habitats on all continents, except for Antarctica [22]. manent and temporary bodies of water, such as lakes, ponds, pools, ditches, temporarily flooded wetlands, canals, rice fields, and retention ponds [26,38,[40][41][42][43]. Few of the existing hydrochemical datasets cover a spectrum ranging from low-impacted waterbodies with total P-concentrations below 20 µg L −1 and total N-concentrations between 0.425 and 1.9 mg L −1 [43] to eutrophic habitats [37]. According to Muller et al. [45], C. zeylanica needs temperatures of approximately 25 • C for fructification.
The only European site (reported here for the first time) where the presence of C. zeylanica has been detected is at Cala Fuili, which is located north of Orosei on the east coast of Sardinia, Italy (coordinates: 40 • 25 03" N, 9 • 46 13" E; coordinate system WGS 84) ( Figure 1). The specimens were found in September 2019 at a depth of about 1 m, mainly in sandy to silty places with stony substrate, in a shallow and probably permanent small stream located close to the beach, or 110 m from the Mediterranean Sea. The specific site where the C. zeylanica specimens were found was situated directly next to a bridge ( Figure 1, left image below), and was therefore disturbed by the structure. By contrast, the neighbouring stream sections and landscape areas can be considered semi-natural habitats. The small population of C. zeylanica was observed to have high fertility, with ripe antheridia, oogonia, and oospores. The nutrient conditions at the sampling date were as follows: NH 4 -N 0.108 mg L −1 , NO 3 -N 0.279 mg L −1 , total N 1.143 mg L −1 , PO 4 -P 0.073 mg L -1 , and total P 0.137 mg L −1 . The water hardness was 26.4 • dH (Ca 62.2 mg L −1 , Mg 76.8 mg L -1 ), pH 8.3. Although the salinity at the sampling date was 1.9, the salinity of the site probably varies because it is close to the coast. At the same site in May 2016, a salinity level of 4.4 was recorded and the Cl concentration was found to be 2819 mg L −1 , instead of 1290 mg L -1 , as measured in September 2019.

Morphological Description
The specimens are 30-60 cm long, erect and straight, stout, fresh to greyish green, and slightly incrusted ( Figure 2). The main axis diameter is 589-1076 µm with a mean value of 844 µm, slightly (0-4) branched. Most of the internodes are 4.4-8.0 cm long, and are usually much longer (up to 4 ×) than the branchlets. The uppermost 1-2 internodes are

Morphological Description
The specimens are 30-60 cm long, erect and straight, stout, fresh to greyish green, and slightly incrusted ( Figure 2). The main axis diameter is 589-1076 µm with a mean value of 844 µm, slightly (0-4) branched. Most of the internodes are 4.4-8.0 cm long, and are usually much longer (up to 4 ×) than the branchlets. The uppermost 1-2 internodes are only 0.5-2.2 cm long, and are generally shorter than the adjacent branchlets. The cortex is usually triplostichous, and is rarely (partly) diplostichous and tylacanthous to isostichous ( Figure 2D). Single, acute, thin, and needle-like spines are observed on the young internodes, and rarely on the older internodes. These spines can vary in length (182-1468 µm long) even on the same plant, and mainly point downwards ( Figure 2D). The stipulodes are acute, elongated, and well developed. They are arranged in two regular tiers with two pairs per branchlet ( Figure 2E). The upper stipulodes are longer than the lower ones. As the upper stipulodes are 515-1045 µm long (a mean value of 760 µm), they are usually longer than the diameter of the axes, and are much longer than the lowermost branchlet segment. The lower stipulodes are sometimes of unequal lengths, at 161-475 µm long, with a mean value of 293 µm. The branchlets are 9-12 in whorl and generally much shorter than the internodes, at 2.0-4.3 cm long. The branchlets of the uppermost 1-2 youngest whorls are even shorter, at just 0.2-2.0 cm long. The lowermost basal segments of the branchlets are ecorticated, and are very short at 208-479 µm long (mean value 343 µm) and 189-470 µm wide (mean value 318 µm). These segments are hidden behind the upper stipulodes ( Figure 2A). The branchlets consist of 7-10 segments, with the lowermost segments always being ecorticated, followed by 4-6 corticated segments and 2-5 ecorticated distal segments with a tiny acute end cell on top, surrounded by a ring of bract cells ( Figure 2C,F). The bract cells (5)(6)(7)(8) are well developed (220-843 µm long), slender, and acute, and are shorter than the bracteoles. The two bracteoles are very long (990-1948 µm), at 1-2.5 × longer than the oogonia and oospores ( Figure 2B). All of the fertile specimens are monoecious with conjoined gametangia ( Figure 2B). Gametangia usually occur only at the nodes of corticated segments, and are rarely observed at the lowest nodes just above the ecorticated segment. The gametangia are mainly solitary, and very rarely geminate. The oogonia are elliptical to elongated oval in shape, are yellow or greenish in colour, and generally have constricted coronulae. The length of the oogonia (without coronula) is (600) 700-850 (900) µm, and the width of the oogonia is 417-575 (600) µm. The length of the coronula is 69-125 (150) µm, and the width of the coronula is (127) 160-200 (250) µm. The oospores are elliptical in shape and black in colour, with a length of (539) 600-685 µm, a width of 375-475 (500) µm, and 10-13 striae. The antheridiae are tetrascutate with a diameter of (300) 350-400 (450) µm. The dried specimens are stored at the herbarium of the University of Rostock (ROST).

Phylogenetic Analysis
The three individuals collected on Sardinia had identical rbcL and matK sequences. The BLAST of the GenBank nucleotide collection under default settings with rbcL from the Sardinian samples as query sequences matched the individuals to C. zeylanica from New Caledonia (AB440257) with 100% identity. One basepair (bp) substitution (99% identity) was detected for two further C. zeylanica (HQ380481: Sri Lanka, AY720934: Taiwan), but also for a sequence belonging to C. hydropitys Rchb. (HQ380464: Puerto Rico).
The BLAST of the GenBank nucleotide collection using matK from the Sardinian samples as query sequences matched the individuals with 99% identity (1 bp substitution) to C. zeylanica from Myanmar (MT739758). Chara guairensis R.M.T.Bicudo (KY656924) and C. hydropitys (KY656921) differed from the Sardinian samples by 15 bp substitutions (98% identity), respectively.
Phylogenetic analyses were performed for rbcL and matK separately to confirm the species identified through the BLAST search. The final rbcL alignment was trimmed to 1051 bp. Within the subsect. Willdenowia, 30 variable sites were identified. In the rbcL tree (Figure 3), the relationships within the subsect. Willdenowia were ambiguous, because several nodes did not have significant supports. The specimens from Sardinia formed a cluster together with C. zeylanica, but only with a low level of support (BS: 50%, PP: 0.6). The phylogeny based on the rbcL gene sequences only could not be resolved, and relationships of C. zeylanica to other species of subsect. Willdenowia were ambiguous.
(990-1948 µm), at 1-2.5 × longer than the oogonia and oospores ( Figure 2B). All of the fertile specimens are monoecious with conjoined gametangia ( Figure 2B). Gametangia usually occur only at the nodes of corticated segments, and are rarely observed at the lowest nodes just above the ecorticated segment. The gametangia are mainly solitary, and very rarely geminate. The oogonia are elliptical to elongated oval in shape, are yellow or greenish in colour, and generally have constricted coronulae. The length of the oogonia (without coronula) is (600) 700-850 (900) µm, and the width of the oogonia is 417-575 (600) µm. The length of the coronula is 69-125 (150) µm, and the width of the coronula is (127) 160-200 (250) µm. The oospores are elliptical in shape and black in colour, with a length of (539) 600-685 µm, a width of 375-475 (500) µm, and 10-13 striae. The antheridiae are tetrascutate with a diameter of (300) 350-400 (450) µm. The dried specimens are stored at the herbarium of the University of Rostock (ROST).

Taxonomical Remarks
The specimens collected at Cala Fuili (Sardinia) were shown to qualify, based on their morphological characters, as a taxon belonging to subsect. Willdenowia because they are diplostephaneous with triplostichous cortication and have ecorticated basal segments of otherwise corticated branchlets [23]. Following the approach of van Raam [28], who distinguished 20 species within subsect. Willdenowia-in contrast to Wood [23], who identified a monospecific subsection-the question of to which species the specimens belong is discussed in detail below.
Van Raam [28] analyzed systematically the problem of gymnopodial (ecorticated first branchlet segment) taxa of the genus Chara L. using a stepwise approach. A total of 37 taxa of the genus Chara were found to share the character of an ecorticated basal branchlet segment. We recall that a taxon is a taxonomic unit of any rank, which can be species, but also varieties and forms. Eight taxa from subsect. Willdenowia can be excluded because they are haplostephaneous (and can thus be assigned to sect. Imahoria J. van Raam). Of the remaining 29 diplostephaneous taxa, C. kenoyeri M.Howe and C. rusbyana can be excluded here because they are dioecious. As a haplostichous species, Chara pseudohydropitys Imahori belongs to section Aghardia R.D.Wood, and can also be excluded here. Similarly, C. foliolosa, C. tenuifolia (Allen ex R.D.Wood) R.D.Wood, C. guairensis, C. haitensis, C. indica Bertero ex Spreng., C. martiana Wallman, and C. paucicorticata Cáceres can be excluded because they have octoscutate antheridia. Unlike the specimens described here, Chara drouetii (R.D.Wood) R.D.Wood, C. michauxii (A.Braun) Kütz., and C. formosa C.B.Rob. are characterised by a sejoined gametangia arrangement. Chara cubensis Allen, C. depauperata Allen, C. oerstediana A.Braun, and C. diaphana (Meyen) R.D.Wood have fewer than four corticated branchlet segments, whereas all the specimens found in Sardinia have at least four corticated segments. According to van Raam [28], the remaining 12 taxa belong to C.

Taxonomical Remarks
The specimens collected at Cala Fuili (Sardinia) were shown to qualify, based on their morphological characters, as a taxon belonging to subsect. Willdenowia because they are diplostephaneous with triplostichous cortication and have ecorticated basal segments of otherwise corticated branchlets [23]. Following the approach of van Raam [28], who distinguished 20 species within subsect. Willdenowia-in contrast to Wood [23], who identified a monospecific subsection-the question of to which species the specimens belong is discussed in detail below.
Van Raam [28] analyzed systematically the problem of gymnopodial (ecorticated first branchlet segment) taxa of the genus Chara L. using a stepwise approach. A total of 37 taxa of the genus Chara were found to share the character of an ecorticated basal branchlet segment. We recall that a taxon is a taxonomic unit of any rank, which can be species, but also varieties and forms. Eight taxa from subsect. Willdenowia can be excluded because they are haplostephaneous (and can thus be assigned to sect. Imahoria J. van Raam). Of the remaining 29 diplostephaneous taxa, C. kenoyeri M.Howe and C. rusbyana can be excluded here because they are dioecious. As a haplostichous species, Chara pseudohydropitys Imahori belongs to section Aghardia R.D.Wood, and can also be excluded here. Similarly, C. foliolosa, C. tenuifolia (Allen ex R.D.Wood) R.D.Wood, C. guairensis, C. haitensis, C. indica Bertero ex Spreng., C. martiana Wallman, and C. paucicorticata Cáceres can be excluded because they have octoscutate antheridia. Unlike the specimens described here, Chara drouetii (R.D.Wood) R.D.Wood, C. michauxii (A.Braun) Kütz., and C. formosa C.B.Rob. are characterised by a sejoined gametangia arrangement. Chara cubensis Allen, C. depauperata Allen, C. oerstediana A.Braun, and C. diaphana (Meyen) R.D.Wood have fewer than four corticated branchlet segments, whereas all the specimens found in Sardinia have at least four corticated segments. According to van Raam [28], the remaining 12 taxa belong to C. zeylanica as varieties or forms based on quantitative characters, such as spine length relative to axis diameter and the length of the stipulodes.
At this stage, we can conclude that the Sardinian specimens fit the character combination of C. zeylanica. Because van Raam [28] failed to provide an adequate description of infraspecific taxa beyond offering a series of tables, the specimens discussed here will not be related to varieties or forms. In any case, the specimens clearly do not belong to C. foliolosa, which can be found in the northernmost distribution range of subsect. Willdenowia in North America [25].
However, a sound comparison between our specimens and specimens described by other authors [39,40,43,54] is still impossible because of the different taxonomic concepts applied. Taking Wood [23] as a basis, many authors [27,45,55] did not take antheridia morphology into account. Thus, it is extremely difficult to compare their data with the recent concept proposed by van Raam [28].
To obtain independent proof of the morphology-based determination, phylogenetic analyses were performed with the regularly used barcode markers rbcL and matK. The analyses classified independently the individuals from Sardinia along with other C. zeylanica. Morphologically similar species such as C. foliolosa, C. haitensis, and C. rusbyana (previously considered to be varieties or forms of C. zeylanica sensu Wood) could be excluded through alignment with GenBank sequences. The phylogenetic analyses showed that C. hydropitys, a haplostephaneous species belonging to sect. Imahoria, is closely related to the abovementioned Willdenowia species, consistent with the findings of previous studies [43,56,57]. The phylogenetic relationships between C. zeylanica and C. hydropitis were not evident based on rbcL sequence data. However, the Sardinian samples were shown to have rbcL gene sequences identical to those of a C. zeylanica individual from GenBank (HQ380481), which made the categorisation unambiguous. The assignment of the specimens to this taxon was supported by the results of a matK analysis, which showed that C. zeylanica obtained from GenBank (MT739758) formed a monophyletic clade together with the Sardinian specimens [43,56,57]. Thus, the genetic classification based on the rbcL and matK sequences clearly supported the morphological determination of the individuals collected at Cala Fuili. The phylogeny of the subsect. Willdenowia was not the main focus of this study. Nevertheless, in order to test the phylogeny of Willdenowia species in future studies, the taxonomic and geographical basis for an analysis should be broadened, and additional molecular data should be gathered.

Status and Threats
Many charophyte species and their habitats are threatened throughout Europe, and are mentioned in several national Red Lists [4]. Sardinia has a key role to play in the conservation of Characeae in the Mediterranean region [30,58,59]. Becker [30] identified numerous Sardinian hotspots for the conservation of charophytes, and proposed specific action plans that mainly focused on Characeae in brackish habitats. The Sardinian site where C. zeylanica has been found is in the hotspot area between Orosei and Capo Comino.
In contrast to rare and threatened taxa, introduced non-native species can become invasive and cause ecological damage, as the example of Nitellopsis obtusa (Desv.) J.Groves in North America shows [60]. However, the examples of two alien charophyte species with mainly intertropical distribution that were previously introduced into Europe have so far not been found to have any serious environmental impacts. Both species, Chara fibrosa (including ssp. benthamii) and Chara c.f. chrysospora, were probably introduced by humans into rice fields in Southern France and Northern Italy through the importation of contaminated rice seeds [32,46,47,61]. Moreover, while the presence of a population of Chara fibrosa ssp. benthamii was recorded on the Greek island of Crete [48], it appears that it has been extinct since 2010 [62].
Chara zeylanica cannot currently be considered an invasive species among the European charophyte flora. For the moment, the Sardinian population is very small, and is limited to a single and relatively isolated location. Although the species has a high rate of fertility in Sardinia, strong dispersal cannot be expected at this stage. Nevertheless, the development of the Sardinian population of C. zeylanica should be monitored.
Although the abovementioned intertropical species Chara fibrosa and C. c.f. chrysospora were probably introduced into Europe by anthropogenic factors [47], this is unlikely to be the case for C. zeylanica. The Sardinian site is situated more than 100 km away from the nearest rice fields. The surrounding land is used primarily for grazing sheep and small-scale tourism. Other anthropogenic dispersal pathways (e.g., fishery, bathing, or diving) also appear to be unlikely. On the other hand, Sardinia is an important interim stop for birds migrating between Europe and Africa. As the nearest previous records of the presence of C. zeylanica are from a Saharan pond in Algeria at least 88 years ago [45 and literature therein], and from Senegal and Egypt [26,27], we assume that the species was introduced into Sardinia by migrating water birds. However, against the backdrop of climate change, future investigations of C. zeylanica and other Characeae should consider whether rice fields in Sardinia and throughout the Mediterranean area play a role in the dispersal of the species.

Hydrochemical and Morphological Analyses
Hydrochemical analyses were conducted in a laboratory according to standard methods and national DIN norms, as published by Wasserchemische Gesellschaft [63][64][65]. The nutrient concentrations (NH 4 -N, NO 3 -N, total N, PO 4 -P and total P) were measured using a photometer (CADAS 200 by Dr Lange). The cation concentrations (Ca, Mg) were determined by means of an atomic absorption spectrometer (SpectrAA 55 by Varian). The pH values were analyzed using WTW Multi 3510 IDS. The conductivity, salinity, and chloride levels were determined using WTW Cond 3130, with the specific probe being applied in each case.
The morphological analysis was done by means of a stereo microscope (SZX16; Olympus, Tokyo, Japan) equipped with a digital camera for recording photographs.

DNA Barcoding
The total genomic DNA was extracted using the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. Partial sequences of the rbcL and matK genes were amplified using the primers rbcL-1a (5 -TCG TGT AAC TCC ACA ACC TG-3 ) and rbcL-1b (5 -TAC TCG GTT AGC TAC AGC TC-3 ), and matK-F2 (5 -GAA TGA GCT TAA ACA AGG ATT C-3 ) and matK-R1b (5 -GCA GCC TTA TGA ATT GGA TAG C-3 ). The PCR tests were performed in a 30 µL reaction volume with a Taq PCR Master Mix (Qiagen, Hilden, Germany) consisting of 2.5 mM MgCl 2 (final concentration), and 0.5 pmol of each primer. The PCR products were extracted from agarose gels following the protocol of the Biometra-innuPrep Gel Extraction Kit (Analytik Jena, Jena, Germany), and were sequenced directly using a 3130×L Genetic Analyser (Applied Biosystems, New York, NY, USA) with sequencing primers identical to the primers that were used for the PCR reaction. The quality of the chromatograms of the generated sequences were checked using the BIOEDIT software [66]. The nucleotide sequences identified in this study have been deposited in the GenBank (MZ648319-MZ648324).
Sequences from three specimens collected in Sardinia were submitted to the National Center for Biotechnology Information's (NCBI) Basic Local Alignment Search Tool (BLAST) [67] to allow them to be checked against the nucleotide collection in the GenBank in order to identify other Chara sequences with high scoring similarity pairs (HSP) in the NCBI web server. The phylogenetic analysis was performed with the sequence data from the Chara specimens collected in Sardinia, and with data on closely related taxa in the GenBank's nucleotide database (https://www.ncbi.nlm.nih.gov/nuccore) for both the rbcL and matK sequences separately, because the sequences available in the GenBank were completely different. Alignments were created and trimmed using BIOEDIT software [66]. Identical sequences were merged into one entry. Sequences differing only in length were also reduced to one genotype. If different taxa had identical sequences, they were retained in the alignment ( Table 1). The rbcL dataset contained 36 sequences belonging to nine species of the subsect. Willdenowia, and seven of haplostephanous species belonging to the sect. Charopsis, Protochara and Imahoria, and to the subsect. Wallmania and Agardhia. In addition, Nitellopsis obtusa was used as the outgroup (Table 1). For the matK dataset, the three Sardinien samples of C. zeylanica were analysed together with 13 sequences belonging to six species of the subsect. Willdenowia and Agardhia, and one species of the sect. Charopsis and Imahoria, respectivly. Nitellopsis obtusa was used as the outgroup (Table 1). Phylogenetic trees were created using the Maximum likelihood (ML) method and Bayesian inference (BI) analysis. The best-fit model of sequence evolution was determined using MEGA v. X [68]. The ML method was applied using MEGA v. X [68], with the HKY+G+I model used as the nucleotide substitution model for the rbcL dataset, and the GTR+G+I model used for the matK dataset. Branch supports were evaluated using 1000 bootstrap replicates (BS). MrBayes 3.2.7 [69] was used for the BI method. Two independent runs with four chains were run for 10 million generations using the MCMC method. Calculations of the consensus tree, including clade posterior probability (PP), were performed based on the trees sampled after the chains converged using Tracer 1.7 [70]. The first 25% were discarded as burn-in.