First molecular analysis of the genus Bryopsis (Bryopsidales, Chlorophyta) from Brazil, with an emphasis on the Pernambuco coast

The green algal genus Bryopsis has simple morphology and high phenotypic plasticity, making it difficult to identify its species based on morphological characteristics alone. This study evaluated the diversity of Bryopsis in northeastern Brazil (the State of Pernambuco), based on morphological and molecular data using the markers tuf A and rbc L. Molecular analyses were incongruent with morphology, demonstrating the existence of cryptic and polymorphic species in the genus. Of the four taxa cited for the area based only on morphological data ( Bryopsis corymbosa , B. pennata , B. plumosa , Bryopsis sp.), only B. pennata was recorded. Typical specimens of B. pennata and B. plumosa were grouped with low genetic divergence, 0–0.21 % for tuf A and no divergence for rbc L, indicating that B. pennata is an extremely plastic species that includes specimens with morphotype B. “ plumosa ”. Bryopsis pennata var. secunda is cited for the first time for northeastern Brazil, with divergence from the typical variety of 0.96–1.57 % for tuf A and 0.4 % for rbc L. This study showed that broader sampling of Bryopsis is necessary in order to confirm the taxonomic status of the species referenced for Brazil, whose phenotypic plasticity may cause overestimation of diversity or reveal cryptic species.


Introduction
The marine macroalgal genus Bryopsis (Bryopsidales, Chlorophyta) is one of the most diverse genera of green macroalgae with 138 species names and 84 infraspecific names, of which only 58 species names, eight varieties and five forms are currently accepted taxonomically (Guiry & Guiry 2020). Its representatives constitute a group with a predominantly marine habit, which develops mainly in areas of rocky shores and coral reefs, with temperatures ranging from five to 27 ºC (Horta et al. 2001), but they also grow in mangroves and estuaries at low salinities (Guiry & Guiry 2020). With a morphology considered simple, species of Bryopsis are characterized by non-calcified coenocytic filaments, with erect uniaxial axes with feather-like branches, called fronds, and prostate axes with a variable extension (Wynne 2005;Cremen et al. 2019). Bryopsis species have ecological importance as primary producers in 2005). However, due to the large number of species and infraspecific categories, and the difficulty of delimiting them owing to high phenotypic plasticity and overlapping of morphological characters among its representatives, the taxonomy of the genus Bryopsis has been problematic and conflicting. Consequently, the species delimitation based on morphological analysis may lead to misidentifications and misapplied names (Krellwitz et al. 2001;Wynne 2005;Morabito et al. 2010). Likewise, life-cycle and chromosomal analyses contribute to resolution at higher taxonomic levels, making them insufficient to solve taxonomic problems at the specific level (Kapraun & Shipley 1990).
Considering the ecological and economic importance of Bryopsis, and the difficulty in properly delimiting its infrageneric and infraspecific categories, besides the recognition of cryptic and polymorphic species, molecular studies using DNA barcode techniques and phylogenetic markers have been fundamental in assessing the diversity of the genus (Krellwitz et al. 2001;Lü & Wang 2011;Hollants et al. 2013;Tufiño-Velázquez & Pedroche 2019). Plastid markers have been widely used in molecular studies to delimit green algal species, including Bryopsis, mainly tufA and rbcL genes (Provan et al. 2004;Händeler et al. 2010;Hall et al. 2010;Leliaert et al. 2014;Leliaert & Lopez-Bautista 2015;Cremen et al. 2019). TufA marker (elongation factor Tu gene) obtained the best results in the universality of the primers and quality of the sequences, being proposed as a DNA barcode for green algae, and as a phylogenetic marker as well (Saunders & Kucera 2010). Its effectiveness and successfully as DNA barcode for green algae, showing good phylogenetic resolution at generic and specific level having proven by Dijoux et al. (2012), Famà et al. (2002), Ximenes et al. (2017;, followed by the rbcL (Rubisco large subunit gene) also considered suitable for phylogenetic reconstruction , having been demonstrated a sufficient level of variation to be informative in intergeneric, inter-and intraspecific studies for green and red algae as well (Freshwater & Rueness 1994;Oliveira-Carvalho et al. 2012;Leliaert et al. 2014;Ximenes et al. 2017;. Both have been widely used for other genera in the Bryopsidales, such as Caulerpa, Codium, and Halimeda (Lam & Zechman 2006;Oliveira-Carvalho et al. 2012;Kazi et al. 2013;Belton et al. 2014;Ximenes et al. 2017;, and also for other green algae (Mccourt et al. 2000;Shimada et al. 2003).
In this context, considering the morphological plasticity, the presence of cryptic species and the lack of molecular studies for the genus Bryopsis in Brazil, this work aimed to contribute to the understanding of the diversity of Bryopsis through the application of molecular analyses of the markers tufA and rbcL combined with the morphological traits of the species.

Sampling and morphological analysis
Specimens of Bryopsis were collected along the coast of Pernambuco state (PE) (northeastern Brazil,7º15'45';9º28'18" S and 34º48'35";41º19'54" W) at 13 sampling sites (Fig. 1), in the intertidal zone during low tide from 2016 to 2017. The area is located in the Western Atlantic Ocean in the Tropical Zone and present warm, transparent and oligotrophic waters (Horta et al. 2001). For morphological/molecular studies, each group of fronds was treated as an individual. Samples were stored in absolute ethanol and silica gel for morphological and/ or molecular analyses. Part of group of fronds from a same individual was separated for molecular analyses; the remaining material was used for morphological study and pressed as herbarium vouchers.
Morphological characters were analyzed using both stereoscopic microscope Leica S6D (Wetzlar, Germany) and an optical microscope Zeiss Axioskop (Oberkochen, Germany). Photographic documentation of whole specimens and details of the different portions of the thallus and branchlets were made using a stereomicroscope Zeiss Discovery-V8 A1 (Göttingen, Germany) coupled to a digital camera Axiocam 105 Color (Göttingen, Germany), with images analyzed by the software Zeiss ZEN ® (Jena, Germany); and an optical microscope Zeiss Axio Scope.A1 (Göttingen, Germany) coupled to a digital camera AxioCam MRc (Göttingen, Germany) using the software Zeiss ZEN ® .
Minimum and maximum measurements were taken for morphometric characters from a set of 10 measurements for each analyzed feature in all specimens from different sampling sites. The specimens were deposited at the herbarium Prof. Vasconcelos Sobrinho of the Universidade Federal Rural de Pernambuco (PEUFR). Additional specimens from the PEUFR were also examined.

Molecular analysis
Total DNA was extracted using CTAB (cetyl trimethyl ammonium bromide) protocol as described by Oliveira-Carvalho et al. (2012). For polymerase chain reaction (PCR), the tufA marker was amplified using the primer pair tufAF (5' TTGTTC KAACATAAA ATT GWGGTC 3') and tufA_alg_up (5' ATGATWACNGGHGCNGCWCAAATG 3') (Händeler et al. 2010), following the c ycle described by Händeler et al. (2010). For the marker rbcL was used the primer pair: F623-603 (5' TCWCAACCHTYTATGCGTTGG 3') (Curtis et al. 2008) and R1396-1372 (5' AATTTCTTTCCAAACTTCACAAGC 3') (Lam & Zechman 2006), following the cycle described by Curtis et al. (2008). PCR was performed using illustra PuReTaq Ready-To-Go PCR Beads kit (GE Healthcare, Buckinghamshire, UK) following the manufacturer's instructions. The fragments were amplified in a Techne TC-4000 thermocycler (Bibby Scientific Ltd, Staffordshire, UK). All PCR products were analyzed by electrophoresis in 1 % agarose to check product size and were purified using the GFXTM PCR DNA and Gel Band Purification kit (GE Healthcare, Buckinganshire, UK), following the manufacturer's instructions. Purified amplicons for both markers were sequenced using the BigDyeTM Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems, Foster City, USA), with the same primers of PCR on an ABI PRISM 3730 Genetic Analyzer (Applied Biosystems). Sequences were checked using the BlastN algorithm, through the NCBI online platform (Altschul et al. 1997). Consensus sequences and multiple alignments for both tufA and rbcL sequences were constructed using the computer program BioEdit v7.0.4.1 software (Hall 1999). For each marker, a matrix was created with the sequences generated in this study plus those available in the GenBank database used in the analyses (Tab. S1 in supplementary material).
Molecular analyzes were performed using the Neighbor-Joining (NJ) distance method in PAUP 4.0b10 program (Swofford 2002) with 2000 boostrap replicates. For phylogenetic analysis, the most appropriated model of sequence evolution for maximum likelihood (ML) and Bayesian inference (BI) was selected using jModeltest v2.1.10 (Darriba et al. 2012) under the Akaike Information Criterion (AIC) implemented on the online server CIPRES Science Gateway v3.3 (Miller et al. 2011). The model selected for tufA and rbcL was the general time-reversible model of nucleotide substitution with invariant sites and gamma distributed rates for the variable sites (GTR + I + G). The maximum likelihood analysis (ML) was performed using IQ-Tree v1.4.3 (Nguyen et al. 2015) with 1000 boostrap replicates on the IQ-Tree web portal. Bayesian Inference (BI) was performed using MrBayes v.3.2.2 program (Ronquist et al. 2012). For BI analysis, two runs with four chains of the Markov chain Monte Carlo (one hot and three cold) were used, sampling one tree every 1,000 generations for 4,000,000 generations, starting with a random tree. We discarded the first 50,000 generations in both runs as the burn-in to build the consensus tree. After discarding the trees associated with "burn-in", a consensus tree was built for both tufA and rbcL markers. In all analyses, gaps were considered as missing data. For both tufA and rbcL matrices, the percentages of intraspecific and interspecific divergences were calculated using uncorrected 'p' distances in PAUP.

Molecular analysis
Thirty-one new Bryopsis sequences were obtained from specimens collected on the Pernambuco coast, 19 sequences for tufA and 12 sequences for rbcL. For the tufA marker, 53 sequences were used with an alignment of 733 bp, and of the 34 sequences obtained from GenBank, three were used as outgroups, Codium tomentosum Stackhouse The tufA analyses recover Bryopsis as monophyletic, with high support for NJ and moderate for ML (Fig. 2). The Brazilian sequences formed two subclades with high ML bootstrap support and high Bayesian posterior probability (PP). The major subclade grouped most of the Brazilian samples, identified as B. "plumosa" and B. pennata, with high support of ML and PP, and in which three other unidentified Bryopsis sequences (Bryopsis sp. from Oahu, USA, and Bryopsis sp.1, collected from a marine aquarium) were grouped. Except for the sample B. "plumosa" (102), all sequences of this subclade were 100 % identical, including Bryopsis sp. and Bryopsis sp.1 from GenBank. The divergence between the sample 102 and others of this subclade ranged from 0.15 % (102 vs. B. sp. and B. sp. 1) to 0.21 % (102 vs. 129). The low divergence of this subclade (0-0.21 %) indicates that all samples belong to the same genetic species. The second subclade was formed by two Brazilian samples of B. pennata var. secunda grouped with an unidentified sample of Bryopsis, named as B. sp. 28 from the Philippines, all 100 % identical, and with high to moderate support. The lack of genetic divergence between the sequences of this subclade supports the conclusion that they all correspond to the same genetic species. These two subclades, B. "plumosa"-B. pennata -B. sp.1 -B. sp. and B. pennata var. secunda -B. sp.28, differed from each other from 0.96 % to 1.57 %. For unidentified samples downloaded from GenBank that were grouped in these two subclades (B sp., B. sp.1 and B. sp.28), there are no descriptions (Leliaert et al. 2014;Wade & Sherwood 2017) that allow morphological comparison with the Brazilian material.
Four tufA sequences of Bryopsis plumosa from GenBank were included in the analyses, and this species was shown to be paraphyletic (Fig. 2). The sample of B. plumosa from Argentina (JQ755423) was not grouped with the others from France (LN810504, FJ432653) and with another whose sampling site was not specified (U09424). These latter three sequences were split into two subclades (Fig. 2). The sequence of France (LN810504, Traon Erch, Brittany) corresponds to the complete sequencing of the chloroplast genome done by Leliaert & Lopez-Bautista (2015), and its sampling site corresponds to a region near to the type locality (Exmouth, Devon, England). The sequence LN810504 formed a subclade with Bryopsis sp. from the USA (HQ610244), being 100 % identical, and with a more divergent sequence of B. corymbosa from France (1.6 %). The other Bryopsis plumosa sequence from France (FJ432653, Pas de Calais, Boulogne-sur-Mer), generated by Verbruggen et al. (2009), is also near the type locality; however, it formed another subclade with B. plumosa (U09424), whose divergence between them was 0.99 %. The two sequences of B. plumosa from France showed high divergence (3.3 %), indicating that that belong to separate genetic species. The sample of B. plumosa from Argentina (JQ755423) proved to be highly divergent from the other sequences of B. plumosa, varying from 7.0 % to 7.2 % (Tab. 1). (Bryopsidales, Chlorophyta) from Brazil, with an emphasis on the Pernambuco coast    The samples identified as B. "plumosa" collected at different sites along the Pernambuco coast grouping with Bryopsis pennata samples (see above) were highly divergent from both B. plumosa from France, varying from 12.2 % to 13.6 % from LN810504, and from 12 % to 13.8 % from FJ432653, although morphologically typical plants of this species have been observed.

First molecular analysis of the genus Bryopsis
The eight samples identified as Bryopsis pennata positioned in this same clade, also with wide distribution on the Pernambuco coast, could not be compared molecularly with any other B. pennata tufA sequence because there are no sequences of this marker available in the databases. With the possibility of citing B. plumosa among our samples being discarded by molecular data, we maintained the identification of these samples as B. pennata, also supported by morphological analysis, with typical plants observed in different samples. Our results showed that B. pennata is extremely polymorphic, whereas the genetic divergence between samples was very low. Samples of B. pennata var. secunda also diverged in high percentage values from the B. plumosa from France (12.5-12.8 %, Tab. 2). The interspecific divergence observed between Brazilian samples and those from GenBank considered in this study ranged from 3.3-16.2 % for tufA.
For rbcL, 66 sequences were used with an alignment of 1284 bp. Of the 54 sequences obtained from GenBank, Pseudoderbesia sp. (LK022434) was used as an outgroup (Fig. 3). The genetic divergences presented for rbcL were calculated on the final alignment of 1284 bp, from which the consensus tree was generated (Fig. 3). A second rbcL alignment was constructed with a larger number of Bryopsis sequences available on GenBank in order to ascertain the phylogenetic position of Brazilian samples using a larger sampling. These alignments consisted of a total of 147 sequences and, due to many sequences being partial for this marker, a shorter alignment with 689 bp was generated, from which an ML analysis was made ( Fig. S1 in supplementary material). The results obtained in the two analyses were similar, with the Brazilian samples showing the same phylogenetic position (Fig. 3, Fig. S1 in supplementary material).
The rbcL consensus tree (Fig. 3) showed that the monophyly of Bryopsis was not supported by this gene. However, Brazilian sequences were grouped similarly to the results obtained with tufA, with high support for all analyses. The samples identified as B. pennata and B. "plumosa" are 100 % identical. The two Brazilian sequences of B. pennata var. secunda (100 % identical) were grouped into a subclade with two GenBank sequences, namely Bryopsis sp. 28, one from Tanzania (HF583394) and another from Kenya (HF583393), with high support for ML and PP. Brazilian samples of Bryopsis pennata-B. "plumosa" diverged from the var. secunda by 0.4 %. The divergence within the subclade formed by B. pennata var. secunda and Bryopsis sp. 28 was very low, ranging from 0 % (between B. pennata var. secunda and B. sp. 28 from Tanzania) to 0.15 % (between B. pennata var. secunda and B. sp. 28 from Kenya), indicating that they are the same genetic species. The genetic divergence between samples of B. pennata-B. "plumosa" and B. pennata var. secunda-B. sp. 28) for rbcL ranged from 0.27-0.4 %, being lower than that observed with tufA (Tab. 3).
Eight rbcL sequences of Bryopsis plumosa from GenBank were included in the analyses (Fig. 3). These sequences formed four distinct clades, indicating that B. plumosa is not monophyletic, corroborating the results of tufA. Among the eight sequences analyzed, one is from Traon Erch, Brittany, France (LN810504), a region near to the type locality (Exmouth, Devon, England). The genetic divergence between replace by this sample from France and the Brazilian ones (B. "plumosa"-B. pennata) was high, 8.0 %, whereas the divergence with the samples of B. pennata var. secunda was slightly higher, 8.6 %. Another sequence of B. plumosa from France near the type locality (Pas de Calais, Boulogne-sur-Mer, FJ432637, Verbruggen et al. 2009) was positioned in a distinct clade, and diverged from B. "plumosa"-B. pennata by 6.7-6.9 % and from B. pennata var. secunda by 7.1-7.2 %. The genetic divergences observed between the Brazilian material and the B. plumosa samples from GenBank from different regions (France, Australia, Argentina and Japan) are shown in Table 3.  First molecular analysis of the genus Bryopsis (Bryopsidales, Chlorophyta) from Brazil, with an emphasis on the Pernambuco coast   An rbcL sequence of Bryopsis pennata, with 562 bp (DQ469323) is available in GenBank. This sample was collected in the Virgin Islands (St. Thomas) corresponding, therefore, to a sequence of the type locality (Caribbean Sea). However, the partially amplified region of this sequence was not coincident with our sequences, not allowing a comparison between them. The intraspecific divergence observed for different taxa included in the analysis with rbcL was low. There was no genetic divergence between the four sequences of B. corymbosa from Spain and France (Fig. 3); clades formed by eight B. myosurioides sequences from South Africa, and four B. vestita sequences from New Zealand showed both 0-0.2 % of divergence; the clade with four B. cf. hypnoides 2 sequences from USA and Netherlands ranged from 0.08-0.2 %, while the four sequences that formed the B. corticulans clade from USA and Canada showed slightly higher intraspecific divergence, 0-0.3 %. However, the range of intraspecific variation for these taxa, collected in different geographical areas (0-0.3 %), remained below, although close to the limit of the divergence observed between the Brazilian sequences of B. pennata and B. pennata var. secunda, whose genetic divergence was 0.4%. Despite the lower divergence observed between these two taxa for rbcL (0.4 %) compared to tufA (0.96-1.57 %), the clades formed by B. pennata and B. pennata var. secunda were the same for both markers. The interspecific divergence observed between Brazilian samples and those from GenBank considered in this study ranged from 6.7-11.6 % for rbcL (Tab. 3).

Morphological analysis
Based on molecular analyses, two Bryopsis taxa were identified for the Pernambuco coast: B. pennata and B. pennata var. secunda. Twenty-one samples were examined; 19 were identified as B. pennata (including the morphotype B. "plumosa") and two as B. pennata var. secunda. A comparison of the morphological characters for these taxa is shown in Table 4 For list of material examined and additional material examined see List S1 in supplementary material.
Description: Erect plants, forming small dense tufts, up to 8.0 cm high (Fig. 4A). Filamentous, coenocytic, cylindrical, flaccid in texture, dark green to light green in color, attached to the substrate by rhizoids. Frond with a linear to lanceolate outline (Fig. 4C-D). Erect filaments usually with an order of lateral branches (branchlets) of approximately uniform length along the axis, gradually smaller towards the apex of the thallus (Fig. 4E-I). Filaments generally naked in the lower portion of the thallus and with opposite branchlets, distichous or slightly alternate in the upper third, in some cases with discontinuous branching (Fig. 4J), or with branchlets arranged unilaterally (Fig.  4M). Cylindrical branchlets, without septa, constricted at the base and with obtuse apex, measuring 402-510 µm in length in the basal portions, 400-480 µm in the middle, and 20-44 µm in the apical portions. Fertile specimens not observed.
Remarks: Plants of Bryopsis pennata showed great phenotypic plasticity. Samples previously identified as B. plumosa (Fig. 5A-O) based on morphological data were grouped in the same clade as B. pennata by the two markers used (tufA and rbcL), with low intraspecific divergence. Plants with morphotype B. "plumosa" (Fig. 5A-O) are densely tufted, up to 12 cm high (Fig. 5A), with more dense branching in the apical portions (Fig. 5C-D), or even scarce of very unequal size. Characteristically, the fronds have a pyramidal outline, caused by the decrease in the length of the branchlets towards the apex (Fig. 5F-G). Erect filaments with one or more orders of branches distichously pinnate to bipinnate, or with slightly alternating, irregular or discontinuous branching. Branchlets are often longer in morphotype B. "plumosa", measuring 462-1040 µm in length in the basal portions, 574-1240 µm in the middle, and 354-480 µm in the apical portions (Tab. 4).   Description: Erect plants forming dense tufts up to 2.0 cm high (Fig. 6A-B). Filamentous, coenocytic, cylindrical, flaccid in texture, dark green in color, attached to the substrate by rhizoids. Frond formed by two rows of branchlets with unilateral arrangement, often with the apices incurved (Fig. 6H). Erect filaments with an order of lateral branches (branchlets) of approximately uniform length, gradually smaller towards the apex (Fig. 6C-F). Filaments commonly naked in the lower third of the thallus and with opposite branchlets, distichous e/or alternate in the upper third of the thallus (Fig. 6E-F Remarks: Bryopsis pennata var. secunda is easily confused with young thalli of B. pennata due to its small size; however, it presents distinctive features such as apices of the fronds incurved and two rows of branchlets overlapping unilaterally. There is no record of this taxon for the north and northeast of Brazil, only for the states of Espírito Santo, Paraná and Santa Catarina (INCT 2020). Our examples agree morphologically with the records of Silva et al. (1996); Russell (2000); Coppejans et al. (2004). This is its first citation for northeastern Brazil.

Discussion
Previous studies on the genus Bryopsis for the Pernambuco coast, as well as for the entire Brazilian coast, were based exclusively on morphological data (Pereira & Accioly 1998). The present work constitutes the first molecular study of Bryopsis for Brazil, in which the first sequences of tufA and rbcL were generated from material collected on the Pernambuco coast. Two species were previously cited for the state of Pernambuco, B. pennata and B. plumosa (e.g., Pereira & Accioly 1998;Carvalho et al. 2013;Barros 2013). Different authors have reported the difficulty in delimiting Bryopsis species due to the high phenotypic plasticity and the overlapping of morphological characters, leading them to quote their specimens without specific identification, as Bryopsis sp. (Ribeiro et al. 2008;Burgos 2009;Barros 2013;Carvalho et al. 2013). The occurrence of B. corymbosa for Pernambuco coast (Gaibu Beach, det. Y. Ugadim, SPF-Algae 9527) may be a misidentification, in that there is only a single record based on morphological data. In addition, this species may be confused with the morphotype B. "plumosa" because they share morphological characteristics such as the irregular size of the branchlets and the branches with irregular growth. However, they differ by the gradual decrease of the branchlets length towards the apex, giving a pyramidal outline to the thallus in B. plumosa, and by the gradual decrease of the branchlets length towards the apex, giving a corymbose outline in Table 4  B. corymbosa (Cormaci et al. 2014), although, according to Lee et al. (1991), intermediate characteristics are common, making it difficult to distinguish these species. In the present study, none of the specimens analyzed morphologically fit the description of B. corymbosa or were grouped with sequences of this species available in the databases. The sequences of B. corymbosa included in the molecular analyses are from the western Mediterranean (France and Spain), regions near the type locality (Livorno, Italy), and they were positioned phylogenetically distant from the Brazilian samples. Thus, the occurrence of B. corymbosa was not confirmed for the Pernambuco coast. Our molecular and morphological results allowed us to register only the occurrence of B. pennata var. pennata and B. pennata var. secunda. Samples of B. pennata analyzed in this study proved to be extremely plastic, and the strictly morphological analyses led us to previously identify some of our specimens as B. plumosa. Table 4 compares the morphological characteristics of the studied material in regard to habit, distribution and length of branchlets. First molecular analysis of the genus Bryopsis (Bryopsidales, Chlorophyta) from Brazil, with an emphasis on the Pernambuco coast Our analyses showed that there is an overlap in most of the compared characteristics; however, B. pennata specimens had an average thallus size smaller than that identified as B. "plumosa"; the latter also presented, in general, branches distributed discontinuously and longer, and often naked axes in its basal region. The diagnostic characteristics that separate B. pennata from B. plumosa are the habit and the size of the branchlets along the frond. In B. pennata, the thallus is lanceolate with branchlets uniformly sized, except near the apex, whereas in B. plumosa the thallus is pyramidal with branchlets non-uniformly sized, gradually smaller towards the apex (Krellwitz et al. 2001). In a study carried out in Praia de Serrambi (Pernambuco), Accioly (1989) observed that B. pennata occurred more frequently in the subtidal zone during the rainy season and B. plumosa in the intertidal zone during the dry season. However, our results showed that specimens of B. pennata and the B. "plumosa" co-occurred in some sampling sites (Cupe, Ilha de Santo Aleixo, Enseada dos Corais) at the same distribution range (intertidal) and essentially in dry season, and it was not possible to establish any distribution pattern, even in the intertidal zonation, or attribute the morphological variations observed to differences in environmental conditions. Although plants morphologically typical of B. pennata and B. plumosa could be recognized in the studied material, our molecular results for the two markers used, clearly refuted the occurrence of B. plumosa for the Pernambuco coast because B. pennata-B. "plumosa" were grouped with very low genetic divergence, 0-0.21 % for tufA, and no divergence for rbcL. Our molecular analyses showed that B. plumosa is polyphyletic with a high genetic divergence (Tabs. 1, 3), corroborating the polyphyletism previously highlighted for this species by Ciancia et al. (2012). None of the Brazilian sequences previously identified as B. plumosa matched with samples of B. plumosa available in the databases, including sequences of the genetic species of B. plumosa from the French coast of the British Channel near the type locality (Devoshire, England) and whose divergence was high (12-13.6 % for tufA and 6.7-8.4 % for rbcL), leaving no doubt that B. plumosa does not occur in the studied area. Thus, due to the phylogenetic positioning, the low genetic divergence observed between the B. pennata and "B. plumosa" and the identification of specimens morphologically typical of B. pennata in the analyzed material, we chose to cite our specimens under the name of B. pennata. Future comparisons with sequences from the type locality of B. pennata are necessary to confirm this taxon for the Pernambuco coast and Brazil. Likewise, the sample from Argentina (JQ755423) with high genetic divergence from others B. plumosa (7.0 % to 7.2 % for tufA and 2.3-2.6 % for rbcL, Tabs. 2, 3) was generated by Ciancia et al. (2012) having been described, illustrated and sequenced for tufA and rbcL genes by these authors, confirming its identification and reinforcing to be B. plumosa, a complex of cryptic species. The recognition of cryptic and polymorphic species for Bryopsis has already been reported in previous molecular studies (Krellwitz et al. 2001;Lü & Wang 2011;Hollants et al. 2013;Tufiño-Velázquez & Pedroche 2019). Similar to that observed in our study in which B. pennata was shown to be polymorphic and included the morphotype B. "plumosa", Tufiño-Velázquez & Pedroche (2019), using the plastid psbB gene, concluded that B. pennata is the only species present in the Mexican Atlantic, and that the citations under the name of B. hypnoides and B. plumosa are names misapplied or uncertain, and even considered the possibility of B. ramulosa and B. halliae as being synonymous with B. pennata.
The analyzed specimens of B. pennata var. secunda showed diagnostic characteristics of this variety, differing from the typical variety by the small size of the frond (1.5 to 2.0 cm high in B. pennata var. secunda and 2.5 to 12 cm in B. pennata var. pennata), incurved apex, and unilateral distribution of branchlets (Tab. 4, Fig. 6C-I). Although there are no sequences of B. pennata var. secunda from the type locality (Key West, Florida, USA), available in the databases for comparison, we consider that the phylogenetic position and the genetic divergence observed between B. pennata var. secunda and the typical variety, mainly for the DNA barcode marker tufA (0.96-1.57 %), support the maintenance of the taxonomic status of Brazilian samples as var. secunda, a decision also supported by the morphological data highlighted above. This is the first citation of B. pennata var. secunda for northeastern Brazil. This variety was previously cited only for the southeast (Espírito Santo state) and southern (Paraná, Santa Catarina states) regions of Brazil (INCT 2020).
Our results from the study of the Brazilian material showed that tufA was more efficient in delimiting species (and infraspecific categories) than rbcL, confirming its use as a DNA barcode for green algae. The molecular data obtained so far for Bryopsis showed that a taxonomic revision for the genus is necessary, including a wide sampling on the Brazilian coast, in order to unveil the true diversity of the genus and clarify the taxonomic entities referenced for Brazil.