The monophyletic Rhodophyta has been divided into seven classes based on multigene data (Yoon et al., 2006). Among the seven classes, the Florideophyceae is the largest and well known red algal class, comprising 95% of currently described species (Guiry and Guiry, 2021). A series of molecular phylogenetic studies by Saunders et al. (2002), Withall and Saunders (2006), and Le Gall and Saunders (2007) provided many improvements for resolving relationships among florideophyte orders. Recently, the close relationships of Halymeniales, Sebdeniales, and Rhodymeniales were confidently resolved based on multigene and mitochondrial genome data (e.g., Yang et al., 2015, 2016). However, to date, there have been no extensive molecular phylogenetic studies of the Halymeniales as a whole.

In 1996 Saunders and Kraft recognized a new name, the Halymeniales, for the red algal order that was previously known as the Cryptonemiales Kylin, based on small-subunit rRNA phylogenetic analyses. A historical account of the modified taxonomic concepts with regard to the Cryptonemiales was provided in detail by Silva and Johansen (1986), Silva (2002), and Krayesky et al. (2009). The emended order Halymeniales first accommodated the families Halymeniaceae Bory and Sebdeniaceae Kylin, and was characterized by taxa with a multiaxial thallus, non-procarpic female reproductive development, outwardly directed carpogonial branches and intercalary auxiliary cells (Saunders and Kraft, 1996). Subsequently, Saunders and Kraft (2002) transferred the genus Tsengia K. C. Fan and Y. C. Fan from the Nemastomatales to the Halymeniales with the establishment of the new monogeneric family Tsengiaceae. The Sebdeniaceae was later removed from the Halymeniales and elevated to a new order, Sebdeniales Withall and Saunders (Withall and Saunders, 2006) as suggested by Gavio et al. (2005). In the current molecular taxonomic framework, the order Halymeniales includes two families, the Halymeniaceae and Tsengiaceae, consisting of 38 genera and about 358 species (Guiry and Guiry, 2021).

The Halymeniaceae, which includes most of the taxa in the Halymeniales except for 12 species of Tsengiaceae, was established based on the genus Halymenia C. Agardh along with four other genera (Bory de Saint-Vincent, 1828). Harvey (1849) suggested the family Cryptonemiaceae based on the tribe Cryptonemieae J. Agardh, including Cryptonemia J. Agardh, Grateloupia C. Agardh, Halymenia, and 21 other genera. The family name Grateloupiaceae was proposed by Schmitz (1889), as a part of the suborder Cryptoneminae, based on 11 genera (Grateloupia, Cryptonemia, Halymenia, Aeodes J. Agardh, Carpopeltis F. Schmitz, Corynomorpha J. Agardh, Dermocorynus P. Crouan and H. Crouan, Pachymenia J. Agardh, Polyopes J. Agardh, Prionitis J. Agardh, and Thamnoclonium Kützing). Later, Papenfuss (1955) synonymized the Grateloupiaceae with the Cryptonemiaceae; however, Guiry (1978) reinstated the Halymeniaceae Bory based on nomenclatural priority. As a consequence, the Grateloupiaceae and Cryptonemiaceae have been considered synonyms of the Halymeniaceae.

A diagnostic feature of the Halymeniaceae is the presence of auxiliary cell ampullae and the special accessory branch system in the female reproductive structure (Chiang, 1970). The details of the female reproductive morphology, such as the type and development of the auxiliary cell ampullae before or after fertilization, have been regarded as a delimiting character among halymeniacean genera (Kawaguchi et al., 2004; De Clerck et al., 2005a; Lin et al., 2008; Gargiulo et al., 2013). The other female reproductive features, such as the absence or presence of a pericarp, location of carpogonial branches, and number of cells in carpogonial branch ampullae, have also been suggested as potentially diagnostic for genera in the Halymeniaceae (e.g., Balakrishnan, 1961a; Hommersand and Fredericq, 1990; Kawaguchi et al., 2004; Rodríguez-Prieto et al., 2018). In addition, monoecy or dioecy of the gametophyte was referred to as one of the noticeable characteristics for inferring the phylogenetic relationships within the Halymeniaceae (Guiry and Maggs, 1982; Womersley and Lewis, 1994; Lee and Kim, 2019). In contrast, vegetative features such as overall habit and thallus texture show little taxonomic utility in the family (Wang et al., 2001; Wilkes et al., 2005; Rodríguez-Prieto et al., 2018).

The genus Grateloupia C. Agardh is the largest and most taxonomically complex group in the Halymeniaceae (Gargiulo et al., 2013). In the early 2000s, several genera were merged into Grateloupia based on rbcL phylogenetic analyses and Grateloupia-type auxiliary cell ampullae (Wang et al., 2001; De Clerck et al., 2005b; Wilkes et al., 2005). However, Lin et al. (2008) suggested two differentiating types of auxiliary cell ampullae (Grateloupia taiwanensis- and G. orientalis- type) and emphasized the importance of careful observation of the developmental stages. Subsequently, Gargiulo et al. (2013) provided a new interpretation of this clade in terms of the type of ampullae and post-fertilization events, suggesting that some species be separated into distinct genera. As a result, the genera, i.e., Dermocorynus, Prionitis, Pachymeniopsis Yamada ex Kawabata, and Phyllymenia J. Agardh, which had been subsumed into Grateloupia by previous authors, were resurrected. Despite the resurrection of several genera, many species still have the generic name “Grateloupia.” In the historical context of this classification, Grateloupia sensu lato includes not only Grateloupia species, but also Dermocorynus, Prionitis, Pachymeniopsis, and Phyllymenia. In addition, Kintokiocolax Tak.Tanaka and Y. Nozawa and Yonagunia S. Kawaguchi and M. Masuda have been regarded as members of Grateloupia sensu lato based on the rbcL phylogenetic analyses (Calderon et al., 2014a; Yang and Kim, 2015). Later, two new genera, Neorubra M. S. Calderon, G. H. Boo, and S. M. Boo and Mariaramirezia M. S. Calderon, G. H. Boo, A. Mansilla, and S. M. Boo, were added to the Grateloupia sensu lato clade (Calderon et al., 2014a, 2016).

Regarding earlier systematic studies, two critical points have yet to be resolved in the Halymeniales. First, molecular studies were conducted primarily using only plastid rbcL gene sequences, a gene that has been shown to be promising for the taxonomic study of Halymeniales (e.g., Lin et al., 2008; Gargiulo et al., 2013; Calderon et al., 2014a, b). Multigene phylogenies are needed to examine deep nodes within the order (i.e., resolving family-level taxonomy), as shown in other groups of red algae (e.g., Dixon et al., 2015; Boo et al., 2016; Entwisle et al., 2018). Second, most studies have focused on few species and genera, so taxon sampling is limited. Further studies involving more species from a suite of genera are required in order to increase our understanding on the phylogeny of the Halymeniales.

In this study, we reconstructed the phylogenies of the Halymeniales by using (i) a taxon-rich rbcL data set with 45 new sequences and (ii) representative taxa using five genes (plastid rbcL, psaA, psbA, mitochondrial cox1, and nuclear LSU rRNA) with 153 new sequences for a combined data set analysis to test the monophyly of the Grateloupia sensu lato clade and to clarify the phylogenetic relationships among the clade and other genera of the Halymeniales. In addition (iii) morphological observations focused on the reproductive structures to clarify the taxonomic delimitation corresponding to our multigene phylogeny, especially to enable the distinction between Grateloupia sensu lato from the Halymeniaceae (exemplified by the generitype Halymenia). Consequently, herein we reinstate the family Grateloupiaceae, which was previously established by Schmitz (1889), and emend the taxonomic concept of the Grateloupiaceae to include nine genera.

We consider that the resurrection of the monophyletic Grateloupiaceae is a first step toward resolving the Halymeniales with the aim of adapting the classification system based on genetic monophyly. In addition to all available rbcL sequences from previous studies (e.g., Wang et al., 2001; Gargiulo et al., 2013; Calderon et al., 2014a, b), we determined 45 new rbcL sequences. Our rbcL data set covers more than 60% (82 species out of 138) of currently described species from Grateloupiaceae. On the other hand, it is noteworthy that the genera of the Halymeniales, except for the Grateloupiaceae and Tsengiaceae, were not resolved as a monophyletic lineage in the present and previous analyses (Manghisi et al., 2014, 2017). We believe that this is due to insufficient taxon sampling from Halymenia-Cryptonemia related genera caused by two reasons: known species that are not analyzed and taxa that are not yet known. Although Cryptonemia (51 spp.) and Halymenia (78 spp.) account for more than 30% within the Halymeniales (358 spp.) and are the second and third largest genera in the order, few studies have been conducted on these genera compared to Grateloupia. In addition, recent studies have continuously discovered new taxa within this group (e.g., Rodríguez-Prieto et al., 2018; Lee and Kim, 2019; Schneider et al., 2019). We expected that extensive studies using taxon-rich sampling focused on Halymenia-Cryptonemia and related genera could resolve the non-monophyly of the Halymeniaceae, but this is beyond the scope of this study. At this point, we tentatively used the “Halymeniaceae sensu lato” for the genera of Halymeniales except for the Grateloupiaceae and Tsengiaceae (Figures 2, 3). A more comprehensive study on the Halymeniaceae sensu lato is pending in a future study.

There have been no extensive molecular phylogenetic studies conducted of the Halymeniales as a whole so far. However, Wang et al. (2001) discussed the phylogeny of Halymeniales based on rbcL sequence analyses. They suggested the Pachymenia-Aeodes clade was basal within the Halymeniaceae followed by the Cryptonemia-Halymenia complex clade, a Polyopes clade, then a Grateloupia-Prionitis clade. This was the first extensive study for the Halymeniales; however, they rather focused on few specific genera (e.g., Prionitis). Although these supergeneric clades have been recovered in subsequent systematic studies on the Halymeniaceae (e.g., Nelson et al., 2014; Manghisi et al., 2017), those studies were based on a limited number of genera, and the Tsengiaceae was used as an outgroup or was not included. Meanwhile, Saunders and colleagues suggested several familial and ordinal level changes based on limited genera (Figures 1A–C). To overcome limitations in previous studies, we extensively included 33 genera out of 38 with 30 type species of the Halymeniales except for five genera of Halymeniaceae. The taxonomic status of these five genera are unstable (Figure 1). For instance, the monotypic genera Blastophye, Hymenophlea, and Zanardinula have been regarded as a synonym of Prionitis, Neoabbotiella, which has been referred to as being part of Cryptonemia (Guiry and Guiry, 2021), and Grateloupiocolax, which has never been studied after the first description from the Caribbean coast (Schnetter et al., 1983).

Based on this taxon-rich rbcL phylogeny, we designated the representative taxa and performed a multi-gene phylogeny of the Halymeniales. Phylogenetic analyses of a five-gene combined data set have shown that the Tsengiaceae diverged first (100% MLB; Figure 3) within the Halymeniales followed by Pachymenia and Aeodes. The basal position of Pachymenia and Aeodes within the Halymeniaceae is largely congruent with the phylogenetic concepts of Chiang (1970) and Fredericq et al. (1996) based on reproductive morphology, as well as previous molecular phylogenies (e.g., Wang et al., 2001). After the divergence of Pachymenia and Aeodes, three monotypic genera have distinct phylogenetic positions (Isabbottia, Norrissia, and Zymurgia; Figure 3). Among the other genera of Halymeniaceae sensu lato, the sister relationship of Polyopes with Glaphyrosiphon, which have been suggested as a sister taxon in previous studies (Hommersand et al., 2010; Manghisi et al., 2014), was fully supported (100% BTS). Lastly, the Grateloupiaceae was recovered as a monophyletic clade with full support (100% MLB; Figure 3). The close relationships among the species of the Grateloupiaceae and their differences to the other species of Halymeniaceae were also supported by several morphological characteristics in common regarding female and tetrasporangial developments: type of auxiliary cell ampullae (Chiang, 1970), pericarp origin (Balakrishnan, 1961a, b), and origin of tetrasporangia (Figures 3, 4).

The members of the Halymeniaceae and Grateloupiaceae share the following reproductive features: 2-celled carpogonial branches, similar ampullae structures producing carpogonial branches and auxiliary cells, formation of a pericarp, and cruciately divided tetrasporangia (Chiang, 1970; Womersley and Lewis, 1994; Rodríguez-Prieto et al., 2018). The monogeneric family Tsengiaceae has no auxiliary cell ampullae (Fan and Fan, 1962), which makes it distinct from the Grateloupiaceae and Halymeniaceae. The auxiliary cell ampullae of the Grateloupiaceae consist of two-or three-ordered ampullary filaments mostly having one or two, sometimes 3 s-order ampullary filaments branched laterally from the first-order ampullary filament (Figures 4A–C; Kawabata, 1962, 1963; Kawaguchi, 1989; Kawaguchi et al., 2004; Lin et al., 2008). Although there are different interpretations for the development of auxiliary cell ampullae in some species groups of Grateloupiaceae (Gargiulo et al., 2013; Calderon et al., 2014a, b), we consider that the type of auxiliary cell ampullae present in the Grateloupiaceae is of the simple lateral branched type regardless of the number of ampullary cell orders and the position of the auxiliary cell, a situation that can be distinguished from the other halymeniacean taxa. In the case of Halymenia floresii, the species has the Halymenia-type auxiliary cell ampullae, 3–4-ordered auxiliary cell ampullae producing two and more second-order ampullary filaments laterally from the first-order ampullary filament, with several short third-order ampullary filaments (Chiang, 1970; Rodríguez-Prieto et al., 2018). Compared to other members of the Halymeniaceae such as Aeodes, Cryptonemia, Glaphyrosiphon, Pachymenia, Polyopes, and Thamnoclonium, the auxiliary cell ampullae of the Grateloupiaceae are distinctly simple enough to be distinguished phylogenetically (Table 1).

The pericarps of the Grateloupiaceae are formed by the fusion of secondary medullary filaments which are initiated from subcortical cells, and of lateral ampullary filaments borne from a fusion cell complex (Figures 4G–I; Balakrishnan, 1961a; Kawabata, 1962, 1963; Kawaguchi, 1989; Lin et al., 2008; Gargiulo et al., 2013). However, although the pericarps of Codiophyllum, Corynomorpha, Pachymenia, Polyopes, and Thamnoclonium, taxa belonging to the Halymeniaceae clade (Figure 2), have been described as formed by elongated ampullary filaments and neighboring medullary filaments (Balakrishnan, 1962; Chiang, 1970; Womersley and Lewis, 1994; Kawaguchi et al., 2002), their involucral networks are looser than those of the Grateloupiaceae. In the members of Cryptonemia, the pericarps are mostly consisting of elongated lateral ampullary filaments, or sometimes of lateral ampullary filaments and neighboring medullary filaments (Kylin, 1925; Chiang, 1970; Lewis, 1994; Acleto and Zúñiga, 2011). The pericarps of the members of Halymenia are produced only by ampullary filaments originating from a fusion cell complex (Rodríguez-Prieto et al., 2018). In other Halymeniaceae sensu lato taxa such as Amalthea, Galene, Glaphyrosiphon, Neofolia, and Nesoia, the pericarps are produced by elongated lateral ampullary filaments only (Hommersand et al., 2010; D’Archino et al., 2014; Rodríguez-Prieto et al., 2018, 2019, 2020; Lee and Kim, 2019).

Tetrasporangia in the Grateloupiaceae are initiated laterally from inner cells of the outer cortex and grow outwardly between dichotomously branched outer cortical filaments such as the other cell files of the outer cortex (Figures 4J–L). Regarding tetrasporangia, type of division (i.e., cruciate or tetrahedral), shape, and size have been used to describe species and classify halymeniacean taxa (e.g., Kawaguchi et al., 2001; Saunders and Kraft, 2002; Rodríguez-Prieto et al., 2018). However, in our study we focused on the initiation of tetrasporangia. The tetrasporangia of Halymenia including Halymenia sensu stricto and Cryptonemia members are ovoid and similar in size with the outermost cortical cells produced from subcortical cells (Chiang, 1970; D’Archino et al., 2014; Rodríguez-Prieto et al., 2018, 2019, 2020; Lee and Kim, 2019). The division type of tetrasporangia is one of the diagnostic characters classifying the family groups in red algae (Kylin, 1932). Although tetrasporangial division is cruciate in all members of Grateloupiaceae and Halymeniaceae (Chiang, 1970; Womersley and Lewis, 1994), the tetrasporangia of the Grateloupiaceae are ellipsoidal, and initiated laterally from the third to the fifth cell, up to the sixth, of outer cortical cell files beneath the thallus surface, whereas their initiation in each group in the Halymeniaceae is different: in Halymenia, including Halymenia sensu stricto, and Cryptonemia, they are ovoid, initiated from a subcortical cell; in Aeodes, Pachymenia, Polyopes, and Glaphyrosiphon, they are ellipsoidal, initiated laterally from a cortical cell (Table 1).

Considering our phylogenetic results and morphological observations on reproductive structures, including the taxonomic history (Wang et al., 2001; Gargiulo et al., 2013), the monophyletic Grateloupia sensu lato clade is clearly distinct from the representative halymeniacean taxa including Halymenia floresii. As a result, we reinstate the family Grateloupiaceae based on the generitype Grateloupia, which was established by Schmitz (1889), but later synonymized with the Halymeniaceae (as Cryptonemiaceae) by Papenfuss (1955). Consequently, we propose the resurrection of the family name Grateloupiaceae Schmitz for the Grateloupia sensu lato clade as a distinct family within the Halymeniales and also provide an emended genus composition. Based on our current study, the emended Grateloupiaceae now includes Grateloupia, Dermocorynus, and Prionitis among the 11 genera of Grateloupiaceae sensu Schmitz, 1889 and an additional six genera, Pachymeniopsis, Phyllymenia, Neorubra, Mariaramirezia, Kintokiocolax, and Yonagunia. Although Dermocorynus, Mariaramirezia, and Neorubra were not included in our multigene analyses, we consider them as members of the Grateloupiaceae based on the rbcL phylogeny and previous studies (Wilkes et al., 2005; Calderon et al., 2014a, b).

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: NCBI (accession: MW690928–MW691096).

HSY and MSK designed the study. SYK and ECY performed molecular experiment, data analyses, and drafted the molecular section. HWL conducted morphological observation and drafted the morphological section. SMB, JL-B, SF, and RD’A contributed the samples collection and revised the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.