Molecular Phylogeny and Barcoding of Caulerpa (Bryopsidales) Based on the tufA, rbcL, 18S rDNA and ITS rDNA Genes

Abstract

The biodiversity assessment of different taxa of the genus Caulerpa is of interest from the context of morphological plasticity, invasive potential of some species and biotechnological and pharmacological applications. The present study investigated the identification and molecular phylogeny of different species of Caulerpa occurring along the Indian coast inferred from tufA, rbcL, 18S rDNA and ITS rDNA nucleotide sequences. Molecular data confirmed the identification of 10 distinct Caulerpa species: C. veravalensis, C. verticillata, C. racemosa, C. microphysa, C. taxifolia, C. sertularioides, C. scalpelliformis, C. serrulata, C. peltata and C. mexicana. All datasets significantly supported the sister relationship between C. veravalensis and C. racemosa var. cylindracea. It was also concluded from the results that the specimen identified previously as C. microphysa and C. lentillifera could not be considered as separate species. The molecular data revealed the presence of multiple lineages for C. racemosa which can be resolved into separate species. All four markers were used to ascertain their utility for DNA barcoding. The tufA gene proved a better marker with monophyletic association as the main criteria for identification at the species level. The results also support the use of 18S rDNA insertion sequences to delineate the Caulerpa species through character-based barcoding. The ITS rDNA (5.8S-ITS2) phylogenetic analysis also served as another supporting tool. Further, more sequences from additional Caulerpa specimens will need to be analysed in order to support the role of these two markers (ITS rDNA and 18S insertion sequence) in identification of Caulerpa species. The present study revealed the phylogeny of Caulerpa as complete as possible using the currently available data, which is the first comprehensive report illustrating the molecular phylogeny and barcoding of the genus Caulerpa from Indian waters.

Introduction

The siphonous green algal taxa, particularly those belonging to the genus Caulerpa, poses considerable difficulty in taxonomic identification at the species level due to the phenotypic plasticity in diagnostic characters. This can be further substantiated by the fact that out of 359 species (including forms and varieties) in the genus Caulerpa, only 85 are taxonomically valid [1]. The previous reports have shown that the assimilators and ramuli that are used as taxonomic keys in identification seems to be under the control of environmental factors such as temperature, irradiance, water movement, etc. [2-4]. Thus, conventional diagnostic characters alone have rather limited application when determining the correct identification and phylogeny of the species.

The classical accounts of the species of Caulerpa were given by Agardh [5] and Weber-van Bosse [6] based on morphological characteristics. Subsequently, Svedelius [7] investigated the biodiversity of Caulerpa from Ceylon (Sri Lanka) following the work of Agardh [8]. The first taxonomically identified species of this genus from the Indian coast were Caulerpa serrulata (as C. freycinetii), C. lessonii and C. racemosa var. turbinata (as C. chemnitzia) by De Toni [9], although subsequent studies added several species. A new species, C. veravalensis, containing narrow, linear, non-overlapping, flat pinnules with a rounded apex was described from Veraval, Gujarat, India [10]. Rao [11] described C. mexicana f. indica from North Andaman. Duraiswamy [8,12,13] prepared a comprehensive account of Caulerpa describing 21 taxa from the Indian shores based on morphological, cytological, anatomical and secondary metabolites (caulerpin, caulerpicin and β-sitosterol). However, these studies largely consisted of collections made from the southern part only. The biodiversity assessment study from the Gujarat coast reported the occurrence of 14 species including five varieties and three forms [14].

The genus Caulerpa has attracted attention recently because of the invasive nature of some species [15]. The secondary metabolites from the Caulerpa were also reported to have various biotechnological and pharmacological applications [16]. Some taxa belonging to the genus Caulerpa showed relatively well-defined morphological characters that can be easily differentiated. Nevertheless, validation of the reported Caulerpa species is necessary because of clear evidence of adaptive variation in their morphology, which has resulted in a series of varieties and forms. De Senerpont Domis et al. [17] suggested the need for detailed study of the C. racemosa complex, which harbours a number of varieties and forms. The recent molecular study on the C. racemosa-C. peltata complex revealed the presence of six different lineages that can be differentiated into species-level entities [18]. Therefore, correct taxonomic identification of the species from this genus is of paramount importance in biodiversity assessment studies.

Although certain chemotaxonomic markers based on secondary metabolites have been considered as an additional tool, their utility in taxonomy is limited. The recent development of molecular-marker-based characterization in several groups of seaweeds has opened up new opportunities for studying phylogenetics and resolving the taxonomic issues of cryptic species. A wide range of molecular markers has been employed in the past to decipher the identification and phylogeny of the genus Caulerpa [18-30]. Phylogeny primarily reflects the evolutionary relationships among organisms. Moreover, the rate of evolution is variable for different molecular markers; therefore to resolve the phylogenetic relationship it usually requires extensive sequencing of multiple molecular markers.

Herbert et al. [31] proposed the use of DNA barcodes, i.e. small DNA sequences amplified and sequenced from a standardized portion of the genome to identify and discriminate species. The Consortium for the Barcode of Life (CBOL) has proposed the RuBisCO large subunit (rbcL) and matK as DNA barcodes for plants [32]. Some genus of brown (e.g. Fucus) and red (e.g. Dilsea and Mazzaella) algae, which have high morphological plasticity and are difficult to identify, were resolved successfully by utilizing the 5’end of the cytochrome c oxidase 1 gene (COI-5P) as a DNA barcode [33-35]. The difficulty in amplification of COI-5P [36] and the absence of matK from green algae (except Charophyte [37]) make them inappropriate candidates for barcoding in Caulerpa. Therefore, there is a need to develop an efficient DNA barcode system based on small DNA sequence amplified and sequenced from a standardized portion of the genome that is able to identify the Caulerpa species, thereby helping to explore its cryptic diversity.

The relatively conserved tufA gene is a preferred marker for identification and phylogeny of green algal taxa [18,23,36]. Saunders and Kucera [36] evaluated several markers for marine green macroalgae, albeit not Caulerpa, and proposed tufA as the preferred barcode. The stability of the rbcL exon, with high amino acid sequence similarity, makes it another useful and reliable marker for such studies [38]. The 18S nuclear rDNA has been widely used in phylogenetic studies since it comprises highly conserved regions among the species and shows a high degree of functional constancy with a slow evolutionary rate. However, the highly conserved nature reduces the genetic distance between species in pairwise distance analysis if the complete locus is utilized [39]. However, a characteristic intronic insertion sequence is present in the Caulerpa 18S rDNA sequence. These insertion sequences were reported by Kooistra [40] in two Caulerpacean specimens, and were utilized by Durand et al. [24] for phylogenetic analysis. We have utilized these 18S introns for character-based barcoding in the present study. The molecular marker ITS rDNA shows high variability in its sequence as well as in its length, which can be exploited for comparing the Caulerpa populations at the inter- and intraspecific levels [19].

The use of molecular markers for identification and phylogenetic studies of the genus Caulerpa from India has not been reported to date. The present study thus investigates the utility of tufA, rbcL and ITS rDNA by standard barcode methods (Neighbour-joining (NJ) analysis and nucleotide-sequence divergences) as described by Hebert et al. [31], the monophyletic association of taxa in phylogenetic trees [41] and 18S rDNA introns by character-based analysis as a DNA barcode to identify Caulerpa species. An extensive phylogenetic reconstruction was also accomplished to elucidate the relationships and the phylogenetic placement of Caulerpa species from Indian waters. Additionally, possible congruence between morphology and molecular data was also analysed in the present study.

Materials and Methods

Ethics Statement

The Chief Conservator of Forest, Marine National Park Jamnagar, Government of Gujarat, India permitted the collection of Caulerpa specimens from Poshitra Rocks and the collection from Krusadai Island was permitted by the Chief Wild Life Warden, Gulf of Mannar Biosphere Reserve, Government of Tamil Nadu, India. The other sampling locations are not the part of any national parks or protected areas and do not require any specific permits. It is further to confirm that the field studies did not involve endangered or protected species.

Collection of samples and morphological identification

Repeated sampling was performed at 16 different locations, which seemed to cover nearly all Caulerpa species reported from India. The species diversity in the genus Caulerpa is concentrated along the northwest (Gujarat) and southeast (Tamil Nadu) coast of India [42]. Therefore, intensive sampling was performed mainly from these areas. The detailed morphological descriptions, images and references used for identification [43-56] for collected specimens are given in the supporting information (Figures S1-S19 in File S1 ). In total, 29 Caulerpa specimens (20 species including seven varieties and three forms and two unidentified taxa) were investigated based on the morphological differences for molecular barcoding and phylogenetic analysis. The collection sites for the specimens included in this study are shown in Figure S1 . Specimens used in the analyses, and the specimen voucher numbers, collection sites and accession numbers, are listed in Table S1 . Samples were cleaned with sterile seawater to remove mud and epiphytes and finally rinsed with distilled water. Specimens were stored at -20°C before genomic DNA isolation. Voucher specimens for individual species were submitted to the Taxonomic Reference Centre for seaweeds at the Council of Scientific and Industrial Research-Central Salt and Marine Chemicals Research Institute (CSIR-CSMCRI).

DNA extraction, amplification and sequencing

Genomic DNA was isolated by a modified CTAB DNA extraction method [57]. Amplification by PCR was performed in a master mix of volume 25 µL containing 5 pmol of each primer; 200 µM of each dNTP; 1X assay buffer; and 1.25 units of Taq DNA polymerase. The details of the molecular markers, primers and amplification conditions utilized in this study are summarized in Table 1 . Amplifications were carried out using a PCR system (Bio-Rad, Hercules, CA, USA). PCR products were purified and subjected to commercial sequencing (Macrogen Inc., Korea).

Table 1. Primer details and PCR conditions used for the amplification of 18S rDNA, ITS rDNA, rbcL and tufA molecular markers used in this study.

Primer	Primer Sequences (5’-3’)	DNA Markers	PCR conditions	Reference
18SF 18SR	CAACCTGGTTGATCCTGCCAGT TGATCCTTCTGCAGGTTCACCTAC	18S rDNA	94°C ⁄ 5 min; 35cyclesof 94°C ⁄ 1min, 52.4°C ⁄ 1 min 72°C ⁄ 2min; final extension 72°C ⁄ 5 min	[73]
ITSF ITSR	CCTCTGAACCTTCGGGAG TTCACTCGCCATTACT	ITS rDNA	94°C ⁄ 5 min; 35 cycles of 94°C ⁄ 1 min, 52.4°C ⁄ 1 min, 72°C ⁄2 min; final extension 72°C ⁄5 min	[20]
rb-F rb-R	GCTTATGCWAAAACATTYCAAGG AATTTCTTTCCAAACTTCACAAGC	rbcL	94°C ⁄ 5 min; 35 cycles of 94°C ⁄ 45 sec, 41.5°C ⁄ 45 sec, 72°C ⁄2 min; final extension 72°C ⁄10 min	[27]
Tuf-F Tuf-R	TGAAACAGAAMAWCGTCATTATGC CCTTCNCGAATMGCRAAWCGC	tufA	94°C ⁄ 5 min; 35 cycles of 94°C ⁄ 1 min, 50°C ⁄ 1 min, 72°C ⁄2 min; final extension 72°C ⁄5 min	[23]

IUPAC nucleotide ambiguity codes: W= A/T, Y= C/T, M= A/C, N= A/T/C/G

Data analysis

Individual sequences obtained in this study were compared with accessions in the National Center for Biotechnology Information (NCBI) (Table S1 ) using BLAST analysis. The additional sequences were retrieved from GenBank in order to compare the inter- and intraspecific nucleotide divergences and to produce the phylogeny of Caulerpa as complete as possible using the currently available data. Caulerpella ambigua was used as an outgroup in tufA and rbcL tree as it was found to be the most basal taxon to all the Caulerpa species [23,28]. Multiple sequence alignment was performed with MAFFT version 6 with Q-INS-i strategy activated (which considers secondary-structure information of RNA for alignment) for ITS rDNA and 18S rDNA insertion sequence alignment [58].

The estimation of nucleotide divergence between sequences was calculated using the Kimura 2- parameter (K2P) for ITS rDNA, Tamura-Nei (TN93) for rbcL and Tamura 3-parameter (T92) for the tufA dataset. The rate variation among sites was modelled with a gamma distribution (shape parameter = 5). Model selection analysis was conducted to calculate the best-fit model of substitution by MEGA v.5 [59]. A neighbour-joining (NJ) tree was constructed by the bootstrap resampling method with 1,000 bootstrap replications in MEGA v.5 [59]. Minimum interspecific distances and maximum intraspecific distances were calculated for each species identified and named by traditional taxonomical features.

Phylogenetic trees were constructed by Bayesian inference (BI) using MrBayes v.3.1.2 [60]. Model selection analysis was conducted to calculate the best-fit model of nucleotide evolution by jModelTest 0.1.1 [61,62]. A codon-based partition strategy was used for rbcL and tufA gene datasets. The best scheme of model substitutions for partitioned data was generated through PartitionFinder v.1.0 [63]. The models were selected based on the Bayesian Information Criterion (BIC) [64] scores for each dataset (Table 2 ).

Table 2. Nucleotide substitution models for respective datasets for Bayesian analysis.

Dataset	No. of sequences	Subset Partitions	Model	BIC	lnL
rbcL	45	p1 = 1-1076\3 p2 = 2-1076\3 p3 = 3-1076\3	HKY+G JC HKY+G	7300.60	-3304.74
ITS rDNA	62		HKY+G	13695.80	- 6445.29
tufA	82	p1 = 1-815\3 p2 = 2-815\3 p3 = 3-815\3	GTR+G F81+G GTR+I+G	9917.05	-4335.13

BIC, Bayesian Information Criterion; lnL, Maximum Likelihood value; +G, Gamma distribution; JC, Jukes-Cantor; GTR, General Time Reversible; HKY, Hasegawa-Kishino-Yano; F81, Felsenstein 1981; p1, p2, p3, partition of dataset based on codon position.

The Markov chain Monte Carlo (MCMC) method was used for Bayesian phylogenetic analyses. Each analysis consisted of three heated and one cold Markov chains. Sample and print frequency was set to 100 and 1,000 respectively for 2,000,000 generations. The 50 per cent majority rule consensus tree was obtained after discarding 25% of sampled trees as burn-in.

Results

Morphological identification

The morphological characters of collected specimens were studied by following the traditional taxonomic keys for the genus Caulerpa. In total, 20 species including seven varieties and three forms were identified (Figures S1-S19 in File S1 ). The cryptic nature of two taxa (Caulerpa sp. C03 and Caulerpa sp. C13) made it difficult to identify them at the species level. Two specimens of C. veravalensis (C10 & C23) resembling the specimen described by Thivy and Chauhan [10] were selected for molecular analysis. These specimens were characterized by a pinnately divided flat broad midrib, opposite to alternate flat ramuli with a rounded apex and occasional bifurcation in apices of ramuli (Figure S9 in File S1 ). The C. scalpelliformis (C21), var. denticulata (C12) and forma dwarkensis (C01) were differentiated following the treatment given by Børgesen [65]. C. scalpelliformis var. denticulata was characterized by denticulation along the outer marginal lobes (Figure S1 in File S1 ). The forma dwarkensis has an alternate arrangement of same-length ramuli throughout, except at the top, on regularly divided assimilators (Figure S1 in File S1 ). The morphology of the specimen C29 agrees very well with C. racemosa var. racemosa f. remota Coppejans described by Coppejans et al. [66] except the length of rachis, which ranged from 1.5 to 5 cm (Figure S17 in File S1 ). Two specimens, C05 (collected from the western coast of India) and C20 (collected from the eastern coast of India), had cylindrical, sometimes laterally compressed, ramuli that were radially arranged on assimilators (Figure S4 in File S1 ). There is a strong resemblance of these specimens to the description of C. racemosa var. laetevirens f. laxa by Børgesen [65] and with C. racemosa var. cylindracea f. laxa described by Coppejans et al. [66]. The specimen C06 was identified as C. microphysa characterized by far fewer vesicles that are not clear in longitudinal series on the axis, short assimilators up to 3.0 cm and spherical ramuli up to 2.0 mm diameter (Figure S5 in File S1 ). Specimen C14, however, with erect assimilators up to 10 cm long, and densely covered with spherical to sub-spherical ramuli and ramuli with constricted pedicels, was identified as C. lentillifera (Figure S11 in File S1 ).

Barcoding Analysis

A total of 82 tufA sequences were aligned to generate the dataset of total length of 815 nucleotide positions in alignment to construct the NJ tree and to calculate pairwise distance. The NJ tree (Figure 1 ) revealed the presence of 19 distinct well-supported clades. The average corrected divergence over all sequence pairs was 0.063. The average intraspecific genetic divergence was 0.003 and the average interspecific divergence was 0.068. The divergence between morphologically different species C. serrulata and C. cupressoides was exceptionally low (from 0.003 to 0.004). The morphologically distinct species C. subserrata (AJ417935) and C. biserrulata (AJ417934) showed interspecific variation of 0.003. The pairwise distance analysis of tufA gene data showed the intraspecific variation ranged from 0.0 to 0.011 whereas interspecific variation ranged from 0.003 to 0.173. The result clearly indicates the overlapping of maximum intraspecific and minimum interspecific genetic distances (Figure 2 ). The highest divergence was observed for the C. verticillata with a mean genetic distance 0.160 (0.147-0.173). The specimen identified previously as C. microphysa and C. lentillifera showed no sequence divergence for all the markers studied. In view of this, these two taxa were excluded from the interspecific variation range.

Figure 1. NJ tree based on tufA gene sequence data.

Support values at nodes correspond to bootstrap proportion (BS). Sample ID for specimens from this study and accession numbers for the reference sequences are given for identification in Table S1. Solid lines on the right indicate possible clades.

Figure 2. Plot of intra- and interspecific genetic distances for the tufA, rbcL and ITS rDNA molecular markers.

A total of 45 sequences of the rbcL gene were used to generate the dataset of total length of 1076 nucleotide positions in alignment. The intron was removed from the rbcL gene sequences before analysis. In total, 10 clusters were recovered in NJ analysis (Figure S2 ). The cluster with C. mexicana (C27) and C. cuppressoides var. lycopodium (AJ512470) was poorly supported among these clusters. The average divergence over all sequence pairs was 0.043. The average intraspecific genetic divergence was 0.0033 and the average interspecific divergence was 0.053. The maximum intraspecific genetic distance (0.010) exceeds the minimum interspecific distance (0.004) in the present dataset. C. verticillata showed the highest divergence with a mean genetic distance of 0.127 (0.121-0.132).

In addition to the above markers, we sequenced the ITS1-5.8-ITS2 region of the rDNA. We tried to align 62 ITS rDNA sequences, but the ITS1 region was virtually impossible to align and was removed before further analysis. In NJ analysis, 13 distinct clusters were recovered with high support values (Figure S3 ). All species studied were clearly differentiated into distinct clades. The average divergence over all sequence pairs was 0.168. The average intraspecific genetic divergence was 0.025 and the average interspecific divergence was 0.17. The nucleotide divergence varied from 0.003 to 0.301 within species whereas interspecific variation ranged from 0.04 to 0.661. The highest divergence was observed for the C. verticillata with a mean genetic distance 0.510 (0.469-0.661).

The length of insertion sequence in 18S rDNA sequence was found to be in the range of 113-115 nucleotides in Caulerpa species. The insertion and deletion pattern in insertion sequence was species specific and could be utilized for species-level identification by a character-based approach. The alignment of 18S rDNA insertion sequences (Figure 3 ) clearly differentiated the taxa at the species level. C. microphysa (C06) and C. lentillifera (C14) shared an identical insertion sequence. C. racemosa var. cylindracea f. laxa and C. veravalensis were separated by a single diagnostic character.

Figure 3. Bayesian phylogenetic tree based on tufA gene sequence data.

Support values at nodes correspond to posterior probabilities (pp). Sample ID for specimens from this study and accession numbers for the reference sequences are given for identification in Table S1, Solid lines on the right indicate possible clades.

Phylogenetic analysis

The Bayesian phylogenetic tree of the tufA gene (Figure 4 ) supported the differentiation of species depicted in the NJ tree. It was observed that C. racemosa var. cylindracea f. laxa (C05 and C20) clustered with C. veravalensis with strong support (posterior probability (pp) =1.0). This clade was a sister lineage to C. racemosa var. cylindracea (pp=1.0). Similarly, C. racemosa var. peltata (AJ417949) and C. racemosa var. laetevirens (AJ512415) clustered with C. peltata (C19 and C28) with very strong support (pp=1.0). No sequence difference was observed for C. microphysa (C06) and C. lentillifera (C14), which clustered together. Furthermore, Caulerpa sp. (C03, C13) showed no sequence difference with C. veravalensis, and these clustered together. The position of C. serrulata (C18) was clearly paraphyletic in the BI tree.

Figure 4. Sequence alignment of 18S rDNA insertion sequences for listed Caulerpa species revealed specific insertion-deletion pattern.

The rbcL gene phylogeny of Caulerpa based on Bayesian analysis depicts the presence of eight well-supported clades with eight separate lineages (Figure 5 ). C. racemosa var. cylindracea f. laxa (C05 and C20), C. veravalensis (C10 and C23) and Caulerpa sp. (C03, C13) clustered together with high support values (pp=0.95). C. microphysa (C06) and C. lentillifera (C14) showed no sequence difference and clustered together. Similarly no sequence difference was observed in C. flexilis (AJ512485) and C. okamurae (AB038484). The C. racemosa and varieties were positioned in four different lineages.

Figure 5. Bayesian phylogenetic tree based on rbcL gene sequence data.

Support values at nodes correspond to posterior probabilities (pp). Sample ID for specimens from this study and accession numbers for the reference sequences are given for identification in Table S1. Solid lines on the right indicate possible clades.

The Bayesian analysis of ITS rDNA resulted in a phylogenetic tree (Figure 6 ) consisting of 12 well-supported and one weakly supported clade. Among these clades, six showed the presence of taxa belonging to the C. racemosa complex underlining the polyphyly of the complex. C. serrulata and C. cupressoides were recovered as sister lineages with strong support (pp=1.0). C. peltata (C19 and C28) formed a separate lineage with the species that were mostly characterized by turbinate, trumpet or peltate ramuli (pp=1.0).

Figure 6. Bayesian phylogenetic tree based on ITS rDNA sequence data.

Support values at nodes correspond to posterior probabilities (pp). Sample ID for specimens from this study and accession numbers for the reference sequences are given for identification in Table S1. Solid lines on the right indicate possible clades.

Discussion

This study aimed to determine the identification and phylogenetics of the Indian Caulerpa species by employing multiple markers. This inclusive multi-gene approach has further improved the reconstruction of the phylogenetic relationship among Caulerpa species.

Barcoding Analysis

In this study, the overlap between levels of intraspecific genetic distance and interspecific genetic distance was observed for all the datasets (Figure 2 ). The lower interspecific genetic distance was observed in some of the taxonomically well-established species. Therefore, it is difficult to define species boundaries using a distance-based approach in Caulerpa. Another approach of delineation of species is through monophyletic association of taxa in a Neighbour-Joining (NJ) tree, which does not depend on the distance-based threshold method [41]. In NJ analysis of tufA and ITS rDNA, most of the clades were recovered as being monophyletic with strong support with few exceptions. The Bayesian phylogenetic tree (Figure 4 and Figure 6 ) supported the differentiation of species, which was also depicted in NJ trees. ITS-rDNA-based phylogenetic analysis was found to be mostly congruent with the tufA gene analysis. In the rbcL-gene-based analysis, the position of certain taxa did not resolve sufficiently and also showed incongruence with other datasets. The rbcL gene was also found to be least variable in comparison with tufA and ITS rDNA (Figure 2 ). Handeler et al. [67] and Saunders and Kucera [36] supported tufA gene as a barcode for green algae. Therefore, monophyletic association of taxa in the tufA-gene-based tree can be utilized for species identification. The character-based identification of the species of Caulerpa by using the 18S rDNA insertion sequence was another useful identification tool that we found in this study. For example, C. peltata (C19 and C28) was clearly separated from C. racemosa and varieties in the 18S rDNA insertion sequence alignment (Figure 3 ). The ITS rDNA (5.8S-ITS2) analysis can be used as an additional supporting tool for identification purpose, but more sequences from species of Caulerpa will need to be analysed before defining the role of ITS rDNA in species identification. For example, C. serrulata and C. cupressoides showed paraphyly in tufA analysis but formed a sister lineage in ITS rDNA analysis. However, there is a single valid ITS rDNA sequence available for C. cupressoides in GenBank dataset and more sequences will be required for differentiating the C. cupressioides and C. serrulata as monophyletic lineages. Similarly, more sequences from additional species of Caulerpa will need to be analysed in order to support the role of 18S insertion sequence in identification of Caulerpa species. Following this molecular barcoding scheme, the identity of 10 distinct species of Caulerpa was confirmed from Indian waters.

Phylogenetic analysis

The tufA-gene-based phylogenetic tree (Figure 4 ) was found to be congruent with those of the findings of Sauvage et al. [18], Fama et al. [23] and Stam et al. [28]. The major addition to these phylogenetic analyses was C. veravalensis, which was recovered as a sister lineage to C. racemosa var. cylindracea. C. veravalensis was considered as a form of C. taxifolia and was later differentiated as a separate species based on morphological characters [68]. The species that was identified as C. racemosa var. cylindracea f. laxa, based on morphology, and two unidentified Caulerpa sp. C03 and C13, consistently placed in the same clade with C. veravalensis in all the phylogenetic trees. These taxa may be considered as part of a new C. veravalensis complex and need further detailed investigations.

The rbcL phylogeny (Figure 5 ) was not consistent with the phylogeny of other datasets. For example, the incongruity was observed in the position of C. flexilis as it formed a separate lineage in tufA gene analysis and clustered with C. okamurae, C. microphysa and C. lentillifera in the rbcL gene phylogenetic tree. The position of C. serrulata also differed in rbcL phylogeny in comparison to tufA and ITS rDNA phylogenetic analyses. In rbcL tree (Figure 5 ) C. serrulata was polyphyletic whereas, it was paraphyletic in tufA analysis (Figure 4 ) and formed sister lineage with C. cupressoides in ITS rDNA analysis (Figure 6 ). Similarly, De Senerpont Domis et al. [17] showed the topological differences between phylogenies inferred from tufA and rbcL genes for the genus Caulerpa. De Senerpont Domis et al. [17] and Fama et al. [23] indicated the possible reasons of incongruence between these two genes such as hybridization, incomplete lineage sorting, and horizontal gene transfer (organismal- level cause) and rate heterogeneity, selection, and base/codon composition biases (genetic-level causes). Fama et al. [22] reported the high levels of intra- and inter-individual polymorphism in the rDNA ITS1 region which can affect the phylogenetic reconstruction. The removal of the ITS1 region from the ITS rDNA dataset and the use of 5.8S-ITS2 in the phylogenetic analysis resulted in a robust and well-resolved phylogenetic tree (Figure 6 ).

The polyphyly of the C. racemosa complex was evident in the analysis of all the datasets. Sauvage et al. [18] also reported the presence of six different lineages in the C. racemosa-peltata complex. Most of these lineages can be resolved into separate species. For example, phylogenetic analysis strongly favoured the separate species position of C. peltata (C19 and C28) having distinctly peltate ramuli, which was evident from the distant placement of taxa from other C. racemosa clades. Furthermore, the results confirmed that C. racemosa var. laetevirens (C25) is a part of the C. racemosa complex and is not a separate species C. laetevirens Montagne.

The results of phylogenetic analysis were consistent with the study of Yeh and Chen [26], as C. microphysa deviated from other species proving its taxonomic distinction. Coppejans and Beeckman [69] considered C. lentillifera and C. microphysa as separate species. From the phylogenetic trees, it was inferred that both these taxa grouped together into a single clade. Furthermore, these two taxa showed no difference in rbcL gene sequence. These two taxa also shared a common 18S rDNA insertion sequence. Therefore, it can be considered that these two morphologically different specimens collected from India belong to the same species. The present results also agree with the study of Olsen et al. [21] wherein the conspecific nature of C. taxifolia and C. mexicana were rejected. These species have formed separately placed clades in the phylogenetic trees.

Morphology vs. molecular analysis

Weber van Bosse [6] classified Caulerpa species into 12 sections on the basis of morphology. Similarly, Calvert et al. [70] studied the phylogeny of these 12 sections based on chloroplast ultrastructure. Duraiswamy [8] categorized the Indian Caulerpa species into five sections. Following this work, the taxa investigated in this study can be grouped into three sections, i.e. Filicoideae, Sedoideae and Charoideae. In the present study, it was observed that clades formed in the phylogenetic trees do not entirely follow the sectional scheme. Similar findings have been reported by Fama et al. [23] for tufA-gene-based phylogenetic analysis in Caulerpa.

Stam et al. [28] reported the polyphyletic nature of C. scalpelliformis. In the present tufA analysis, all C. scalpelliformis specimens from India (C01, C12 and C21) clustered with C. scalpelliformis var. denticulata (AJ417972) from Lebanon but away from the Australian specimen (AJ417971). On the other hand, morphologically different species C. racemosa var. cylindracea f. laxa (C05 and C20) and C. veravalensis (C10 and C23) clustered together. Thus, there was no consistent pattern observed in the relationship between morphological characters and placement in the phylogenetic tree of taxa based on the molecular markers investigated. Similarly, C. serrulta and C. cupressoides are clearly differentiable on morphological characteristics but formed paraphyletic lineages in tufA analysis. Therefore, for truly understanding the placement in phylogenetic tree and proper identification of different populations of C. scalpelliformis, C. serrulta and C. cupressoides more detailed study with other molecular markers will be required. The weak supports at internal nodes in clades restrict further distinction at a lower taxonomic level. However, these sub-specific ranks can be delineated by using longer nucleotide sequences [71] or detailed transcriptome analyses [72].

Conclusion

In the present study, we present a comprehensive phylogeny of Caulerpa using most of the currently available data. The findings of this study showed the phylogenetic position of the Indian Caulerpa population vis-à-vis other parts of the world. The study supports the use of the tufA gene as a preferred marker with the monophyletic association of taxa as the main criteria for identification at the species level. The ITS rDNA (5.8S-ITS2) analysis could also be used as an additional supporting tool for identification purposes, coupled with character-based identification by 18S rDNA insertion sequences at the species level. Our molecular analyses eventually led to establishment of 10 distinct Caulerpa species from Indian waters. Further, the taxa identified as C. racemosa var. cylindracea f. laxa and two unidentified taxa showed close proximity with C. veravalensis at a molecular level despite the distinct morphological variations indicating the presence of a new C. veravalensis complex, which needs further detailed investigation.

Supporting Information

File S1

Includes Figures S1-S19. Morphological description and images of Caulerpa specimens collected for this study.

(PDF)

Figure S1

Sample collection sites from India.

(PDF)

Figure S2

NJ tree based on rbcL gene sequence data. Support values at nodes correspond to bootstrap proportion (BS). Sample ID for specimens from this study and accession numbers for the reference sequences are given for identification in Table S1. Solid lines on the right indicate possible clades.

(PDF)

Figure S3

NJ tree based on ITS rDNA gene sequence data. Support values at nodes correspond to bootstrap proportion (BS). Sample ID for specimens from this study and accession numbers for the reference sequences are given for identification in Table S1. Solid lines on the right indicate possible clades.

(PDF)

Table S1

Specimens used in analyses, specimen voucher number, collection locations, collection date and accession numbers.

(PDF)

Funding Statement

The financial assistance received from Council of Scientific and Industrial Research (www.csir.res.in), New Delhi (CSC0116: BioEn) is duly acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.