•The origin of the Ectocarpus strain used for genome sequencing (the ‘genome strain’) was Peru, where no Ectocarpus had been reported previously. To study the genetic diversity in the region and to increase the number of individuals from this area available for genetic experiments, 119 new Ectocarpus strains were isolated from eight localities along the 3000 km of coastline from central Peru to central Chile.
•Internal transcribed spacer 1 (ITS1) genotyping revealed nine different genotypes, five of which were endemic to the area studied and three of which were previously unknown.
•Individuals of the same genotype as the genome strain occurred from Peru to northernmost Chile, representing 61% of the samples in this area, from which five more genotypes were isolated. Further south, down to central Chile, most individuals belonged to Ectocarpus siliculosus, Ectocarpus fasciculatus and Ectocarpus crouaniorum. In sexual crosses, the genome strain and the new isolates of the same genotype were fully compatible.
•Sequences from four nuclear and cytoplasmic genetic markers (ITS1, ITS2, Rubisco spacer and Cytochrome-c oxidase subunit 3 (cox3)) separated the genome strain from the known species of Ectocarpus. It may in future be recognized as a separate species.
The Ectocarpus strain chosen for genome sequencing (Cock et al., 2010) is a male gametophyte which has the designation ‘Ec32’ in the Ectocarpus strain collection at Roscoff (France). Henceforth it will be referred to as the ‘genome strain’. It was selected because it displays an alternation of two morphologically distinguishable generations (Peters et al., 2008) and produces unilocular sporangia on parthenogenetic sporophytes, both of which facilitate studies of the complex brown algal life history (for more information and for the terminology used in morphological and life history descriptions, see Peters et al., 2004a; Coelho et al., 2007; Charrier et al., 2008). The genome strain was obtained from a meiospore of a field sporophyte collected by the first author (A.F.P.) in November 1988 at San Juan de Marcona, Peru (site 3 in Fig. 1). It was identified as Ectocarpus siliculosus (Dillwyn) Lyngbye in an initial molecular phylogeny (lineage 1c in Stache-Crain et al., 1997; strain designation ‘SAm120h’), but it showed post-zygotic incompatibility (Peters et al., 2004b) when crossed with a strain of E. siliculosus sensu stricto (genetically close to those used by Müller (1967) for the description of the life history; lineage 1a in Stache-Crain et al., 1997) or with a strain of Ectocarpus from New Zealand (Müller, 1991; lineage 4 in Stache-Crain et al., 1997): hybrids possessed a normal morphology, but were unable to form meiospores. Further crosses of the genome strain with an Ectocarpus from northern Chile, as well as with Ectocarpus crouaniorum Thuret in Le Jolis from France, gave similar results (A.F. Peters, unpublished data). Although sisters of the genome strain (female gametophytes from the same field sporophyte from Peru) were available for genetic experiments, for instance to test sex-linkage of a mutation (Peters et al., 2008), no female gametophyte of sufficient genetic distance was available for experiments requiring outcrossing (e.g. to produce a genetic map; Heesch et al., 2010).
Apart from the genome strain there was no unambiguous published record of Ectocarpus from Peru. Howe (1914: p. 50) mentioned ‘small specimens of Ectocarpus’ on Desmarestia peruviana Montagne at Ancón (13 February 1907) and on Lessonia nigrescens Bory de Saint-Vincent at Chincha Island (18 June 1907); however, he regarded the material as ‘too meagre to justify an attempt at determ-ination or description’. The nearest published record of Ectocarpus was from Iquique, northern Chile, at c. 800 km distance from site 3 (Ramírez & Santelices, 1991).
The aim of this study was to isolate Ectocarpus strains that were fully compatible with the genome strain. Individuals were collected along the South American Pacific coast from central Peru to central Chile and characterized based on sequence comparisons with 43 strains of Ectocarpus and seven of Kuckuckia (the sister genus of Ectocarpus) from all continents except Antarctica (Stache-Crain et al., 1997). Our experiments revealed a surprisingly high genetic diversity in Ectocarpus in the studied area, indicating that additional species will possibly have to be established within this genus in the future. Isolates of one of the nine genotypes found were genetically similar to the genome strain and proved to be fully interfertile with it, providing a polymorphic outcrossing line that has been used elsewhere for genetic analyses.
Materials and Methods
Field collection and isolation of cultures
Nine sites (four in Peru and five in Chile) spanning a distance of c. 3000 km were visited for collection, most during one visit in February–March 2006 (Fig. 1, Table 1). Sampling at site 8 was performed for studies on the response of Ectocarpus to copper pollution (A.D. Mann et al., unpublished data). From each field thallus macroscopically resembling Ectocarpus, filaments were inoculated into a 2-ml Eppendorff tube containing autoclaved sea water. After transfer to the Biological Station at Roscoff, isolation and cultivation of clean unilalgal clonal cultures were undertaken as described previously (Peters et al., 2010). A number of thalli from site 8 were isolated by filtering nearshore surface seawater and cultivating filaments that developed on filter paper, or by inoculating field macrothalli of Scytosiphon into culture and isolating all Ectocarpus thalli developing on them. Only thalli possessing ribbon-shaped plastids (which are diagnostic for the genera Ectocarpus and Kuckuckia) were retained for further study.
La Lancha and Palito (close to copper discharge point)3
Intertidal or developed in culture on Scytosiphon
514–520, 527, 608, 610, 612–613, 616–619
sil + cro (2)
La Lancha and Palito (close to copper discharge point)3
Intertidal or developed in culture on Scytosiphon
2 October 1990
5 October 2006
Marine Station of Universidad Andrés Bello
Large tide pool, saxicolous
Strain designations beginning with ‘CCAP’ are from the Culture Collection of Algae and Protozoa (held at the Scottish Marine Institute; Oban, UK); strain designations beginning with numbers or ‘Ec’ are isolates housed in the Ectocarpus strain collection (held at Roscoff), and are maintained by the first author (A.F.P.).
Characterization of strains
Ectocarpus taxonomy is currently being updated, with major modifications being made to the existing classification system. The two-species concept, recognizing only E. siliculosus and Ectocarpus fasciculatus Harvey (Russell, 1966, 1967; Stache-Crain et al., 1997), has recently been rejected following the reinstatement of a third species, Ectocarpus crouaniorum; additional lineages so far included in E. siliculosus or E. fasciculatus may equally deserve species rank (Peters et al., 2010). In the present paper the taxa E. siliculosus and E. fasciculatus are referred to in a narrow sense; that is, corresponding to lineages 1a and 5b, respectively, in Stache-Crain et al. (1997).
For the identification of isolates from Peru and northern Chile, we used DNA sequences. DNA was extracted from living cultures as described in Peters et al. (2004b). Nuclear ribosomal internal transcribed spacer (ITS) data for 50 Ectocarpus and Kuckuckia strains were available as reference sequences (Stache-Crain et al., 1997). The first part of ITS1, which is 134–701 bp long and flanked by the motifs GATCATTACCGA and TTATYGGTYGGG, is particularly useful to distinguish genotypes because it is highly variable and indel-rich. Initial characterization of all strains was based on ITS1 length (as in Peters et al., 2010), followed by sequencing of ITS1 in 30% of the isolates or by a diagnostic polymerase chain reaction (PCR) using a primer specific for the ITS1 sequence of the genome strain (for details and oligonucleotide primers employed, see Supporting Information Notes S1 and Table S1, respectively). Blast searches (Altschul et al., 1997) were carried out utilizing the first parts of newly generated ITS1 sequences as queries. This allowed classification of the sequences into different ITS1 genotypes (GTs): sequences presenting ≥ 95% identity were considered to belong to the same GT. The different GTs were numbered based on their ITS1 length, the lowest number corresponding to the shortest length. For at least one member of each GT discovered, three additional markers (ITS2, 3′-Ribulose-bisphosphate carboxylase (rbcL) + Rubisco spacer and Cytochrome-c oxidase subunit 3 (cox3) were sequenced to provide data for phylogenetic analyses. Sequences were deposited in the European Molecular Biology Laboratory (EMBL)/GenBank/DNA Data Bank of Japan (DDBJ) database (Table 2).
Table 2. DNA sequences utilized in the present study
ITS1 + 2
Rubisco spacer2 (515 bp)
cox3 (665 bp)
Newly generated sequences are shown in bold.
Acc, GenBank/European Molecular Biology Laboratory (EMBL) Nucleotide Sequence Database/DNA Data Bank of Japan (DDBJ) accession; bp, base pairs (length); –, not applicable; ITS, internal transcribed spacer; cox3, cytochrome-c oxidase subunit 3.
1Limits: ITS1: start atcattaCCGA, end GTTGTAaaacttt; ITS2: start gtctgttGACACC, end TTTCGTTcggacct (Stache-Crain et al., 1997).
2Including flanking ribulose-bisphosphate carboxylase large subunit (rbcL) and ribulose-bisphosphate carboxylase small subunit (rbcS) gene sequences.
3Only ITS1; in addition to the sequence of strain Ec147, the slightly different ITS1 sequence of a particularly copper-resistant strain from site 8 (CCAP1310/333; Ritter et al., 2010) was deposited (FN564444).
Sequences were aligned manually using Se-Al v. 2.0a11 (Rambaut, 2002). Sequences of the four markers were concatenated, excluding more variable parts of ITS, which could not be aligned with confidence. Sequences of reference strains (nine from Europe, and one each from Pacific South America and New Zealand; Table 2) were then added, and phylogenetic analyses undertaken. Phylogenetic trees were constructed using the various programs in phylip version 3.69 (Felsenstein, 1995), and the robustness of the alignments was tested with the bootstrapping option (SeqBoot). Genetic distances, applicable for distance matrix phylogenetic inference, were calculated using the Dnadist program in the phylip package. Phylogenetic inferences based on the distance matrix (Neighbor), maximum likelihood (Dnaml) and parsimony (Dnapars) algorithms were applied to the alignments. In all cases, the best tree or majority rule consensus tree was selected using the consensus program (Consense). The trees were visualized and drawn using the treeview software version 2.1 (Page, 1996).
Thalli of Ectocarpus occurred at eight out of the nine sites surveyed, usually as epiphytes on other macroalgae such as Macrocystis, Lessonia, Desmarestia and Gracilaria. Occasionally they were saxicolous. The Ectocarpus individuals were present from sheltered to exposed conditions but were not found at a highly exposed site (Table 1). The maximum thallus size was 14 cm but several samples were only minute felts or dark spots; other thalli appeared in culture after inoculation with their substratum or they developed on paper that was used to filter sea water (Figs 2a–c; Table 3). A total of 120 strains, including strain CCAP1310/40 isolated previously at site 9, were available for identification of genotypes.
Table 3. Genotypes (GTs) and internal transcribed spacer 1 (ITS1) length types (LTs) of Ectocarpus/Kuckuckia from Peru and northern Chile and results of blast against sequences from previous studies
ITS1 length (bp)2
Length of PCR product3
Sites in Peru4
Sites in Chile4
Best hits in blast of variable part of ITS16
Identity according to blast
sil, Ectocarpus siliculosus; cro, Ectocarpus crouaniorum; nd, no data; –, not applicable.
1For strains isolated from filtered water or developed in culture on inoculated substratum the field morphology is unknown.
Five different ITS1 length types were clearly distinguishable on agarose gels, and one putative hybrid (Table 3, Fig. 3). Sequencing of 36 individuals (Table 1) revealed that two of the length types were heterogeneous and contained more than one ITS1 GT. In total we obtained nine different GTs representing seven Ectocarpus genotypes, one Kuckuckia genotype and one hybrid between two of the Ectocarpus genotypes (Table 3).
Individuals with the same GT as the genome strain (GT4) occurred from site 1, where they were the only Ectocarpus collected, to site 5 at the northern border of Chile (Tables 1,3). The ITS1 regions of the seven sequenced strains of this GT were identical to that of the genome strain (n =2; sites 2 and 3), differed by a single nucleotide substitution (n =4; three from site 1 and one from site 5; see Table 2 for the sequence accession of the latter strain), or had an ambiguity at the same position (n =1; site 5).
We collected eight additional genotypes in the study area (Table 3). GT1, which was not known previously, was collected at sites 2, 6 and 8. It had the shortest ITS1 found so far in any Ectocarpus isolate (302–306 bp). GT2 was common at site 3, but also present at the most southern locality, site 10. It did not match well with any published ITS1; the best blast hit was an isolate from Puerto Deseado, Patagonia, Argentina (91% identity). However, when we sequenced ITS1 of CCAP1310/40 from site 9 (for which only the Rubisco spacer sequence had hitherto been available; U38736) its variable part showed 98% identity with that of our isolates of GT2. GT3 was collected once (at site 6); its variable part of ITS1 presented 85% identity to that of GT2. It also resembled (93% identity) that of the isolate from Argentina mentioned above, which again gave the best blast hit. Additional hits with reasonably high levels of identity were strains from New Zealand (U38761; 88%) and Isla Robinson Crusoe, Chile (U38763; 84%). GT5 was present from site 3 to site 10. It showed 97–98% identity to sequences of E. fasciculatus from Isla Robinson Crusoe and from France. GT6 was the only Ectocarpus collected at site 7, and was also common at site 8. It matched published sequences of E. siliculosus sensu stricto (98–99% identity). GT7, obtained from a single isolate from site 5, was not known previously; the most similar sequence available was from a Kuckuckia from Isla Robinson Crusoe, but it showed only 75% identity. In culture, the isolate of GT7 developed phaeophycean hairs which are characteristic for Kuckuckia and absent in Ectocarpus (Hamel, 1939). As the ITS1 of GT7 had the same length as that of E. siliculosus, all isolates of this ITS1 length were examined for the presence of hairs, but they were only observed in the isolate from site 5. ITS1 of two isolates of this ITS1 length from site 7 and three from site 8 were sequenced and their ITS1 matched that of E. siliculosus. We concluded that all isolates with an ITS1 length of 714–716 bp, apart from the one from site 5, belonged to E. siliculosus. GT8 was isolated only from site 7; it was highly similar (95% identity of the variable ITS1 part) to strains from Valdivia, Southern Chile, and from the salt-polluted German Werra river. Because of its long ITS1 GT8 was tentatively classified as E. crouaniorum despite the fact that it shared only 90% sequence identity with E. crouaniorum individuals from Western Europe. Phylogenetic analyses (see end of Results section) confirmed this classification. GT9 presented a double band with lengths that corresponded to E. siliculosus and E. crouaniorum. Isolates of these two species have been shown to produce hybrids, but such hybrids were unable to form meiospores (Peters et al., 2010). Two strains of this putative hybrid genotype were isolated from site 7, where they co-occurred with E. siliculosus and E. crouaniorum.
The concatenated, alignable sequences from ITS1, ITS2, Rubisco spacer and cox3 (Table 2) had a length of 1533 bp. The different calculation methods used to generate phylogenetic trees gave comparable topologies and, consequently, only one is shown (Fig. 4). The three isolates of Kuckuckia, including the isolate of GT7 from site 5, formed a sister group to Ectocarpus, which was split into two major lineages. The first consisted of E. fasciculatus and an additional genotype from Europe, and the second comprised the remaining isolates. Within this latter clade, there were four subclades with strong and one subclade with moderate (> 80%) statistical support, but the hierarchy among the subclades varied and none of the hierarchies had strong statistical support in any analysis. In all trees, the genome strain clustered strongly with GT1. It was separated from E. siliculosus, E. crouaniorum, the reference strain from New Zealand and a clade formed by the genotypes 2 and 3. Strain Ec608 from site 7 (GT8), which had been tentatively classified with E. crouaniorum based on variable ITS1 sequences, strongly clustered with the reference strains of E. crouaniorum from Western Europe.
Our study revealed a genetically diverse Ectocarpus flora in Peru and northern Chile (Fig. 5). The three species recognized in Western Europe were all present, including the recently reinstated E. crouaniorum (Peters et al., 2010). In addition there was a Kuckuckia genotype and four genotypes of Ectocarpus which have not been recorded outside the study area, and for which we can only propose provisional designations, that is genotypes (GTs) 1–4. Identification of strains based on ITS1 length (Peters et al., 2010) proved insufficient in the present study as several genetically different isolates had similar ITS1 lengths. Sequencing of more (if not all) samples or PCR with genotype-specific primers will be required in similar studies in future.
Kuckuckia had been reported previously from Isla Robinson Crusoe but, based on sequence data, our isolate was different from both that strain and from six other previously sequenced isolates from Europe and South Africa which form two lineages within Kuckuckia (Stache-Crain et al., 1997). Our new strain from northern Chile adds a third lineage. Based on culture studies, all Kuckuckia were merged into a single species, Kuckuckia spinosa (Kützing) Kuckuck (Pedersen, 1989); the molecular data and our new isolate suggest that the systematics of the genus may need to be reassessed.
Our data suggest a separation of the study area into two parts. The northern half, extending from Ancón, central Peru to Pisagua, northern Chile (sites 1–6), was dominated by the genotype of the genome strain (GT4) which represented 61% of the strains isolated in this half but was absent further south. The southern half, ranging from Antofagasta to Quintay (sites 7–10), was dominated by E. siliculosus, E. crouaniorum and their hybrids. They represented 65% of the samples in the southern half but were missing in the north. This apparent split may reflect adaptation to different oceanographic environments (see the next paragraph), but could also be an artefact of the different sampling methods: in the northern region we collected at the end of summer from drift or subtidal material, while in the southern region we collected in spring and summer in the intertidal zone and from spores in the water column. The sampling zone and season certainly influence which species of Ectocarpus are encountered; for instance, in Brittany, E. crouaniorum occurs only in the high intertidal zone and macrothalli are seasonal in spring and early summer (Peters et al., 2010). More complete sampling would be required for an accurate quantification of the occurrence of the different genotypes in the area.
Isolates similar to the genome strain were the northernmost samples (Table 1). They represented two-thirds of the isolates from Peru. The distribution area of this genotype is characterized by continuous upwelling of cool water (16–20°C at the sea surface) irregularly interrupted by El Niño events which may result in several weeks of > 10°C higher sea surface temperatures as a result of the southward incursion of warm waters (Peters & Breeman, 1993). To date no Ectocarpus has been collected from similar low latitudes. It remains to be determined whether this genotype extends even further north, perhaps up to the deep-water kelp forests around Galápagos where its frequent host, Desmarestia, is present (Graham et al., 2007). We have not found GT4 south of Arica. Apart from the two individuals from Arica, which were collected in situ at 3–5 m depth, all samples of this genotype were from drift material, and their original habitat and depth distribution are not known. We have also only limited data on the seasonality of this genotype. In November 1988 it was present as a minute epiphyte or endophyte in Desmarestia peruviana. Our collections of macroscopic thalli of GT4 in 2006 were from late summer. The Museo de Historia Natural, Lima, houses seven herbarium specimens which are thalli of Ectocarpus, collected by César Acleto close to Lima between September and February from 1964 to 1976. If we assume that they belong to the same genotype, macroscopic thalli appear to be present in the field from September to March.
Our phylogenetic analyses, based on the genetic markers used in Stache-Crain et al. (1997) plus cox3, confirmed most lineages obtained in that pioneer study. However, the lineages 2a, 2d and 3 in Stache-Crain et al. (1997) were not represented in our taxon set. Our analyses confirmed that GT4 groups within the Ectocarpus subclade ‘siliculosi’ (Fig. 4). In a previous analysis involving fewer taxa (Peters et al., 2010), it had clustered with Stache-Crain et al.’s (1997) lineage 1a (E.siliculosus sensu stricto). The addition of more South American strains led to its separation from that lineage. The genome strain was also genetically distant from E. crouaniorum and from an Ectocarpus from New Zealand. The latter had been referred to as E. siliculosus in previous studies (Stache, 1989; Müller, 1991; Peters et al., 2004b). However, phylogenetic analyses (Stache-Crain et al., 1997; and the present paper) as well as crossing studies (Stache, 1989) indicated that it is separated from E. siliculosus sensu stricto.
Three other genotypes of Ectocarpus found in Peru and Chile did not cluster with any of the recognized species. GT1 was found in both parts of the study area. It formed minute dark spots on Macrocystis and Lessonia, was new to science, and in phylogenetic analyses clustered with GT4. Cross-fertility of these two GTs has not yet been examined. Thalli of GT2 were large (Fig. 2c) and saxicolous or epiphytic on different red and brown algae, such as Gracilaria and Desmarestia. GT2 co-occurred with GT4 at site 3. In phylogenetic analyses, GT2 did not cluster with GT4. A cross between the genome strain and a member of GT2 from site 8 produced viable hybrids with weakly growing erect thalli which did not form meiosporangia (A.F. Peters, unpublished data). Together these results suggest that GT2 and GT4 belong to different species. GT3, found once as a minute epiphyte at an exposed site, was previously unknown. In phylogenetic analyses it was separated from GT4 but it clustered with GT2; its variable part of ITS1 showed 84–93% identity with those of previous isolates from Atlantic South America, New Zealand and Isla Robinson Crusoe (Stache-Crain et al., 1997). Possibly GT2 and GT3 belong to a species that is more widely distributed in the Southern Hemisphere. GT3 has not so far been involved in cross-fertility tests.
In summary, the genome-sequenced Ectocarpus appears to be genetically separated from all genotypes with recognized names and also from all but one (GT1) of the genotypes endemic to Pacific South America or the Southern Hemisphere; it could therefore be regarded as a further species of Ectocarpus.
A proportion of the seaweeds that are reported from Peru and Chile also occur in the northeast Pacific (Santelices, 1989); this is the case for Ectocarpus acutus Setchell et Gardner reported in Chile from Coquimbo and Cartagena (Santelices, 1989; Ramírez & Santelices, 1991). It is possible that the lineages GT1+4 and 2+3 belong to this or another one of the species of Ectocarpus described from the Pacific coast of North America (Setchell & Gardner, 1922, 1925) or northeast Asia (Yoshida et al., 1995; Yoshida, 1998). So far, four isolates of Ectocarpus from this vast region have been sequenced. Strains from San Francisco and Santa Barbara (California, USA) and from Kanagawa, Enoshima Prefecture, central Japan, belonged to E. siliculosus sensu stricto, while a strain from Akkeshi, northeast Hokkaido, Northern Japan, by contrast, was more closely related to GT4 (Stache-Crain et al., 1997; Tanaka et al., 2010). The Akkeshi strain was nevertheless fully compatible in crosses with E. siliculosus sensu stricto, but it has not been tested against GT4 (Müller & Kawai, 1991). More comprehensive sampling in the North Pacific may be helpful for the taxonomic revision of the genus and the nomenclature of Pacific South American Ectocarpus.
Another aim of our work was to find strains that were sexually compatible with the genome strain. Seven out of our 36 new strains of GT4 produced unilocular sporangia in culture; female gametophytes obtained from meiospores of two sporophytes from sites 2 and 5 were used for crosses with the genome strain. In both combinations the zygotes developed into sporophytes capable of meiosis, and one of these sporophytes was selected for further experiments leading to the production of the genetic map of the Ectocarpus genome (Heesch et al., 2010).
We thank César Acleto for permission to study specimens in the phycological herbarium at the Museo de Historia Natural, Lima, and Matt Hall for advice during the molecular experiments. A.F.P. was supported by a travel grant from Université Pierre et Marie Curie for the collections and by a Ray Lankester Investigatorship by the Marine Biological Association of the United Kingdom for the molecular work. J.B. was supported by a Darwin Initiative Pre-project grant (Darwin Reference: EIDPR35). Collections and isolations in Chile were supported by FONDAP 1501 0001 (CONICYT) awarded to the Center for Advanced Studies in Ecology & Biodiversity (CASEB) Program 7. Additional support to J.A.C. was provided by an ICA grant.