Abstract
The haptophyte genus Pseudohaptolina (formerly Chrysochromulina clade B1-3) currently harbors two species: Pseudohaptolina arctica and Pseudohap-tolina sorokinii. In addition, Chrysochromulina birgeri is expected to belong to this genus due to its morphological similarity to P. sorokinii, but has not yet been genetically characterized. A strain belonging to Pseudohaptolina was brought into culture from Arctic waters, characterized by 18S and 28S rRNA gene sequencing as well as optical and transmission electron microscopy, and deposited in the Roscoff Culture Collection with the code RCC5270. Molecular and morphological data from RCC5270 were compared with those from previously described Pseudohaptolina and Pseudohaptolina-like species. Strain RCC5270 showed strong phylogenetic affinity to P. sorokinii, but TEM observations showed that RCC5270 possesses three types of organic body scale, rather than two as originally described in P. sorokinii. We found that the occurrence of three scale types is likely to have been overlooked in the original descriptions of both P. sorokinii and C. birgeri. We also found that environmental metabarcodes identical to the sequence of RCC5270 were abundant in the location from which C. birgeri was initially described (Gulf of Finland). We conclude that P. sorokinii and C. birgeri are conspecific and P. sorokinii is therefore synonymous with C. birgeri. Based on its phylogenetic placement and nomenclatural priority we propose the new combination Pseudohaptolina birgeri and emend the description of this species.
Introduction
Haptophyte identification is based on both molecular phylogeny and comparison of morphological features such as cell shape, length and movement of the haptonema, and ornamentation of organic body scales. The genus Pseudohaptolina was erected from the former Chrysochromulina B1-3 clade (Edvardsen et al., 2011). Like most haptophytes, Pseudohaptolina are solitary, flagellated and photosynthetic, with two species currently described: the type species Pseudohaptolina arctica Edvardsen & Eikrem (Edvardsen et al., 2011) and Pseudohaptolina sorokinii Stonik, Efimova & Orlova (Orlova et al., 2016). Both of these Pseudohaptolina species were described from high latitude northern hemisphere marine waters, P. sorokinii having been collected during an under-ice algal bloom in Amurskiy Bay in the northwestern Sea of Japan (Orlova et al., 2016). A new representative strain from the genus Pseudohaptolina was brought into culture from Canadian Arctic waters in 2016 (Gérikas Ribeiro et al., 2020) allowing comparison to previously described Pseudohaptolina species using morphological and genetic features.
Material and Methods
Strain RCC5270 was isolated into clonal culture from Canadian Arctic waters in 2016 (Gérikas Ribeiro et al., 2020), more specifically from Baffin Bay close to the Inuit village of Qikiqtarjuaq, Nunavut on Baffin Island (67◦28’ N, 63◦47’ W). The strain was identified by 18S rRNA gene sequencing and optical microscopy and deposited in the Roscoff Culture Collection (http://roscoff-culture-collection.org) with the code RCC5270. Strain RCC5268 was recovered from the same sample than RCC5270 and its 18S rRNA sequence (MH764749) shares 100% similarity of with that of RCC5270.
The nearly complete 18S rRNA gene was amplified using the primers 63F (5’-ACGCTTGTCTCAAAGATTA-3’) and 1818R (5’-ACGGAAACCTTGTTACGA-3’) (Lepère et al., 2011) and sequenced using the same primers and the internal primer 528F (5’-CCGCGGTAATTCCAGCTC-3’) (Zhu et al., 2005). The 28S rRNA gene was amplified and sequenced using primers D1R (5’-ACCCGCTGAATTTAAGCATA-3’) and D3Ca (5’-ACGAACGATTTGCACGTCAG-3’) (Lenaers et al., 1989). Sequencing was performed at Macrogen Europe (https://dna.macrogen-europe.com). Consensus sequences were generated using de novo assembly in Geneious® 10 (Kearse et al., 2012). The RCC5270 18S and 28S rRNA gene sequences were deposited in GenBank under accession numbers MT311519 and MT311520, respectively. For phylogenies, sequences from strain RCC5270 were aligned to closely related Haptophyta sequences from Genbank using the Muscle plugin in Geneious® 10 (Kearse et al., 2012).
Samples for transmission electron microscopy (TEM) were prepared as whole mounts fixed with osmium vapor following Eikrem (1996) with slight modifications (cooling of all equipment). Observations were made using a Jeol JEM-2010 FEG at the Imaging Core Facility at the Station Biologique de Roscoff, France. The size of more than 100 scales from RCC5270 and RCC5268 was measured from TEM micrographs using the imaging software ImageJ (https://imagej.nih.gov/ij/). Representative images are available at http://www.roscoff-culture-collection.org/rcc-strain-details/5270.
In order to determine the oceanic distribution of the species corresponding to RCC5270, we examined a large set of publicly available metabarcode datasets (Table 1) covering the V4 and V9 region of the 18S rRNA gene. Twenty-one oceanic 18S rRNA metabarcode datasets were downloaded and reprocessed with the dada2 R package (Callahan et al., 2016) following the standard operating procedure https://benjjneb.github.io/dada2/tutorial.html in order to produce amplicon single variants (ASVs). The taxonomy of each ASV was assigned using the dada2 assignTaxonomy function against version 4.12 of the PR2 database (Guillou et al., 2013) available at https://github.com/pr2database/pr2database/releases/tag/v4.12.0. Twenty datasets corresponded to the V4 of the 18S rRNA gene, and one to the V9 region (Tara Oceans). ASVs with a 100% match to the sequence of RCC5270 were selected and the number of reads in each sample determined using the R library dplyr. Maps and figures were drawn using the R libraries ggplot2, sf and cowplot.
Results and Discussion
The 18S rRNA gene sequence from RCC5270 was compared with similar sequences in GenBank including those from previously described Pseudohaptolina species. The best match of the sequence was to the two P. sorokinii 18S rRNA sequences in GenBank (KF684962 and KU589286), both linked to its original description, although only KF684962 is cited in the text of the original description. The 18S rRNA gene sequence of strain RCC5270 differs from sequence KF684962 by five base pairs (four substitutions and one deletion) in a 1,655 bp alignment and by only one base pair deletion when compared to KU589286 (1,213 bp alignment). The divergences from KF684962 seem to originate from sequencing errors in the P. sorokinii description, since they occur in well conserved positions (Figure 1) and when there is a base variation within these positions in related haptophytes, they do not match with those in the P. sorokinii sequence (Figure 1). Furthermore, the two sequences linked to the original description of P. sorokinii do not share the same substitutions.
The 28S rRNA gene sequence from RCC5270 has a six base pair difference to the only P. sorokinii 28S rRNA sequence available in GenBank (KU589284), which did not originate from the same isolate used for the description of P. sorokinii, and is not mentioned in Orlova et al. (2016). Both RCC5270 28S rRNA and KU589284 best hits in GenBank correspond to the environmental clone KU898784 from a sea ice sample in the Barrow Sea (Hassett et al., 2017), with 100% and 98% similarity, respectively.
The shape, size and ornamentation of the organic body scales are taxonomically important characters in Haptophyta, and usually more than one type of body scale occurs per species. Chrysochromulina birgeri Hällfors & Niemi (Hällfors and Niemi, 1974) was described before the genus Pseudohaptolina was erected, but is expected to be incorporated within Pseudohaptolina based on its morphological similarity to members of this genus. The discrimination between C. birgeri and other Pseudohaptolina species is only possible through morphological examination, since no molecular data or culture strains are available from its first description (Hällfors 1974). C. birgeri, P. arctica and P. sorokinii were all described as possessing two types of body scale (Hällfors and Niemi, 1974; Edvardsen et al., 2011; Orlova et al., 2016), usually referred to as ‘small’ and ‘large’ scales. For the P. sorokinii description (Orlova et al., 2016), three morphological features of the organic body scales are indicated as distinctive enough to assign it to a new species: horn morphology, shape of the connecting bridge and density of radial ribs. However, apart from the feature ‘number of radial ribs arranged in quadrants’ present in the so-called small scales, all other measurements overlap to some extent with those recorded for C. birgeri (see Table 1 from Orlova et al., 2016, page 511).
In general, the scale morphology of RCC5270 corresponds closely to that described for C. birgeri and P. sorokinii, including a radial pattern of ribs arranged in quadrants that coincide with the two orthogonal axes of the scale, and two horn-like projections connected by a straight or slightly curved bridge (Figure 2). However, both morphometric data and observations of TEM images of RCC5270 indicate that at least three types of organic scales can be differentiated (Table 2, Figure 2) using scale length, width and distance between the horns, and number of radial ribs per quadrant (Figure 4). Small scales of strain RCC5270 have 37-39 ribs on each quadrant (Figure 2B), as in the description of C. birgeri (Hällfors & Niemi, 1974), whereas the medium scales have 54-56 and large scales have 63-68 radial ribs per quadrant (Table 2). The distinction between small and medium scales is, however, most readily visible when comparing scale length versus width (Figure 4A). Medium and large scales have somewhat overlapping sizes, so their separation is better achieved by comparing distance between the horn bases versus width (Figure 4B), due to a clear distinctive horn bridge structure, with large scales presenting bigger and usually slightly curved bridges (Figure 2).
When measurements are conducted on the images displayed in the original descriptions, we found that the three types of scales can be distinguished for P. sorokinii (Figure 3A, Figure 4) and most likely also for C. birgeri, as shown in Figure 3D. Two P. sorokinii organic scales, identified as ‘small scales’ in the original description (Figure 3B and C, see also Orlova et al., 2016, page 510, figures 9 and 11), fall in the same size range as the ‘medium’ scales identified here (Figure 4), which impacts the number of ribs counted. In addition, independent measurements of small scales depicted in figure 8 of the original paper (Figure 3A in the present work), which are true small scales, fall outside the size range of small scales described by Orlova et al. (2016) (Figure 4). Unfortunately, the resolution of available P. sorokinii images is not sufficient to perform an independent count of the ribs in the small scales. The size of the connecting bridge was used by Orlova et al. (2016) as a distinctive feature of large scales, so small and medium scales were probably grouped together, which might have led to the discrepancies observed in the number of ribs per quadrant reported in the P. sorokinii description. In contrast, in the C. birgeri description medium and large scales with evident differences in the connecting bridge structure were grouped together as ‘large’ (Figure 3E and F). It is noteworthy that neither P. sorokinii nor RCC5270 scale measurements correspond precisely to the size limits described for C. birgeri (Hällfors & Niemi, 1974), particularly for small scales (Figure 4).
Other morphological characteristics used to differentiate P. sorokinii from C. birgeri by Orlova et al. (2016) are horn length and the shape of the connecting bridge. Orlova et al. (2016) reported long horn projections and curved connecting bridges, in contrast to the description of C. birgeri, although long horn-like projections connected by a curved bridge in large scales have already been reported for C. birgeri (Takahashi, 1981; Hällfors and Thomsen, 1979). The horn projections of large scales of RCC5270 are in general smaller than observed by Orlova et al. (2016), but are somewhat superimposed within their size range (Table 2). We also observed curved connecting bridges in the large scales (Figure 2A). There is therefore considerable overlap but some variability in the size and features of scales of RCC5270, P. sorokinii and C. birgeri which might reflect morphological plasticity within a single species, since heteromorphic life cycles have been observed within the Prymnesiales (Paasche et al., 1990; Edvardsen and Vaulot, 1996).
The metabarcode datasets used to determine the oceanic distribution of RCC5270 correspond to more than 2,200 samples included in large scale surveys such as Ocean Sampling Day (OSD) and the Tara Oceans and Malaspina expeditions that sampled a wide range of coastal and oceanic waters as well as more limited studies from polar waters and the Baltic Sea. We did not retrieve any V9 metabarcodes identical to the RCC5270 sequence. We did, however, retrieve six V4 metabarcodes (ASVs) that were 100% identical to the RCC5270 sequence (Figure S1). In contrast, no exact match was found to either KF684962 or KU589286 P. sorokinii in any of these datasets, which further corroborates the assumption that the mismatch between 18S rRNA P. sorokinii and RCC5270 sequences are due to sequencing errors. The RCC5270 metabarcodes were only observed in the Arctic Ocean and in the Baltic Sea from ice and water samples as well from algal aggregates collected from the deep-sea floor (Figure 5A-B). Metabarcodes identical to the sequence of RCC5270 were particularly abundant in three datasets (Table 1) from the Polarstern expedition in the Central Arctic Ocean (Rapp et al., 2018), from the Nares strait, the northernmost outflow gateway of Baffin Bay (Kalenitchenko et al., 2019) and from the Gulf of Finland (Baltic Sea) (Enberg et al., 2018). At the latter location, which corresponds to the region from which C. birgeri was initially described, metabarcodes identical to the RCC5270 sequence first appeared in February in the ice where they peaked in early March and then increased massively in the water column one month later, representing up to 70% of the metabarcodes at the time the ice melted in mid-April (Figure 5C). These data indicate that RCC5270 is an ice alga that can seed and proliferate in the water column and even accumulate on the deep-sea floor.
Conclusions
We isolated a culture strain from the Arctic which was genetically affiliated to P. sorokinii. Morphological data indicate that a third scale type was overlooked in the original description of P. sorokinii (Orlova et al., 2016), impacting the number of radiating ribs described for each scale type. We also found that C. birgeri cells have three types of organic body scale, not two as reported in the original description (Hällfors and Niemi, 1974). Metabarcode data indicates that sequences identical to that of RCC5270 were abundant near the type locality of C. birgerii. We conclude that P. sorokinii is conspecific with the formerly described C. birgeri and we therefore transfer C. birgeri to the genus Pseudohaptolina and emend its description. P. birgeri is the valid name for this species due to nomenclatural priority over P. sorokinii.
Taxonomic appendix
Pseudohaptolina birgeri (Hällfors & Niemi) Ribeiro and Edvardsen comb. nov. emend. Ribeiro and Edvardsen
BASIONYM
Chrysochromulina birgeri Hällfors & Niemi in Hällfors & Niemi (1974). Memoranda Societatis pro Fauna et Flora Fennica 50. Drawing Fig. 4.
SYNONYM
Pseudohaptolina sorokinii Stonik, Efimova & Orlova.
EMENDED DESCRIPTION
Scaly covering composed of three round to oval scale types. Small scales have width x length c. 0.6-1.4 x 1.1-1.7, medium scales c. 1.1-2 x 1.5-2.4 and large scales c. 1.1-2.1 x 1.9-2.8 nm. All scales with radial ribs on both distal and proximal faces. Small scales have 37-39 radial ribs per quadrant, medium scales 54-60 and large scales 63-68. Medium and large scales have two horns on the distal face. The distance and form of the horns are different in medium and large scales.
Contributions
Contributed to conception and design: CGR, IP, DV, BE
Contributed to acquisition of data: CGR, ALS, IP, DV, BE
Contributed to analysis and interpretation of data: CGR, ALS, IP, DV, BE
Drafted and/or revised the article: CGR, ALS, IP, DV, BE
Approved the submitted version for publication: CGR, ALS, IP, DV, BE
Funding information
Financial support for this work was provided by the Green Edge project (ANR-14-CE01-0017, Fondation Total), the ANR PhytoPol (ANR-15-CE02-0007) and TaxMArc (Research Council of Norway, 268286/E40). ALS was supported by FONDECYT grant PiSCOSouth (N1171802). CGR was supported by the FONDECYT project 3190827.
Competing interests
The authors have no competing interests.
Data accessibility statement
Supporting data have been deposited to GitHub https://github.com/vaulot/Paper-2020-Ribeiro-Pseudohaptolina.
Supplementary material
Supplementary Data
Supplementary data are available on GitHub at https://github.com/vaulot/Paper-2020-Ribeiro-Pseudohaptolina
Supplementary Data S1: Alignment of sequences for 18S rRNA gene (fasta file).
Supplementary Data S2: Alignment of sequences for 28S (fasta file).
Supplementary Data S3: Scale measurements (xlsx file).
Supplementary Data S4: Number of P. sorokinii reads in each of the metabarcode samples analyzed (xlsx file).
Supplementary Data S5: Alignment of the V4 region of the 18S rRNA for Pseudohaptolina reference sequences and metabarcodes (fasta file).
Supplementary Figures
Acknowledgments
We are grateful to Sophie Le Panse from the Merimage microscopy platform at the Roscoff Marine Station for assistance with the transmission electron micrographs and to the Roscoff Culture Collection for maintenance of the algal strain.