ABSTRACT

INTRODUCTION

Several algal species are reported to cause the macroscopic phenomenon of orange and red to pinkish snow, the latter often referred to as ‘watermelon snow’ or ‘blood snow’ when occurring as mass developments, including Chlamydomonas nivalis (Wille 1903; Kol 1968; Czygan 1970; Fjerdingstad et al. 1974; Kawecka 1981; Gradinger and Nürnberg 1996; Leya 2004), Chloromonas nivalis (Remias et al. 2010; Matsuzaki, Nozaki and Kawachi 2018), Cr. brevispina (Hoham, Roemer and Mullet 1979), Cr. polyptera (Remias et al. 2013), Cr. rubroleosa (Ling and Seppelt 1993), Cr. nivalis subsp. tatrae (Procházková et al. 2018a), Chlorosarcina antarctica (Ling 2002), Smithsonimonas abbotii (Kol 1942), Chlainomonas kolii (Novis 2002) and Chlainomonas rubra (Hoham 1974).

One of the first field samples of red snow was collected during the Northern Expedition under Captain Ross in August 1818 to Baffin Bay (75°54' N 67°15' W) between northern Greenland and Canada (Bauer 1819) and brought back to England to be investigated. Red snow has been reported from almost all polar and alpine regions (Kol 1968; Marchant 1982; Mataloni and Tesolín 1997; Thomas and Broady 1997). The first phylogenetic studies that identified a separate clade comprising solely red spherical cysts were based on 18S rRNA gene data (Leya 2004; Remias et al. 2010, 2013), but a deeper molecular identification based on nuclear ribosomal DNA internal transcribed spacer 2 (ITS2) sequence data so far has been missing for this taxon except for a recent study by Segawa et al. (2018) investigating samples from both poles. Red spherical cysts have also been the focus of a few metagenomic studies carried out in the USA (Brown, Ungerer and Jumpponen 2016), Japan (Terashima et al. 2017), the Arctic (Lutz et al. 2016; Segawa et al. 2018) and the Antarctic (Segawa et al. 2018). However, detailed investigations of red snow from mid-latitude mountain ranges like those in Europe are lacking.

Though modern microscopes are much better equipped than 200 years ago, a persistent problem is that the red spherical cysts abundantly found in melting snow simply do not reveal enough intracellular details that can be used for a deep morphological description and thus, identification. The unicellular, thick-walled cyst stages [‘cysts’ = unicellular life cycle stage protected from unfavourable environmental conditions with a wall, which was built exogenously or endogenously according to Ettl and Gärtner (2014)] dominate snow surfaces during summer. The characteristic dark blood-red pigment astaxanthin is accumulated in similarly large quantities only in Haematococcus lacustris, Chlainomonas rubra and some others (Remias, Lütz-Meindl and Lütz 2005). In many cases astaxanthin covers the cysts' chloroplast structure almost entirely, making light microscopy hardly illuminating. A first revision of Chlamydomonas by Pröschold et al. (2001) revealed how important chloroplast morphology is for consistency with the phylogeny of the Chlamydomonadaceae. Many snow algal species are held as live cultures in culture collections (e.g. CCCryo, UTEX, CCAP), but none of these ‘green’ species ever formed a single phylogenetic clade together with red spherical cysts from red snow. Instead, all 18S rRNA gene phylogenetic studies have revealed a single clade comprising immotile spherical red cysts (Leya 2004; Leya, Müller, Ling et al.2004; Remias et al. 2010, 2013). This ‘red snow’ clade is clearly separated from other clades of Chlamydomonas, Chloromonas, Lobochlamys, Oogamochlamys, Haematococcus, Carteria, etc. Most likely, green, flagellated isolates from red snow field samples often were misinterpreted as having resulted from red spherical cysts or even vice versa; e.g. Novakovskaya et al. (2018) thought to have observed such a phenomenon when linking green isolates of Cr. reticulata from snow samples to red cysts in the same sample. The lack of knowledge about other life cycle stages, whether vegetative (e.g. green actively dividing cells with a well-visible chloroplast) or sexual (e.g. gametes), and the lack of sufficient molecular data from red spherical cysts, have been the most hindering in identifying their true taxonomy.

In this study we used a polyphasic approach to study field material composed of spherical red and orange cysts sampled from alpine habitats in Europe, North America, South America and both polar regions. Our aims were (i) to reveal their molecular diversity based on ITS2 rDNA and rbcL gene sequences, (ii) to describe their morphological and ultrastructural cellular details using light, transmission and scanning electron microscopy, and (iii) to characterise their habitat conditions (pH, conductivity, elevation). Furthermore, we wanted (iv) to describe the geographic distribution of haplotypes detected by establishing a haplotype network, and (v) to test the existence of spatial patterns (Mantel test, sPCA). Based on the results, a new genus-level taxonomy of the examined species is proposed with a description of two new species: the new genus Sanguina Leya, Procházková et Nedbalová with Sanguina nivaloides Procházková, Leya et Nedbalová and Sanguina aurantia Leya, Procházková et Nedbalová.

MATERIALS AND METHODS

Sampling localities

Patches of clearly visible red- or orange-coloured snow caused by spherical red or orange cells were collected from 42 localities in melting seasons between the years 1999 and 2018 (Fig. 1, Supplementary Tables 1 and 2): 22 samples originated from alpine sites in continental Europe, three from sites in North America and one from South America. Moreover, 14 samples were collected at polar sites on the island of Spitsbergen (Svalbard archipelago) and two from polar sites on the Antarctic Peninsula.

Open in a separate window

Figure 1.

Sampling locations of Sanguina nivaloides (red circles), Sanguina aurantia (yellow circles) and regions where both species were sampled (red-yellow circles) in continental Europe (right), on the archipelago of Svalbard (left), in North and South America (middle) and in Antarctica (middle). Numbering corresponds to the sample origin from the following countries: Austria (1–10; 1-Haferkarlespitze, 2-High Tauern, 3–5 Kühtai, 6–10 Ötztal Alps), Italy (11–14; 11–13 Dolomites, 14-Sarntal Alps), Norway (15–19), Slovakia (20-High Tatra Mountains), Slovenia (21-Julien Alps), Switzerland (22-Urner Alps), Svalbard (Norway) (23–36; 23-Amsterdamøya, 24-Ny Ålesund, 25–29 Longyearbyen, 30-Raggfjellet, 31-Midterhuken, 32-Doktorbreen, 33–35 Hornsund, 36-Chomjakovbreen), Colorado (USA) (37–39 Rocky Mountains), Argentina (40-Patagonia), Antarctica (41–42 Antarctic Peninsula). Habitat description of sampling localities including geographical data are shown in Supplementary Table 1.

To ensure the selection of ‘clean’ (mono-specific) bloom spots and to avoid a taxonomically diverse sample, a field microscope was used to check the sample in the field before collection. In the lab, samples were further thoroughly examined using laboratory light microscopy (see below) prior to molecular analyses. If the sample appeared to be too diverse, it was discarded. In addition, Alanagreh et al. (2017) reported that the dominant haplotype was usually evident from the Sanger sequencing, similar as in cloning approach.

Measurements of pH and conductivity and snow water equivalent (referred to as ‘snow water content’ in the paper cited) were carried out according to Procházková et al. (2018a).

In addition we were also able to investigate red cysts collected in 1851, namely from the desiccated sample ‘MB_ES_1781c’, further labelled ‘Rother Schnee Crimsen [sic] Cliff Beverly 1851, 76°3'N.B. 1781c’ obtained from the Ehrenberg Collection at the Museum für Naturkunde in Berlin. This historical material was collected during the Grinnel Arctic Expedition (1850–1851) under Lieut. De Haven of the U.S. American Marines from red snow on scarlet-coloured cliffs (Crimson Cliffs) in Baffin Bay on the Greenland coast (Beverly Greenland), though a bit further north of the location, from where the Ross type sample of 1818 originated (75°54' N 67°15' W; Ross 1820). Unfortunately, the latter material was not traceable and could not be included in this study.

RESULTS

Cytological observations

Light microscopy (Fig. 2) and electron microscopy (SEM: Fig. 3, TEM: Fig. 4) showed that field samples of Sanguina nivaloides consisted of red spherical cells exhibiting different accumulation levels of secondary carotenoids in abundant lipid droplets, which in mature cysts fully masked the chloroplasts making any detailed structure unrecognisable by light microscopy (Fig. 2A) and thus unusable for identification. A single parietal chloroplast and one or two large, unpigmented vacuoles were observed in young cells in their early stage of development only (Fig. 2B–D and Fig. 4C and D). The typical mature red cyst was smooth-walled (Figs 2A and 3A), but in two cases the red snow also consisted of populations of cysts with nipples (Figs 2E, 3B and 4F). A few scattered cysts with short blunt nibs (Fig. 3C) or pronounced papillae (Figs 2F and 3D) or cells carrying uniformly distributed small smooth papillae (Fig. ​(Fig.3E)3E) were noticed in a population of smooth-walled cells . Sanguina aurantia field samples (Fig. 5) consisted of smaller spherical cysts with a striking orange-coloured cell content, the cells had an outer distant cell layer or the distant layer was absent. A few scattered cells of ‘ruby’-like cysts (the term ‘ruby’ is frequently used for this type of cyst as the dark red pigmented cyst content and the wall structure appear like a ruby encased in mother-of-pearl) were found in two red snow samples (Supplementary Fig. 3A-F, marked with an asterisk in Supplementary Table 4). We are certain that no sequences of those ruby cysts were used in any of the analyses in this study, as they would have been greatly outnumbered by the dominating smooth-walled cysts, even assuming they were equally amenable to DNA extraction in the first place. As far as the intracellular ultrastructure of our samples is concerned, mature cysts had a naked pyrenoid in a single chloroplast, which is forced into a central position by surrounding lipid globules (Fig. ​(Fig.4A)4A) and cell walls consisted of two layers (Fig. 4B). Early stages of red cysts differed in the presence of a clearly visible single parietal chloroplast (Fig. 4C and D) and a three-layered cell wall, the outermost being fuzzy (Fig. 4E). Orange-coloured cysts of Sanguina aurantia differed from red cysts of Sanguina nivaloides in the structure of the second layer of their cell wall, being multilayered and slightly undulating on the surface (Fig. 5C and E). This layer most likely reflects the maturity of the cysts, as orange-coloured cysts of the same cell diameter but with smooth surfaces also existed in this population of Sanguina aurantia (Fig. 5D). Cytosolic electron-dense vacuoles filled with a crystalline structure were observed in both red cysts of Sanguina nivaloides and in orange-coloured cysts of Sanguina aurantia (Fig. 5G).

Open in a separate window

Figure 2.

Light microphotographs of Sanguina nivaloides. (A) Typical mature red cyst with smooth cell wall (holotype specimen CCCryo RS_0015–2010), abundant lipid droplets entirely masking the cell interior and structure of the chloroplast. (B-D) Young red cysts in early stage of development, (B) showing visible parts of a single parietal chloroplast (paratype sample WP123), (C-D) depicting the large unpigmented vacuole (arrowhead) (samples: Sva4 and Saddle). (E-F) Mature red cysts of less frequent haplotypes causing red blooms, (E) visualises nipples on the cell surface (arrowheads, paratype specimen CCCryo RS_0003–2004). (F) shows pronounced papillae on cell surface (arrowheads, sample 4D). Scale bar 20 µm.

Open in a separate window

Figure 3.

Scanning electron microphotographs of Sanguina nivaloides. (A) Typical mature red cyst with smooth cell wall (holotype specimen RS_0015–2010). (B-D) Mature red cysts of less frequent haplotypes causing blooms, (B) showing nipples on cell surface (paratype specimen RS_0003–2004), (C) depicts the smooth surface with short blunt nibs, which may represent the compressed remnants of the presumably transient primary cell wall, this cell wall layer may decompose over time (note: on the left side of the cell 'papillae' are present, on the right side they are absent, specimen RS_0011), (D) shows the pronounced papillae on the cell surface (sample 4D). (E) Scattered cells with small papillae on their cell surfaces in a population of smooth-walled cells (sample redCol). Scale bars 5 µm.

Open in a separate window

Figure 4.

Transmission electron microphotographs of Sanguina nivaloides. (A, B) Mature red cyst (holotype specimen RS_0015–2010), (A) shows the centrally located plastid with a naked pyrenoid, surrounded by abundant lipid droplets. (B) depicts the electron-dense thick cell wall covered with a very thin transparent layer. (C-E) Young red cyst (paratype sample WP123), (C) shows the large cytoplasmatic-free area (arrowhead) between the cytoplasmatic membrane and the cell wall, the nucleus and the parietal plastid, as well as parts of the Golgi aparatus, no lipid bodies visible, (D) depicts the parietal plastid with a naked pyrenoid, lipid bodies and vacuoles, (E) shows a detail of the cell wall consisting of three layers: the innermost—thick electron dense (CW1), the middle one—very thin electron transparent (CW2; trilaminar sheath) and the fuzzy upper layer (CW3; presumably the primary cell wall). (F) A detail of a cell wall nipple of less frequent haplotypes (paratype specimen RS 0003–2004). C = chloroplast, CY = cytoplasm, CW1 = cell wall 1, CW2 = cell wall 2, G = Golgi stacks, L = lipid globules, N = nucleus, P = pyrenoid, PM = plasmatic membrane, V = vacuole. Scale bars = 2 µm for (A, C, D). Scale bars = 0.5 µm for (B, E, F).

Open in a separate window

Figure 5.

Light and electron micrographs of Sanguina aurantia (holotype specimen RS_0017–2010), (A) depicts the distant outer cell layer. (B) shows that the distant cell layer is absent. (C-D) outermost layer is wrinkled and when it is decomposed, the smooth cell wall below may appear. (E) shows details of the cell wall with the electron-dense layer (CW1) being covered by several more or less undulating thin transparent layers (CW2, CW3). (F) shows a cross-section of the cyst showing the nucleus and the centrally compressed chloroplast. (G) depicts the cytosolic electron-dense vacuoles filled with crystalline structures. C = chloroplast, CW1 = cell wall 1, CW2 = cell wall 2, CW = cell wall 3, L = lipid globules, N = nucleus, V = vacuole. Scale bars = 10 µm for (A, B), 5 µm for (C, D), 0.5 µm for (E) and 2 µm for (F, G).

Size ranges of the orange-coloured cysts and red cysts overlapped, with 6.2–15.5 µm and 7.3–39.0 µm, respectively. The variability of cell size ranges in field samples of each haplotype of both snow algal species (for description see below) is shown in Supplementary Fig. 4. The material from Baffin Bay (sample ‘MB_ES_1781c’ from the Ehrenberg Collection at the Museum für Naturkunde in Berlin) contained cells in the size range typical for red cysts (Fig. 6).

Open in a separate window

Figure 6.

Cell size ranges of cysts of the most common haplotypes of Sanguina nivaloides (H1 and H2 = red cysts with smooth cell wall, H3 = red cysts with nipples) and Sanguina aurantia (HA1 = small orange cysts) in the dataset studied compared with cell sizes from the Crimson Cliffs in Baffin Bay on Greenland (specimen MB_ES_1781c) (Ehrenberg 1851; Leya 2008).

Molecular identification of genus Sanguina and its closest relatives

According to our 18S rRNA gene and rbcL phylogenies (Figs 7 and 8), the genus Sanguina forms a monophyletic lineage within the Chlamydomonadaceae containing samples from nearly all regions of the Earth (Greenland, North America, South America, Africa, Europe, Asia, Svalbard and Antarctica). The closest relatives to the genus Sanguina are organisms found on glaciers or glacier-associated environments, such as the strain CCCryo 086–99 (cf. Ploeotila sp.) from a glacier surface or uncultured eukaryote clones from glacial debris in Alaska and periglacial environments on the top of Mount Kilimanjaro in Tanzania. Slightly distant relatives represent strains from the Arctic permafrost in close vicinity to glaciers and snow fields, e.g. CCCryo 101–99 and CCCryo 133–01 (both cf. Sphaerocystis sp.) or Arctic snow fields, e.g. CCCryo 147–01 (cf. Coenochloris sp.), and from soil such as ‘uncultured Haematococcaceae clones’ isolated from an aspen rhizosphere and Gloeocystis sp. from desert soil in Washington State. Phylogeny based on rbcL data (Fig. 8) showed an increased molecular variability in comparison with 18S rRNA gene data among specimens of the genus Sanguina because of the degeneracy of the genetic code, which accounts for the existence of synonymous mutations. Interestingly, the small orange cysts of Sanguina aurantia formed a small subclade within the red cyst samples of Sanguina nivaloides. The rbcL phylogeny confirmed the positions of cf. Sphaerocystis and cf. Coenochloris as slightly distant relatives to Sanguina nivaloides. Additionally, it revealed that these genera form closely related lineages (Fig. 8) to Achoma brachiatum isolated from an alpine herbfield soil in New Zealand (Novis and Visnovsky 2012). Whether any of them could be assigned to the genus Achoma is discussed further below, but remains to be investigated. To sum up, the monophyly of the Sanguina-clade and the phylogenetic positions of its closest known culturable relatives were confirmed for both markers (Supplementary Figs 5 and 6). Moreover, our phylogenetic analyses demonstrate that the positions of strains SAG 26.86 of Cr. typhlos (= UTEX 1969 = CCCryo 122–00), and Cr. reticulata SAG 29.83 (= UTEX 1970 = CCCryo 213–05) in the Chloromonas-clade based on 18S rRNA gene and rbcL phylogenetic trees clearly differed from the position of S. nivaloides (Supplementary Figs 5 and 6). These strains were isolated from snow samples and were formerly believed to cause red snow, and thus, they originally were (mis)identified as Cd. nivalis and Cd. yellowstonensis (Sutton 1970), respectively. The former is still listed under two different names, as Cr. typhlos (SAG) and Cd. augustae (UTEX), as is the latter, Cr. reticulata (SAG) and Cr. clathrata (UTEX); in both cases the designations at SAG are correct. Moreover, strain CCAP 11/128 (= UTEX 2824 = CPCC 528 = CCCryo 214–05; accession number HQ404886.1) is listed as Cd. nivalis at UTEX (and others), but in our phylogeny proved to be Cr. typhlos based on the comparison of its ITS2 sequences (pairwise blast search), with 100% identity and no CBC for this marker with CCCryo 122–00 of Cr. typhlos (= SAG 26.86 = UTEX 1969 = CCAP 11/51B = IAM C-263 = NIES 2226 = IU 1969 [original deposition]; accession number HQ404869.1).

Open in a separate window

Figure 7.

18S ribosomal RNA gene-based Bayesian phylogenetic tree. The Sanguina-clade with Sanguina nivalis and Sanguina aurantia is highlighted in a grey box. Posterior probabilities (≥0.95) and bootstrap values from maximum likelihood analyses (≥75%) are shown. Full statistical support (1.00/100) is marked with an asterisk. Thick branches represent nodes receiving the highest posterior probability support (1.00). Newly obtained sequences are in bold. Accession numbers, strain or field sample codes are indicated after each species name. Species names of the closest species that are available as laboratory cultures are underlined. Abbreviation: unc, uncultured.

Open in a separate window

Figure 8.

rbcL gene-based Bayesian phylogenetic tree. The Sanguina (S.)-clade with Sanguina nivaloides and Sanguina aurantia is marked in a grey box. Posterior probabilities (≥0.95) and bootstrap values from maximum likelihood analyses (≥75%) are shown. Full statistical support (1.00/100) is marked with an asterisk. Thick branches represent nodes receiving the highest posterior probability support (1.00). Newly obtained sequences are in bold. Accession numbers, strain or field sample codes are indicated after each species name. Species names of the closest species that are available as laboratory cultures are underlined.

The CBC species concept applied to Sanguina nivaloides and Sanguina aurantia

The network analysis of field samples of Sanguina nivaloides from Europe, North America, South America and Antarctica (35 specimens) applying a 95% connection limit resulted in a total of 18 ITS2 haplotypes interconnected in one network (Fig. 9). Haplotypes differed by one up to seven nucleotide changes out of 205 base pairs (Table 1). The network analysis of Sanguina aurantia field samples from Svalbard (seven specimens) resulted in two ITS2 haplotypes differing by one nucleotide change out of 213 bp. Based on the comparison of ITS2 rDNA secondary structures, one CBC was found between Sanguina nivaloides (the red cysts) and Sanguina aurantia (the small orange cysts) at the 13th position in helix II (Fig. 10). However, there is no consensus as to whether the CBC detected at this position can be used for a species separation within the Chlorophyceae. In other words, some authors treat this position as a variable one (e.g. Demchenko et al. 2012), others as a conservative one (e.g. Caisová, Marin and Melkonian 2013). Most recently, CBCs in the 5´ end of helix III were only treated as relevant for evaluating species boundaries between closely related taxa (Segawa et al. 2018). Sequence identities between red cysts and small orange cysts were ≥95.1%, whereas sequence identities within red cysts were ≥96.5% (Supplementary Table 7), suggesting that both types of cysts investigated in this study are very closely related. Among the red cysts of Sanguina nivaloides no CBC was found in the conserved parts of this marker structure, suggesting that according to the CBC species concept as proposed by Coleman (2000), all our samples of red snow from Europe, North America, South America and Antarctica were caused by one species. The presence of the one CBC and the number of base differences between ITS2 haplotypes of Sanguina nivaloides (red cysts) and Sanguina aurantia (orange cysts) are summarised in Table 1. Interestingly, two haplotypes of red cysts (H4 and H5) lack the CBC at the 13th position in helix II in their ITS2 secondary structure when compared to small orange cysts (Table 1). When comparing three molecular markers, the lowest intraspecific variability between S. nivaloides and S. aurantia haplotypes was detected in 18S (≤0.2%), was higher in rbcL (≤4.4%) and the highest was in ITS2 (≤4.9%) (Supplementary Tables 7–9).

Open in a separate window

Figure 9.

Geographic distribution of ITS2 rDNA haplotypes detected in this study: (A) Sanguina nivaloides, (B) Sanguina aurantia. Each haplotype network was constructed by a statistical parsimony method with a 95% connection limit. The Sanguina nivaloides network includes haplotypes from polar regions (Antarctica, Svalbard) and alpine regions in Europe (Austria, Italy, Norway, Slovakia, Slovenia, Switzerland), North America (Colorado) and South America (Argentina). The Sanguina aurantia network includes haplotypes from Svalbard. Each ball represents a haplotype (i.e. in this study's field samples, which have identical ITS2 sequences). The size of the ball is proportional to the number of specimens sampled, which belong to that specific haplotype. The colour represents a country of sampling. Lines connect each haplotype with its most similar relative. Open dots represent mutation steps between haplotypes, one dot indicates a change of one base pair.

Open in a separate window

Figure 10.

Comparison of the secondary structure of the ITS2 rDNA transcripts of 18 haplotypes of Sanguina nivaloides (H1-H18) and two haplotypes of Sanguina aurantia (HA1, HA2). Helices are labelled with Latin numbers (I-IV). Nucleotide differences of all haplotypes are shown outside the structure and are linked by dotted lines. Numbers 2–18, A1 and A2 indicate in which haplotype the difference was detected when compared to haplotype number 1 of Sanguina nivaloides (represented by the holotype specimen CCCryo RS_0015–2010, accession number MK728599). One compensatory base change between Sanguina nivaloides and Sanguina aurantia is indicated by the bold rectangle. The dashed rectangle at the end of helix II is showing an expansion segment (length is not conserved). Note: the U–U mismatch in helix II (arrows). The 5´ and 3´ ends of the ITS2 transcript are indicated.

Table 1.

Amount of compensatory base changes in the ITS2 secondary structure (lower left) and the number of base differences/number of indels (each of two bp length) (upper right; e.g. ‘3/4’ means that 3 differences and 4 indels were found) between 18 haplotypes of Sanguina nivaloides (H1-H18) and two haplotypes of Sanguina aurantia (HA1, HA2).

	H1	H2	H3	H4	H5	H6	H7	H8	H9	H10	H11	H12	H13	H14	H15	H16	H17	H18	HA1	HA2
H1		1	2	1	5	4	2	2	3	1	2	3	2	3	3	4	3	3	3/4	4/4
H2	0		1	2	4	5	3	1	2	2	1	2	3	2	2	3	2	2	2/4	3/4
H3	0	0		3	5	6	4	2	3	3	2	3	4	3	3	4	3	3	3/4	4/4
H4	0	0	0		6	5	3	3	4	2	3	4	3	4	3	4	4	4	4/4	5/4
H5	0	0	0	0		7	5	3	6	6	5	6	7	6	6	7	3	6	4/4	5/4
H6	0	0	0	0	0		2	4	5	5	5	6	6	6	6	7	5	6	6/4	5/4
H7	0	0	0	0	0	0		2	5	3	3	4	4	4	4	5	3	4	5/4	6/4
H8	0	0	0	0	0	0	0		3	3	2	3	4	3	3	4	1	3	3/4	4/4
H9	0	0	0	0	0	0	0	0		4	3	4	5	4	4	5	4	4	4/4	3/4
H10	0	0	0	0	0	0	0	0	0		3	4	3	4	4	5	4	4	4/4	5/4
H11	0	0	0	0	0	0	0	0	0	0		1	4	1	1	2	3	1	3/4	4/4
H12	0	0	0	0	0	0	0	0	0	0	0		5	2	2	3	4	2	4/4	5/4
H13	0	0	0	0	0	0	0	0	0	0	0	0		5	5	6	5	5	5/4	6/4
H14	0	0	0	0	0	0	0	0	0	0	0	0	0		2	3	4	1	4/4	5/4
H15	0	0	0	0	0	0	0	0	0	0	0	0	0	0		1	4	2	4/4	5/4
H16	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0		5	3	5/4	6/4
H17	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0		4	4/4	5/4
H18	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0		4/4	5/4
HA1	1	1	1	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1		1
HA2	1	1	1	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	0

Open in a separate window

DISCUSSION

The first collections of red snow date back to British expeditions to the far north looking for a navigable North-West passage (N.N. 1818; Bauer 1819; Ross 1819; Bauer 1820; Ross 1820). In our study many samples of red snow cysts from around the world have yielded sequences that fell in the new clade Sanguina, so one is tempted to assume that the historical sample of 1818 also probably represents the red snow-causing species S. nivaloides; however, this cannot be proven yet. A genetic comparison was not possible as the historic sample of 1818 could not be located in museums or herbaria in either London or Berlin. Intact DNA could also not be isolated from the similar sample collected during the Grinnell Arctic Expedition at Beverly Greenland [76°3' N in 1851 (Ehrenberg 1851; Leya 2008) a bit further north of the original location of 1818–1819 at the Crimson Cliffs (75°54' North, 67°15' West)]. Thus, we do not attempt to synonymise S. nivaloides with Cd. nivalis. In this study we have suggested that all our investigated red and orange spherical cysts belong to Sanguina nivaloides and Sanguina aurantia, respectively. We showed that the genus Sanguina forms one independent lineage in the Chlamydomonadaceae. The conservative nuclear marker 18S was 100% identical for these two species. Strains CCCryo 086–99, CCCryo 101–99, 133–01 and 147–01 represent the most closely related and culturable isolates/species and appear to be distributed into several lineages that are sister genera to the clade of Sanguina, some of them possibly representing the genus Achoma (Novis and Visnovsky 2012). The revised genus Chloromonas (Pröschold et al. 2001) so far contains the majority of other culturable snow algal species involved in coloured snow, but is robustly separated from the Sanguina-clade.

This study is the first record of ITS2 sequence data of spherical cysts forming red snow from Argentina and many mid-latitude alpine sites in Europe. Moreover, our data will increase the knowledge about this taxon due to several new findings on Svalbard. The ITS2 marker was previously used in this context only in studies of snow samples from the USA (Brown, Ungerer and Jumpponen 2016), Greenland, Alaska, Japan and Tajikistan (referenced in accession numbers only, see Materials and Methods) and also most recently from Antarctica and Svalbard (Segawa et al. 2018). The set of different molecular data used in this study has proved to provide valuable insights into the relationship of this taxonomically complex algal group. Haplotype networks have been used to demonstrate the intraspecific diversity (i.e. within a species) (Škaloud et al. 2015). Based on our results, S. nivaloides represents a diverse species containing 18 different ITS2 genotypes and S. aurantia has two ITS2 genotypes. According to the metagenomic study of Brown, Ungerer and Jumpponen (2016) in Colorado and Washington (USA), sequences of this locus of cysts in blooming snowfields were distributed into 30 haplotypes classified as ‘Coenochloris’ and 27 haplotypes termed as ‘Chlamydomonas’. Further samplings in other mountain regions might reveal additional unique genotypes as shown by Segawa et al. (2018). Similarly, 15 haplotypes of ITS2 rDNA were resolved for the common ‘blood rain alga’ Haematococcus lacustris (referred to as H. pluvialis in the paper cited) in temperate Europe (Allewaert et al. 2015). Intraspecific diversity within the widespread lichen photobiont Asterochloris lobophora in Europe, America and Asia accounted for 18 haplotypes of ITS2 rDNA, while half that amount of intraspecific diversity was found for other Asterochloris species (Škaloud et al. 2015).

Although the overall mean distance among the ITS2 red cysts haplotypes of S. nivaloides was rather low [p-distance of 1.7%; this value was similarly low for Asterochloris lineages described by Škaloud et al. (2015)], a divergence between haplotypes can be observed with regard to the red cysts of Sanguina nivaloides and orange cysts of S. aurantia in their variable ITS2 marker (sequence differences up to 5%), which might indicate a currently ongoing genetic diversification. Such a process was described for the soil alga Bracteacoccus bullatus by Fučíková and Lewis (2012). Further detailed pigment studies, physiological experiments with cysts, sequencing of further highly variable molecular markers (Škaloud et al. 2015) and cultivation attempts, possibly using strategies so far unconsidered, are clearly necessary to bring a richer understanding of morphological and molecular variability within the genus Sanguina.

When comparing red snow samples of Baffin Bay (Greenland) from 1851 with our recent samples, this showed that cell sizes corresponded well to the cell size ranges determined in this study (Fig. 6). Spherical red globules as well as green globules 'mostly with a small stalk and fluffy attachment, being the thallus or the root and leaf layer’ were reported from this sample [quotation translated into English from Ehrenberg (1851)]. While we could identify the former in the desiccated sample, the latter were not visible. However, we have also occasionally observed such stalks in recent samples [Fig. 2b in Leya (2008)]. We assume that they are fungal mycelia, which are not rare in field samples of red snow (Brown, Olson and Jumpponen 2015).

Cysts of Sanguina nivaloides are usually smooth-walled, and other specific variations in cell surface structure (e.g. cysts with nipples) may either reflect a level in the process of cyst maturation reached at the time of sampling or a causality between intraspecific variability and morphology (as e.g. all samples of H3 are expected to be represented by cysts with nipples, which is the case for both specimens of H3 in this study). To test this hypothesis, more specimens of the same morphologically different haplotype are needed (i.e. for H7, H14). Additionally, the observation of populations of smooth-walled cysts with the rare presence of cells with small papillae in two of our samples is in good correspondence with the findings reported from red snow in North America by Brown, Ungerer and Jumpponen (2016): they point out that very low haplotype diversity exists within the same algal bloom sampled: each algal patch was dominated by only one haplotype, with a very minor proportion of subdominant haplotypes. In the past, blooms of other snow algal taxa producing red cysts were often wrongly assigned to ‘Chlamydomonas nivalis’, e.g. most often Chlainomonas was not recognised [see Figs 23 and 24 in Stein and Amundsen (1967), in Kawecka (1981), and in Fig. 4 in Kobayashi and Fukushima (1952)]. The first comprehensive cell ultrastructure and secondary pigment profile comparisons between cysts of Chlainomonas and ‘Chlamydomonas nivalis’ were performed by Remias et al. (2016).

In our study, cell sizes of S. nivaloides red cysts correspond to the findings of Mosser et al. (1977) and Remias et al. (2005). The sizes of S. aurantia orange-coloured cysts from our study are in the range reported from Arctic pack ice floes (Gradinger and Nürnberg 1996) and from Spitsbergen (Müller et al. 1998; Leya 2004; Stibal et al. 2007). We assume that cell size can be highly plastic during their seasonal development, which would also explain the variable cell sizes among the haplotypes of red cysts we found, since our locations were sampled at different times during the snow melting season and were also varying in aspects related to their topographical properties (snowfield's slope angle, cardinal orientation of the site to the sun, elevation) and snow texture. Interestingly, the largest mean cell sizes were observed in samples collected either late in the season (e.g. samples ‘2RON’, ‘NOR4–1’) or on glacier-based snowfields (e.g. samples '4HT', 'Foxfona', 'Retten'). Indeed, cell sizes may also reflect culture conditions as in the lichen photobiont Asteriochloris (Škaloud et al. 2015). An alternative hypothesis could be that cell sizes of S. nivaloides and S. aurantia haplotypes are genetically determined with some variability. Interestingly, we observed narrow cell size variations when comparing genetically very uniform small orange cysts of S. aurantia from the different locations on Svalbard (two haplotypes only). More observations of orange cysts of S. aurantia from other geographic regions are needed to test this hypothesis, e.g. a bloom in North America was genetically identical for ITS2 (accession number KX063716; Brown, Ungerer and Jumpponen 2016) to S. aurantia orange cysts sampled in Svalbard (this study), but for the former no morphology data were collected. Interestingly, some haplotypes of S. nivaloides red cysts [i.e. H9 (‘WP123'), H18 (‘Esperanza-8')] were as small as S. aurantia orange cysts. Considering this fact, the sole parameter of cell size only should not be used for distinguishing between S. nivaloides and S. aurantia by light microscopy. Other aspects of mature cysts also should be considered such as cell wall morphology, intracellular pigmentation reflecting the astaxanthin to chlorophyll-a ratio and the macroscopic colour of the blooming snowfield.

Due to their small size (Wilkinson et al. 2012) and ability to resist drought and freezing stress, the cysts of S. nivaloides and S. aurantia are suitable propagules for long distance dispersal by prevailing air streams (Brown, Larson and Bold 1964; Schlichting Jr. 1969), birds (e.g. polar birds migrating between the two polar regions, such as the Arctic tern Sterna paradisaea) and ocean currents (Gillespie et al. 2012). Our data showed a cosmopolitan distribution of S. nivaloides in alpine and polar snowfields in both hemispheres, which supports the theory of a trans-equatorial dispersal of microbes (Hodač et al. 2016). No population structure was detected when analysing the ITS2 rDNA data, as there was no phylogeographic signal. Metagenomic analyses have shown red pigmented snow algae to be cosmopolitans based on the analysis of partial sequences of the 18S rRNA gene (Lutz et al. 2016) as well as of the ITS2 rDNA (Segawa et al. 2018). The latter study also detected endemic snow algal phylotypes being distributed in one of the two polar regions. Antarctica was regarded for a long time as a continent in biological isolation until large amounts of vascular plant pollen were collected on Signy Island (South Orkney Islands), and a parallel observation of the daily synoptic weather patterns revealed that roughly 1.5 times per year a corridor of strong winds develops between Tierra del Fuego (Argentina) and Antarctica (Marshall 1996). This explains our finding that the South American sample of S. nivaloides (‘La Hoya’, H8) was nearly identical in its ITS2 rDNA sequence with the haplotype from maritime Antarctica (H18) and with the most common haplotype in Europe (H1). The distribution of these small microorganisms could be controlled by selection due to local environmental factors as well as dispersal limitations. Moreover, a strong founder effect from only few propagules and intense kin competition is assumed to take place (Brown, Ungerer and Jumpponen 2016).

Concerning local environmental factors at the moment of sampling of encysted cells, the usually slightly acidic pH and low conductivity of snow patches with algal blooms resembled those of pristine, white snow presumably free of any algae. However, Brown and Jumpponen (2019) showed that snow algae are seemingly omnipresent, though in low abundances, at the surface of melting subpolar and alpine ‘white snow’ that are void of a visible algal bloom. Nitrate, ammonium and phosphate were not limiting nutrients for Arctic algal communities in a study by Spijkerman et al. (2012). On the other hand, atmospheric trace gases were recently recognised to play a role in supporting primary production in oligotrophic habitats in polar regions (Ji et al. 2017). None of these findings conclusively explains the patchy occurrence of red snow, which often can be observed in the same pattern year after year at the same location. Spijkerman et al. (2012) suggested that the highly patchy distribution of algal blooms in snow could be partly a result of small-scale topographical and geological characteristics influencing snowfield persistence or disappearance due to enhanced or reduced snow melt or snow accumulation. Similarly, inputs of mineral impurities can result in significantly earlier snow melt (Di Mauro et al. 2018). The slower the melting, the more prolonged is the period that is available for a snow algal population to develop. This is realised in alpine regions at higher elevations (above 1900 m a.s.l. in this study). With regards to the availability of liquid water, the algal cells seem to thrive during the growth season in the thin but persistent aqueous reservoir that by the laws of physics remains liquid around and between snow and ice crystals (Jones et al. 2001). To find a conclusive answer about which factors on a small scale actually determine the development of snow blooms such as those of S. nivaloides and S. aurantia or whether the cysts are just randomly dispersed there, a time series sampling would be necessary to observe the development of a blooming snow field while concurrently measuring various parameters.

Red spherical cysts from the European Alps morphologically identical to S. nivaloides cysts were shown to tolerate short periods of elevated UV-B radiation (Remias et al. 2010) due to the massive production of secondary carotenoids, which act mainly as shading pigments, but also as very powerful radical scavengers. This is especially true of ketocarotenoid astaxanthin, which absorbs energy in the near UV-range. Sanguinanivaloides shows high photophysiological plasticity with a highly efficient photosystem under low irradiation and no photoinhibition up to 2000 μmol m⁻² s⁻¹ (Procházková et al. 2018b). Additionally, inorganic rock dust particles covering mature cysts seem to provide additional shading protection (Lutz-Meindl and Lutz 2006). Sanguina nivaloides cysts are required to survive over summer when snow has melted, and provided they still contain chlorophyll and are not photoinhibited, the photosynthetic apparatus should continue to function (Remias, Lütz-Meindl and Lütz 2005). After total snow melt, cysts most likely ‘oversummer’ the rest of the year, generally in a dried state, on rocks, permafrost or moss-covered soil, all being subject to drastic diurnal temperature fluctuations; non-vegetated soils in particular can experience temperature changes of 30°C within 24 h (Ley, Williams and Schmidt 2004). Therefore, S. nivaloides cysts can be regarded as being eurythermic and desiccation-resistant. The question remains whether the vegetative stages (gametes or other green stages if they exist) are dependent on cold temperatures, which would characterise them as cryophiles.

Key triggers to a successful S. nivaloides cultivation have not been elucidated yet. Indeed, many green proliferating strains have been isolated from snowfields, but molecular methods and phylogeny have always shown that these isolates are taxa other than S. nivaloides. A well-known example of such a misinterpretation was strain UTEX 1969, which for years was listed as Chlamydomonas nivalis and frequently used in studies addressing cold adaptation, just as other strains erroneously were assigned to this taxon (Lu et al. 2012a,b, 2016). When, in 1990, H. Ettl and U.G. Schlösser (pers. comm.) re-identified many UTEX, CCAP and SAG strains of Chlamydomonas, they found that many were incorrectly named—among them also Cd. nivalis strains, which were first assigned to Chlamydomonas augustae, but later partly renamed again on the basis of the phylogenetic analyses as Chloromonas typhlos (Matsuzaki, Hara and Nozaki 2012) or Cr. reticulata (CCALA 753; CCALA 754). At present, it must be accepted that we do not have any living strain on hand to enable the studies on the life cycle of S. nivaloides. All diagrams describing the life cycle of this species are thus just hypothetical without any empirical proof (e.g. in terms of molecular data) (Müller, Leya and Fuhr 2001; Sattler et al. 2010; Remias 2012). Any conclusions can be made only very cautiously and indirectly based on the knowledge of life cycles of more distantly related species. The sexual life cycle of several Chlamydomonadacean species usually include the formation of zygotes into zygospores with a thick cell wall, which represents the dormant stage in the life cycle (Žárský, Kalina and Sulek 1985). After the formation and growth of the secondary cell wall, the primary cell wall of the zygospore is shed from the surface (VanWinkle‐Swift and Rickoll 2008) and the ornamented structure of the mature zygote becomes apparent (Pröschold et al. 2001). We observed a similar cell wall organisation in the ruby cysts type, which have been found so far on the Northern hemisphere in continental Norway (sample ʿN4’ and ʿNOR4–1’ in this study), in Svalbard [sample RS_0011 in this study and in Müller et al. (1998)], on Iceland (D. Remias, pers. comm.), on the western coast of Canada (L. Quarmby, pers. comm.), in Washington State in the USA (K. Thomas, pers. comm.), Oceania (P. Novis, pers. comm.), as well as in the southern hemisphere on the Antarctic King George Island (T. Leya, pers. comm.). Wille (1903) was referring to this distinctive cellular morphology in his study of Cd. nivalis from Djupvatshytten (Norway) (plate III, Figs 44 and 45). Their identity needs to be revealed by single cell sequencing. The first insights into their 18S rRNA gene phylogeny suggest that the ruby-type cysts might in fact be members of the Chloromonas-clade (Segawa et al. 2018).

The CBC species concept was successfully applied for several algal groups, e.g. snow-dwelling Chloromonas species (Matsuzaki et al. 2019). This concept can be used also in a metagenomic study: Segawa et al. (2018) showed that their ‘Chlamydomonas’ snow group B (conspecific with S. nivaloides) differs by one CBC in helix III only compared with the closely related ‘Chlamydomonas’ snow group A (see their Supplementary Figures S2 and S8). On the other hand, it was shown that the CBCs are not diagnostic in some cases at the species level and that even genera, families and orders of green algae can lack CBCs in such regions (Caisova, Marin and Melkonian 2011; Škaloud and Rindi 2013). Therefore, the presence and number of CBCs are probably direct consequences of the accumulation of mutations during the evolutionary process, simply reflecting the genetic distance among organisms. In our study, two field samples of Sanguina nivaloides (‘NOR4–1’ and ‘RS_0014–2013’) lack any CBC in their secondary structure when compared with Sanguina aurantia. Both two newly described species are closely related: they have identical 18S rRNA gene and more than 95% identical ITS2 (one CBC found). In the rbcL phylogeny, S. aurantia has been placed within Sanguina nivaloides field samples; this might be linked to unusual gene substitutions in rbcL, which was reported to result in artefacts in the phylogeny (Nozaki, Onishi and Morita 2002). Species boundaries between S. nivaloides, S. aurantia and other related species should be estimated by combination of substantial morphological, ecological and genetic difference (investigation of several molecular markers, the level of ITS2 rDNA sequence similarity should be also taken into account). S. aurantia has uniquely different ITS2 haplotypes to S. nivaloides, orange spherical cysts and red spherical cysts are clearly distinguishable, both microscopically and macroscopically, and the area of their geographic distribution is only partly overlapping. No strain for these species is available yet. According to our 18S rRNA gene and rbcL phylogenies, the closest known and culturable relatives of Sanguina nivaloides are strains CCCryo 133–01/101–99 (currently assigned to cf. Sphaerocystis sp.), strain CCCryo 147–01 (cf. Coenochloris sp.) and CCCryo 086a-99 (cf. Ploeotila sp.). Some of these relatives are known for their production of astaxanthin (Leya et al. 2009), though in less dominating amounts than in S. nivaloides, but none of them has ever been reported to cause striking snow colouring. They have been isolated from different habitats on Svalbard such as a moss and a snowfield as well as from a glacier. These strains may rather be regarded as permafrost algae than snow algae. Demchenko (2013) investigated the above strains and described them all as having coccoid cells. When studying their life cycles he observed zooid formation in strains CCCryo 101–99 and CCCryo 086b-99. Motile stages so far have not been described for the only known species of Ploeotila (P. ramosa T.Mrozinska-Webb) and since he also found slightly different characteristics in the other two strains in comparison to Sphaerocystis or Coenochloris, their attribution to these genera remains unclear (hence the ‘cf.’). Our rbcL phylogeny (Fig. 7) confirmed that these tentative genera form closely related lineages to Achoma brachiatum isolated from an alpine herbfield soil in New Zealand (Novis and Visnovsky 2012). However, as outlined above, CCCryo 101–99 and 133–01 clearly differ from strain CCCryo 147–01, the former two producing zoids and considerable amounts of secondary carotenoids and the latter lacking such cell stages and ability. Which of them actually might have to be assigned to the genus Achoma remains to be investigated. The three strains CCCryo 086a-99, 101–99 and 147–01 further markedly differ in the following aspects: cells are single or organised in clusters of tetrads in CCCryo 133–01, in microscopic colonies in an extracellular polymeric substance (EPS) in CCCryo 147–01, and in cells or a group of cells sitting on a mucilaginous stalk in CCCryo 086a-99. Chloroplasts are parietal (CCCryo 133–01, CCCryo 086a-99) and pot-like (CCCryo 147–01), and a pyrenoid is present. Yet S. nivaloides shares several features with the above mentioned three CCCryo strains: its origin from cold habitats, its parietal chloroplast with a pyrenoid (at least observable in young cysts) and its prevailing morphologic organisation as a coccoid life form.

Our study has shown that sequencing of the field-collected cysts was a suitable strategy and also the only option to decipher their phylogenetic position, as establishing actively growing isolates from them has yet to be successful. This newly described genus Sanguina with two closely related species S. nivaloides and S. aurantia represents a single monophyletic lineage, independent from other Chlamydomonadacean algae. Using molecular methods we showed that S. nivaloides has a cosmopolitan distribution in polar and alpine regions, having detected the same ITS2 haplotypes on several continents. It is possible to distinguish the cysts of S. nivaloides from red-coloured cysts of other algae based on the cell size, number of cell walls, plastid organisation and habitat preference. Further studies could involve single-cell sequencing methods, e.g. with a further focus on the ruby-type cysts. The genetic variability below species level might be detected using a microsatellite-based approach (Nagai et al. 2007) developed from single cell genomics (Muramoto et al. 2010). Additionally, sampling at known locations early in the season might help to discover vegetative stages of this species, enabling us to study this interesting alga in much more detail using viable cultures/isolates. The molecular identity of other red coccoid species from snow reported by Kol (1968) such as Chlamydomonas antarcticus (red cysts with a broad mucilage layer) or Trochiscia spp. (red cysts with spikes) might also yet be revealed.

Supplementary Material

ACKNOWLEDGEMENTS

REFERENCES