Sediment surface samples were taken from 28 January to 15 February 2020 from four study sites near the Argentinian research station Carlini Base (S 62°14′17.45”, W 58°40′2.19”) at Potter Cove on King George Island, Antarctic Peninsula, and used for benthic diatom isolation (Figure 1, Table 1). The marine culture D288_003 originated from sample location D288 (62° 14′ 30.55” S, 58° 40′ 54.96” W) and was isolated from a biofilm of an intertidal rock pool. D296_001 was isolated from a marine sample at the inner part of Potter Cove (62° 13′ 43.61“ S, 58° 39′ 49.36” W), at 15 m depth, from an epipsammic community. The marine culture D323_018 originated from a sample location closer to the glacier front in the inner part of the bay at 10 m water depth from an epipsammic community (62°13′25.68” S, 58°38′33.50” W). Both limnic isolates (D300_015 and D300_025) were established from biofilms on top of stones in a freshwater drinking reservoir (62° 14′ 16.30” S, 58° 39′ 44.10” W).

At Carlini Base, various abiotic parameters were recorded. The minimum water temperature was −1.69°C in the inner part of Potter Cove and −1.4°C in the outer part, while the maximum temperature at both sites was 2.89 and 1.98°C, respectively. Furthermore, the salinity of the outer cove was stable at ~33.5 S_A, while the salinity in the inner cove can decline to 29.6 S_A because of meltwater run-off (Hernández et al., 2019). Light conditions are diurnally and seasonally highly variable, and more recent data were published by Hoffmann et al. (2019).

The diatom cells were isolated from aliquots of environmental samples to establish clonal cultures. Under an inverse light microscope (Olympus, Hamburg, Germany), single cells were transferred using microcapillary glass pipettes onto microwell plates containing culture medium [Guillard's f/2 medium (Guillard and Ryther, 1962) or Walne's medium (Walne, 1970), 34 S_A for marine samples and 1 S_A for freshwater samples]. All samples and isolated diatom cells from Antarctica were maintained at 5–7°C at the Botanische Museum Berlin. Irradiation was accomplished by white LEDs with 5,000 K under a 16/8 day/night cycle with several dark phases during the day to prevent photo-oxidative stress. After the successful establishment of clonal cultures, they were separated into subsamples for DNA extraction, morphological analysis, and ecophysiological, biochemical, and cell biological experiments. For all experimental approaches, diatom cultures were transferred to the University of Rostock and cultured in sterile-filtered Baltic Sea water, enriched with Guillard's f/2 medium (Guillard and Ryther, 1962) and metasilicate (Na₂SiO₃ • 5 H₂O; 10 g 100 mL⁻¹) to a final concentration of 0.6 mM (further referred to as culture media). The salinity of 33 S_A for the marine cultures was adjusted by adding artificial sea salt (hw-Marinemix^® professional, Wiegandt GmbH, Germany), while 1 S_A for the limnic cultures was achieved by dilution with deionized water. The media were regularly changed to replenish nutrients.

All stock cultures at the University of Rostock were kept at 8–9°C and 15–20 μmol photons m⁻² s⁻¹ under a 16/8-h light/dark cycle [Osram Daylight Lumilux Cool White lamps L36W/840 (Osram, Munich, Germany)].

In order to obtain clean diatom frustules for species identification, material harvested from the clonal cultures was treated with 35% (v/v) hydrogen peroxide at room temperature to oxidize the organic material, followed by several washing steps with distilled water. To prepare permanent slides for light microscopy (LM) analyses, the cleaned material (frustules and valves) was dispersed on cover glasses, dried at room temperature, and embedded with the high refraction index mounting medium Naphrax^®.

For each culture, 15–20 cells were measured for morphometry. Observations were conducted with a Zeiss Axioplan Microscope equipped with differential interference contrast (DIC) using a Zeiss 100x/1.30 Plan Apochromat objective, while microphotographs were taken with an AXIOAM MRc camera. Aliquots of cleaned sample material for scanning electron microscopy (SEM) observations were mounted on stubs and observed under a Hitachi FE 8010 scanning electron microscope operated at 1.0 kV.

After cultivation for 4–6 weeks to achieve sufficient biomass for the experiments, each strain was divided into different Erlenmeyer flasks. From the control condition (T0, control), three replicate samples were harvested for evaluation, and an additional three replicates were transferred to 3 months of darkness (T3) at 5°C. Cells from both T0 and T3 treatments were investigated concerning physiological, biochemical, and cell biological traits.

Five strains from the four sample locations were investigated in this study. The morphological and molecular analyses confirmed the assignment to four distinct species. The taxonomic identity of some strains was already published before (D288_003, D296_001, D300_015; Prelle et al., 2022; Schimani et al., 2023).

D288_003 was identified as Navicula criophiliforma Witkowski, Riaux-Gob. and Daniszewska-Kowalczyk. This taxon was first published by Witkowski et al. (2010) from the Kerguelen Islands coastal area, Southern Ocean, and more recently also reported from Livingston Island, north of the Antarctic Peninsula (Zidarova et al., 2022). LM and SEM pictures of this strain can be found in Prelle et al. (2022).

D296_001 was identified as Chamaepinnularia gerlachei Van de Vijver and Sterken. This species was first published by Van de Vijver et al. (2010) from dry soil samples from James Ross Island, near the northeastern extremity of the Antarctic Peninsula, and has been observed until now just from maritime Antarctica (Kopalová et al., 2012; Sterken et al., 2015; Zidarova et al., 2016). A thorough examination of this strain can be found in Schimani et al. (2023).

D323_018 (Figure 2) could only be identified at the genus level as Melosira sp. Cells are subcylindrical with a diameter ranging between 18.6 and 20.9 μm and a pervalvar axis ranging between 19.4 and 22.0 μm (Figures 2A–E). The hypovalve is more rounded than the epivalve (Figure 2H), which exhibits at the mantel a more cylindrical shape. Our strain resembles Melosira moniliformis C. Agardh in Crawford (1977), who examined the type and new material. However, in M. moniliformis, ridges on the outer mantle surface fuse to form bigger spines, which are not visible in our strain. Therefore, unambiguous taxonomic assignment of this strain to a particular Melosira species is not possible at this stage.

D300_015 was only identified at the genus level as Planothidium sp. in Prelle et al. (2022). Morphological and molecular investigations showed that strain D300_025 is the same species. With the new data generated from those strains and two other strains obtained from the same location (D300_019, D300_020), this species is described as new.

Planothidium wetzelii Schimani, N. Abarca et R. Jahn sp. nov.

Description:

Living material (Figures 3Q–AB, AP–BA): One plate of the C-shaped plastid lies appressed to the valve face. The concavity of the plastid is visible on the valve side lacking the cavum. Pyrenoids are not clearly visible, but Figures 3AR, AT suggest that there is one pyrenoid located in the vicinity of the cavum. In the girdle view, especially in the larger cells before cell division (Figure 3BA), two lobes of the chloroplast are visible appressed to the valves, and connected in the center of the cell.

LM (Figures 3A–P, AC–AO, BB–BN): The outline ranged from elliptic with slightly protracted rostrate apices in smaller valves to elliptic or lanceolate with protracted, rostrate to capitate apices in larger valves. One valve side—especially strain D300_020—appeared more convex than the other. The length of the valves varied between 10.9 and 18.8 μm and the width between 5.6 and 6.7 μm (n = 43). Sternum valve (SV): Axial area narrow, becoming wider toward the center of the valve. Central area asymmetrical with an oblong cavum with parallel sides, extending slightly over the axial area on one side and with shortened striae on the other side. The parallel margins of the cavum widen just shortly before the mantle, and its aperture is seen as a roundish line close to the mantle. Striae parallel or weakly radiating in the center, becoming more radiate near the apices (16–18 in 10 μm in smaller valves and 14–16 in 10 μm in larger valves). Raphe valve (RV): Axial area narrow. The central area widened due to three to four shortened striae on both sides. Raphe straight with enlarged proximal endings. Distal endings are not visible in light microscopy. Striae are parallel or weakly radiating in the center, becoming more radiating near the apices (16–18 in 10 μm in smaller valves and 14–15 in 10 μm in larger valves).

SEM (Figures 4A–W) SV: The striae are composed of three to four rows of small rounded areolae exceeding the width of the virgae. Striae near the center of the valve become pointed with one or two areolae near the axial area. Externally, striae reach over the mantle with up to four rows of areolae. Irregularly rounded depressions are externally present along the axial area and concentrated in the central area. Internally, areolae are covered by hymenate occlusions. The broad cavum had a tight hood opening close to the mantle. Cingulum is composed of unperforated girdle bands. RV: The striae are composed of usually three to four rows of rounded areolae exceeding the width of the virgae. Similar to the central striae on the SV, the shortened central striae on the RV gradually narrow toward the axial area. At each stria, one to three areolae reach over the edge of the valve face. Internally, areolae are covered with the same hymenate occlusions as in SV. Externally, the raphe straight with drop-like expanded proximal ends and distal raphe fissures unilaterally deflected, not or just shortly continuing onto the valve mantle. Internally, proximal raphe endings are slightly bent to opposite sides, and raphe endings terminate on a small helictoglossae not continuing on the valve mantle.

Holotype: Slide B 40 0045341a, Botanic Garden and Botanical Museum, Berlin, Figure 3BJ for SV and Figure 3BK for RV from the strain D300_025 illustrate the holotype. SEM-stub deposited as B 40 0045341b. For molecular material and data, see Methods.

Type locality: Drinking water reservoir of Carlini Station, Potter Cove, King George Island, South Shetland Islands, collected by J. Zimmermann on 01 February 2020, coordinates: S 62.237861, W 58.662250.

Registration: http://phycobank.org/103793.

For INSDC Accession numbers see Table 1.

Habitat: Freshwater biofilm on stones.

Etymology: With this name, we would like to acknowledge the morphological and taxonomic work on the genus Planothidium done by Carlos Wetzel (Luxembourg Institute of Science and Technology, Luxembourg) in recent years, as well as his help in finding the identity of this species.

Differential diagnosis: Planothidium wetzelii shares similarities with several cavum-bearing species and species with external shallow rounded depressions on the rapheless valve (as defined by Wetzel et al., 2019): The smaller valves of P. wetzelii share similarities in shape with the small valves of P. frequentissimum (Lange-Bertalot), Lange-Bertalot as illustrated from type material in Wetzel et al. (2019). However, they can be distinguished by the form of the cavum, which is oblong with parallel sides and extending slightly over the axial area, vs. a round cavum in P. frequentissimum. The shape of the central area on the RV is much smaller in P. wetzelii compared to the type of P. frequentissimum. In SEM, the virgae in the SL and to some extent on the RV are wider in P. wetzelii compared to images of the type of P. frequentissimum. Additionally, the striae on the RV do not continue over the mantle in P. frequentissimum as they do in P. wetzelii. Larger valves of P. frequentissimum are lanceolate with less developed rostrate apices than those of P. wetzelii. Valves of P. victorii Novis, Braidwood et Kilroy (syn. P. caputium) are in contrast to P. wetzelii lanceolate to weakly elliptic-lanceolate with protracted but never capitate apices (Jahn et al., 2017; Wetzel et al., 2019). The species can be differentiated by the form of the cavum too as it is rounder or has a V-shape in P. victorii. In addition, the tight hood opening in P. wetzelii (SEM), seen as a roundish line close to the mantle (LM), differs from the wider hood opening (SEM) in P. victorii, which is present as an almost straight line distant from the mantle (LM). The axial area is narrower in this species and the cavum has a wider aperture toward the mantle than P. wetzelii. Some of the medium sized valves could be confused with P. naradoense R.Jahn et J.Zimmermann but differ from P. wetzelii by the outline, as P. naradoense is lanceolate to elliptic-lanceolate and somewhat asymmetric with slightly rostrate and rounded apices. The axial area is narrow on both valves (Jahn et al., 2017). Larger valves of P. wetzelii share similarities with P. tujii and P. gallicum C. E. Wetzel et Ector (Wetzel et al., 2019). However, the valves of P. tujii are shorter, while the width is almost in the same range. The central area in the RV of P. tujii is bowtie-shaped instead of the smaller rounded central area in P. wetzelii. The virgae are wider in P. tujii and have almost the same width as the striae; the helictoglossa is continuing onto the valve mantle. P. gallicum has clear broadly elliptic-lanceolate shaped valves with markedly rostrate apices. Furthermore, P. gallicum has a more prominent aperture of the cavum toward the mantle, and the striae located in the central area are not becoming narrower toward the axial areal as in P. wetzelii. P. straubianum C. E. Wetzel, Van de Vijver et Ector has elliptic-lanceolate valves with slightly parallel margins and round, obtuse ends with smaller dimensions than P. wetzelii (length 6.0–14.0 μm vs. 10.9–18.8 μm and width 4.0–5.5 μm vs. 5.6–6.7 μm; Wetzel et al., 2019). Striae are composed of four to five rows of areolae, in contrast to three to four in P. wetzelii. P. biporomum (M. H. Hohn et Hellerman) Lange-Bertalot can be differentiated from P. wetzelii in SEM by the striae on the rapheless valve (Wetzel et al., 2013). The striae start externally with one, rarely two, areolae at the axial area and end in two or three rows toward the valve face/mantle junction. Additionally, they are interrupted at the junction with the valve mantle. Furthermore, the cavum aperture is wide with its borders linked to the neighboring striae. P. alekseevae Gogorev et E. K. Lange has smaller dimensions (length 9.5–13.5 μm, width 4.5–5.0 μm) and wider apices in relation to the valve width. This feature is more apparent in smaller valves (Wetzel et al., 2019).

Molecular results: The four strains of P. wetzelii showed no intraspecific variability in the 18S V4 and the rbcL sequences (Supplementary Tables 1, 2). Compared to the available sequences of other Planothidium species, genetic distances are apparent in both marker genes. It is placed within the F2 subclade (Jahn et al., 2017, p. 84, Figure 1): strains of P. victorii (e.g., type strain from New Zealand), P. straubianum (strain B86_3 from Lake Baikal, renamed by Wetzel et al., 2019), and P. tujii show the smallest genetic divergence: in 18S V4 0.5–1.1, 0.7–0.9, and 0.9–1.1%, respectively, which corresponds to 3–5 bp differences, and in rbcL 0.5–0.8, 0.5, and 0.5% (3–7 bp). Higher genetic differences are evident for P. naradoense and P. frequentissimum: in 18S V4 2.5% and 2.5–3.2% (11 bp) and in rbcL 2.2 and 1.8–2.3%, respectively (8–22 bp). Species with a sinus instead of a cavum [P. lanceolatum (Brébisson ex Kützing) Lange-Bertalot, P. cf. subantarcticum, P. taeansa R. Jahn and N. Abarca, P. cryptolanceolatum R. Jahn and N. Abarca, and P. suncheonmanense R. Jahn and J. Zimmermann] exhibit the highest divergence with a p-distance of 4.3–7.8% (19–38 bp) in 18S V4 and 3.0–6.2% (30–61 bp) in rbcL.

The values for Y(II)_max at T0 and T3 are shown in Table 2 (mean values ± SD, n = 3). All results were between 0.53 (P. wetzelii (D300_025) at T3) and 0.63 (N. criophiliforma and C. gerlachei at T0) and therefore within a range reflecting “good” physiological activity of all diatom cells during the dark incubation. Small, but significant differences in Y(II)_max between T0 and T3 were found only in both cultures of the limnic species P. wetzelii (Table 2).

The results of the chlorophyll a measurements, which were referenced to cell counts, showed species-specific responses (Table 2). In both strains of P. wetzelii, the chlorophyll a content per cell did not significantly change during 3 months of dark treatment. In contrast, N. criophiliforma and C. gerlachei exhibited a pronounced chlorophyll a decline up to 74% after dark incubation compared to the control, while the Melosira sp. decreased chlorophyll a by ~46% (Table 2).

All measured P–I curves exhibited a typical shape without photoinhibition, as already reported for some of the species in detail in Prelle et al. (2022). The key parameters NPP_max and respiration were selected for all species before and after dark incubation (Table 2), and species-specific responses could be outlined. At T0, the highest value for NPP_max was determined in the marine species N. criophiliforma (80.0 ± 20.2 μmol O₂ mg⁻¹ Chl a h⁻¹), while P. wetzelii (D200_015) showed the lowest NPP_max (16.7 ± 4.0 μmol O₂ mg⁻¹ Chl a h⁻¹). While some species (N. criophiliforma, both P. wetzelii isolates) did not exhibit any significant NPP_max decline (p < 0.05) during dark treatment at T3, C. gerlachei and Melosira sp. showed a strong and significant decrease (p < 0.05) in NPP_max by 39.9 and 44.6%, respectively (Table 2).

Although respiration rates were also species-specifically different and variable, we could not detect any significant difference in respiration between T0 and T3 among all benthic diatom species (Table 2). While the highest respiration values occurred in the marine species N. criophiliforma (−72.2 ± 70.3 O₂ mg⁻¹ Chl a h⁻¹), the lowest respiration rates were measured in the limnic culture P. wetzelii (D300_015, −7.8 ± 6.2 O₂ mg⁻¹ Chl a h⁻¹). The NPP_max: Respiration ratios were relatively low for all isolates between 0.8 ± 0.1 and 2.9 ± 4.0 at T0 and T3, except for Melosira sp. which exhibited higher ratios of 6.6 ± 5.0 and 5.0 ± 2.2, respectively, due to proportionally lower respiration rates (Table 2). All diatom cultures were not axenic, and the bacterial abundance ranged from a very low 1.7% bacteria volume when compared with the diatom volume in Melosira sp. to 23.7% in P. wetzelii (D300_25; Table 2). All other species showed bacterial contamination between these extreme percentages.

SYTOX Green staining was applied as a cell biological approach to evaluate membrane integrity during dark incubation. The percentage of damaged cells increased significantly in four out of the five tested benthic diatom species (Figure 5). In N. criophiliforma, C. gerlachei, and Melosira sp. at T0 only <5% of the cells were damaged pointing to highly viable cell populations. After 3 months, dark incubation membrane integrity decreased in all three species, i.e., cells with compromised membranes accounted for between 16.4 and 39.0% (Figure 5). Both P. wetzelii strains were already at T0 and less viable as reflected in approximately one-third of the damaged cells. While isolate D300_025 exhibited 71.5% compromised membranes, P. wetzelii (D300_015) maintained membrane integrity over 3 months of dark incubation (Figure 5).

Figures 6, 7 show the first transmission electron micrographs (TEMs) of benthic diatom species from Antarctica under control (T0, Figure 6) conditions and after 3 months of dark incubation (T3, Figure 7). The three pennate (N. criophiliforma, C. gerlachei, and P. wetzelii) and one-centric diatom species (Melosira sp.) exhibited the typical cellular ultrastructure of a diatom cell. All cells contained a central nucleus, several chloroplasts, and, most conspicuously, one or more large lipid droplets that occupied large parts of the cell volume (Figure 6). After the dark treatment, ultrastructural changes could be detected (Figure 7). The most conspicuous observations were degraded chloroplasts with plastoglobules (PGs) in all species and partially more endoplasmic reticulum (ER). In addition, the volume or number of lipid droplets drastically decreased (Figure 7). In N. criophiliforma, lipid droplets were virtually absent (Figures 7A, B) and only observed in a few cases. Similarly, in Melosira sp., no or only very small lipid droplets occurred at T3. In contrast, in C. gerlachei and P. wetzelii (D300_015), well-developed lipid droplets could be identified. However, in Figure 7C of C. gerlachei beginning, the degradation of a lipid droplet was observed (marked by an arrow). In addition, N. criophiliforma and P. wetzelii (D300_015) exhibited multivesicular bodies (MVBs) after 3 months of darkness (Figures 7A, H), a typical sign of apoptosis. Additionally, in P. wetzelii, several larger electron-dense bodies were observed (Figures 7G, H). While the origin of these bodies remains unknown, they might resemble the degradation products of the chloroplasts.

The visual differences of the species-specific lipid content after Nile red staining of the diatom cells before and after 3 months of dark incubation are summarized in Figure 8. In this study, a clear decrease in the size of the stained lipid droplets can be seen. The corresponding calculated lipid volume changes in the examined diatom cells are shown in Figure 9 as boxplots of 25 measured cells each. A significant decrease in the lipid content over 3 months of dark incubation could be documented for all species. The marine isolates N. criophiliforma, C. gerlachei, and Melosira sp. exhibited >95% the strongest decrease in the total lipid content per cell within 3 months of darkness, and in the first and third species only traces of lipids remained detectable (Figure 9). In contrast, in both P. wetzelii strains, the lipid content changed much less, with a decrease to only 58.7–60.7% of the control, i.e., these cells contained at T3 still detectable amounts of lipids.

The chemical GC-MS analysis of the total lipid content supports the results shown in Figures 8, 9, i.e., the marine isolates N. criophiliforma, C. gerlachei, and Melosira sp. exhibited concentration declines between 90.0 and 94.5% (Table 3). In both limnic P. wetzelii strains, the lipid content decreased by 20.6 and 35.2% after 3 months of darkness (Table 3). Saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), and polyunsaturated fatty acids (PUFAs) before and after 3 months of dark incubation were also evaluated, but a clear trend is not visible, since all species showed different response patterns. The quantitatively most abundant fatty acids in all benthic diatoms were the saturated 16:0 and the monosaturated 16:1n7 (Table 3). While the percentage proportion of 16:0 in relation to all measured fatty acids increased in N. criophiliforma (from 43.6 to 58.4%) and C. gerlachei (from 38.3 to 48.1%) after 3 months of darkness, it strongly decreased in the remaining species (e.g., in Melosira sp. from 36.2 to 16.9%). Reversely, the percentage proportion of 16:1n7 in relation to all fatty acids strongly declined in N. criophiliforma, C. gerlachei, and Melosira sp., while both P. wetzelii strains exhibited a slight increase (Table 3).

However, the first step, prior to ecophysiological experiments, is always to carefully address the taxonomic position of freshly collected field material. Two of the investigated diatom species were morphologically and molecular-genetically identified as species only known from maritime Antarctica (Chamaepinnularia gerlachei) or the Southern Indian Ocean (Navicula criophiliforma). Two additional cultures belong to the newly described freshwater species Planothidium wetzelii, which is so far unknown in any other region of the world. Finally, one centric species was investigated in this study, which does not fit morphologically and genetically to any described Melosira species, and most probably represents a new taxon, which was not further investigated in the present study.

Due to a thorough identification of the investigated cultures, a restricted geographical distribution in the southern hemisphere could be verified for all of them, and some might even be endemic to the Antarctic region. Morphological variability of the newly described species Planothidium wetzelii shows that valves from environmental samples need to be complemented with molecular and morphological information gained by cultures. This case highlights once again the importance of integrative taxonomy for the investigation of diatom biodiversity. In addition to an integrated taxonomy, the establishment of clonal cultures is the prerequisite for a combined taxonomical and physiological investigation of benthic diatoms. The two P. wetzelii strains D300_015 and D300_025 exhibit some variability in valve morphology. They might be considered two separate species when just examined by light microscopy, based on their differences in shape and size. However, diatoms can exhibit phenotypic plasticity due to environmental changes (Andrejić et al., 2018) or when observed over long time in culture (Mohamad et al., 2022). Both strains represent the same species based on molecular markers and micromorphology, and the results from this study show very similar response patterns during dark incubation.

As a result of the inaccessibility of Antarctica, the diversity of benthic diatoms is still poorly known. Considerable taxonomic work needs to be done in polar regions, which can then be used as a baseline for physiological, biochemical, and cell biological studies, allowing important conclusions, such as, for example, if geographic boundaries and environmental conditions led to many endemic taxa.

Phototrophic organisms in the polar regions such as benthic diatoms cope with strong diurnally and seasonally fluctuating light conditions. The review of Pavlov et al. (2019) comprehensively summarized the underwater light climate and all main controlling factors in the Arctic Kongsfjorden, one of the best studied polar field sites, which is discussed here as a representative high latitude coastal habitat since adequate data for Antarctica are missing. In addition to the seasonal changes from midnight sun to polar night, clouds, sea ice conditions, and the optical seawater properties further strongly influence dose and spectral composition of solar radiation penetrating into the water column, thus defining the underwater light conditions. In addition, input of local run-off and glacial meltwater introducing inorganic and organic matter into the system, and phytoplankton blooms at diverse times and locations. Together, these factors result in a complex underwater light climate with high variability in time and space (Pavlov et al., 2019). Although not comprehensively investigated, similar processes and physico-chemical properties can be assumed for Potter Cove (Hoffmann et al., 2019) and other bays around the Antarctic Peninsula. Polar benthic diatoms are additionally affected by huge amounts of particles released into the water column due to coastal glacier retreat and melt under global change conditions, resulting in increased turbidity and hence less available light (Hoffmann et al., 2019).

Photosynthesis is the essential mechanism for the energy metabolism and, thus, not only responsible for viability and survival of benthic diatoms, but also primarily dependent on light availability. All benthic diatom species examined in this study showed Y(II)_max of 0.57–0.63 before dark incubation as physiological marker for photosynthetic activity. Other studies on Antarctic benthic diatoms reported similar maximum Y(II)_max values between 0.6 and 0.7, indicating that the strains were in good physiological conditions prior to dark incubation (Longhi et al., 2003; Wulff et al., 2008a).

Prelle et al. (2022) used for their P–I curves various Antarctic benthic diatoms, including the species also investigated in the present study: Navicula criophiliforma, Chamaepinnularia gerlachei, and P. wetzelii strain D300_015, and demonstrated generally low light requirements, as light-saturated photosynthesis was reached at <70 μmol photons m⁻² s⁻¹. A similarly low light demand was observed for Melosira sp. and P. wetzelii strain D300_025 (data not shown). Other polar benthic diatoms are also known for their shade acclimation with low light compensation (I_c) and light saturation points (I_k) as well as steep initial slopes (α value) in their P–I curves (Wulff et al., 2008a). At the same time, they are also described as phototrophic microorganisms with a pronounced photophysiological plasticity, allowing them to cope with high solar radiation (Prelle et al., 2022).

Remarkable are the relatively high respiration rates, which were about as high as the net photosynthetic rates in some of the studied benthic diatom species, resulting in low ratios of NPP_max: respiration, and which can be explained by co-occurring bacteria in the cultures. Overall, we cannot rule out that diatom-associated bacteria could affect the results of the P–I curves through their respiration, particularly when present in high cell numbers. However, bacteria and diatoms are known to have complex biotic interactions that affect each other's viability, and diatoms can even control their own phycosphere (microbiome; Amin et al., 2012). Therefore, the often-applied antibiotic treatment prior to experiments could potentially not only reduce essential bacteria but also affect diatom viability with unforeseen consequences for dark survival.

Three months of darkness did not negatively affect Y(II)_max values in the studied benthic diatom species, which provides at least a crude measure for still-existing photosynthetic viability, which is confirmed by the oxygen data. Previous studies on the dark survival of polar diatoms showed a decrease in various photosynthetic parameters during dark incubation (Wulff et al., 2008a; Reeves et al., 2011; Karsten et al., 2012; Lacour et al., 2019). The only study on Antarctic benthic diatoms was performed by Wulff et al. (2008b) on semi-natural cultures with a dark period of 64 days. Their results revealed significant decreases in the parameters F_v/F_m, rETR (relative electron transport rate), and α (light-using efficiency).

All species revealed oxygen production after the dark period but to different degrees. While C. gerlachei and Melosira sp. showed a moderate decline in NPP_max (39.9 and 44.6%, respectively), the remaining species (N. criophiliforma, both P. wetzelii isolates) exhibited unaffected NPP_max values. Furthermore, the chlorophyll a data pointed to species-specific responses. While chlorophyll a decreased in N. criophiliforma, C. gerlachei, and Melosira sp. after 3 months of dark incubation, it remained more or less stable in both P. wetzelii strains. This observation is further supported by the TEM observations, which showed that some of the thylakoid membranes remained intact in this species. In contrast, in N. criophiliforma, much more degradation of the thylakoid system was observed with an increased number of plastoglobules. The lower NPP_max values can be related to the decline in chlorophyll a content as discussed by Wulff et al. (2008a) and Reeves et al. (2011), who explained the decrease in photosynthetic potential in Antarctic diatoms with a progressive degradation of the antenna complex and the reaction center. The decline in chlorophyll a concentration in some of the benthic diatom species can also be explained by chloroplast degradation after 3 months of darkness, but remarkably, this did not seem to have an effect neither on the physiological performance nor on the viability. Apparently, parts of the thylakoid membranes are degraded in darkness, but a functional part seems to remain intact and thus contributes to a basal level of photosynthetic activity. Kennedy et al. (2019) found similar results in Antarctic sea ice diatoms during 4 months of dark incubation and reported a rapid reduction of light-harvesting complexes and photosystems while maintaining photosynthetic capacity.

Applying SYTOX Green to the Arctic benthic diatoms Surirella cf. minuta and Navicula directa during 5 months of darkness indicated that an increasing number of cells exhibited damaged membranes over time in the first species (Karsten et al., 2019a), while the latter one was almost unaffected (Karsten et al., 2019b). In the present study, SYTOX Green staining was applied for the first time to Antarctic benthic diatoms, and the results are comparable to those on both Arctic species. While the percentage of damaged cells increased significantly in N. criophiliforma, C. gerlachei, Melosira sp., and P. wetzelii (D300_025) after 3 months of dark incubation, P. wetzelii (D300_015) maintained membrane integrity over this period of darkness. Although some diatom cells could not cope with longer darkness, others maintained viability and hence guaranteed the survival of the population after re-irradiation.

These are the first transmission electron micrographs of Antarctic benthic diatoms after dark treatment. They showed degraded chloroplasts in all benthic diatom species after 3 months of dark incubation with a markedly increased amount of plastoglobules (PGs) in the stroma of the plastids. PG appearance is a common response of plant cells to high light stress (Meier and Lichtenthaler, 1981) or any other stress, leading to a reduction of thylakoid membranes, as PGs contain the building blocks for thylakoids, including the enzymatic setting. The increase in PG in the Antarctic benthic diatoms can thus also be interpreted as a stress response; in this case, however, it is a response to long-term darkness. Schwarz et al. (2017) reported in the green microalga Micrasterias sp. how lipid droplets are released from the chloroplast into the cytoplasm during starvation, where they are degraded by autophagy. In addition, multivesicular bodies (MVBs) were also found in the cytoplasm after 7 weeks of starvation in darkness in Micrasterias sp., confirming autophagy (Schwarz et al., 2017). In the literature, it has been discussed that apoptosis and autophagy are both distinct processes but strongly interlinked (Mariño et al., 2014). In our study, the observable TEM structures are MVBs, which were present in N. criophiliforma and P. wetzelii (D300_015). Along with the numerous PGs and the degradation of the chloroplasts, this could indicate a mobilization of energy reserves by the autophagy of chloroplast components to survive the polar night. Furthermore, the chlorophyll a content per cell after 3 months of darkness strongly decreased in N. criophiliforma, C. gerlachei, and Melosira sp., supporting the ultrastructural results.

The cell biological data on Antarctic benthic diatoms confirm the few results on their Arctic pendants, which also reported, for example, a 30–50% decrease in chloroplast length (Karsten et al., 2012, 2019b). In addition, Wulff et al. (2008a) demonstrated condensed chloroplasts after dark treatment, which recovered within hours after re-exposure to light. Thus, no long-term damage appears to have been caused by the partial decomposition of chloroplasts during the prolonged darkness. According to Karsten et al. (2019b), the degradation of chloroplast compounds seems to be a key mechanism in Arctic benthic diatoms to survive the polar night and generate energy for their maintenance metabolism. The results of this study confirm that this is also a survival strategy for the examined Antarctic benthic diatoms.

Both the light microscopic and the TEM images indicated significant differences in the lipid content of all species between control and after 3 months of dark incubation, which was additionally supported by the GC-MS analysis. N. criophiliforma and Melosira sp. consumed most of their lipid reserves in darkness, C. gerlachei depleted ~85% of their lipids, and both P. wetzelii strains still exhibited lipid droplets even after 3 months of darkness, which could suggest that this species might survive additional months in darkness. Enhanced lipid content, as well as the consumption of storage substances during darkness to maintain the cellular metabolism, have been described before for Arctic diatoms (Zhang et al., 1998; McMinn and Martin, 2013; Karsten et al., 2019b). Only a few studies on Arctic benthic diatoms quantified the decrease of lipid storage compounds during darkness (Schaub et al., 2017; Karsten et al., 2019a). Investigations on Antarctic benthic diatoms in this regard are entirely lacking so far.

Diatoms carry two pathways of β-oxidation for the breakdown of lipid compounds, the plant-like and the animal-like metabolic capabilities (Armbrust et al., 2004; Schaub et al., 2017). In the darkness, the mechanism of animal-like β-oxidation, located in the mitochondria, is upregulated and thus can provide energy from lipid storages during the polar night (Chauton et al., 2013), which probably allows diatoms to survive long periods of darkness, thereby maintaining their metabolism and the fundamental functions of their organelles (Armbrust et al., 2004). Thus, the results of the present study confirm the described metabolic pathway in Armbrust et al. (2004) and Chauton et al. (2013). In addition, in the more recent review of Leyland et al. (2020), the authors describe the storage of lipids in droplets and characterize these structures as organelles composed of a core of neutral lipids, mostly triacylglycerol (TAG), surrounded by a polar lipid monolayer. Most interestingly, lipid droplets can store not only reserves of energy but also membrane components, carbon skeletons, carotenoids, and even proteins (Leyland et al., 2020, and references therein).

While the membrane lipids of polar diatoms are mainly composed of polyunsaturated fatty acids in order to maintain membrane fluidity at low temperatures (Murata and Los, 1997), the storage lipids are mainly formed by triacylglycerol (TAG), which has three fatty acids that are mainly saturated or monounsaturated (Schaub et al., 2017). The latter authors carefully investigated qualitatively and quantitatively the lipid classes of the Arctic benthic diatom Navicula perminuta during 2 months of dark incubation. Schaub et al. (2017) demonstrated by the ratio of SFA and MUFA to PUFA that, indeed, lipid reserves were depleted, while membrane lipids remained unchanged, including thylakoid membranes. In the Antarctic C. gerlachei, Melosira sp., and both P. wetzelii strains, a decline in the ratio of fatty acids groups was observed, indicating enhanced degradation of storage lipids. In contrast, N. criophiliforma showed a slight increase in the ratio of fatty acids after 12 weeks, which points to a stronger increase in membrane degradation compared to lipid consumption. Apparently, both the consumption of storage lipids and the partial degradation of chloroplasts took place in the investigated Antarctic benthic diatoms during prolonged darkness in order to obtain energy for the maintenance metabolism.