Differential Responses to UV-A Stress Recorded in Carotenogenic Microalgae Haematococcus rubicundus, Bracteacoccus aggregatus, and Deasonia sp.

UV-A is the main ultraviolet component of natural (solar) radiation. Despite it, its effect on phototrophs is studied less than UV-B. Effects of UV-A on photosynthetic apparatus of three carotenoid-producing microalgae were elucidated. Photosynthetic activity was studied using chlorophyll fluorescence analysis. Cell extracts were evaluated by absorbance spectroscopy. On the one hand, there were some common features of three strains. In all cases the changes involved PSII primary photochemistry and antennae size. All strains accumulated UV-absorbing polar compounds. On the other hand, some responses were different. Upregulation of non-photochemical quenching was observed only in B. aggregatus BM5/15, whereas in other cases its level was low. H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 accumulated secondary carotenoids, whereas B. aggregatus BM5/15 accumulated primary ones. Microscopic features of the cultures were also different. H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 were represented mostly by solitaire cells or small cell clusters, lacking their green color; the cells of B. aggregatus BM5/15 formed aggregates from green cells. Cell aggregation could be considered as an additional UV-protecting mechanism. Finally, the strains differed by their viability. B. aggregatus BM5/15 was most resistant to UV-A, whereas massive cell death was observed in two other cultures.


Introduction
Solar radiation is the primary source of energy for the biosphere. In the process of photosynthesis, it is captured and converted into energy-rich metabolites and reducing power. Although the photoassimilates are required for fixation of inorganic carbon by photoautotrophic organisms such as higher plants and (micro)algae, excessively captured light energy can kill the phototrophic cell. Photosynthetic apparatus (PSA) is the site of light energy assimilation in the cell and hence the main target of the photodestruction [1]. The spectrum of the sunlight covers the whole visible, infrared, and UV ranges. The latter contains high-energy photons mediating the damage caused by unattenuated solar radiation. Although the spectral quality of solar radiation undergoes annual and diurnal variations and affected by weather conditions and geographical location, it commonly includes UV-A radiation [2,3]. Near the surface of the Earth, solar spectrum includes 8-9% UV including 6.3% UV-A (320-400 nm) and 1.5% UV-B (280-320 nm) [2,3].
Mechanisms of destructive effects of UV-B are relatively well understood. UV-B quanta are absorbed by aromatic rings, S-S bonds, and peptide bonds of biomolecules. UV-B induces damage to nucleic acids and proteins, facilitates the generation of reactive oxygen species (ROS) in the presence of photosensitizers [2][3][4][5][6][7][8]. Despite a larger proportion of UV-A in solar spectrum, molecular basis of its effect on phototrophic cell are less elucidated [1]. Typical of UV-A is its bi-lateral effect on photoautotrophs: low and moderate acids and the fractions of different groups of membrane lipids [25]. It should be noted that most of data were obtained either for UV-B or natural/mixed UV. Hence, collection of new information about the effect of UV-A on photosynthesis is a topical current issue.
Carotenogenic algae are considered as model objects for studies of the response of PSA to various stress factors, especially UV radiation. However, the data on the effect of UV-A on different groups of carotenogenic microalgae are poor. The response of PSA is especially interesting as it is the main target of damage by harmful radiation. The present study aims at elucidating the effects of UV-A on the PSA of three carotenogenic microalgae from different families of Chlorophyta originated from contrasting habitats, namely H. rubicundus Allewaert et Vanormelingen (an aeroterrestrial alga), Bracteacoccus aggregatus Tereg (a freshwater alga), and Deasonia sp. (a soil alga). Toward this end we exposed algal cultures to artificial light with adding of UV-A. Then, water-methanol (polar) and chloroform (non-polar) fractions of the extracts of the microalgae were analyzed by absorbance spectroscopy. It provided the data on changes in the composition of hydrophobic and hydrophilic compounds which absorb radiation, i.e., photosynthetic pigments and photoprotectants. Photosynthetic primary photochemistry also was evaluated by the analysis of chlorophyll fluorescence induction kinetics during the UV-A treatment.
In the Results section we provide the data of visual observations of algal cultures as well as microscopic observations of algal cells during 21 day of UV-A treatment (Section 2.1), describe changes of light absorbing properties and pigment indices based on the analysis of the chloroform extracts (Section 2.2), describe the changes of UV absorption by the water-methanol extracts (Section 2.3), and showed the changes of the parameters of PSII primary photochemistry based on the chlorophyll fluorescence induction kinetics analysis (Section 2.4). Obtained data are discussed and compared with previously published results in the Section 3. Experimental details of the work are explained in the Section 4. Obtained conclusions are summarized in the Conclusion (Section 5).

Microscopic Observations
The microphotographs from each experiment stage were obtained for all studied strains: Deasonia sp. NAMSU 934/2 (Figure 1a-d)  served. They contained from 4 to 32 autospores. The culture consisted of either single cells or, more frequently, tetrahedral (or occasionally ring-like) colonies. In the culture of B. aggregatus BM5/15 small spherical coccoid cells 5-10 μm in diameter persisted (Figure 1e). Occasionally, large cells up to 25 μm diameter also were observed. As a rule, B. aggregatus cells formed large clusters. H. rubicundus BM7/13 was presented by individual spherical vegetative cells up to 50 μm in diameter (Figure 1i).  In the beginning of the experiment all cultures were characterized by their typical cell morphology. Thus, vegetative cells of Deasonia sp. NAMSU 934/2 were immotile, coccoid, 5-40 µm in diameter ( Figure 1a); asexual reproduction via autosporangia was observed. They contained from 4 to 32 autospores. The culture consisted of either single cells or, more frequently, tetrahedral (or occasionally ring-like) colonies. In the culture of B. aggregatus BM5/15 small spherical coccoid cells 5-10 µm in diameter persisted ( Figure 1e). Occasionally, large cells up to 25 µm diameter also were observed. As a rule, B. aggregatus cells formed large clusters. H. rubicundus BM7/13 was presented by individual spherical vegetative cells up to 50 µm in diameter (Figure 1i).
The most pronounced feature of the H. rubicundus BM7/13 and B. aggregatus BM5/15 cultures was the mass production of bi-flagellated zoospores at the 7th d of UV-A treatment (Figure 1f,j). Particularly, zoosporangia with motile zoospores were observed occasionally ( Figure 1f). Later, the zoospores disappeared abruptly and were not detected after 14 (Figure 1g

Cell Pigment Composition
To evaluate the changes in the pigment composition of the studied strains during UV-A-treatment, their chloroform extracts were studied. absorbance spectra of the extracts were obtained in the visible and near infrared (NIR) range for all three carotenogenic microalgae: H. rubicundus BM7/13 (Figure 2a), B. aggregatus BM5/15 ( Figure 2b) and Deasonia sp. NAMSU 934/2 ( Figure 2c). cell extracts (Figure 2a-c). UV-A treatment of H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 (Figure 2a,c) cultures led to an increase in the extract absorbance in the blue-green region of the spectrum reflecting the accumulation of secondary ketocarotenoids. Similarly pronounced changes lacked in the spectra of B. aggregatus BM5/15 cell extracts (Figure 2b). Based on the spectral features, this culture did not accumulate ketocarotenoids. Statistically significant decrease of chlorophyll content was observed in all cultures during the treatment by UV-A (Supplementary File S2).
The calculated Car/Chl ratio served as a proxy to the extent of carotenoid accumulation (Figure 3a-c). The strains H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 were characterized by a sharp rise of this parameter after seven days of UV-A treatment (Figure 3a,c). Then, these cultures were characterized by high dispersion of the parameter (Figure 3a-c). Most likely, it was attributed to the massive cell death and destruction of the pigments. The Car/Chl of the B. aggregatus BM5/15 cultures did not change so sharp. However, statistically significant difference was observed between 7th and initial days (Figure 3b).

The Spectra of Water-Methanol Extracts
To evaluate the changes in the pigment composition of the studied strains during UV-A-treatment, their chloroform extracts were studied. absorbance spectra of the extracts were obtained in the visible and near infrared (NIR) range. While in the chloroform extracts pronounced changes was observed in the light absorbance in the visible range, the water-methanol fractions from different days of incubation were different in the UV part of the spectrum. The differential spectra of water-methanol extracts were analyzed to indicate accumulation of non-polar UV-absorbing compounds during UV-A treatment. The spectra were calculated by subtraction of the absorbance spectra of the extract made at the beginning of the experiment (D(λ) 0 day ) from the spectra of extracts of the cells sam-

The Spectra of Water-Methanol Extracts
To evaluate the changes in the pigment composition of the studied strains during UV-A-treatment, their chloroform extracts were studied. absorbance spectra of the extracts were obtained in the visible and near infrared (NIR) range. While in the chloroform extracts pronounced changes was observed in the light absorbance in the visible range, the water-methanol fractions from different days of incubation were different in the UV part of the spectrum. The differential spectra of water-methanol extracts were analyzed to indicate accumulation of non-polar UV-absorbing compounds during UV-A treatment. The spectra were calculated by subtraction of the absorbance spectra of the extract made at the beginning of the experiment (D(λ) 0 day ) from the spectra of extracts of the cells sampled at subsequent days of the experiments (∆D(λ) = D(λ) − D(λ) 0 day ). The spectra were obtained for all studied strains: H. rubicundus BM7/13 (Figure 4a) ANOVA of the data on the absorbance at 280 nm, certain groups of statistically different datasets (corresponded to days of cultivation) were distinguished (p < 0.05) only for Deasonia sp. NAMSU 934/2. The data on the extracts from H. rubicundus BM7/13 ( Figure  5d) and B. aggregatus BM5/15 cells were statistically homogeneous (Figure 5e). The absorbance at 280 nm increased slightly in Deasonia sp. NAMSU 934/2 (Figure 5f) on the 7th day of UV-A treatment. The decrease on the 2nd and 3rd weeks of culturing could be explained by cell death and release of UV-absorbing compounds to the culture medium.   ANOVA of the data on the absorbance at 280 nm, certain groups of statistically different datasets (corresponded to days of cultivation) were distinguished (p < 0.05) only for Deasonia sp. NAMSU 934/2. The data on the extracts from H. rubicundus BM7/13 ( Figure  5d) and B. aggregatus BM5/15 cells were statistically homogeneous (Figure 5e). The absorbance at 280 nm increased slightly in Deasonia sp. NAMSU 934/2 (Figure 5f) on the 7th day of UV-A treatment. The decrease on the 2nd and 3rd weeks of culturing could be explained by cell death and release of UV-absorbing compounds to the culture medium.   The absorbance at 332 nm, also potentially indicative of MAA presence, increased gradually throughout the experiment in the case of B. aggregatus BM5/15 (Figure 5b). In H. rubicundus BM5/15 (Figure 5a) extracts it increased after 7 days of culturing and then changed insignificantly. Its sharp rise was also observed in Deasonia sp. NAMSU 934/2 (Figure 5c), then the absorbance changing was not statistically significant. In the one-way ANOVA of the data on the absorbance at 280 nm, certain groups of statistically different datasets (corresponded to days of cultivation) were distinguished (p < 0.05) only for Deasonia sp. NAMSU 934/2. The data on the extracts from H. rubicundus BM7/13 ( Figure 5d) and B. aggregatus BM5/15 cells were statistically homogeneous (Figure 5e). The absorbance at 280 nm increased slightly in Deasonia sp. NAMSU 934/2 (Figure 5f) on the 7th day of UV-A treatment. The decrease on the 2nd and 3rd weeks of culturing could be explained by cell death and release of UV-absorbing compounds to the culture medium.
The amplitude of the Chl fluorescence transient curves decreased sharply after first seven days of UV-A treatment in the cultures of H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 (Figure 6a,c). This decrease was less pronounced in the B. aggregatus BM5/15 culture (Figure 6b). The characteristic shape of the fluorescence transient curve was retained until 14 days of incubation (Figure 6b). The UV-A treatment led to a decrease of photosynthetic activity measured as the maximal PSII photochemical quantum yield in the dark-acclimated state (Fv/Fm) in all cases. In H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2, Fv/Fm decreased sharply after the first seven days of the incubation under UV-A (Table 1) approaching zero by last days of the experiment. In the B. aggregatus BM5/15 culture, Fv/Fm was as low as 0.4, i.e., lower than in the other cases (Table 1) but its decrease was less abrupt. The photosynthetic activity was retained in B. aggregatus BM5/15 cultures until the last day of the experiment. As in the case of Fv/Fm, H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 were characterized by a sharp decline of the probability of electron transport beyond QA (Ψ0) during UV-A treatment (Table 1). It was most pronounced in the Deasonia sp. NAMSU 934/2 reflect likely an effect of UV-A on PSA at the level of electron transport in the PSII RC. At the same time, B. aggregatus BM5/15 retained relatively high values of this parameter, because there were statistically different groups of data based on the ANOVA analysis (p < 0.05) (  The UV-A treatment led to a decrease of photosynthetic activity measured as the maximal PSII photochemical quantum yield in the dark-acclimated state (Fv/Fm) in all cases. In H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2, Fv/Fm decreased sharply after the first seven days of the incubation under UV-A (Table 1) approaching zero by last days of the experiment. In the B. aggregatus BM5/15 culture, Fv/Fm was as low as 0.4, i.e., lower than in the other cases (Table 1) but its decrease was less abrupt. The photosynthetic activity was retained in B. aggregatus BM5/15 cultures until the last day of the experiment. As in the case of Fv/Fm, H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 were characterized by a sharp decline of the probability of electron transport beyond Q A (Ψ 0 ) during UV-A treatment (Table 1). It was most pronounced in the Deasonia sp. NAMSU 934/2 reflect likely an effect of UV-A on PSA at the level of electron transport in the PSII RC. At the same time, B. aggregatus BM5/15 retained relatively high values of this parameter, because there were statistically different groups of data based on the ANOVA analysis (p < 0.05) (

Discussion
Overall, UV-A is less harmful for microalgae than UV-B [47][48][49]. Different groups of phototrophs are characterized by different responses of their PSA to UV [50]. To further understand the diversity and specificity of the UV-A effect on the algal PSA, it was studied in the three ecologically and phylogenetically distant strains of carotenogenic microalgae.
In two of the studied strains, an increase in Car/Chl ratio was observed. In case of primary carotenoids, its value is strictly defined by conserved stoichiometric relationships between the carotenoids and Chl in PSA pigment-protein complexes [54,55,57]. UV-Ainduced rise of Car/Chl has been reported in other algae [22,24,49,52,56]. Its moderate increase can be due to the synthesis of additional primary carotenoids, such as those participation in xanthophyll cycles. Expectedly, a pronounced accumulation of secondary carotenoids was recorded in the two of three studied microalgal strains (H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2).
Another carotenogenic microalga, D. salina, under a moderate UV-A exposure showed an increase in the content of major cell carotenoids, especially lutein and β-carotene, as well as in the pigments of violaxanthin cycle [49,52,56]. At the same time, high doses of UV-A promoted only the accumulation of β-carotene as the secondary carotenoid [49]. Similar effect was also shown for C. rubescens [22] and Ch. krienitzii [15]. At the ultrastructural level, UV-A treatment in the carotenogenic algae C. rubescens and Ch. krienitzii was accompanied by an expansion of the hydrophobic (sub)compartments of the cell including cytoplasmic oil bodies and plastoglobuli harboring the secondary carotenoids [15,22]. Accumulated secondary carotenoids absorb excessive UV-A and shield PSA. Thus, B. aggregatus BM5/15 demonstrated an effect typical of moderate UV-A treatment, the same doses of UV-A exerted a deteriorative effect on H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2.
In phototrophs, the actual photosynthetic performance is determined by the balance of UV-A and visible light irradiation. When the visible light is limiting, UV-A is favorable for photosynthesis, but under over-saturating visible light intensities UV-A exposure leads to photoinhibition [5]. UV-A has a bi-phasic effect on PSA [1,2,5,[9][10][11]. It has been shown previously especially for microalgae. In N. oceanica moderate UV-A doses promote culture growth, as well as upregulating overall metabolic activity, estimated as O 2 photosynthetic emission, and nitrate assimilation [11]. The moderate doses also increase the rate of the O 2 production in D. salina [49,52]. In P. globosa moderate increase of the UV-A dose leads to an increase of photosynthetic activity and decrease of NPQ [53]. At the same time, in the same species of microalgae high UV-A downs out photosynthesis [11,23,49].
In all studied cases, the UV-A treatment led to a decrease of the potential photochemical efficiency of PSII (measured as Fv/Fm), this effect was also documented in the present work and in C. rubescens under the same experimental conditions [22]. Mechanistically, the harmful effect of UV-A on PSA is thought to be implemented mainly by PSII RC damage [9,13,52,56,58,59]. The Fv/Fm parameters provides only information about the potential maximal photochemical efficiency of PSII, which is often not sufficient; such cases require additional parameters such as Ψ 0 to be considered. Thus, a pronounced difference of Ψ 0 was observed in the studied strains. While H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 demonstrated a sharp decrease of Ψ 0 , B. aggregatus BM5/15 retained moderately high values of this parameter. Ψ 0 reflects the efficiency of the electron transport between Q A and Q B in PSII [60]. Indeed, UV-A predominantly affects the OEC, Q A and Q B binding sites in PSII [9,10,14]. Accordingly, the data obtained in this study reinforce the current view of the UV effect on PSA, although in B. aggregatus BM5/15 this effect was less pronounced.
In all studied cultures, the UV-A irradiation promoted the decrease in Chl a/b. Since Chl b is contained exclusively in the light-harvesting antenna (LHC), this effect suggests a relative increase of the antenna size. However, the effective PSII antenna size (ABS/RC) calculated from the Chl fluorescence transient curves decreased in all cases. Most likely, the algal cells responded to the UV-A stress by reducing the amount of the absorbed PAR and hence by lowering the excitation pressure on PSII and subsequent ROS formation. At the same time, Chl a/b decrease could reflect increasing could be explained by the decreasing of the PSI antennae size. PSI upregulation might be important to weaken over-reduction of the plastid electron transport chain. Ivanova et al. [10] have shown the decrease of antenna size facilitates the protection of PSA against UV-A damage in a legume Pisum sativum. Uncoupling of LHCII complexes and their oligomerization is considered as an additional UV-protective mechanism [10].
Non-photochemical quenching of the excited Chl states is an omnipresent photoprotective mechanism in photosynthetic organisms [61]. It redirects the excessively absorbed light energy from destructive photochemical reactions to safe thermal dissipation. Under our experimental conditions, H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 were not characterized by the increase of the Stern-Volmer NPQ parameter under the UV-A exposure. This finding suggests a downregulation and/or inhibition of non-photochemical quenching. By contrast, in B. aggregatus BM5/15, NPQ increased suggesting upregulation of the thermal dissipation. Under the UV-A treatment, P. globosa [53], N. oceanica, and C. rubescens also decreased their NPQ [22,49,53]. This effect was observed despite of the expansion of the violaxanthin cycle pigment pool responsible, at least partially, for NPQ development [56,59]. However, in chlorophytes the xanthophyll cycle seems to be less important for the NPQ engagement than in higher plants [62,63]. In view of this, the response of B. aggregatus BM5/15 to UV-A can be regarded as unusual. It could be explained by the upregulation of NPQ to protect its highly functional PSA, whereas two other strains did not need it since their photosynthetic activity declined dramatically.
Under the UV-A treatment, an increase in the absorbance at 332 nm was observed in all studied strains. This response was most pronounced in B. aggregatus BM5/15. Most likely, absorbance at this wavelength together with shape of the spectra could be attributed to the accumulation of MAA. These compounds, widely distributed among microalgae, are wellknown as efficient screening agents [4][5][6][18][19][20]. Their protective role, especially against UV, has been shown in vivo [4,18,64]. As a rule, MAA synthesis is UV-inducible [19,20,22]. It is in line with the data obtained in current work.
Although mechanisms of regulation of MAAs synthesis in algae are poorly understood [64], it mediated most likely by the cryptochromes absorbing UV-A [8]. One can speculate that the higher viability and photosynthetic activity of B. aggregatus BM5/15 was attributed, inter alia, to a higher MAA level. Nevertheless, it might only be a side effect because the cultures were irradiated by UV-A in the range of 380-415 nm (see Materials and Methods), whereas the absorption band of putative MAA is located in the range of 300-350 nm. However, MAA also have antioxidant effect [4,18] so they could contribute to ROS detoxication. Furthermore, a pronounced increase of the culture extract absorption was observed at 280 nm in Deasonia sp. NAMSU 924/2. It could reflect accumulation of simple phenolic compounds such as in some other chlorophytes [65].
Microscopic observations could reveal the relationship between UV radiation and its effect on the microbial physiology and structure. The most spectacular feature under our experimental condition in the H. rubicundus and B. aggregatus cultures was the mass zoospore production observed on the seventh day of UV-A exposure. Production of motile cells in the carotenogenic H. lacustris can be induced by a sharp change of cultivation conditions [66]. Light intensity and spectral quality play a significant role in both sporogenesis and gametogenesis in green algae. Short-wavelength (blue) PAR radiation is favorable for massive formation of flagellated cells, whereas long-wavelength (red) PAR blocks it [26]. In the macrophyte Bryopsis plumosa (Ulvophyceae), a positive effect of UV-A on motile stages production has been reported [27]. Thus, zoospore production in the studied alga is another typical response UV-A. From the practical standpoint, UV-A irradiations could be used for the induction and study of motile stages in the cultures of carotenogenic algae.
Another important conclusion from obtained microscopic data is the relationship between UV-A response of cells and their ability to form aggregates. B. aggregatus BM5/15 formed large cell clusters by the end of the experiment, moreover autospores in sporangia stayed together under the mother envelope. The external cells of these clusters were bleached or intensely redcolored due to carotenoid accumulation, whereas internal cells remained green. It is likely that the external cells were exposed to high doses of UV-A blocked it by either their photosynthetic pigments or by secondary carotenoids or by UV-absorbing compounds in their walls, such as rigid sporopollenin or algaenan layer [15,22]. Even if damaged or killed by the UV-A exposure, they would continue to shield the internal cells of the cluster against the radiation. Thus, formation of aggregates and autosporangia could be considered as a mechanism for protection against the harmful impact of UV-A. Such protection by self-shielding has been proposed previously for microalgae of biological soil crust and biofilms [3,19].
Notably, the effect of radiation strongly depends on the size of an organism. Larger organisms receive a lower dose than smaller ones under the same radiation flux. Therefore, unicellular algae with larger cell size are more resistant to UV-A and UV-B than those with a smaller size [5]. Accordingly, the colonies of B. aggregatus BM5/15 could mimic the elevated UV-A radiation resilience of larger organisms. This ability might be critical to retain the functional PSA and increased viability. By contrast, H. rubicundus BM7/13 cultures lacked massive cell aggregates so this might be the explanation of its low viability and inhibited photosynthesis. A high variation of the parameters recorded in the experiment might also reflect a high functional heterogeneity of the cells in the UV-treated cultures. Further research is required to deeper understand the relationships between the species-specific and general patterns of the UV-A response in the photosynthetic microorganisms, especially carotenogenic microalgae, and their cultivation condition.

Strains and Culture Conditions
The following strains of green carotenogenic microalgae were studied: Haematococcus rubicundus BM7/13, Bracteacoccus aggregatus BM5/15 (IPPAS C-2045), and Deasonia sp. NAMSU 934/2. They were isolated and identified previously [33,40]. The cells were taken from the stationary growth phase, because at this growth stage the cultures are characterized by relatively low division rate, and, thus, more resistant to UV radiation [67]. Initial optical density of the cultures was 0.4 as was measured at 660 nm in standard quartz civets in an Agilent Cary 300 spectrophotometer (Agilent, Santa Clara, CA, USA) with an integrative sphere (150 mm diameter) of the same manufacturer. They were cultured autotrophically on the mineral medium BG-11 [68] under continuous illumination by the cold-white LED COB-X544-8mm 24V White6000 (Arlight, Moscow, Russia) (60 µmol/m 2 /s) at 25 • C in the 250 mL T-75 TC-treated cell culture flasks (Eppendorf, Hamburg, Germany) for three weeks. Photon flux density of the visible light was controlled at the level of cell suspensions by a LI-COR LI-250A quantum meter (LI-COR Inc., Lincoln, NE, USA). UV-A treatment was performed by LED 5M 033220 (Arlight, Moscow, Russia) in the spectral range of 380-415 nm with the power of 2.9 W/m 2 . Emission spectrum of the radiation is given in supplementary (Supplementary File S3). All experiments were carried out in three replicates, the measurements were carried out in three replicates. As a control experiment the cultures of microalgae also were grown under the same conditions listed above, but without UV-A treatment (Supplementary File S1).

Microscopy
The cultures were monitored by bright-field microscopy on a Leica DM2500 microscope (Leica Microsystems, Wetzlar, Germany) with the attached Leica DFC 700T camera.

Absorbance Spectroscopy
Extraction of hydrophilic and hydrophobic compounds and registration of their absorbance spectra were carried out as was previously described by Zaytseva et al. [22]. The spectra were registered on an Agilent Cary 300 in the UV-visible-NIR range. Differential absorbance spectra (∆D(λ)) in the UV range (250-400 nm) were presented as ∆D(λ) = D(λ) − D(λ) 0 day , where D(λ) 0 day is the optical density at the wavelength λ in the control (initial day of cultivation). The standard spectral indices (Car/Chl and Chl a/b) were calculated as was previously described [36].

Chlorophyll Fluorescence Analysis
The Chl fluorescence induction curves were registered by the FluorPen FP100s PAMfluorometer (Photon System Instruments, Drásov, Czech Republic) as described previously [62] using the protocols for Chl fluorescence transient (OJIP curve) and NPQ induction.
For the OJIP curves the maximal photochemical quantum yield of PS II (Fv/Fm) and the probability of electron transport beyond Q A (Ψ 0 ), PS II absorbance cross section, or PS II effective size (ABS/RC) [60,69], were calculated as, where Fm and Fo are maximal and minimal Chl fluorescence intensity, respectively, registered in the dark-acclimated state; where F J is the Chl fluorescence intensity in the point J (at 2021 µs); The Stern-Volmer non-photochemical quenching parameter (NPQ) was calculated from the stationary Chl fluorescence curves registered according to the NPQ induction protocol [62] as, where Fm is the maximal Chl fluorescence intensity in the light-acclimated state [69]. Minimal values of Fm were taken to acquire maximal NPQ for the analysis.

Statistical Treatment
All measurements were performed in total of six replicates (three biological replicates with two analytical replicates in each biological one). The data on each estimated parameter were divided into three independent datasets corresponded to studied strains of microalgae. For analysis of the data of absorbance spectroscopy the standard one-way ANOVA (p = 0.95) was applied for the datasets of each parameter for each algal strain. After confirming the hypothesis of differences in certain groups of data, the pairs of samples corresponded to the days of cultivation, were compared by the post-hoc Tukey test. The Kolmogorov-Smirnov test was used to determine normality distribution.
Since the data of chlorophyll fluorescence analysis are not approximated by the normal distribution [70], the non-parametric Kruskal-Wallis H test was applied for the datasets (p = 0.99). The standard Mann-Whitney U test commonly used for this type of data [71,72] than was carried out to compare pair of samples.
The analysis was carried out in STATISTICA v. 13.3 (StatSoft, Dell, Round Rock, TX, USA).

Conclusions
We presented the first work on the effect of UV-A on three phylogenetically distant green ketocarotenoid-producing microalgae, namely Haematococcus, Deasonia, and Bracteacoccus. The genera Deasonia and Bracteacoccus are poorly studied in terms of their stress physiology.
There were some common features of response to UV-A of three studied strains of green carotenogenic microalgae. (1) In all cases the changes involved PSII primary photochemistry: the decrease of Fv/Fm and Ψ 0 (reflecting the transfer from Q A to Q B ), PSII antennae size also decreased. (2) The Chl a/b ratio decreased, most likely, due to the enlarging og the PSI antennae size. However, this hypothesis should be further tested. (3) All strains accumulated UV-absorbing polar compounds, most likely phenolics and MAAs. The increase of the absorbance at 332 nm (corresponding to MAAs) was most pronounced in B. aggregatus BM5/15. At the same time, the strains were different in terms of their response to UV-A. (4) NPQ upregulation was observed only in B. aggregatus BM5/15, whereas in other cases its level was low. (5) In all cases carotenoid content increased, but in H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 they were secondary carotenoids, whereas in B. aggregatus BM5/15 formed primary ones. In the last cases they could contribute to the NPQ build up. (6) Most pronounced difference was observed in microscopical features of the cultures. H. rubicundus BM7/13 and Deasonia sp. NAMSU 934/2 were represented mostly by solitaire cells or small cell clusters, which lacked their green color. At the same time, B. aggregatus BM5/15 formed aggregates from green cells. Formation of such aggregates could be considered as an additional UV-protecting mechanism. (7). Finally, the strains differed by their viability. B. aggregatus BM5/15 was most resistant to UV-A, whereas massive cell death was observed in two other cultures.