Deoxygenation in global marine ecosystems is increasing in scope and severity due to anthropogenic activities and climate change (Keeling et al., 2010). According to Breitburg et al. (2018), the total dissolved O₂ (DO) content in the ocean has decreased by 2% in recent decades (Breitburg et al., 2018) and will decrease by a further 1% to 7% by the end of this century (Long et al., 2016; Schmidtko et al., 2017), leading to a sharp increase in the frequency and severity of marine hypoxia (dissolved O₂ <2.0 mg L⁻¹) (Schmidtko et al., 2017; Rabalais and Turner, 2019; Wang et al., 2023). Hypoxia affects most marine aerobic organisms through inhibiting visual function, reproduction, and population development (e.g., McCormick and Levin, 2017; Wang et al., 2016) and the photoautotrophic macroalgae through photosystem (e.g., Peckol and Rivers, 1995; Alamoudi et al., 2022) and seagrasses through chlorophyll (e.g., Che et al., 2022). Recently, many field studies have shown that marine phytoplankton that dwell in hypoxic zones are also experiencing the low O₂ level occasionally (Li et al., 2018; Sun et al., 2022) or permanently (Gomes et al., 2014; Xiang et al., 2019); therefore, many laboratory experiments have been conducted to explore how hypoxia affects phytoplankton (Chen et al., 2021; Zhao et al., 2022; Eom et al., 2024; Chen et al., 2024; Tong et al., 2024). One of the reasons for this widespread interest is that phytoplankton, apart from their important role as the main primary producer in the ocean, require exogenous O₂ to maintain respiratory metabolism at night (Li et al., 2016), although this challenge can be met by photosynthesis during the day.

Although marine phytoplankton contribute to about half of the global photosynthetic O₂ (Field et al., 1998), the death of extensively bloomed phytoplankton often accelerates the formation of hypoxic environments in the water column through O₂ depletion by increasing respiration themselves and the nitrification of microbes supported by organic matter released by the phytoplankton cells (Chen et al., 2007). In this scenario, phytoplankton are also affected by the lowered O₂ level when they fall into the O₂-deficient layer due to their own sinking or vertical mixing of the water column, especially in the shallower coastal or estuarine regions (Li et al., 2018; Zhai et al., 2021). Theoretically, lowering O₂ level benefits photoautotrophic organisms as it increases the likelihood of CO₂ molecules binding to ribulose-1, 5-bisphosphate carboxylase/oxygenase (RubisCO) (Gao and Campbell, 2014). It is therefore generally believed that lowering O₂ can promote the photosynthetic production of phytoplankton cells (Pruder and Bolton, 1980; Raven and Larkum, 2007) and decrease the consumption of the photosynthates by mitochondrial respiration (Reinfelder et al., 2000; Giordano et al., 2005), allowing them to store more products for growth (Sun et al., 2022). On the other hand, lowering O₂ levels is known to decrease the production of harmful reactive oxygen species (ROS) which often inactivate many cellular enzymes and thus block the cell growth (Pérez-Pérez et al., 2012). However, many studies demonstrated the negative effects of lowered O₂ on phytoplankton, such as decreasing the cell size and cell growth of diatoms (Wu et al., 2012; Zhao et al., 2022), decreasing the cell carbon content of coccolithophores (Tong et al., 2024) and increasing the cell mortality of dinoflagellates (Eom et al., 2024). An inhibitory effect of lowering O₂ was also observed on cellular nitrogen metabolism (Chen et al., 2023) and calcification (Tong et al., 2024). These results indeed illustrate the effects of lowering O₂ on phytoplankton physiology but are highly contradictory.

Phytoplankton, like heterotrophic organisms, require exogenous O₂ to sustain their respiration at night. Therefore, the duration of light exposure together with light intensity has been found to influence the responses of phytoplankton to a decrease in O₂, such as the increase in growth of the diatom Thalassiosira pseudonana at short photoperiod or low light intensity, while the growth decreases inversely at long photoperiod or high light intensity (Chen et al., 2021, 2023, 2024). These studies mainly focused on the effects of lowered O₂ in the light state and neglected the effects in the dark that phytoplankton often experience in nature, including the rhythmic dark state due to the Earth’s rotation or the long-term darkness when sinking out of the euphotic layer. To address this gap, we cultured a representative diatom, Thalassiosira pseudonana, under three O₂ levels (8.0, 4.0, and 1.3 mg L^-1) and comparatively studied its physiochemical differences between the light and dark periods, with emphasis to the changes in biochemical composition and sinking rate of cells during a long-term darkness (84 h). We selected this diatom as a target species because it is one of the dominant groups in hypoxic zones (Gomes et al., 2014; Xiang et al., 2019; Sun et al., 2022), and it can also serve as food to enhance the survival of heterotrophs in hypoxic environments (Eom et al., 2024). On the other hand, the diatoms have siliceous shells that allow them to settle more easily into the deep layer in nature and encounter low O₂ (Bannon and Campbell, 2017).

The growth rate of T. pseudonana of total and during the light and dark periods is shown in Figure 1 . The total growth rate under ambient O₂ was 0.60 ± 0.02 day^-1 and showed no significant difference from that under low O₂ (p = 0.29), while it was decreased by 15% by hypoxia. Under ambient O₂, the growth rate during light period was about two times the total growth rate, but the growth rate during dark period was only 6% of the total growth rate. Comparing to ambient O₂, the growth rate during light period was increased by 10% by low O₂ (p < 0.05) but showed no significant effect by hypoxia, while the growth rate during dark period was significantly decreased by both low O₂ and hypoxia (p < 0.01).

During the light period, cellular Chl a content under ambient O₂ was 0.41 ± 0.02 pg cell^-1 and was decreased by 7% and 20% by low O₂ and hypoxia, respectively ( Figure 2A ). Chl a content under ambient O₂ during the dark period was 20% lower than that during the light period and was further decreased by 12% and 26% by low O₂ and hypoxia, respectively ( Figure 2A ). Car content was 0.41 ± 0.02 pg cell^-1 under ambient O₂ and showed a similar trend of lowered O₂ effect as Chl a during both light and dark periods ( Figure 2B ). POC content during the light period was 10.7 ± 0.14 pg cell^-1 under ambient O₂, being increased by 15% by low O₂ but decreased by 15% by hypoxia. The POC in the dark period was 20% lower than that in the light period and was increased by 21% by low O₂ but decreased by 17% by hypoxia ( Figure 2C ). By contrast, the PON content during light (i.e., 1.30 ± 0.09 pg cell^-1) and dark periods (i.e., 0.95 ± 0.09 pg cell^-1) showed no significant difference between ambient and low O₂ treatment (light period: p = 0.27; dark period: p = 0.33) ( Figure 2D ). However, the PON was decreased by 21% and 33% by hypoxia during the light and dark periods, respectively.

The maximum photochemical quantum yield (F_V/F_M) of PSII of T. pseudonana was 0.65 ± 0.01 and 0.73 ± 0.01 during the light and dark periods, respectively, and showed no significant effect of lowering O₂ (one-way ANOVA, F _2,6 = 2.33, p = 0.18) ( Figure 3A ), although the negative effects by lowering O₂ on the RLC-derived light utilization efficiency (α), maximum relative electron transport rate (rETR_max), and saturation light (E_K) occurred under both light and dark conditions ( Table 1 ). However, the photosynthetic O₂ evolution rate (P_n, 2.13 ± 0.09 fmol O₂ cell^-1 min^-1) under growth light was reduced by 11% and 26% by low O₂ and hypoxia, respectively ( Figure 3B ). The dark respiration rate (R_d) was 0.58 ± 0.04 and 0.45 ± 0.02 fmol O₂ cell^-1 min^-1 under growth light and dark conditions, respectively, and decreased with lowering O₂ in both light and dark conditions ( Figure 3C ).

To study the metabolism of diatoms that had dropped out of the euphotic zone, we cultured T. pseudonana in the dark for 84 h ( Figure 4 ). At ambient O₂, the cell density decreased from 1,678 ± 56 to 513 ± 95 cell mL^-1 from T0 to T84, with a cell mortality rate of 14 cell mL^-1 h^-1 ( Figure 4A ). Low O₂ significantly reduced the mortality rate by 30%, while hypoxia exacerbated it by 40%. At ambient O₂, cellular carbon content (POC) decreased from 10.4 ± 0.97 to 4.34 ± 0.29 pg cell^-1 from T0 to T84, with a C loss rate of 0.066 pg cell^-1 h^-1 ( Figure 4B ), while nitrogen content (PON) decreased from 1.14 ± 0.05 to 0.33 ± 0.03 pg cell^-1, with a N loss rate of 0.011 pg cell^-1 h^-1 ( Figure 4C ). Lowering O₂ showed no significant effect on the C loss rate (F = 0.39, p = 0.68) but significantly decreased the N loss rate (F = 3.65, p = 0.031) ( Figure 4C ). In addition, a greater decrease in C loss rate than N loss rate led to an increase in the POC/PON ratio with the duration of dark time ( Figure 4D ). At T0, T48, and T84, cells at ambient O₂ settled in a rate of 4.88 ± 0.07, 5.12 ± 0.08, and 5.88 ± 0.19 cm day^-1, respectively ( Figure 5 ). Low O₂ showed no significant effect on sinking rate at T0 and T48 but decreased it by 10% at T84; hypoxia, however, increased the sinking rate by 13%. In addition, there was an interactive effect of lowered O₂ and dark duration on the cell sinking rate (two-way ANOVA, F _4,12 = 15.68, p < 0.01).

The effects of lowering O₂ levels on the physiological and metabolic processes of marine aerobic organisms are well documented (e.g., Wang et al., 2016; McCormick and Levin, 2017). Many studies have also shown the different photophysiological responses of diatoms to a decrease in O₂ level at different light intensities (Pruder and Bolton, 1980; Wu et al., 2012; Zhao et al., 2022; Sun et al., 2022; Chen et al., 2023, 2024) and at different light exposure times, i.e., photoperiod (Chen et al., 2021). In this study, we further showed that the diatom T. pseudonana responds differently to lowered O₂ level during the day and at night.

Lowering O₂ was generally thought to benefit the photosynthetic carbon fixation of phytoplankton (Raven and Larkum, 2007), as the carboxylation efficiency of ribulose-1, 5-bisphosphate carboxylase/oxygenase (RubisCO), a rate-limiting enzyme of carbon fixation, was directly relevant to the O₂/CO₂ ratio (Gao and Campbell, 2014). Low O₂ means a fewer O₂ molecules that compete for the binding site of RubisCO with CO₂ and thus less energy cost to maintain CO₂ concentrating mechanisms (CCMs) (Reinfelder et al., 2000; Giordano et al., 2005), which could make cells allocate more carbon ( Figure 2C ) to support growth ( Figure 1 ). It is consistent with previous reports on the field phytoplankton communities (Sun et al., 2022) and laboratory individual species (Chen et al., 2021; Sun et al., 2022; Chen et al., 2024). However, previous studies have also noted the negative effects of lowered O₂ on Photosystem II (PS II), especially at a longer photoperiod and a higher light intensity (Chen et al., 2021, 2024). During the transformation of photosynthetic electron, oxidization of harmful reactive oxygen species (ROS), and inactivation of photosynthetic components such as the PsbA protein (Pérez-Pérez et al., 2012), diatoms had to additionally allocate by up to 15% total nitrogen to repair the damaged PS II (Li et al., 2015). In this situation, the inhibited nitrogen metabolism by lowered O₂ might further disrupt the balance between damage and repair of PSII (Chen et al., 2023), resulting in a more significant decrease in photosynthetic capacity during the day ( Figure 3A ; Table 1 ).

Aerobic metabolic processes of diatoms, such as mitochondrial respiration and the tricarboxylic acid cycle (TCA cycle), are vital for cells to obtain and allocate energy that is normally required to support other metabolic processes, such as nutrient uptake and assimilation (Lomas and Gilbert, 1999; Clark et al., 2002) and fatty acid synthesis (Chen et al., 2024), whose activities are directly related to available O₂ (Chen et al., 2021; Zhao et al., 2022). These metabolic processes are therefore more sensitive to lowered O₂ levels compared to photosynthesis ( Figure 3 ) and are additionally inhibited and even blocked at night, leading to increased degradation of cellular organic matter ( Figures 4B, C ) and even cell death ( Figure 1 ). Furthermore, our results showed that the decreased degradation of organic C and N caused by lowered O₂ is disproportionate, such that the C/N ratio increases with the duration of darkness ( Figure 4D ). According to Wu et al. (2012), hypoxia can also decrease the adenylate energy charge of diatom [AEC = (ATP + 0.5 ADP)/(ATP + ADP + AMP)] by ~50%, suggesting the inhibition of ATPase and thus a decrease in cellular protein and fatty acid content (Chen et al., 2023, 2024) as well as a decrease in chlorophyll and organic N and C content ( Figure 3 ). Such a decrease in chlorophyll content and Chl a/Chl b ratio due to hypoxia was also observed in the seagrass Enhalus acoroides at night (Che et al., 2022).

In the dark, diatoms normally decomposed the organic substances of, e.g., sugars and fatty acids via the TCA cycle first to obtain energy for their survival, followed by precious proteins (Cheng et al., 2022; Chen et al., 2023). Thus, how low O₂ affects cell growth depends on the net accumulation of cell composition. Our findings showed that low O₂ decreases dark respiration ( Figure 3C ), simultaneously promotes cellular C accumulation ( Figure 4B ), and decreases the mortality of cells ( Figure 4A ). However, under hypoxia, the cellular fatty acid content of diatom T. pseudonana was reduced by up to ~40% (Chen et al., 2024), which may have forced the cells to obtain more energy by decomposing functional proteins (Chen et al., 2023), thus leading to a linear increase in POC/PON ratio ( Figure 4D ). Such a phenomenon of changing metabolic substrate preference in response to environmental changes was also observed in the diatom S. dohrnii (Cheng et al., 2022).

The sinking speed of T. pseudonana was estimated to be ~5.0 cm day^-1, which is comparable to other studies (Waite et al., 1997; Bannon and Campbell, 2017), and increased with duration of darkness ( Figure 5 ). Low O₂ decreased the sinking rate while hypoxia accelerated it, which can be explained by the lower mortality of cells in the former and higher mortality in the latter condition compared to the ambient O₂ condition. On the other hand, diatoms normally aggregate multi-cellularly in nature and accumulate fatty acids to decrease gravity and avoid sinking from the euphotic zone (Ge et al., 2014; Janssen et al., 2014). The increased sinking rate under hypoxia could be related to the decrease in cell activity and cell composition as found here ( Figures 2 , 4 ) or in other studies (Chen et al., 2024). Such an increase in sinking rate combined with increasing mortality under hypoxic conditions suggests that the diatom-derived carbon burial may be accelerated under the marine deoxygenation scenario in the future ( Figure 6 ). In addition, our results from individual diatom species grown in the laboratory should be taken with caution, although they provide a potentially mechanistic understanding of the effects of decreasing O₂. Further studies with a phytoplankton community or multiple species closer to field conditions need to be conducted.