Lowering pO2 Interacts with Photoperiod to Alter Physiological Performance of the Coastal Diatom Thalassiosira pseudonana

Exacerbating deoxygenation is extensively affecting marine organisms, with no exception for phytoplankton. To probe these effects, we comparably explored the growth, cell compositions, photosynthesis, and transcriptome of a diatom Thalassiosira pseudonana under a matrix of pO2 levels and Light:Dark cycles at an optimal growth light. The growth rate (μ) of T. pseudonana under a 8:16 L:D cycle was enhanced by 34% by low pO2 but reduced by 22% by hypoxia. Under a 16:8 L:D cycle, however, the μ decreased with decreasing pO2 level. The cellular Chl a content decreased with decreasing pO2 under a 8:16 L:D cycle, whereas the protein content decreased under a 16:8 L:D cycle. The prolonged photoperiod reduced the Chl a but enhanced the protein contents. The lowered pO2 reduced the maximal PSII photochemical quantum yield (FV/FM), photosynthetic oxygen evolution rate (Pn), and respiration rate (Rd) under the 8:16 or 16:8 L:D cycles. Cellular malondialdehyde (MDA) content and superoxide dismutase (SOD) activity were higher under low pO2 than ambient pO2 or hypoxia. Moreover, the prolonged photoperiod reduced the FV/FM and Pn among all three pO2 levels but enhanced the Rd, MDA, and SOD activity. Transcriptome data showed that most of 26 differentially expressed genes (DEGs) that mainly relate to photosynthesis, respiration, and metabolism were down-regulated by hypoxia, with varying expression degrees between the 8:16 and 16:8 L:D cycles. In addition, our results demonstrated that the positive or negative effect of lowering pO2 upon the growth of diatoms depends on the pO2 level and is mediated by the photoperiod.


Introduction
Anthropogenic marine eutrophication and warming are exacerbating the deoxygenation in the water column through, for example, unbalancing O 2 production and consumption, lowering O 2 solubility, and hindering the exchange with atmospheric O 2 [1,2]. Over the past five decades, the global oceans have had an estimated loss of~2% of their oxygen [3]; the degree and area of O 2 -deficiency or O 2 -limitation have extended quickly [1,4]. At present, the number of hypoxia zones (dissolved oxygen, DO < 2.5 mg L −1 ,~25% saturation) have exceeded 500 worldwide, with the hypoxic area reaching over 1,000,000 km 2 [4]. Dissolved O 2 in seawater is important in marine ecosystems as the water must contain sufficient O 2 to maintain aquatic biota. The reduced available O 2 is known to threaten the health and lives of marine aerobic organisms that dwell in water or sediments, such as fish, shrimp, and shellfish, etc. [4,5]. Apart from affecting these aerobic organisms, the lowered available O 2 also influences the photosynthetically O 2 producing organisms [6][7][8], as these O 2 producers, including phytoplankton, need surrounding O 2 to maintain mitochondrial respiration under the limited-light or dark conditions, which can supply energy

Culture Protocol and Growth Rate
A temperate marine centric diatom T. pseudonana (CCMP 1335), originally obtained from the Provasoli-Guillard National Center of Marine Phytoplankton (NCMP), was semicontinuously cultured with sterilized enriched artificial seawater (ESAW) in 500 mL conical flasks with 400 mL culture at 18 • C in an incubator (ZQZYCGF8, Shanghai, China). Light in the incubator was provided by a panel of LED (light-emitting diode) lamps and automatically turned on at 08:00 and turned off at 16:00 (8:16, Light:Dark cycle) or at 24:00 (16:8, L:D cycle). Light intensity was measured with a microspherical quantum sensor (QRT1, Hansatech, UK) in a culture flask filled with medium. Cultures were gently bubbled with the commercially prepared air (Qingdao Jinpeng Gas, China) with three pO 2 levels [21% v/v (ambient pO 2 , Amb pO 2 ), 10% v/v (Low pO 2 ), and 3% v/v (Hypoxia)], which maintained the dissolved O 2 (DO) in cultures at 310 ± 25, 150 ± 10, and 40 ± 7.5 µM, respectively. The DO concentration was monitored with an optode sensor controlled through Oxygen Logger software (OXR230, PyroScience Tech, Aachen, Germany). Before bubbling into cultures, the air streams filtered through a 0.2 µm microfilter were bubbled through another medium that was kept in the incubator, to eliminate the shock effects of varied pO 2 and temperature during the medium replacement. No notable disruption of growth of T. pseudonana was detected by this bubbling when compared to no bubbling. In total, 18 separate semicontinuous cultures were grown under the combination (three replicates for each) of 3 pO 2 levels and 2 photoperiods. Considering the seasonal photoperiod changes in temperate regions, we set the photoperiods at 8:16 and 16:8 L:D cycles, and we set the growth light at the approximately optimal intensity of 150 µmol photons m −2 s −1 , according to [30,31].
The growth of T. pseudonana was estimated using chlorophyll a (Chl a) fluorescence with excitation of 440 nm and emission at 680 nm measured with a molecular device spectrofluorometer (CYT5M, BioTech, CA, USA) every morning (10:00 a.m., 2 h after lights on), before and after the dilutions with fresh medium. The Chl a in cultures was maintained at 0.86 ± 0.10 µg mL −1 during the cultivation period, and the specific growth rate (µ, d −1 ) was calculated as: where N1 and N0 denote the OD680 value at time t1 and t0, respectively. After the cultures went through over 9 generations, duplicate 5 mL cultures were taken at 10:00 a.m. from each flask after gently shaking and fixed with glutaraldehyde to a final concentration of 1%, for cell counting with an Accuri C6 flow cytometer (Becton-Dickinson, AZ, USA). After this, the aliquots of samples were taken for determination of cell compositions, physiological performance, and transcriptome as described below.

Chlorophyll Fluorescence
To measure the maximum photochemical quantum yield (F V /F M ) of Photosystem II (PS II), 5 mL culture was taken from each flask and dark-acclimated for 15 min at growth temperature (18 • C). After this, the maximal (F M ) chlorophyll fluorescence was measured with a fluorometer (AquaPen-C 100, Photon Systems Instruments, Prague, Czech Republic) under saturating blue-light pulse (3000 µmol photons m −2 s −1 , 1 s), and the minimal fluorescence (F O ) was measured in the presence of a weak modulated measuring light. The F V /F M was calculated [39] as: Meanwhile, the relative electron transport rate (rETR) was measured 0, 10, 20, 50, 100, 300, and 500 µmol photon m −2 s −1 actinic lights, to obtain the rapid light curve (RLC) [40] as: where F M ' and Ft are maximal and instantaneous fluorescence under each of 7 actinic lights (PAR, µmol photons m −2 s −1 ). The RLC-derived photosynthetic parameters, i.e., light utilization efficiency (α), maximal rETR (rETRmax), and saturation irradiance (E K , µmol photons m −2 s −1 ) were obtained [41] as: where a, b, and c are adjusted parameters.

Photosynthetic and Respiratory Rate
To measure the photosynthetic rate and dark respiratory rate of T. pseudonana, 300 mL cultures were collected from each of the combined pO 2 and photoperiod treatments at the end of cultivation, followed by a 5-min dark-acclimation in a photosynthetic chamber where the temperature was maintained at 18 • C with a thermal cooler. Then, the increase in dissolved O 2 concentration in the chamber under growth light and the decline in the dark were tracked with the optode oxygen sensor. The photosynthetic oxygen evolution rate (Pn) and respiratory rate (Rd, fmol O 2 cell −1 min −1 ) were obtained by dividing the oxygen increase and decline rate by {cells mL −1 × measuring time}, respectively [31]. Each measurement lasted for~15 min. After measuring the photosynthetic and respiratory rate, the culture was used for the following cell composition and transcriptome measurements.

Cell Compositions
To measure cellular Chl a concentration, a 50 mL culture from each treatment was vacuum filtered onto a Whatman GF/F glass fiber filter (25 mm in diameter), extracted in 4 mL 90% acetone (v/v) saturated with magnesium carbonate overnight at 4 • C in the dark. After 10 min centrifugation (10,000× g) at 4 • C, the optical absorption of the extraction was scanned with a spectrophotometer (Shimadzu model UV 1800-PC, Kyoto, Japan). Chl a concentration (µg mL −1 ) was calculated according to [42].
To measure protein concentration, a 50 mL culture was vacuum filtered onto a GF/F glass fiber filter and extracted in 1.0 mL pre-cooling buffer (pH 8.0, 20 mM Tris, 1 mM EDTA, 10 mM MgCl 2 , 50 mM NaHCO 3 , and 5 mM β-mercaptoethanol). The cells on filters were then broken through oscillating for 20 min with grinding beads at 4 • C using a vortex mixer (G560E, Scientific Industries, New York, NY, USA). After centrifugation (10,000× g, 10 min, 4 • C), the supernatant was used to quantify the protein using a protein assay kit (A045-2, Nanjing Jiancheng Biological Engineering Co., Nanjing, China) following the manufacturer's protocol with the bicinchoninic acid (BCA) method [31,32]. The malondialdehyde (MDA) in the protein solution, a product of membrane lipid peroxidation, was determined using an assay kit (A003-1, Nanjing Jiancheng Biological Engineering Co., Nanjing, China) [32], as well as the superoxide dismutase (SOD) activity with an assay kit (A001-3, Nanjing Jiancheng Biological Engineering Co., Nanjing, China) [43] following the protocol of the manufacturer.

Transcriptome Sequencing and Analysis
At the end of cultivation, 80 mL cultures from each flask of each pO 2 and photoperiod treatment were collected for transcriptome analysis. The total RNA of T. pseudonana was extracted with Trizol (Takara Bio. Inc., Shiga, Japan), and the degradation and purity of RNA were assessed by 1% agarose gels using a NanoPhotometer ® spectrophotometer (IMPLEN, Westlake Village, CA, USA) and an Agilent Bioanalyzer 2100 system (Agilent Tech., Santa Clara, CA, USA). Sequencing libraries were generated using NEBNext ® Ultra™ RNA Library Prep Kit for Illumina ® (NEB, BioLabs Inc., Ipswich, MA, USA). The library preparations were sequenced using the Illumina HiSeq platform to produce clean reads.
Transcriptome assembly was accomplished with the protocols of Trinity [44] and Corset [45]. Gene function was annotated on the base of following databases: Nr (NCBI non-redundant protein sequences), Nt (NCBI non-redundant nucleotide sequences), Pfam (Protein family), KOG/COG (Clusters of Orthologous Groups of proteins), Swiss-Prot (A manually annotated and reviewed protein sequence database), KO (KEGG Ortholog database), and GO (Gene Ontology). The RSEM (RNA-Seq by Expectation Maximization) was used to estimate the gene transcription levels [46]. Differential transcription analysis among the different pO 2 and photoperiod treatments was performed with DEGseq package [47], and the p value was adjusted using the q-value, with the q-value of <0.005 and |log2(foldchange)| of >1 as significant threshold [48]. Metabolic pathway analysis of differentially expressed genes (DEGs) was conducted according to KEGG pathway database (http://www.genome.jp/kegg/, accessed on 10 November 2021), and functional enrichment analysis was performed using KOBAS (corrected p < 0.05) [49]. All the detailed information was supplied in supplemental materials.

Data Analysis
Data were shown as mean and standard deviations (mean ± sd). Paired t-test, oneway ANOVA with Tukey post-tests (Prism 5, Graphpad Software, San Diego, CA, USA), and comparisons of linear curve fits were used to detect the significant difference among cultures of each pO 2 and photoperiod treatment. Two-way ANOVA with Tukey post-tests were used to detect the interactions of the pO 2 level and photoperiod. The confidence level for the statistical tests was set at 0.05.

Results
The growth of T. pseudonana under the light intensity of 150 µmol photonsm −2 s −1 differed greatly among different pO 2 and photoperiod treatments ( Figure 1). The specific growth rate (µ) was 0.49 ± 0.05 d −1 at ambient pO 2 , with no significant effect of photoperiod. The µ was enhanced by~34% by low pO 2 under the 8:16 L:D cycle but decreased bỹ 22% with further decreased pO 2 . Under the 16:8 L:D cycle, however, the µ decreased from 0.53 ± 0.02 to 0.46 ± 0.03 d −1 with decreasing pO 2 . Prolonged light duration did not affect the µ at ambient pO 2 but enhanced it by~21% at low pO 2 and reduced it by~20% at the hypoxia condition. manually annotated and reviewed protein sequence database), KO (KEGG O tabase), and GO (Gene Ontology). The RSEM (RNA-Seq by Expectation M was used to estimate the gene transcription levels [46]. Differential transcrip among the different pO2 and photoperiod treatments was performed with D age [47], and the p value was adjusted using the q-value, with the q-value o |log2(foldchange)| of >1 as significant threshold [48]. Metabolic pathway an ferentially expressed genes (DEGs) was conducted according to KEGG pathw (http://www.genome.jp/kegg/, accessed on 3 December 2021), and functiona analysis was performed using KOBAS (corrected p < 0.05) [49]. All the de mation was supplied in supplemental materials.

Data Analysis
Data were shown as mean and standard deviations (mean ± sd). Paired way ANOVA with Tukey post-tests (Prism 5, Graphpad Software, San Diego and comparisons of linear curve fits were used to detect the significant differ cultures of each pO2 and photoperiod treatment. Two-way ANOVA with Tuk were used to detect the interactions of the pO2 level and photoperiod. The con for the statistical tests was set at 0.05.

Results
The growth of T. pseudonana under the light intensity of 150 μmol photo fered greatly among different pO2 and photoperiod treatments ( Figure 1). growth rate (μ) was 0.49 ± 0.05 d −1 at ambient pO2, with no significant effect iod. The μ was enhanced by ~34% by low pO2 under the 8:16 L:D cycle but d ~22% with further decreased pO2. Under the 16:8 L:D cycle, however, the μ dec 0.53 ± 0.02 to 0.46 ± 0.03 d −1 with decreasing pO2. Prolonged light duration d the μ at ambient pO2 but enhanced it by ~21% at low pO2 and reduced it by hypoxia condition. Cellular Chl a content of T. pseudonana under the 8:16 L:D cycle decreased from 0.42 ± 0.02 to 0.36 ± 0.01 pg cell −1 from ambient pO 2 to hypoxia conditions (Figure 2A). The prolonged light duration reduced the Chl a by~36% at ambient pO 2 and 23% and 29% at low pO 2 and hypoxia, respectively. The protein content (6.78 ± 0.97 pg cell −1 ) under the 8:16 L:D cycle showed no significant difference among the three pO 2 levels but decreased from 18.4 ± 0.25 to 12.3 ± 0.35 pg cell −1 from ambient pO 2 to hypoxia ( Figure 2B). The prolonged light duration enhanced the protein by~150% at ambient pO 2 and~200% and 87% at low pO 2 or hypoxia, respectively.  (Figure 2A). The prolonged light duration reduced the Chl a by ~36% at ambient pO2 and 23% and 29% at low pO2 and hypoxia, respectively. The protein content (6.78 ± 0.97 pg cell −1 ) under the 8:16 L:D cycle showed no significant difference among the three pO2 levels but decreased from 18.4 ± 0.25 to 12.3 ± 0.35 pg cell −1 from ambient pO2 to hypoxia ( Figure 2B). The prolonged light duration enhanced the protein by ~150% at ambient pO2 and ~200% and ~87% at low pO2 or hypoxia, respectively.  Figure 3A). Under the 16:8 L:D cycle, however, the FV/FM decreased from 0.52 ± 0.02 to 0.42 ± 0.03 from ambient pO2 to hypoxia. The prolonged photoperiod reduced the FV/FM by ~19% at ambient pO2 and by ~33% at low pO2 or hypoxia. In parallel, the RLC-derived light utilization efficiency (α) and maximal rETR (rETRmax) exhibited no significant difference among all three pO2 levels under both L:D 8:16 and 16:8 cycles; while the saturation irradiance (EK) was enhanced by low pO2 at the L:D 8:16 cycle ( Table 1). The prolonged photoperiod significantly reduced the α, EK, and rETRmax among all pO2 treatments (p < 0.05) ( Table 1).  Figure 3A). Under the 16:8 L:D cycle, however, the F V /F M decreased from 0.52 ± 0.02 to 0.42 ± 0.03 from ambient pO 2 to hypoxia. The prolonged photoperiod reduced the F V /F M by~19% at ambient pO 2 and by~33% at low pO 2 or hypoxia. In parallel, the RLC-derived light utilization efficiency (α) and maximal rETR (rETRmax) exhibited no significant difference among all three pO 2 levels under both L:D 8:16 and 16:8 cycles; while the saturation irradiance (E K ) was enhanced by low pO 2 at the L:D 8:16 cycle ( Table 1). The prolonged photoperiod significantly reduced the α, E K , and rETRmax among all pO 2 treatments (p < 0.05) ( Table 1) Figure 3C). The prolonged photoperiod significantly reduced the photosynthetic rate (p < 0.01) but enhanced the respiratory rate (p < 0.01) among all three pO 2 treatments.
The photosynthetic oxygen evolution rate (Pn) under the 8:16 L:D cycle decreased from 2.72 ± 0.06 to 2.49 ± 0.05 fmol O2 cell −1 min −1 from ambient pO2 to hypoxia, while such a decreasing trend did not occur under the 16:8 L:D cycle ( Figure 3B). The dark respiration rate (Rd) decreased from 0.71 ± 0.19 to 0.30 ± 0.07 fmol O2 cell −1 min −1 with decreasing pO2 under the 8:16 L:D cycle and from 1.16 ± 0.24 to 0.55 ± 0.08 fmol O2 cell −1 min −1 under the 16:8 L:D cycle ( Figure 3C). The prolonged photoperiod significantly reduced the photosynthetic rate (p < 0.01) but enhanced the respiratory rate (p < 0.01) among all three pO2 treatments.   The cellular MDA content was 0.21 ± 0.01 fmol cell −1 at ambient pO 2 under 8:16 L:D cycle,~25% lower than that at low pO 2 but similar to hypoxia ( Figure 4A). The prolonged photoperiod enhanced the MDA by~71%,~95%, and~56% at ambient pO 2 , low pO 2 , and hypoxia, respectively. The SOD activity was (0.60 ± 0.08) × 10 −6 U cell −1 under 8:16 L:D cycle, with no significant effect of pO 2 levels ( Figure 4B). The prolonged photoperiod did not affect the SOD activity at ambient pO 2 but enhanced it by~50% at low pO 2 and reduced it by~18% at hypoxia. Note: Numbers indicate the mean and standard deviation from measurements of 3 independent cultures (n = 3); and different letters at the top-right of the numbers indicate significant differences (p < 0.05, one-way ANOVA).
The cellular MDA content was 0.21 ± 0.01 fmol cell −1 at ambient pO2 under 8:16 L:D cycle, ~25% lower than that at low pO2 but similar to hypoxia ( Figure 4A). The prolonged photoperiod enhanced the MDA by ~71%, ~95%, and ~56% at ambient pO2, low pO2, and hypoxia, respectively. The SOD activity was (0.60 ± 0.08) × 10 −6 U cell −1 under 8:16 L:D cycle, with no significant effect of pO2 levels ( Figure 4B). The prolonged photoperiod did not affect the SOD activity at ambient pO2 but enhanced it by ~50% at low pO2 and reduced it by ~18% at hypoxia. Transcriptome analysis showed that a total of 26 differentially expressed genes (DEGs) (|log2(foldchange)| > 8) that relate to photosynthesis, respiration, and metabolism were more sensitive to varying pO2 levels and photoperiods, based on the KEGG pathway enrichment (Table S1). Most of these identified DEGs were downregulated by lowered pO2 or hypoxia to varying degrees under the 8:16 and 16:8 L:D cycles, as compared to ambient pO2 ( Figure 5). Of these DEGs, 11 and 7 transcripts were, respectively, related to the functions of metabolism (citrate cycle and fatty acid metabolism) and oxidation (oxidative phosphorylation), and 1 transcript was related to photosynthesis (light reaction). The transcriptional pattern seemed to be more sensitive to low pO2 in the citrate cycle, Transcriptome analysis showed that a total of 26 differentially expressed genes (DEGs) (|log2(foldchange)| > 8) that relate to photosynthesis, respiration, and metabolism were more sensitive to varying pO 2 levels and photoperiods, based on the KEGG pathway enrichment (Table S1). Most of these identified DEGs were downregulated by lowered pO 2 or hypoxia to varying degrees under the 8:16 and 16:8 L:D cycles, as compared to ambient pO 2 ( Figure 5). Of these DEGs, 11 and 7 transcripts were, respectively, related to the functions of metabolism (citrate cycle and fatty acid metabolism) and oxidation (oxidative phosphorylation), and 1 transcript was related to photosynthesis (light reaction). The transcriptional pattern seemed to be more sensitive to low pO 2 in the citrate cycle, with 7 key nodal enzymes in four comparable strategies being downregulated (6.7-fold [avg.]) ( Figure 6). Under the 16:8 L:D cycle, 7 DEGs of oxidative phosphorylation were downregulated by 6.3-fold [avg.] by low pO 2 and by 4.5-fold [avg.] by hypoxia. Moreover, under the 8:16 L:D cycle, the expression of the ATP-targeted enzyme gene ATP1 was upregulated by 9.6-fold and 8.1-fold by low pO 2 or hypoxia, respectively, as compared to ambient pO 2 ( Figure S1, Table S1). The prolonged photoperiod mainly downregulated the expression of the photosynthesis-related genes, as well as the DEGs of the citrate cycle ( Figure S1, Table S2). The downregulation of the RubisCO gene expression (1.47-fold [avg.]) by prolonged photoperiod was similar among all three pO 2 treatments but different for that of C 4 metabolism.
with 7 key nodal enzymes in four comparable strategies being downregulated (6.7-fold [avg.]) ( Figure 6). Under the 16:8 L:D cycle, 7 DEGs of oxidative phosphorylation were downregulated by 6.3-fold [avg.] by low pO2 and by 4.5-fold [avg.] by hypoxia. Moreover, under the 8:16 L:D cycle, the expression of the ATP-targeted enzyme gene ATP1 was upregulated by 9.6-fold and 8.1-fold by low pO2 or hypoxia, respectively, as compared to ambient pO2 ( Figure S1, Table S1). The prolonged photoperiod mainly downregulated the expression of the photosynthesis-related genes, as well as the DEGs of the citrate cycle ( Figure S1, Table S2). The downregulation of the RubisCO gene expression (1.47-fold [avg.]) by prolonged photoperiod was similar among all three pO2 treatments but different for that of C4 metabolism.  Table S1.  Table S1.
Microorganisms 2021, 9, x FOR PEER REVIEW 9 of 13 with 7 key nodal enzymes in four comparable strategies being downregulated (6.7-fold [avg.]) ( Figure 6). Under the 16:8 L:D cycle, 7 DEGs of oxidative phosphorylation were downregulated by 6.3-fold [avg.] by low pO2 and by 4.5-fold [avg.] by hypoxia. Moreover, under the 8:16 L:D cycle, the expression of the ATP-targeted enzyme gene ATP1 was upregulated by 9.6-fold and 8.1-fold by low pO2 or hypoxia, respectively, as compared to ambient pO2 ( Figure S1, Table S1). The prolonged photoperiod mainly downregulated the expression of the photosynthesis-related genes, as well as the DEGs of the citrate cycle ( Figure S1, Table S2). The downregulation of the RubisCO gene expression (1.47-fold [avg.]) by prolonged photoperiod was similar among all three pO2 treatments but different for that of C4 metabolism.  Table S1.  Table S1.

Discussion
Most previous studies reported the individual effect upon phytoplankton physiology by lowering pO 2 levels [6,7,18] or varying photoperiods [30][31][32][33][34]. In this study, we showed the coupling effects of lowered pO 2 and photoperiods on the physiological performance of the diatom T. pseudonana. Lowering pO 2 interacted with longer photoperiod to reduce the light-harvesting pigments and photosynthetic rate through downregulating the genes related to photosynthesis and metabolism and, consequently, reduced the growth of T. pseudonana (F (2,12) = 6.46, p < 0.05). Moreover, at lowered pO 2 , the reduction in respiration and photosynthesis occurred under both the 8:16 and 16:8 L:D cycles, as compared to ambient pO 2 ; while the enhancement of the growth rate was present under the former but reduced under the latter L:D cycle, which indicated the alteration of light duration on the balance of the lowered pO 2 -induced savings of consumption and the reduction in photosynthetic production.
Longer light duration reduced the cellular pigment content of T. pseudonana (Figure 2A). Similar to previous studies [30,31], the light intensity of 150 µmol photons m −2 s −1 was expected to saturate the cell growth under a shorter photoperiod and may have oversaturated under the longer photoperiod, as indicated by an insignificant difference of the µ between the 8:16 and 16:8 L:D cycles (Figure 1). It is easily understood that phytoplankton cells usually lower the synthesis and accumulation of Chl a in oversaturated light, to lessen the excessive energy harvesting [43] and thus to alleviate the excessive light energy-caused photodamage or photoinhibition [30,50]. Moreover, the decreased pO 2 interacted with the prolonged photoperiod to aggravate the decrease in Chl a (F (2,12) = 7.53, p < 0.01), as also supported by the transcriptome results ( Figure 6). This can be explained by the fact that the decreased pO 2 has pulled the RubisCO-catalyzed biochemical reaction towards the photosynthetic process [13] thus saving the energy derived from the light-harvesting complex. Such energy savings may feedback to adaptively cause cells to lower the capacity for light harvesting through lowering the Chl a containing photosystems. If considering the function of CCM for concentrating CO 2 in diatoms [10][11][12], however, the positive effect of reduced pO 2 may be tempered as indicated by the reduction in growth in the diatom T. pseudonana at hypoxia condition ( Figure 1) or in Skeletonema costatum [6] and C. reinhardtii [21]. Furthermore, an increase in cellular proteins occurred under the combined hypoxia and longer photoperiod status ( Figure 2B), which may help to further save energy; more cellular proteins usually indicate more enzymes and, thus, more activity of biochemical reactions within cells [10,13,51].
The longer photoperiod suppressed the photosynthetic efficiency of T. pseudonana but enhanced the respiration (Figure 3), indicating the growth light here was oversaturated under the 18:6 photoperiod. Such an adverse effect was aggravated by the extremely low pO 2 (F V /F M , F (2,12) = 4.14, p < 0.05; and Pn, F (2,12) = 3.65, p < 0.05), consistent with more products of membrane lipid peroxidation, i.e., higher cellular MDA content at hypoxia status ( Figure 4A). Under such an adverse circumstance, phytoplankton cells often adaptively promote the ROS-scavenged ability, including enhancing cellular SOD activity to protect themselves against oxidative damage [52]. Our findings supported this with higher SOD activity at low pO 2 ( Figure 4B); at hypoxia, however, the SOD activity was depressed, suggesting the cellular O 2 level was too low to maintain the ROS-scavenging capacity, thus, leading to the enhancement of MDA content, although the lowered pO 2 circumstance theoretically goes against the ROS production and MDA accumulation [53]. Supporting the decrease in photosynthetic capacity, the transcriptome results indicated the downregulation of most DEGs, including those related to the photosynthetic process ( Figure 6). Compared to photosynthesis, the lowered pO 2 reduced the respiration even more, leading to an even lower respiration rate at hypoxia ( Figure 3C), supported by the severe downregulation of DEGs related to the key enzymes in the citric acid cycle (Figures 5 and 6). Such downregulation also occurred in Chlorella vulgaris that had overcome the metabolic constraint through lowering respiration and allocating more fixed C to maintain growth [54], as well as in T. pseudonana at a low pO 2 state (Figures 1 and 3C). Mechanically, the dissolved O 2 most likely affects the ATP synthesis in mitochondria [55] and RubisCO-catalyzed photorespiration in plastid [56]. It is the case in this study, as the decreased pO 2 downregulated a total of 25 genes that relate to ATP synthesis (Figure 6), which may have lowered the biochemical activities within cells through inhibiting, e.g., PPDK [57] and, thus, reduced the respiration rate ( Figure 3C). Moreover, the expression of C 4 genes such as PEPC, PPDK, and PEPCK, showed no linear correlation with decreasing pO 2 ( Figure S1), consistent with previous results [11,12]. Finally, the expressions of GAP, GDC, and GLY genes that relate to photorespiration were also downregulated by lowering pO 2 (Figure 6), as well as the decreased photosynthetic efficiency, the mechanical contradiction of which needs to be studied further.

Conclusions
In this study, we found the hypoxia and longer light period interactively reduced the growth of diatom T. pseudonana through reducing the photosynthetic capacity by downregulating the related genes' expression. Growing under lowered pO 2 condition, both cellular Chl a and protein contents were lower, but the MDA content and SOD activity were higher and the lowered pO 2 -induced effect was mediated by the light duration. Moreover, our results demonstrated that whether there is a reduction in respiration by the lowered pO 2 over that of photosynthesis or not determines the positive and negative effects of lowering pO 2 on the growth of diatoms, which depends on the pO 2 level and is mediated by the photoperiod.

Data Availability Statement:
The data presented in this study are available within the article.