Transcriptome landscape of Synechococcus elongatus PCC 7942 for nitrogen starvation responses using RNA-seq

The development of high-throughput technology using RNA-seq has allowed understanding of cellular mechanisms and regulations of bacterial transcription. In addition, transcriptome analysis with RNA-seq has been used to accelerate strain improvement through systems metabolic engineering. Synechococcus elongatus PCC 7942, a photosynthetic bacterium, has remarkable potential for biochemical and biofuel production due to photoautotrophic cell growth and direct CO2 conversion. Here, we performed a transcriptome analysis of S. elongatus PCC 7942 using RNA-seq to understand the changes of cellular metabolism and regulation for nitrogen starvation responses. As a result, differentially expressed genes (DEGs) were identified and functionally categorized. With mapping onto metabolic pathways, we probed transcriptional perturbation and regulation of carbon and nitrogen metabolisms relating to nitrogen starvation responses. Experimental evidence such as chlorophyll a and phycobilisome content and the measurement of CO2 uptake rate validated the transcriptome analysis. The analysis suggests that S. elongatus PCC 7942 reacts to nitrogen starvation by not only rearranging the cellular transport capacity involved in carbon and nitrogen assimilation pathways but also by reducing protein synthesis and photosynthesis activities.

Global concerns about energy security and environmental issues affecting climate changes have focused attention on engineering photosynthetic organisms that are able to sequester and convert CO 2 to organic materials using solar energy 1,2 . In order to improve the strains, system-wide engineering and optimization is required to reduce the time-, cost-, and labor-intensive process of strain development 3 . In particular, systems biology including omics analysis has been integrated to strain development.
Nitrogen is an essential element of all the complex macromolecules in a bacterial cell. Anthropogenic nitrogen supply in the presence of CO 2 is one of the critical factors in the occurrence of cyanobacterial blooms mainly caused by Microcystis spp. 4 . On the other hand, a nitrogen limitation strategy is commonly used to maximize production of the desired product such as triacylglycerol content in microalgae in the carbon partitioning mechanisms 5 and to trigger polyhdroxyalkanoate (PHA) accumulation as carbon storage in Ralstonia eutropha 6 and in engineered Escherichia coli 7 . In addition, improvement of isoprenoid production (astaxanthinin Haematococcus pluvialis 8,9 and amorpha-4,11-diene in E. coli 10 ) has been reported under nitrogen limitation. Thus, it is important to understand the positive and negative mechanisms of nitrogen regulation in a cell for system-wide strain development.
Systems biology-based analyses have revealed the cellular mechanisms for nitrogen regulation in industrially-relevant bacteria. Nitrogen assimilation and amino acid biosynthesis of Corynebacterium rubrerythrin (a non-heme iron protein) that has been related to oxidative stress protection 26 and a gene encoding a hypothetical protein.
In order to find functional categories of DEGs, we performed a GO enrichment analysis over biological processes, molecular functions, and cellular locations. As shown in Fig. 2, many cellular transporter activities were up-regulated for nitrogen starvation responses, i.e. ion, nitrate, and nitrogen compound transport, whereas photosynthesis and protein translation process were down-regulated.
Transcriptomic analysis in specific metabolisms of S. elongatus PCC 7942 for nitrogen starvation response. Stringent responses of nitrogen starvation have been shown in the cellular transporters to adapt environmental changes 12 . Here we described the transcriptional outlook of specific metabolisms over nitrogen starvation responses in S. elongatus PCC 7942 with RNA-seq analysis ( Fig. 3 and Supplementary Table S1).
Nitrogen assimilation. For ammonium uptake, three cyanobacterial ammonium transporters (Amt1, AmtB, and Amt1-like) were annotated in S. elongatus PCC 7942. The expression of gene encoding NtcA-regulated AmtB (encoded by locus Synpcc7942_2279) is responsible for nitrogen limitations, which was up-regulated by 1.7 of logFC after 24 h of nitrogen starvation. On the other hand, transcriptional levels of genes encoding for Amt1 (encoded by locus Synpcc7942_0442) and Amt1-like (encoded by locus Synpcc7942_0965) were not differently altered. Besides ammonium, nitrate and urea are commonly used by cyanobacteria as a nitrogen source. An ABC-type transporter constituted by NrtABCD (encoded by locus Synpcc7942_1239-1236), which is involved in nitrate-nitrite uptake 27 , were up-regulated (1.1 of logFC). Also, a NtcA-regulated ferredoxin-nitrite reductase NirA (encoded by locus Synpcc7942_1240) was up-regulated (1.9 of logFC) to the nitrogen starvation response. However, a ferredoxin-nitrate reductase NarB (encoded by locus Synpcc7942_1235) was not up-regulated. Genes encoding for urease and urea transporters meanwhile have not identified in S. elongatus PCC 7942 and their orthologous genes predicted by the orthologs of Synechocystis sp. PCC 6803 were not differently expressed upon nitrogen starvation. Interestingly, an ABC-type cyanate transporter and a cyanate lyase (encoded by the cynDBAS genes; Synpcc7942_2107-2104) were up-regulated (1.1 of logFC). Thus, nitrogen-starved cyanobacterial cells transport cyanate and irreversibly catalyze bicarbonate-dependent conversion of cyanate to ammonium 28 .
Further assimilation of ammonia into glutamine and glutamate, the GS/GOGAT pathway was used to subsequently distribute nitrogen to other cellular components in S. elongatus PCC 7942. Expressions of the glnN (Synpcc7942_0169) gene encoding for glutamine synthase (Type III GS) were differentially up-regulated by 1.5 of logFC due to transcriptional activation by NtcA. However, expressions of the glnA (Synpcc7942_2156) gene encoding for the type I GS and the glsF (Synpcc7942_0890) gene encoding for a ferredoxin-dependent glutamate synthase (GOGAT; Glutamine oxoglutarate aminotransferase) were not differentially up-regulated by 0.9 and 0.4 of logFC, respectively. Nitrogen control is regulated by a transcriptional regulator NtcA (encoded by locus Synpcc7942_0965). As a RNA-seq result, NtcA was positively auto-regulated (1.4 of logFC) to the nitrogen starvation response and regulated transcriptional expressions of the amtB, nirA, glnN, and cynDBAS genes by activating NtcA-promoters. Another regulon of NtcA is the glnA (Synpcc7942_2156) and glnB (Synpcc7942_0321) genes encoding for glutamine synthase (GS) and P II protein, respectively 27 . However, RNA-seq showed that the amt1, glnA, and glnB genes were not transcriptionally activated after 24 h of nitrogen starvation. Gene expression of NtcB (LysR family of transcriptional regulator), encoded by locus Synpcc7942_1242) and PipX (encoded by Synpcc7942_2061), a co-activator of NtcA, was not differentially changed.
Scientific RepoRts | 6:30584 | DOI: 10.1038/srep30584 regardless of inorganic carbon conditions. Besides bicarbonate transporters, the multiple NADH:ubiquinone oxidoreductase complex is essential to the CO 2 uptake system in cyanobacteria 30 . In S. elongatus PCC 7942, a small NADH:ubiquinone oxidoreductase (NDH-1S) encoded by the ndhF3-ndhD3-cupAS (Synpcc7942_2091-2094) gene clusters plays a CO 2 uptake system, which were not down-regulated in nitrogen starvation (N−) compared with nitrogen (N+ ), although the genes were down-regulated in nitrogen starvation (N−) compared with a nitrogen starvation control (N− Ctr ) (see the details in Supplementary Table S1). Furthermore, genes encoding for NDH-1L and NDH-1M were not altered.
Carbon assimilation pathway and central metabolisms. Nitrogen starvation in S. elongatus PCC 7942 did not alter significant expression of genes encoding for β -carboxysome 31 , which acts as a microcompartment to supply CO 2 to Rubisco along with carboxysomal CA activity. No particular changes were shown for expression of carbonic anhydrase encoded by ccaA (Synpcc7942_1447), ecaA (Synpcc7942_1388), and ccmM (Synpcc7942_1423). The carbon fixing enzyme of the Calvin-Basham-Benson (CBB) cycle, Rubisco encoded by the cbbLS (Synpcc7942_1426-1427) genes, is crucial for carbon assimilation in cyanobacteria. Expression of the cbbLS genes was not differentially expressed for the nitrogen starvation. Unlike CbbLS, expression of the prk (Synpcc7942_0977) encoding for phosphoribulokinase and fbpI (Synpcc7942_0505) encoding for the type I fructose-1,6-bisphosphatase in the CBB pathway was down-regulated in both nitrogen starvation (N−) compared with nitrogen (N+ ) and nitrogen starvation (N−) compared with a nitrogen starvation control (N− Ctr ) by logFC ≤ −1. However, expressions of the remaining genes in the CBB pathway were not differentially changed.
In the glycolytic pathway, the pentose phosphate pathway, and the TCA cycle, no differentially expressed genes were found except genes encoding enzymes that involved a reaction with G3P as either a substrate or a product such as Gap3 (encoded by locus Synpcc7942_1939), Eda (encoded by locus Synpcc7942_0017), or Tal (encoded by locus Synpcc7942_2297). Furthermore, expressions of genes in glycogen biosynthesis were not changed.
Secondary metabolic pathways. S. elongatus PCC 7942 possesses the methylerythritol phosphate (MEP) pathway, which utilizes glyceraldehyde-3-phosphate and pyruvate to produce IPP and DMAPP for terpenoids biosynthesis (i.e. geranylgernaly diphosphate; GGPP). The terpenoid pathway in cyanobacteria is linked to chlorophyll a biosynthesis 32 . Geranylgernaly reductase converts GGPP to phytyl diphosphate (Pdh), which is used as a substrate for chlorophyll synthase (ChlG encoded by locus Synpcc7942_2084) to produce a chlorophyll a for photosynthesis. Another co-substrate for ChlG is a chlorophyllide a, which is produced from L-glutamate via both chlorophyllide a biosynthesis and tetrapyrrole biosynthesis.
As a result of RNA-seq for nitrogen starvation response, expressions of most genes involved in the MEP pathway were not changed except the dxs (Synpcc7942_0430) encoding for deoxyxyulose-5-phosphate synthase, which is the first enzyme of the MEP pathway. Dxs was down-regulated (logFC ≤ −1) in S. elongatus PCC 7942 for the nitrogen starvation condition. On the other hand, genes in involved in chlorophyllide a and tetrapyrrole biosynthesis were significantly changed by logFC ≤ −1.
Photosynthetic and respiratory electron transport pathways. Cyanobacteria utilize redox-active components in thylakoids for both photosynthesis and respiration 33 . For oxygenic photosynthesis, both photosystem I and II are required. In photosystem II (PSII), allophycocyanin (APC) and phycocyanin (PC) are a light-harvesting pigment-protein complex of the phycobiliproteins of cyanobacteria. Type-I NADPH dehydrogenase oxides NADPH and reduces plastoquinone (PQ) with PS II. Electrons are transported from the PQ pool to the cytochrome b 6 f complex. Soluble electron carriers such as plastocyanin or cytochrome c reduce the oxidized photosystem I (PSI) and a single electron is transferred to NADP. Reduced NADP can be used for CO 2 fixation. A proton gradient from the electron transport pathway is used for ATP synthesis.
RNA-seq for the nitrogen starvation response revealed that genes encoding for PSII and its phycobilisome in S. elongatus PCC 7942 were differentially expressed by logFC ≤ −4.2. Also, several genes encoding for F 0 F 1 ATPase were altered by logFC ≤ −1. Genes encoding for NDH-1L and NDH-1M complexes, cytochrome b 6 f complex, PS I, and cytochrome c oxidase were not significantly altered under nitrogen starvation. Interestingly, NblA, a phycobilisome degradation protein encoded by the nblA (Synpcc7942_0430), was significantly up-regulated by 1.96 of logFC, resulting in chlorosis in S. elongatus PCC 7942 under nitrogen starvation.
Biological evidences for down-regulated photosynthesis and up-regulated bicarbonate transporter of S. elongatus PCC 7942 for nitrogen starvation response. Based on the transcriptomics analysis, two distinct features were shown for nitrogen starvation responses in S. elongatus PCC 7942. First, genes encoding for APC and PC and its biosynthesis pathway (chlorophyll a biosynthesis) were significantly down-regulated. On the other hand, genes encoding for bicarbonate uptake transporters were up-regulated for nitrogen starvation. Thus, we performed biological experiments to support the findings from the RNA-seq analysis.
To confirm whether the levels of Chlorophyll a (Chl a) and phycobiliproteins (APC and PC) were decreased under nitrogen starvation, we determined the levels of Chl a and PC content in S. elongatus PCC 7942 using a spectrophotometer. As a result, both levels of Chl a and PC content of S. elongatus PCC 7942 were gradually decreased after nitrogen starvation stress was induced (Fig. 4), consistent with previous results 25 . Eventually, chlorosis of the cyanobacterial cells occurred. Subsequently, we investigated whether the CO 2 fixation rate was changed in S. elongatus PCC 7942 under a nitrogen starvation condition. Thus, we performed the measurement in a controlled photobioreactor supplied with bubbled air containing 5% CO 2 (Fig. 5). The results showed that the CO 2 fixation rates for the nitrogen starvation condition were statistically higher than for the nitrogen repletion Scientific RepoRts | 6:30584 | DOI: 10.1038/srep30584 condition (p-value < 0.01), which corresponds to the experimental conditions for the transcriptomics. However, the difference in the CO 2 fixation rates diminished over the cell culture.

Discussion
Although there are many questions with respect to the regulation of nitrogen control, there is evidence that regulation of the expression of genes involved in nitrogen assimilation in cyanobacteria involves interactions of P II -protein X (PipX) with the global regulator NtcA and the specific DNA binding of NtcA to target sites 20,21 . Consistently, the Ntc target genes such as the ntcA, amtB, nirA, glnN, and cynDBAS genes were up-regulated for nitrogen starvation conditions in our results. However, some target genes listed 12 were not significantly regulated. The reason remains unclear, although they conserved the DNA binding motif of NtcA. Our RNA-seq analysis did not reveal that the nitrogen starvation response is regulated by PipX alone in an NtcA-independent manner. Thus, a high resolution ChIP-seq analysis for NtcA or PipX could be useful to provide accurate information on transcriptional regulation. Nonetheless, up-regulation of an AmtB transporter, ferredoxin-nitrite reductase, and Type II glutamine synthase (GlnN) in the nitrate assimilation system must be critical for S. elongatus PCC 7942 to respond to nitrogen starvation for 24 h.  Moreover, regulation of NtcA is also involved in ammonium assimilation into carbon metabolism 27 . Nitrogen starvation is perceived as an increase in the concentration of intracellular 2-oxoglutrate, which acts as a signal molecule for intracellular sensor P II protein. Accumulation of the levels of 2-oxoglutarate (2-OG) modulates the phosphorylation status of P II protein 34,35 to interact with PipX for further transcriptional regulation with NtcA. Besides serving as a substrate for the GS/GOGAT pathway, 2-OG is a product or substrate in the 2-OG decarboxylating TCA cycle. When expressions of metabolic enzymes that consume or produce 2-OG were investigated, expressions of genes encoding for isocitrate dehydrogenase (Synpcc7942_1719) and 2-OG decarboxylase (Synpcc7942_1435) were not changed. Thus, increased levels of 2-OG must be due to high expression of the GS/GOGAT activity in nitrogen starved S. elongatus PCC 7942, which was also confirmed by our RNA-seq analysis. Moreover, NtcA regulates OmpR-type regulator Rre37, which is involved in nitrogen-expression of sugar catabolic genes in Synechocystis sp. 6803 36 . Homology of Rre37 in S. elongatus PCC 7942 is annotated in a transcriptional regulator but it has not been characterized yet.
Furthermore, protein P II co-regulates inorganic carbon uptake in cyanobacteria 37,38 . Given that phosphorylated P II is related to high CO 2 concentration (5% CO 2 ) and low nitrogen availability, the mutant ΔglnB (encoding for P II of Synechocystis sp. 6803) showed high affinity to high CO 2 concentration. In addition, decreased phycocyanin content and increased glycogen content have been shown in the mutant ΔglnB 39 . However, the effect is limited on the metabolome and the inorganic carbon uptake in S. elongatus PCC 7942 40 . Similarly, expression of the glnB genes in this study was not significantly changed for the nitrogen starvation condition. Thus, the role of P II signaling protein is important to decipher carbon and nitrogen metabolisms with the status of phosphorylation, which could be useful for metabolic engineering of cyanobacteria.
The global regulatory model for carbon and nitrogen assimilation involves the levels of 2-OG in expression of the high affinity carbon concentrating mechanism. It has been reported that CcmR of Synechocystis sp. 6803 negatively regulates expression of genes for SbtA and NDH-1S transporters in the presence of co-repressors (NADP+ and 2-OG) 41 . Interestingly, our RNA-seq study showed that gene expressions for carbon uptake transport (BicA, SbtA, and BCT1) were up-regulated for the nitrogen starvation condition in S. elongatus PCC 7942. CmpR of S. elongatus PCC 7942 has been shown to be a transcriptional activator for the cmp operon encoding for a BCT1 transporter in the presence of 2-phosphoglycolate as a co-activator 42 . Up-regulation of SbtA under nitrogen starvation remains unclear because CcmR could be missing in S. elongatus PCC 7942. These up-regulations of bicarbonate transporters under nitrogen starvation led us to investigate the CO 2 uptake rate, and thereupon a significant increase of the CO 2 uptake rate was measured under a nitrogen starvation condition in a controlled photobioreactor (Fig. 5). Interestingly, genes of a distinct cyanate transporter of S. elongatus PCC 7942 were highly up-regulated, which might also be involved in an increase of the CO 2 uptake rate by catalyzing a bicarbonate-dependent conversion of cyanate to supply ammonium. Thus, an increased CO 2 uptake rate could be an initial nitrogen starvation response.
The stringent response in photosynthesis is regulated with C and N metabolism in S. elongatus PCC 7942 under a nitrogen starvation condition. Previously, accumulation of intracellular glycogen has been observed in nitrogen-starved cyanobacteria, in contrast with decreased levels of total lipid and total protein contents. However, our experiment did not show the DEGs for glycogen biosynthesis. However, genes coding for Gap3, Eda, and Tal (G3P consuming or producing enzymes) were highly up-regulated, which might be related to accumulation of glycogen under nitrogen starvation. Thus, a comprehensive gene expression analysis could be required to understand the genetic regulation of accumulation of glycogen under nitrogen starvation responses in cyanobacteria. We speculated that a signal protein P II -PipX complex could be related to the genetic regulation of biosynthesis of glycogen under the nitrogen limitation stress responses in cyanobacteria, which is currently being investigated. Our study also confirmed that expression of the nblA gene was positively regulated by NtcA under nitrogen starvation. Phycobilisome as 30% of total proteins in a cell 18 can be degraded to supply an alternative nitrogen source by the increased activity of NblA. Its degradation process has been shown to be related to acclimation and long-term survival for cyanobacterial cells 14,16 . In addition, a decrease of the first enzyme (Dxs) of the MEP pathway under nitrogen starvation is related to shutting down biosynthetic metabolism for photosynthesis as a stringent response.

Conclusion
A transcriptome analysis of S. elongatus PCC 7942 under a nitrogen starvation condition was provided in the point of the global regulatory network of carbon and nitrogen assimilation. Genes involved in carbon and nitrogen assimilation were highly up-regulated. Photosynthesis and its related metabolic genes were significantly down-regulated. Detailed information of RNA-seq for nitrogen starvation in S. elongatus PCC 7942 could be useful for metabolic engineering to accelerate the development of biosolar cell factories to re-direct carbon flux of CO 2 to desired products under a nitrogen limiting condition.

Methods
Cyanobacterial strain and growth conditions. S. elongatus PCC 7942 strain was grown at 30 °C while shaking under constant illumination light (100 μ mol photons m −2 s −1 ) in BG-11 medium as described previously 2 . For nitrogen starvation experiments, cyanobacterial cells were grown in 50 ml BG-11 for 24 h as a control (N+) and the cells at OD 730 of 1.0 were harvested by centrifugation at 13 000 rpm for 5 min. Then, cells pellets were re-suspended either in fresh BG-11 medium (N+) or nitrogen-free BG-11 medium (N−) after washing twice (refer 'arrow 1' in Fig. 1a). The cells were cultivated for 24 h and the biological duplicated samples were prepared for RNA-seq (refer 'arrow 2' in Fig. 1a) and stored at − 80 °C until they were analyzed.
RNA preparation and RNA-seq. 5 mL of Trizol reagent (California, Carlsbad, USA) was added to the cyanobacterial cell pellet from 50 mL culture under liquid nitrogen. To break the cell wall, the quick frozen cell was grinded using a mortar and pestle and aliquot into 1 mL tube. After addition 200 μ l of 1-bromo-3-chloropropane in a 1 mL tube and incubation for 3 min at room temperature, the extraction mixture was centrifuged at 13 000 rpm for 15 min 4 °C. Retrieved aqueous phase was mixed with the equal volume of isopropanol and incubated for 10 min at room temperature and centrifuged to concentrate the precipitated RNA at 4 °C. The RNA pellet was washed using 75% (v/v) ethanol (1 mL), air dried and finally dissolved in 30 μ L RNase-free water.
The quantity and quality of total RNA were evaluated using RNA electropherograms (Agilent 2100 Bioanalyzer, Agilent Technologies, Santa Clara, USA). Total RNA (10 μ g) from each sample was used as a starting material to prepare sequencing libraries. The Ribo-Zero rRNA removal kit (Epicentre, USA) was used for ribosomal RNA depletion according to manufacturer instructions. Libraries for Illumina sequencing were made with the TruSeq Stranded mRNA sample prep kit (Illumina, USA) following the manufacturer's protocol. RNA sequencing was performed on the Illumina HiSeq 2500 platform using single-end 50 bp sequencing. All RNA-sequencing and alignment procedures were conducted by ChunLab (Seoul, South Korea). All data sets have been uploaded to the Gene Expression Omnibus under accession (GSE79726) and all sequencing reads were submitted to NCBI Sequence Read Archive (SRP072154).

Data analysis and Statistical analysis.
The quality of raw reads was checked by Fast QC and the reads with low quality (Q < 30) were eliminated using HTQC 17 . Bowtie2 17 software was used to align the reads with the S. elongatus PCC 7942 reference genome. The relative transcript abundance was quantified as raw read counts and reads per kilobase per million mapped reads (RPKM). edgeR 43 , a Bioconductor components in R packages was used for differential gene expression analysis. Cyanobacterial gene ontology analysis was performed by assigning COG (Clusters of Orthologous Groups) categories with KEGG (Kyoto Encyclopedia of Genes and Genomes) database. In addition, cyanobacterial clusters of orthologous groups of proteins (CyoOG) 44 categories to each gene of S. elongatus PCC 7942 were assigned (Supplementary Table S1). Heat maps generated by MeV (MutiExperiment Viewer ver. 4.8) showed differential expressed genes for nitrogen starvation responses in Fig. 3. For gene ontology enrichment analysis, a GO term annotation file obtained by IPR2GO (http://genome. microbedb.jp/CyanoBase) was used to perform an hypergeometic test with unsupported model organisms using the GOstat 45 of R package. Then, a custom data structure of the GO ontologies was created.
Measurement of Chlorophyll a and phycocyanin contents using spectrophotometer. For the measurement of Chlorpholly a (Chl a) and phycocyanin (PC) contents, cyanobacterial cell cultures (1 mL) were harvested at the cyanobacterial culture time and the pigments of cell pellets were extracted in 100% acetone at 50 °C for 15 min. After centrifugation (13 000 rpm for 3 min), the pigment extract was analyzed for Chl a (mg per mL) by following the previous spectrophotometric assay 14 .
The supernatant was obtained and the absorbance was measured at 620, 680, and 750 nm on Cary 60 UV-Vis spectrophotometer (Agilent technologies, CA, USA). For PC measurement, harvested cell suspension (1 mL) was treated either with or without heating at 75 °C for 8 min. PC was determined by measuring the absorbance at 620 nm and 750 nm and PC content (mg per mL) was calculated by following the previous spectrophotometric assay 14 .
Calculations for carbon fixation rate using CO 2 analysis. Cyanobacterial cells were inoculated to . 5% CO 2 (v/v) and 95% air (v/v) was supplied directly into the medium at flow rate of 140 mL min −1 (0.08 vvm) by gas bubbling and 5 N NaOH was used to adjust pH 7 in the medium. CO 2 analysis was performed by continues monitoring CO 2 concentration from off-gas line before and after cell inoculation. CO 2 concentration (ppm) was digitally recorded every per second by CO 2 analyzer (Q-S153, Qubit systems, Kingston, ON, Canada).
The carbon fixation rate in a unit of mg of CO 2 per liter per day was calculated from the previous calculation 1 . Briefly, CO 2 value was converted from ppm into μ mol CO 2 L −1 using an equation of , where C refers to the temperature in °C ant T refers to the absolute temperature (273K). Based on the flow rate (in L min −1 ) (i.e., 140 mL min −1 ) and the volume of the culture medium (i.e., 1.8 L), the CO 2 fixation rate can be obtained with units of mg l −1 d −1 .