1. Introduction
Efficient recycling and sustainable production of valuable compounds are major challenges of the future. Development of carbon neutral production systems and improvement of recycling could help to prevent further global warming and reduce pollution. Municipal, piggery, and dairy wastewaters, as well as anaerobic digestion reject waters, are typically nutrient rich, containing plenty of ammonium, nitrate, and phosphate [
1,
2]. Cyanobacteria are efficient in collecting nutrients [
3] and many valuable products, including ethanol, isopropanol, butanol, biohydrogen and alkanes, have already been produced in cyanobacteria [
4]. In the most tempting scenario, cyanobacteria are engineered to produce high value compounds, not only biomass, and simultaneously remediate wastewaters. For biotechnical applications, robust strains with high acclimation capacity are beneficial. Keeping up constant optimal environmental conditions, such as optimal temperature and constant light, is expensive, and using wastewater as a growth medium might cause additional problems; wastewaters typically contain harmful chemicals, and their nutrient balance is variable and not optimal.
Acclimation of cyanobacteria to suboptimal conditions is strongly dependent on changes in gene expression. In cyanobacteria, regulatory σ subunits of the RNA polymerase (RNAP) play major roles in acclimation responses. In optimal growth conditions, the RNAP core mainly recruits the primary σ factor that guides RNAP to transcribe housekeeping genes [
5]. In suboptimal conditions, RNAP holoenzymes bearing stress-condition-specific alternative σ factor(s) become abundant and change the transcriptome to allow acclimation to that particular stress [
5,
6]. The cyanobacterium
Synechocystis sp. PCC 6803 contains nine different σ factors. SigA is a primary σ factor that is mainly responsible for transcription of housekeeping genes during exponential growth; SigB, SigC, SigD, and SigE are group 2 σ factors that play major roles when cells acclimate to suboptimal conditions, and SigF, SigG, SigH, and SigI are group 3 σ factors [
6]. Construction of robust cell factory strains might be possible via modification of the σ factor content of cells [
7], and highly similar group 2 σ factors SigB and SigD [
8,
9] are of particular interest in the construction of robust production strains.
The expression of the SigB gene is typically transiently up-regulated when
Synechocystis cells are transferred from standard conditions to stress conditions, such as high temperature [
10,
11], high salt [
8,
12,
13], or from darkness to light [
14]. The expression of the
sigD gene is enhanced in high light [
9,
14], and cells do not acclimate properly to high light without the SigD factor [
8,
15].
The amounts of SigB and SigD proteins are highly dependent on environmental conditions and so is the formation of the transcription initiation competent RNAP-SigB and RNAP-SigD holenzymes. The amount of the RNAP-SigD holoenzyme decreases in darkness and increases in high light, compared to the standard conditions, whereas formation of the RNAP-SigB holoenzyme is induced by high temperature or salt treatments [
5]. Furthermore, inactivation of the other group 2 σ factors influences the amount of the remaining group 2 σ factor, and the amounts of SigB and SigD proteins increase two-fold if the other group 2 σ factors are knocked out [
16].
To get a more comprehensive picture of the roles of SigB and SigD factors, the performance of ΔsigB, ΔsigD, ΔsigCDE, and ΔsigBCE strains were studied in heat and high light conditions. In addition, the transcriptomes of the ΔsigB, ΔsigD, ΔsigCDE, and ΔsigBCE strains were compared in standard growth conditions and after heat and high light treatments, in order to characterize the specific regulons of SigB and SigD in these stress conditions.
2. Material and Methods
2.1. Strains
The glucose tolerant strain of
Synechocystis sp. PCC 6803 (CS) was used as a control and host strain [
8]. The construction of ΔsigB, ΔsigD [
8], ΔsigBCE, and ΔsigCDE [
17] strains has been described earlier. Single and triple inactivation strains were maintained on BG-11 plates in standard growth conditions in the presence of appropriate antibiotics, as described earlier [
8,
17], but antibiotics were omitted from liquid cultures.
To study the
sigB and
sigD promoters, P
sigB- and P
sigD-lux strains were constructed. First, a nourseotricine resistance cassette (NAT) with the
rrnBT1 transcription terminator sequence of
E. coli was released as a HindII fragment from plasmid pGEM-NATter (generous gift from L. Lopez-Maury), and the single stranded ends of the DNA fragment were filled with Klenow enzyme. This NAT fragment was then cloned into the SmaI site after the
luxCDABE operon into the pTETlux plasmid (generous gift from Marko Virta) to obtain the pTETluxNAT plasmid. Either the 402 bp long fragments of upstream from the coding region of the
sigB gene (
Figure 1A) or 284 bp long upstream fragment of the
sigD gene (
Figure 1B) was used to control the expression of the
lux operon. For insertion to the
Synechocystis chromosome, the 200 bp long upstream and downstream regions of the
psbA1 gene were inserted to the both ends of the construct containing the lux operon under the sigB or sigD promoter and NAT cassette (
Figure 1A,B). The
psbA1 gene was selected, as it is practically silent in all conditions used in this study and only expressed in low oxygen conditions [
18,
19]. Synthetic DNA fragments were purchased from GenScript Biotech (New Jersey, NJ, USA). Then plasmids pP
sigBlux and pP
sigDlux were transformed to the glucose tolerant control strain of
Synechocystis sp. PCC6803. Colonies appeared after one week on BG-11 plates supplemented with nourseothricin (10 μg/mL), and they were streaked on new selective plates once a week for eight weeks. Thereafter, the segregation of PsigB- and PsigD-lux strains was verified with PCR, as described in [
10], using the primers 5′-CCGCTACCACCTGTTTTATTA-3′ and 5′- TCGGCCCAGGTGCTCACG- 3′. Finally, the modified areas of the genome in PsigB- and PsigD-lux strains were sequenced to confirm that they were as planned.
2.2. Growth at Standard, High Temperature and High Light Conditions
The cells were grown as 30-mL batch cultures in 100-mL Erlenmyer flasks in BG-11 medium that was buffered with 20 mM Hepes, pH 7.5. For standard growth conditions, cells were grown at 32 °C under continuous illumination at the photosynthetic photon flux density (PPFD) of 40 μmol m−2s−1 in ambient air. The light source was a mixture of fluorescent tubes, light colors 865 and 840 (Osram (Munich, Germany)/Philips (Amsterdam, The Netherlands)). The flasks were shaken at 90 rpm.
To induce long-term, high temperature stress, the OD730 of the cell culture was set to 0.06, and 30-mL cell cultures in 100-mL Erlenmyer flasks were grown in Hepes-buffered (pH 7.5) BG-11 medium at 42 °C or at 40 °C, as indicated, PPFD 40 μmol m−2s−1, ambient air. Cells were grown in Climatic Chamber KK1200 (Pol-Eko Aparatura®, Wodzisław Śląski, Poland), the same light source as in standard conditions was used, and the temperature was monitored during experiments. Three independent biological replicates were tested.
To test growth in high light, the OD730 of the cell culture was set to 0.06, and 30-mL cell cultures in 100-mL Erlenmyer flasks were grown in Hepes-buffered (pH 7.5) BG-11 medium, PPFD of 750, 500, or 120 μmol m−2s−1 at 32 °C in ambient air. Cells were grown in a Novotron® TR-225 chamber (Infors AG, Bottmingen, Switzerland), and illuminated using Heliospectra L4A lamp (Heliospectra, Göteborg, Sweden). The lamp setting were 400 nm = 200, 420 nm = 400, 450 nm = 450, 530 nm = 1000, 630 nm = 1000, 660 nm = 900, 735 nm = 0. Neutral density filters 0.3 ND and 0.6 ND (LEE Filters, Hampshire, UK) were used to adjust light intensity. Temperature was monitored during experiments. Three independent biological replicates were tested.
Growth was monitored by measuring OD
730 once a day. Dense cultures were diluted, so that the measured OD
730 did not exceed 0.4, and the dilutions were taken into account when the results were calculated. In vivo absorption spectra were measured from 2-day old cells, as described in [
16].
2.3. Imaging Cell Cultures
Images of cultures were taken with a Canon EOS 250D camera (Canon, Tokio, Japan) equipped with a Canon Compact-Macro EF 50 mm lens (Canon, Tokio, Japan), after 2 and 3 days of growth. For single cell imaging, cells were diluted to OD730 of 0.0175, 100 µL of each culture was transferred to a 96-well plate, and images were taken with a Nikon Eclipse Ti2-E-microscope (Nikon, Tokio, Japan) and Nikon DS-Fi3-camera (Nikon, Tokio, Japan).
2.4. Activities of sigB and sigD Promoters
The lux reporter gene operon was expressed under the sigB or sigD promoters. The construction of these strains is described above. Bacterial strains containing the whole lux operon are expected to produce light without the addition of a substrate. However, in Synechocystis, we found that the addition of the substrate decanal greatly enhanced the light signal. The CS, PsigB-lux, and PsigD-lux cultures were grown in standard growth conditions for three days. Prior to measurements, 200 μL of cell cultures were supplemented with 0.1 mM decanal in microplate wells (black, flat, clear bottom polystyrol, Corning), after which cells were incubated in standard, high temperature (40 °C) or high light (PPFD 380 μmol m−2s−1) conditions for 15 min, and then luminescence and OD730 were measured with a plate reader (Infinite Pro200, Tecan, Männedorf Switzerland). Luminescence values were calculated relative to OD730.
2.5. RNAseq Analysis
Thirty-ml cell cultures were grown for three days in standard conditions (to OD
730 0.6). Cells were then collected from the standard conditions or treated for 1 h at the PPFD of 750 μmol photons m
−2 s
−1 or 1 h at 42 °C, as indicated, maintaining the other conditions same as in the standard conditions. Three biological replicates were prepared for each treatment. A total of 15 mL of cell suspension was poured to a pre-frozen centrifugation tube, and a cell pellet was collected by centrifugation at 7000×
g for 5 min at 4 °C. Cell pellets were frozen in liquid nitrogen. Isolation of total RNA was done with the hot phenol method [
20]. Sequencing libraries for total RNA samples were prepared using Illumina TruSeq Stranded mRNA kit protocol. The quality of the library was confirmed using the Advanced Analytical Fragment Analyzer (Advanced Analytical Technologies, Heidelberg, Germany), and the concentrations of the libraries were quantified with Qubit
® Fluorometric Quantitation (Life Technologies, ThermoFisher, Waltham, MA, USA). Sequencing was done with a HiSeq2500 Next Generation Sequencing platform at the Finnish Functional Genomics Centre. The quality control of raw sequencing reads was performed with FastQC (
http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ accesed on 20 November 2020). Reads were aligned to the
Synechocystis sp. PCC 6803 (Moscow strain) reference genome (Ensemble release 35) using HISAT2 [
21]. HTseq-count [
22] was used to calculate summarized read counts for each gene, except that the two ribosomal RNA operons were excluded from the analysis. DESeq2 [
23] was used to identify differentially expressed genes. Wald test statistic is used for hypothesis testing in DESeq2. Genes with a false discovery rate, <0.05, were considered to show significantly changed expression. Data analysis was done with the CSC’s Chipster v.4 platform [
24]. We considered the gene to be up- or down-regulated in a mutant strain if its expression was at least two-fold up-regulated (log
2 fold change ≥ 1) or down-regulated to one half or less (log
2 fold change ≤ −1), compared to the CS strain in the same conditions, respectively.
For regulon analysis, we chose genes showing a log
2 fold change ≥ 1 or ≤ −1 in at least one of the mutant strains, compared to the control strain in the same conditions. Genes were assigned to the SigB regulon if they were significantly down-regulated both in ΔsigB and ΔsigBCE strains, but not in the other strains (up-regulation in ΔsigCDE was allowed), or significantly up-regulated both in ΔsigB and ΔsigBCE strains, but not in the other mutant strains (down-regulation in ΔsigCDE was allowed). Genes were assigned to the SigD regulon if they significantly down-regulated both in ΔsigD and ΔsigCDE strains, but not in the other mutant strains (up-regulation in ΔsigBCE was allowed), or significantly up-regulated both in ΔsigD and ΔsigCDE strains, but not in the other mutant strains (down-regulation in ΔsigBCE was allowed). Genes simultaneously significantly down- or up-regulated in both triple mutant strains, but not in the single mutants, were considered to belong the SigC and/or SigE regulons. Since the ΔsigC and ΔsigE strains were not included in the study, we cannot tell whether a gene belongs to the SigC or the SigE regulon. To visualize the results, heat maps were created using a web tool [
25]. RNAseq data are available in GEO (accession GSE192357).
4. Discussion
The construction of robust cyanobacteria strains is one of the current challenges for biotechnology applications. Inactivation of the three other group 2 σ factors leads to two-fold overproduction of the remaining SigB or SigD factor in the ΔsigCDE and ΔsigBCE strains, respectively [
16]. As a previous study [
16] has shown that ΔsigCDE and ΔsigBCE strains tolerate chemically-induced oxidative stress better than the CS strain, we considered the possibility that extra SigB or SigD factors in the triple inactivation strain would protect cells in high light or temperature conditions. However, that was not the case; actually, our results point to the involvement of all group 2 σ factors in different stress conditions. To test if the fitness of the cells in some stress conditions could be improved by overexpressing a particular σ factor should, therefore, be done with a wild-type strains as the host strain.
The comparison of transcriptomes of ΔsigB, ΔsigD, ΔsigCDE, and ΔsigBCE strains, in standard conditions and after 1 h of treatment in high light (750 μmol photons m
−2s
−1) or at temperature (42 °C), revealed new features and confirmed old findings of gene regulation in
Synechocystis. Group 2 σ factors indeed play only marginal roles in standard growth conditions but are important for acclimation to suboptimal conditions [
6]. Our results revealed that although SigB and SigD factors are close homologs [
8], their regulons do not overlap (
Figure 3,
Figure 5 and
Figure 7). Furthermore, the comparison of up- and down-regulated genes after high light and temperature treatments showed that both SigB and SigD regulate different sets of genes in high light and temperature (
Figure 7). Although our study did not focus on SigC or SigE regulons, transcriptomes of triple mutants contained numerous sugar catabolism and other genes belonging to SigE regulon [
26,
27], and our results indicate that the SigE regulon does not show similar high variation between the studied environmental condition as SigB and SigD regulons. As both triple inactivation strains grew slowly at high temperature, our results suggest that the SigE-dependent regulation of sugar catabolic genes and energy production by respiration might play an important role in high temperature stress, when temperature-sensitive photosynthesis [
28] might not be fully functional. A detailed analysis of the photosynthetic and respiration reactions in heat stressed σ factor mutant strains would be interesting.
Both triple inactivation strains grew more slowly than the CS or single inactivation strains at 42 °C (
Figure 4A), but this disadvantage of triple mutants disappeared at 40 °C (
Figure 4B). In triple mutants, this growth behavior might mainly be due to the missing SigC factor, as the ΔsigC strain has been shown to have serious growth defects at 43 °C, but not at 38 °C [
28]. SigC is a growth restricting σ factor [
29,
30,
31], and the formation of RNAP-SigC holoenzymes at high temperature occurs more regularly than at optimal temperatures [
5]. In accordance with enhanced formation of growth restricting RNAP-SigC holoenzymes, the growth rate of the CS strain is very slow at 42 °C, in comparison to the optimal growth temperature (
Figure 2 and
Figure 4A). RNAP-SigC induced adjustments of gene expression are important for heat acclimation. At high temperatures, numerous household genes, including transcription, translation, pigment synthesis, and photosynthesis genes, are not normally down-regulated upon temperature upshift in triple inactivation strains, and cells show early growth cessation at high temperature (
Figure 4). Obviously, the SigC controlled growth restriction is a central process for high temperature acclimation and retaining a functional SigC is a prerequisite for engineering temperature robustness of
Synechocystis strains.
The amount of the RNAP-SigB holoenzyme is highly up-regulated at high temperature, whereas the amount of RNAP-SigD remains constant [
5]; accordingly, more genes belong to the SigB regulon than to the SigD regulon at high temperature (
Figure 3 and
Figure 7). The ΔsigB strain shows a low survival rate after a short term lethal temperature treatment (48 °C) and defects in the acquisition of acquired thermotolerance [
10], but growth differences were not detected in milder heat stress (
Figure 4; [
10,
11]). The overproduction of SigB in wild-type backgrounds has shown to increase heat resistance of
Synechocystis cells [
32], but we do not see the same effect in the ΔsigCDE strain that contains an abnormally high amount of SigB [
16], but is missing the other group 2 σ factors.
Our results point to high flexibility and variability of high light acclimation. The growth rate of
Synechocystis cells, at standard growth conditions (PPFD 40 μmol photons m
−2s
−1) and in all tested high light conditions (120, 500 and 750 μmol photons m
−2s
−1), remained similar in all mutant strains. One consequence of high light treatment is an accelerated rate of PSII damage. However, all our strains contained either the SigB or SigD factor, and it has been previously shown that the presence of either SigB or SigD can guarantee the normal high light-specific up-regulation of
psbA genes that is required for efficient repair of photodamaged photosystem II centers [
15].
The
sigD gene is induced in high light [
9,
14], and the RNAP-SigD holoenzyme is accumulated in high quantities in high light [
5], which explains why the SigD regulon is big in high light (
Figure 5 and
Figure 7). However, cells without the SigD factor did not show clear growth rate reductions in high light (
Figure 6). The ΔsigBCE strain shows up-regulation of RNAP-SigD holoenzyme formation, compared to the CS strain (Hakkila et al., 2019), and ΔsigBCE cells grow faster in the beginning of high light treatment, but this growth advantage disappeared after long high light treatment (
Figure 6A). The ΔsigBCE strain remains greener in high light than the other strains because the phycobilins and chlorophyll a are not degraded as much as in the other strains and the carotenoid content remains low. Previous studies have shown that the ΔsigBCE strain is more resistant against chemically-induced singlet oxygen and H
2O
2 stresses than the control strain [
3,
16]. We have previously shown that the amount of the SigD protein is two times as high in the ΔsigBCE strain as the CS strain [
16], suggesting that overproduction of SigD might offer protection against some reactive oxygen species. Obviously, the extra protection in the ΔsigBCE strain is not due to carotenoids, as the carotenoid content is lower in ΔsigBCE strain than in the other strains in high light (
Figure 6G). A high tendency of ΔsigBCE strain to form cell aggregates might function as a strategy to escape too high light. Both strains without the SigD factor (ΔsigD and ΔsigCDE) were characterized by high accumulation of carotenoids (
Figure 6G). Carotenoids are especially important protective pigments, and
Synechocystis cells are not viable in light without carotenoids [
33]. Hence, the accumulation of carotenoids could be seen as another strategy to survive in high light.
The
sigB gene is induced in high light (
Figure 1), although to a lesser extent than the sigD gene, and similar induction of the RNAP-SigB holoezyme is not seen in high light as is seen for RNAP-SigD [
3,
16]. Our previous results have shown that high amounts of carotenoids, as well as Flv2, Flv4, and Sll0218 proteins, in ΔsigCDE reduce the production of singlet oxygen and confer resistance of PSII against light-induce damage; simultaneously, reduced contents of glutathione and Flv1/3 proteins led to more pronounced oxidative damage in ΔsigCDE than in the control strain. Due to these opposing effects, the ΔsigCDE grew like a control strain in high light but showed superior growth in H
2O
2 stress [
16]. Our results show that, actually, triple inactivation strains containing only SigD or SigB managed quite well in high light because their acclimation to some features of high light stress was superior, compared to the CS strains, and compensated for problems to acclimate other features of high light stress.