Bacterial strains used in this work are in Table 1. For cloning, E. coli strains were grown in lysogeny broth medium (10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl) at 37°C. When appropriate, medium was supplemented with 25 μg/mL chloramphenicol, 50 μg/mL kanamycin, 100 μg/mL spectinomycin. S. elongatus PCC 7942 (wild type). Derivatives were cultivated at 30°C in 100 mL glass bottles under continuous fluorescent light (100 μmol photons/m²/s) in BG-11 medium (1.5 g/L NaNO₃, 0.006 g/L ferric ammonium citrate, 0.001 g/L Na₂EDTA·2H₂O, 0.039 g/L K₂PO₄, 0.075 g/L MgSO₄·7H₂O, 0.020 g/L Na₂CO₃, 0.036 g/L CaCl₂·2H₂O, 0.006 g/L citric acid, 2.860 mg/L H₃BO₃, 1.810 mg/L MnCl₂·4H₂O, 0.222 mg/L ZnSO₄, 0.39 mg/L Na₂MoO₄·2H₂O, 0.079 mg/L CuSO₄·5H₂O, 0.0494 mg/L Co[NO₃]₂·6H₂O) supplemented with 10 mM MOPS (pH 7.5). Supplied at a constant flow rate of 10 mL/min into medium was 5% (v/v) CO₂ gas and 95% (v/v) air. Selection was with 10 μg/mL spectinomycin, 10 μg/mL kanamycin or 3 μg/mL chloramphenicol. Isopropyl-β-D-1-thiogalactopyranoside (IPTG) was added into culture medium at inoculation for induction. Cell growth was monitored by OD₇₃₀ with a UV spectrophotometer (Agilent Technology, CA, USA).

BglBrick vectors (e.g., pBbE1k-RFP; Lee et al., 2011) and pSe1Bb1s-GFP (NCBI accession number, KJ814971; Chwa et al., 2016) are standard vectors for construction of SyneBrick vectors. Their four parts are promoter, antibiotic selection marker, gene of interest, and integration site (Table 1). Standard SyneBrick vectors were constructed using circular polymerase extension cloning (CPEC; Quan and Tian, 2009) and conventional restriction-ligation cloning methods with specific primers (Table S1). Using the BglBrick cloning method and a linearized PCR product (eYFP) from p426-TEF-eYFP using eYFP-FW/RV primers, the gfp gene was changed to the eyfp gene in SyneBrick vectors, yielding pSe1Bb1s-eYFP, and pSe2Bb1k-eYFP from pSe1Bb1s-GFP and pSe2Bb1k-GFP, respectively.

For construction of additional SyneBrick vectors (pSe3 vectors; NSIII), DNA fragments NSIIIa and NSIIIb were replaced with NSIIa and NSIIb in pSe2Bb1k-GFP and the kanamycin resistance gene was exchanged to chloramphenicol resistance gene. This resulted in pSe3Bb1c-GFP, a gene expression platform in NSIII that was compatible with pSe1Bb1s-GFP and pSe2Bb1k-GFP. For TetR-pTetA based gene expression, pSe1Bb1s-GFP was modified with a functional genetic element of TetR-pTetA from pBbE2c-RFP, yielding pSe1Bb2s-GFP. For T7-based gene expression, pSe2Bb1k-eYFP was modified with a DNA fragment using oligonucleotides containing the T7 sequence (T7-FW/RV), yielding pSe2Bb7k-eYFP. To construct SyneBrick vector 2.0 versions, T7.3-FW and T7.4-FW (containing variant sequences of the original T7 sequences; Temme et al., 2012) were used to generate pSe2Bb7.3k-eYFP and pSe2Bb7.4k-eYFP, respectively.

Transformation of S. elongatus PCC 7942 was as described previously (Golden et al., 1987). Cyanobacterial strains were transformed by incubating at mid-log phase (OD₇₃₀ 1 to 2) with 100 ng plasmid DNA for 24 h in the dark. Cell-DNA mixed cultures were spread on BG-11 plates supplemented with appropriate antibiotics for selecting recombinants (10 μg/mL spectinomycin, 10 μg/mL kanamycin, or 3 μg/mL chloramphenicol). Single colonies were sub-cultured to prevent chromosomal segregation. PCR was used to verify chromosomal integration of targets into either NSI, NSII, or NSIII (Figure 1). DNA sequences were verified using oligonucleotides in Table S1. Genotypes of recombinant S. elongatus strains are listed in Table 1.

S. elongatus mutants with either gfp or eyfp chromosomally integrated using SyneBrick vectors were cultivated in BG-11 medium overnight at 30°C. To induce GFP or eYFP expression, IPTG for Se1Bb1s-GFP or aTC for Se1Bb2s-GFP was added when cyanobacterial cells were inoculated into BG-11 medium. After 1, 4, or 7 days, 200 μL cell suspensions (diluted to OD₇₃₀ of 1) were transferred to flat bottom, 96-well solid black microtiter plates (Corning, USA) for GFP or eYFP measurements. Fluorescence intensity were measured using an automatic microplate reader (Tecan Infinite M200 pro, Tecan Group Ltd., Switzerland) and optical density was measured by UV-Vis spectrometry. Fluorescence intensities of GFP were measured at excitation/emission wavelengths of 485/535 nm and eYFP was measured at 505/535 nm. Specific fluorescence was calculated by normalizing fluorescence intensity by optical density (730 nm). BG-11 medium was the blank.

To engineer S. elongatus PCC 7942 for gene expression, we developed SyneBrick vectors capable of expressing multiple genes controlled by different promoters and transcriptional regulators. An chromosomal integration plasmid with BglBrick cloning features, termed SyneBrick vector, was constructed based on BglBrick (Lee et al., 2011)/CoryneBrick vectors (Kang et al., 2014), using the CPEC cloning method (Quan and Tian, 2011; Figures 1A,B). The set of SyneBrick vectors used three different chromosomal integration sites (Neutral site I, II, and III) where no genetic interference occurred. Although many combinations of neutral sites with antibiotic selection markers (Spec^R, Km^R, and Cm^R) were possible, we chose three pairs of SyneBrick vectors, NSI-Spec^R, NSII-Km^R, NSIII-Cm^R, to generate pSe1Bb1s-GFP, pSe2Bb1k-GFP, and pSe3Bb1c-GFP, respectively. These three vectors can be used for the seven combinatorial construction of the cyanobacterial strains: single integrations, Se1Bb1s, Se2Bb1k, Se3Bb1c; double integrations, Se12Bb11sk, Se13Bb11sc, Se23Bb11kc; and a triple integration, Se123Bb111skc. Additional neutral sites could be identified using transcriptome sequencing analysis (Ng et al., 2015).

The initial SyneBrick vectors have been successfully applied to metabolic engineering (Choi et al., 2016; Chwa et al., 2016; Lee et al., 2017). To expand a selection of the SyneBrick vectors, LacI-pTrc in the standard SyneBrick vectors pSe1Bb1s-GFP and pSe2Bb1k-eYFP were replaced with transcriptional regulator TetR-pTetA and pT7 promoter, respectively. This constructed strains Se1Bb2s-GFP and Se2Bb7k-eYFP. For interpreting the features of SyneBrick vectors, the nomenclature is Se# for S. elongatus and neutral sites (#) and Bb for BglBrick vector-originated (Figure 1B). Annotations of BglBrick vectors are one for the trc promoter, two for the tetA promoter, seven for the T7 promoter, s for spectinomycin, k for kanamycin, and c for the chloramphenicol resistant gene. Name of genes of interest are followed by a dash.

SyneBrick vectors share the features of BglBrick/CoryneBrick vectors. A PCR step is not necessary for gene cloning and multiple gene assembly is possible with repeated enzyme digestions and ligations without choosing unique restriction enzyme sites. The details of the cloning steps have been described (Lee et al., 2011; Kang et al., 2014). In summary: A target gene of interest digested at EcoRI/BamHI sites is ligated to EcoRI/BglII sites of SyneBrick vectors (Table 1) by replacing GFP or eYFP. Each target gene with a BglBrick compatible sequence is sequentially inserted into the SyneBrick vectors in multiple gene clonings. Alternatively, multiple genes built into other BglBrick (Lee et al., 2011) or CoryneBrick vectors (Kang et al., 2014), or BglBrick-formatted genes used in Synechocystis sp. PCC 6803 (Huang et al., 2010) or S. elongatus PCC 7002 (Markley et al., 2015) can be cloned into SyneBrick vectors by one-step digestion and ligation. Target genes must be checked for the presence of EcoRI, BglII, BamHI, and XhoI sites prior to multiple-gene assembly in the vector.

To test for controllable gene expression, GFP-expressing, or eYFP-expressing S. elongatus PCC 7942 strains (Se1Bb1s-eYFP, Se1Bb1s-GFP, and Se1Bb2s-GFP) were constructed after natural transformation with SyneBrick vectors. We tested the wild-type strain with IPTG up to 10 mM and found no growth inhibition (Figure 2). Subsequently, engineered strains were cultivated in BG-11 medium with 5% (v/v) CO₂ bubbling in the presence of IPTG or aTc, depending on the strain. GFP or eYFP fluorescent levels were measured.

IPTG induction up to 10 mM and aTc up to 1 mM did not cause growth inhibition of the engineered strains although slight growth inhibition with 10 mM IPTG or 1 mM aTC was observed in the engineered strains compared to the wild type. Strain Se1Bb1s-None was constructed as a control for calibrating values for gene expression of engineered strains because the wild type shows auto-fluorescence in the range of measurements (McEwen et al., 2013). Compared with a control, strains Se1Bb1s-eYFP, Se1Bb1s-GFP, and Se1Bb2s-GFP showed low levels of leaky GFP/eYFP expression in the absence of inducers (Figure 2B). When 0.01, 0.1, 1, and 10 mM of IPTG was used, higher IPTG concentrations resulted in higher induction of the reporter gene. Increased induction levels over cultivation days were also observed. Compared with 0 mM IPTG, induction of Se1Bb1s-GFP with 1 mM resulted in a 10-fold change in gene expression and 10 mM resulted in an 11-fold change. This result was consistent with the 16-fold induction of β-glucuronidase activity in a reporter assay using an IPTG-inducible pTrc gene expression system in S. elongatus PCC 7942 (Geerts et al., 1995). Compared with Se1Bb1s-GFP, Se1Bb1s-eYFP had a 24-fold higher induction with 1 mM IPTG and 26-fold with 10 mM. This discrepancy could be due to differences in signal intensities between intracellular GFP and eYFP in cyanobacteria. At least 10-fold induction was achieved with either 1 or 10 mM IPTG in S. elongatus PCC 7942 using SyneBrick vectors. For controllable gene expression with SyneBrick vectors with the 1 promoter, 1 mM IPTG induction is recommended.

Se1Bb2s-GFP with the TetR-pTetA transcriptional system was tested with four concentrations of aTc: 10, 100, 200, and 1000 nM in 5% (v/v) CO₂. The highest induction of 19-fold occurred when 100 nM aTc was used, compared to without aTc (Figure 2C). However, induction was decreased dramatically when more than 100 nM aTC was used. Thus, 100-nM aTC induction was recommended with SyneBrick vectors with the 2 promoter and 24-h culture. In previous studies, an aTc-dependent induction system for Synechococcus sp. PCC 7002 had induction ranges of 6- and 32-fold with 1000 ng/mL (≅ 2200 nM) aTc (Zess et al., 2016) until 48 h. Dynamic ranges of inductions up to 290-fold have been reported for Synechocystis sp. PCC 6803 under light-activated heterotrophic growth with 10 μg/mL (≅ 22,000 nM) aTC for 48 h (Huang and Lindblad, 2013). Consistent with these results, our TetR system showed decreased gene expression after 48 h due to aTc degradation in light. However, our TetR-pTetA systems required much less aTC to induce gene expression, compared with previous studies (Huang and Lindblad, 2013; Zess et al., 2016). Thus, time-dependent gene expression in Se1Bb2s-GFP under light conditions could be possible by adding aTc in a pulse-type control mode. Engineering of our TetR-pTetA transcriptional system in S. elongatus PCC 7942 with light wavelength (i.e., red light or white light) and intensities carefully considered for SyneBrick applications with aTC as a light-sensitive inducer. Sequence modification of the pTetA promoter with tetO operators could alter intrinsic expression and induction of target genes. This could be easily incorporated into SyneBrick platform vectors.

To narrow the dynamic range of IPTG induction with tunable gene expression in S. elongatus PCC 7942, T7 RNA polymerase (RNAP) and its cognate T7 promoter were developed in SyneBrick vectors (pSe2Bb7k-eYFP and pSe3Bb1c-T7RNP; Figure 3). We tested the possibility of T7 promoter-driven transcription in the absence of T7 RNAP. No gene expression was observed in Se2Bb7k-eYFP under control of the T7 promoter, indicating that the T7 promoter was not recognized by cyanobacterial RNA polymerases (data not shown). Thus, the T7 RNAP gene 1 encoding for T7 RNAP from E. coli BL21(DE3) was integrated into Se2Bb7k-eYFP using pSe3Bb1c-T7RNP, yielding strain Se23Bb7kc-eYFP/P.

Compared with the LacI-pTrc gene expression system of Se1Bb1s-eYFP, the T7RNAP-T7 gene expression system of Se23Bb7kc-eYFP/P showed efficient induction of target gene with IPTG. Specific fluorescence intensities of Se23Bb7kc-eYFP/P increased 6-fold with 0.01 mM IPTG and 1.5-fold with 0.1 mM compared to Se1Bb1s-eYFP (Figure 4). Thus, induction of the T7RNAP-T7 gene expression system with 0.1 mM IPTG resulted in the same levels of induction as 1 mM IPTG in the LacI-pTrc gene expression system. The efficient induction with the low IPTG concentration could be due to an inherently fast mRNA elongation rate of the T7 RNA polymerase. Decreased gene expression was detected when more than 1 mM IPTG was used. The reason that expression was reduced with high concentrations of IPTG was because the RNA polyadenylation and degradation mechanisms in cyanobacteria are different from E. coli (Rott et al., 2003). Thus, orthogonal expression of the T7 polymerase gene 1 can be alternatively applied by replacing the pTrc promoter with the pTetA promoter to prevent interference with expression from the T7 promoter. The T7RNAP-T7 gene expression system induced well with 0.01 or 0.1 mM IPTG.

In addition, variants of T7 promoters (T7.3 and T7.4) in which a 5-bp strength-determining region was modified (Temme et al., 2012) were replaced with the original T7 promoter to alter promoter strength without RNA specificity (Figure 3). This generated strain Se23Bb7.3kc-eYFP/P with 82% eYFP expression compared to Se23Bb7kc-eYFP/P and Se23Bb7.4kc-eYFP/P, with 90% expression. These reductions in gene expression in cyanobacteria are different from gene expression levels in E. coli, which are 28% for T7.3 and 17% for T7.4 from T7 (Temme et al., 2012). The order of promoter strength using a T7 library in cyanobacteria was consistent with E. coli. Using a library of T7 promoter SyneBrick vectors, the process of refactoring of a gene cluster encoding for complex proteins such as α-carboxysomes would be possible as rewriting the genomes to decipher their functions in cyanobacteria. T7 lysozyme can also be used as an inhibitor for fine-tuning gene expression of the T7 promoter by controlling T7 RNA polymerase activity.

The development of clustered regulatory interspaced short palindromic repeats (CRISPR) interference technology has allowed control of gene expression by inducing dCas9 encoding genes with aTc. This method has successfully been applied to lactate production from 1% CO₂ (Gordon et al., 2016) and to repress multiple genes in the cyanobacterial genome (Yao et al., 2016). Thus, SyneBrick vectors could be further expanded to precisely control gene expression by integrating CRISPR technology for metabolic engineering of S. elongatus PCC 7942.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.