A chimeric vector for dual use in cyanobacteria and Escherichia coli, tested with cystatin, a nonfluorescent reporter protein

Strains and growth conditions

Cultures of the cyanobacterium Synechocystis sp. PCC 6803 (S. sp. PCC 6803, obtained from Pasteur Culture Collection of Cyanobacteria) were grown on BG11 agar plates or in BG11 liquid medium (Stanier et al., 1971) at room temperature under continuous white light of 10–30 μmol photons m⁻² s⁻¹. Liquid cultures were shaken at 100 rpm. For engineered strains transformed with pMJc01 or pPMQAK1 plasmid or their derivatives, the growth medium was supplemented with 50 μg/ml kanamycin.

Escherichia coli XL1-Blue was used for cloning and characterization of vector constructs. All strains were cultured in liquid LB medium with shaking at 140 rpm at 37 °C or on LB agar plates. Cells transformed with pMJc01, pSB3K3 or pPMQAK1 were grown in a medium containing 50 μg/ml kanamycin and those transformed with the RSFmob-I plasmid were cultured in the presence of 50 μg/ml streptomycin.

For the transformation of E. coli, we used chemically competent cells and for the transformation of S. sp. PCC 6803, we used electroporation. In brief, cells in the exponential growth phase were washed three times in ice-cold sterile Milli-Q water (followed by centrifugation for 10 min at 4 °C and 6,000g) and mixed with 0.5–2 µg of plasmid dissolved in 40 µl of sterile Milli-Q water. The suspension of cells and plasmid was transferred to electroporation cuvettes with a gap width of two mm and subjected to an electrical pulse of 2,500 V for five ms (with the capacitance set to 25 µF and the resistance set to 200 Ω). After pulse treatment, cells were immediately resuspended in five ml of BG11 medium without kanamycin and left to recover overnight at lower light intensity. The next day, the liquid cultures were poured onto BG11 agar plates supplemented with five µg/ml kanamycin. The resulting colonies were transferred to BG11 agar plates supplemented with gradually increasing concentrations of kanamycin (10, 25 and finally 50 µg/ml).

Construction of the broad host range shuttle vector pMJc01

A DNA region comprising sequences required for autonomous self-replication was PCR-amplified with primers RSFmobF and RSFmobR (Table 1) from the template plasmid RSFmob-I (Katashkina et al., 2007) (available via Russian National Collection of Industrial Microorganisms). A DNA fragment containing a synthetic biology RFC[10] standard cloning site and a kanamycin resistance cassette was PCR-amplified from the pSB3K3 template plasmid using primers pSB3K3F and pSB3K3R (Table 1). Primers RSFmobR and pSB3K3F contained an 18 bp long complementary sequence that allowed overlap extension PCR for splicing of the 3,771-bp fragment from plasmid RSFmob-I and the 2,238-bp fragment from plasmid pSB3K3 to plasmid pMJc01. The corresponding linear 5,999-bp product was phosphorylated at the 5′ ends, circularized with T4 DNA ligase, and used to transform competent E. coli cells. The sequence of the new pMJc01 vector was verified by Sanger sequencing (GATC Biotech, Eurofins Genomics).

Construction of GFPmut3b or cystatin reporter-expressing plasmids based on pMJc01, pPMQAK1 and pSB3K3

For comparative characterization of the new vector and evaluation of cystatin as a reporter, we constructed plasmids pMJc01, pPMQAK1, and pSB3K3 without reporter-expressing sequences (EV) and with 3 different expression cassettes (Table 2). For the first cassette, we chose BBa_I20260 (all BBa codes refer to the Registry of Standard Biological Parts, http://parts.igem.org/Main_Page, from which they can be retrieved as sequences or as physical DNA inserted into a plasmid backbone), a reference standard construct for relative promoter activity measurements based on GFPmut3b (Kelly et al., 2009). Plasmids pMJc01 and pSB3K3 lacking BBa_I20260 were obtained by XbaI/SpeI restriction followed by ligation. For plasmid pPMQAK1, we started from the pPMQAK1_EV plasmid (BBa_J153000), which was previously prepared in our laboratory. To make the pPMQAK1_BBa_I20260 plasmid, we cut the pPMQAK1_EV backbone with EcoRI. and PstI. The same restriction endonucleases were used to excise BBa_I20260 from pMJc01 and clone it into the pPMQAK1 plasmid backbone.

Since the regulatory sequences in BBa_I20260 (BBa_J23101 promoter and BBa_B0032 RBS) show low activity in Synechocystis sp. PCC 6803 (Huang & Lindblad, 2013; Englund, Liang & Lindberg, 2016), we additionally prepared reporter constructs under the control of the L21_RBS*regulatory element. The designed regulatory element was made from the complementary oligonucleotides L21_RBS*_F and L21_RBS*_R and comprises the −35 element (5′-TTGACA-3′), the −10 element (5′-TATAAT-3′), and two tetO2 operators: one upstream of the −35 element and one between the −35 and −10 elements (Table 1). We chose L21 (Huang & Lindblad, 2013) because the expression of the reporter in Synechocystis was about 10-fold stronger compared to J23101 and because it did not respond to anhydrotetracyline, which was desirable in our case because we were looking for a constitutive reporter.

The coding sequence of the cystatin reporter was designed based on the amino acid sequence deposited in the UniProt database (chicken cystatin, accession number P01038). We included residues 24–139, which constitute a functional protein with 2 disulfide bridges, and reverse-translated them into the nucleotide sequence. An XbaI restriction site with start codon (5′-TCTAGATG-3′) was added to the 5′ end and a hexahistidine tag with stop codon and restriction sites was added to the 3′ end of the cystatin coding sequence (cystatin6803). The sequence was codon-optimized and synthesized by Synbio Technologies.

The GFPmut3b coding region was PCR-amplified from a recombinant pSB3K3 derivative with primers GFPmut3bF and VR (Table 1) and cut with XbaI/PstI. The same restriction endonucleases were used to cut cystatin6803 inserted by the supplier into the pUC57 plasmid. Complementary oligonucleotides containing the regulatory sequence L21_RBS* were hybridized and cut with SpeI. The reporter coding regions and the regulatory fragment were ligated with complementary XbaI/SpeI overhangs and the product was cut with XbaI. The obtained reporter gene expressing cassettes were inserted into plasmid backbones pSB3K3, pMJc01, and pPMQAK1, which were previously linearized with XbaI/PstI restriction endonucleases. E. coli cells were transformed with the constructed plasmids and colonies were verified by colony PCR using primers VF2 and VR after overnight incubation (Table 1).

Characterization of the pMJc01 vector in E. coli

The relative reporter expression from pMJc01, pSB3K3, and pPMQAK1 in E. coli was determined using GFPmut3b fluorescence at the stationary growth phase in cell lysates and in liquid cultures (similarly to Huang et al., 2010). For measurements in liquid cultures, overnight cultures of cells transformed with each of the three plasmid backbones without insert and with BBa_I20260 or L21_RBS*_GFPmut3b reporter constructs were diluted 8-fold to the final volume of 2 ml in LB medium, supplemented with kanamycin (final Abs₆₀₀ 0.10–0.14). Alternatively, for measurements of cell lysates, cells were pelleted from 2 ml overnight cultures and resuspended in TE (10 mM Tris pH 8.0, 1 mM EDTA) buffer, with volume adjusted to achieve equal cell densities in all cell suspensions. Cells were lysed by sonication for 3 × 10 s and 35 μl of the cell lysate was resuspended in two ml TE buffer. Fluorescence emission spectra (505–530 nm) were measured in glass cuvettes with 1 cm light path in Perkin Elmer LS50B fluorimeter with the excitation wavelength of 485 nm. The average of five spectra was recorded for each sample and the fluorescence intensity at the emission maximum (518 nm for cultures and 514.5 nm for cell lysates) was read. The values of fluorescence intensity in the cultures were normalized to cell density (F/Abs₆₀₀) and the average F/Abs₆₀₀ ratio of the biological triplicates of the samples with empty plasmids was subtracted from the F/Abs₆₀₀ ratios of the biological triplicates of the same plasmid with reporter construct and the average was calculated from the differences. For cell lysates, the fluorescence intensity value of cells with empty plasmid was subtracted from the intensity values of the triplicates of plasmids with reporter constructs and the average was calculated from the differences.

Quantification of restriction fragments in the agarose gel was used to evaluate the absolute copy numbers of pMJc01, pPMQAK1, and pSB3K3. This approach to plasmid copy number determination is based on the established densitometric technique described by Projan, Carleton & Novick (1983), but we adapted it for comparison of plasmids with different sizes so that only the fluorescence of a restriction fragment of the same size was quantified in all the plasmids. We isolated pMJc01, pPMQAK1, and pSB3K3 plasmids with the L21_RBS*_GFPmut3b insert from approximately 5 × 10⁹ E. coli cells from overnight cultures and used NotI for digestion, resulting in linear plasmid backbones and a 943-bp fragment encoding the reporter. Restriction products were loaded alongside various amounts of DNA standards (100 bp DNA ladder from New England Biolabs and GeneRuler 1 kb DNA ladder from Thermo Scientific, Waltham, MA, United States), separated in a 1.5% agarose gel with ethidium bromide, and visualized under UV light. The gel image was analyzed using GelAnalyzer software, and the 943-bp fragment was quantified using a calibration curve generated with known amounts of 1,000-bp fragments in the ladders used. For pMJc01 and pSB3K3, we also performed restriction with BspHI, resulting in a 2,128-bp fragment from both plasmids, and used known amounts of the 2,000-bp marker for quantification.

Measurements of GFPmut3b fluorescence at different growth stages were performed on E. coli transformed with pMJc01_L21_RBS*_GFPmut3b and pMJc01_EV. We took culture samples from the liquid cultures (grown in biological triplicates) at different time points and diluted them to Abs₆₀₀ of 0.1. Fluorescence measurements were performed as described above. Since all samples were diluted to equal cell densities, no normalization of fluorescence values was performed.

Characterization of cystatin reporter in E. coli

To assess the expression of recombinant chicken cystatin in E. coli, we used an adapted papain inhibition assay (Kopitar et al., 1978). After overnight cultivation, cells were pelleted from 1.5 ml of bacterial culture and suspended in TE buffer, with the volume adjusted to achieve equal cell densities in all samples. Cells were lysed by sonication for 3 × 10 s and lysates were centrifuged at 14,000g for 2 min. For the inhibition assay, 35 µl of the soluble fraction was used together with 175 µl of 100 mM phosphate buffer (pH 6.0), 105 µl of TE buffer and 35 µl of papain solution (35 µg/ml) in phosphate buffer. Two controls were used: a blank without enzyme to determine background absorbance (used to set zero absorbance on the spectrophotometer) and a positive control without cell lysate to confirm the enzymatic activity of the papain used. After incubation of the samples at 37 °C for 5 min, 35 µl of the substrate BANA (N-benzoyl-DL-arginine naphthylamide hydrochloride) was added (6.7 mg/ml, freshly diluted 1:5 in phosphate buffer from DMSO stock solution) and incubated at 37 °C for a further 10 min. The enzyme reaction was terminated by adding 385 µl of a 1:1 mixture of Fast Garnet solution in 4% Brij (0.3 mg/ml) and reagent III (50 mM 4-chloromercuribenzoic acid, 60 mM NaOH, 60 mM EDTA, pH 6). After 5 min at room temperature, the absorbance at 550 nm was measured for all samples and the average values for biological triplicates of cells with cystatin expression vector were subtracted from the average values for samples of cells with empty vector.

Characterization of pMJc01 vector and chicken cystatin reporter in Synechocystis sp. PCC 6803

Three biological replicates of Synechocystis sp. PCC 6803 transformed with empty pMJc01 or pMJc01 with two GFP reporter constructs (BBa_I20260_GFPmut3b or L21_RBS*_GFPmut3b) were grown to exponential phase (Abs₇₃₀ = 0.4–0.8) under standard conditions. Depending on the Abs₇₃₀ values, 1–2 ml of the cultures were centrifuged and the pelleted cells resuspended in 300 µl TE buffer. Cells were lysed by sonication (4 × 1 min) and 52.5 µl of the cell lysate was suspended in two ml BG11. This resulted in an equivalent number of cells as used for fluorescence measurements in cultures, with all S. sp. PCC 6803 cultures diluted to Abs₇₃₀ = 0.1. Fluorescence measurements were performed as described for E. coli.

Measurements of GFPmut3b fluorescence at different growth stages were performed using S. sp. PCC 6803 transformed with pMJc01_L21_RBS*_GFPmut3b and pMJc01_EV as for E. coli. Cells were diluted to equal cell densities (Abs₇₃₀) at all growth stages.

For the papain inhibition assay, cells were prepared as for the GFP fluorescence measurements in cell lysates and 100 µl of the lysate was used for the assay, which was performed as described above for E. coli, except that the reaction mixture contained 40 µl of TE buffer instead of 105 µl.

Sequence submission

Nucleotide sequence of pMJc01 with inserted codon-optimized cystatin expression cassette and L21_RBS* regulatory element was deposited in the GenBank database under accession number MN201591.

Construction of pMJc01, a new shuttle vector for cyanobacteria

A broadhost range vector pMJc01 was successfully constructed using overlap extension PCR of fragments from plasmids pSB3K3 (Shetty, Endy & Knight, 2008; cloning site and kanamycin resistance cassette) and RSFmob-I (Katashkina et al., 2007; oriV and rep regions), as shown in Fig. 1. The originally constructed vector also contained the BBa_I20260 BioBrick with GFPmut3b reporter derived from pSB3K3 and was subsequently used to prepare pMJc01 empty vector (EV) and two additional pMJc01 reporter vectors with either GFPmut3b or cystatin reporters under the transcriptional control of L21_RBS*. We prepared this additional regulatory sequence because the BBa_J23101 promoter and BBa_B0032 RBS in the BBa_I20260 BioBrick show low activity in Synechocystis sp. PCC 6803 (Huang & Lindblad, 2013; Englund, Liang & Lindberg, 2016). The L21 promoter is based on the TetR-repressible promoter BBa_R0040, which previously showed constitutive expression in S. sp. PCC 6803 at approximately 9-fold increase in strength compared to BBa_J23101 (Huang & Lindblad, 2013). RBS* contains nucleotides complementary to the anti-Shine-Dalgarno sequence of S. sp. PCC 6803 and exhibits higher activity in both E. coli and S. sp. PCC 6803 compared to BBa_B0032 (Heidorn et al., 2011; Englund, Liang & Lindberg, 2016). For comparison with pMJc01, we also prepared pSB3K3 and pPMQAK1 with the same three reporter cassettes as in pMJc01. For the full list of vectors used, see Table 2.

Sequencing of pMJc01 revealed two same-sense point mutations in the kanamycin resistance gene and a single-base deletion in the His terminator downstream of the cloning site, compared to the sequence of pSB3K3 (retrieved from Registry of Standard Biological Parts). For the mutation detected in the His terminator, we used the RNAfold tool (Hofacker, 2003) to determine the possible effects on the secondary structure and function of the terminator. The deleted nucleotide formed a bulge in the double-stranded stem of the original sequence, which led us to conclude that its absence should not affect the functionality of the region. Comparison of the pMJc01 sequence with the RSFmob-I sequence (GenBank ID: EF467360.1) revealed a two-base substitution in oriV (CC to AA, positioned between two iterons) and a single-nucleotide deletion and insertion spaced 6 bp apart in the repC region, causing a short frameshift and substitutions of 3 amino acids (E107D, C108A, H109M). The same two point mutations were observed by Huang and coworkers (Huang et al., 2010) in the sequence of the pPMQAK1 plasmid, whose repC region was derived from pAWG1.1 (another RSF1010 derivative plasmid). This indicates that these are most likely not PCR-generated artefacts, but the exact sequence that was originally incorrectly determined for the RSF1010 plasmid.

pMJc01 is stable in E. coli and allows approximately twofold higher reporter protein expression compared to pPMQAK1

To determine the replication capacity of pMJc01 in E. coli, we used the reporter vector pMJc01_L21_RBS*_GFPmut3b (Table 2). We followed GFPmut3b fluorescence at different time points of cell growth in liquid cultures (Fig. 2A), as described in Methods. Fluorescence was detected at all growth stages with an increase at stationary phase. Over the course of 5 months, we occasionally isolated pMJc01 from E. coli and checked its size and integrity by agarose gel electrophoresis. Reporter expression assays were also repeated several times during this period. We observed no changes in the structure of pMJc01 or in the expression levels driven by this plasmid. On the other hand, pMJc01 showed some structural instability when transferred into the E. coli DH5α strain, a phenomenon that is detailed and discussed in Supplemental Article S1.

Reporter protein expression in E. coli was compared between pMJc01, pPMQAK1, and pSB3K3 by measuring GFPmut3b fluorescence in the stationary growth phase (Fig. 3). We used two GFPmut3b expression cassettes, which allowed validation of the regulatory sequence L21_RBS* and its comparison with the regulatory sequence J23101_B0032 from BBa_I20260, a reference standard construct in promoter strength measurements (Kelly et al., 2009).

Results of relative fluorescence intensities for different plasmids were similar when measurements were performed on bacterial cultures or cell lysates. Also, the results were comparable when GFPmut3b was expressed under the control of L21_RBS* or J23101_B0032 regulatory sequences (see Heidorn et al., 2011 for the description of ribosome binding sites and Huang & Lindblad, 2013 for promoters). Based on the fluorescence intensities, it can be concluded that pMJc01 allows about twofold higher expression compared to pPMQAK1. The expression from pSB3K3 was slightly higher than with pPMQAK1. As judged from fluorescence of cell lysates, the expression ratios pPMQAK1:pSB3K3:pMJc01 were 1:1.3:2.5 for reporter constructs with J23101_B0032 regulatory sequence and 1:1.4:1.9 for reporter constructs with L21_RBS* regulatory sequence. Expression ratios based on cell culture fluorescence were 1:0.5:1.7 (for the reporter construct with J23101_B0032 regulatory sequence; the lower value for pSB3K3 fluorescence was attributed to measurement artefact due to greater differences in sample turbidity, see Supplemental Article S1) and 1:1.1:2.3 (for the reporter construct with L21_RBS* regulatory sequence), respectively.

Since we expected the higher expression levels with pMJc01 to be explained mainly by the higher plasmid copy numbers in E. coli, we decided to roughly determine the absolute average copy numbers of pMJc01 per E. coli cell. For this purpose, we used the quantification of restriction fragments after agarose gel electrophoresis. As a control and for comparison, we also included pSB3K3 in the analysis. Digestion of pMJc01 and pSB3K3 with NotI using two different amounts of isolated plasmids revealed the presence of approximately 18 copies for pSB3K3 and 26 copies for pMJc01. In addition, we verified these results with BspHI. restriction of isolated plasmids, resulting in 17 and 25 copies for pSB3K3 and pMJc01, respectively. We also attempted to include pPMQAK1 in the analysis, but the amount of plasmid isolated from the same number of cells as for the other two plasmids was always too low for quantification. Presumably, this was not due to lower copy number but rather to loss of the plasmid during alkaline lysis (see “Discussion”).

Fluorescence measurements also show that the regulatory sequence L21_RBS* in E. coli is about 3-fold stronger than J23101_B0032 with reporter fluorescence ratios in cell lysates of 1:3.2, 1:3.4, and 1:2.4 when measured with pPMQAK1, pSB3K3, and pMJc01, respectively, and 1:2.8, 1:6.6, and 1:3.7 in cell cultures. The fluorescence determined with pSB3K3 in cultures differed significantly from the other values, which we attributed to a bias in the fluorescence measurement of samples with different turbidity, as mentioned above and shown in Supplemental Article S1.

Plasmid characterization with cystatin as reporter gave comparable results to GFPmut3b in E. coli

To evaluate cystatin as a reporter for E. coli and cyanobacteria, a new expression construct with L21_RBS* regulatory sequence was prepared. The complete sequence of the pMJc01 vector with the reporter construct is deposited in GenBank under accession number MN201591. We used this construct to compare the relative copy numbers of pMJc01, pSB3K3, and pPMQAK1 in E. coli (Fig. 4A). Two separate experiments were performed, both resulting in ratios comparable to those obtained with GFPmut3b reporter (1:1.0:2.0 and 1.0:1.3:2.5 in the first and second experiments, respectively, with cystatin as reporter). In addition, we examined the linearity of the response of the papain inhibition assay used (Fig. 4B) and obtained results that were in good agreement with the expected ratio of the signals.

Since EV controls were also included in the experiment, we were able to evaluate the inhibitory effect of the chassis on the papain inhibition assay. The A₅₅₀ values of the samples containing lysates of cells with EV were only slightly lower compared to the control without the addition of cell lysate, therefore we concluded that there are no endogenous inhibitors in E. coli that could interfere with the cystatin reporter assay.

pMJc01 is stable in Synechocystis sp. PCC 6803 and enables strong expression of GFPmut3b

To determine the replication ability of pMJc01 in Synechocystis sp. PCC 6803 at different growth stages, we used the reporter vector pMJc01_L21_RBS*_GFPmut3b (Fig. 2B). The fluorescence signal showed quite stable intensity values, which slightly decreased in the exponential growth stage. After 3 weeks of cultivation, no decrease in fluorescence was detected, demonstrating stable replication of pMJc01 in Synechocystis sp. PCC 6803 under antibiotic selection pressure. We also examined the amount of expressed GFPmut3b reporter in the stationary growth phase using SDS-PAGE and Coomassie Brilliant Blue staining and saw a distinct band of the appropriate size (26.9 kDa, Fig. 5B), indicating strong reporter expression under the regulatory control of L21_RBS*.

In Synechocystis sp. PCC 6803, the regulatory sequences analysed with pMJc01 showed similar reporter fluorescence values when measured in cultures or cell lysates (Fig. 5A). The relative activity of L21_RBS* compared to J23101_B0032 was even higher than in E. coli, with approximately 14-fold higher activity of L21_RBS* (12.4-fold when measured with cell lysates and 15.4-fold when measured with cultures).

Cystatin is a suitable reporter protein for Synechocystis sp. PCC 6803 and showed approximately 5-fold stronger expression from pMJc01 compared to pPMQAK1

The cystatin expression construct was also analysed in Synechocystis sp. PCC 6803 (Fig. 4C) with vectors pMJc01 and pPMQAK1. We were able to implement the papain inhibition assay for use with Synechocystis sp. PCC 6803 without significant modifications. We only had to optimise the cell lysis step, as this can be more challenging compared to E. coli. Several different approaches were tested (see Supplemental Article S1), but we found that sonication was best for this purpose. Although chemical methods also enabled effective cell lysis or facilitated sonication by decreasing the sonication times required, the chemicals used interfered with the inhibition assay as controls without cell lysate but with lysis buffer added also showed strong papain inhibition. The sonication time was optimised so that after the last sonication interval and centrifugation for 5 min at 14,000 g there was no green pellet left, indicating complete lysis. We again used EV controls to determine the background papain inhibition of the chassis and found that it was negligible, although slightly higher than in E. coli. The designed assay was used to compare pMJc01 and pPMQAK1 in Synechocystis sp. PCC 6803. The amount of cystatin detected was approximately 5-fold higher in cells transformed with pMJc01 than in cells transformed with pPMQAK1 (Fig. 4C).