Nickel-inducible lysis system in Synechocystis sp. PCC 6803

Abstract

We designed and constructed a controllable inducing lysis system in Synechocystis sp. PCC 6803 to facilitate extracting lipids for biofuel production. Several bacteriophage-derived lysis genes were integrated into the genome and placed downstream of a nickel-inducible signal transduction system. We applied 3 strategies: (i) directly using the phage lysis cassette, (ii) constitutively expressing endolysin genes while restricting holin genes, and (iii) combining lysis genes from different phages. Significant autolysis was induced in the Synechocystis sp. PCC 6803 cells with this system by the addition of NiSO₄. Our inducible cyanobacterial lysing system eliminates the need for mechanical or chemical cell breakage and could facilitate recovery of biofuel from cyanobacteria.

Photosynthetic microorganisms, including eukaryotic algae and cyanobacteria, are being optimized to overproduce numerous products of value (1). Because of the global energy shortage and climate change caused by greenhouse gas emission, the scientific community has focused on developing renewable biofuels from photosynthetic microorganisms (2). Cyanobacteria are excellent organisms for biofuel production. Unlike algae, their genomes are relatively easy to manipulate. They are efficient at converting solar energy, and, unlike energy crops, they can be grown on non-arable land (3). We thus have selected Cyanobacterium Synechocystis sp. PCC 6803 (hereafter referred to as Synechocystis 6803) as a model organism to develop methods for easy recovery of lipids for use in biofuel production.

The first goal of our research was to disrupt the cyanobacterial cell envelope to facilitate lipid recovery from biomass. Seog et al. (4) used various methods, such as sonication, French press, bead-beater, and lyophilization, to disrupt the cell envelops of the alga Botryococcus braunii. Extraction was 1.96 times more efficient using the bead-beater method than by using solvents alone, suggesting that a proper method of cell disruption could facilitate lipid extraction by damaging the cell wall. However, the bead-beater method is not economical for large amounts of biomass. There are some alternative cell-breakage methods, e.g., pulsed electric field (5) and hydrolytic enzymes (6), but all these methods add additional cost and reduce the overall utility of the process. Our strategy is simply to make the cyanobacteria lyse at the appropriate time.

The cyanobacterial cell envelope is composed of 4 layers (7) (Fig. 1): the external surface layers (such as S-layers and carbohydrate structures), the outer membrane, the polypeptidoglycan layer (8), and the cytoplasmic membrane. Despite the overall gram-negative structure, the peptidoglycan layer found in cyanobacteria is considerably thicker than that of most gram-negative bacteria (8). In addition, the degree of crosslinking between the peptidoglycan chains within the cell wall layer of cyanobacteria (56–63%) is far higher than that in most gram-negative bacteria (20–33%) (9).

Fig. 1.

The envelope layers of wild-type Synechocystis cell. The S-layer (asterisk), outer membrane (white arrowhead), peptidoglycan layer (arrow), and cytoplasmic membrane (black arrowhead) are indicated. (Scale bar, 50 nm.) (This picture was kindly provided by Robert Roberson, School of Life Sciences, Arizona State University, Tempe, AZ).

To break up the peptidoglycan layer (8), we applied the holin-endolysin lysis strategy used by bacteriophages (10) to exit bacterial cells. For most phages, the infection cycle terminates with programmed lysis of the host by phage-encoded proteins: endolysin (also called “lysin”), and holin, a small membrane protein that activates endolysin (11). Endolysins are peptidoglycan-degrading enzymes that attack the covalent linkages of the peptidoglycans that maintain the integrity of the cell wall (12). Four enzyme activities are associated with endolysins: muramidase (EC 3.2.1.17; e.g., T4 E or P22 gp19), which hydrolyzes the 1–4 glycosidic bonds (13); transglycosylase (EC 2.4; e.g., λ R or P2 K), which attacks the same bond but forms a muramic acid product (14); amidase (EC 3.5.1.4; e.g., T7 gp3.5), which hydrolyzes the amide bond in the oligopeptide crosslinking chains (15); and endopeptidase (EC 3.4.21–24; e.g., phi11 lysin), which attacks the crosslinking peptide bonds (16). In addition to endolysins, some auxiliary lysis factors, such as P22 gp15 and λ Rz, are involved in cleaving the oligopeptide linkages between the peptidoglycan and the outer membrane lipoprotein (17). Holins are small membrane proteins that produce nonspecific lesions (holes) in the cytoplasmic membrane from within, allow the endolysins and auxiliary lysis factors to gain access to the polypeptidoglycan layers, and trigger the lysis process (Fig. 2) (18). By using a holin as a molecular timer for lysis, the phage ensures that the membrane integrity of the host cell is maintained while the progeny virions are being synthesized and assembled, and at the same time an excess of endolysin is accumulated (19, 20).

Fig. 2.

The functions of holin and endolysins in degrading the cell wall. Holin (H) makes non-specific lesions in the cytoplasmic membrane so that the endolysins (A, amidase; M, muramidase; T, transglycosylase; and E, endopeptidase) can reach the peptidoglycan layer and break the glycosidic bonds between N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM) (by M or T), and also break the peptide cross-bridges (P-P) that link the rows of sugars together (by A or E).

A nickel sensing/responding signal system (21) was used to control the timing of the expression of phage lysis genes in Synechocystis 6803. In Synechocystis 6803, the nrsBACD operon encodes 4 proteins involved in Ni²⁺ resistance. Upstream from nrsBACD, 2 genes, nrsR and nrsS, encode a 2-component signal transduction system involved in Ni²⁺ sensing and induction of the nrsBACD operon (22). It was suggested that the presence of Ni²⁺ stimulates the kinase activity of NrsS which transfers a phosphate group to NrsR. Phosphorylated NrsR binds to the nrsRS-nrsBACD intergenic region, activating the transcription of nrsBACD genes and positively autoregulating its own synthesis. The amount of nrsBACD mRNA increased about 20-fold within 4 h after Ni²⁺ addition.

The phage lysis genes for controlled bacterial lysis have been adapted previously for accelerated cheese ripening (23). In this study, we genetically programmed a cell wall disruption process by controlling the phage lysis genes introduced into the Synechocystis 6803 genome. Different strategies for expression of holin and endolysin genes were designed to optimize the lysis efficiency of the system.

Results

Construction of Synechocystis 6803-Inducible Lysis Systems.

To test the feasibility of the lysis system, we initially introduced the Salmonella phage P22 lysis cassette (13 19 15) with a kanamycin-resistance selection marker (K_m^R) downstream of the promoter P_nrsB, resulting in strain SD101 (Fig. 3). On the basis of the successful inducible lysis of SD101, 3 strategies (Fig. 3 and Table S1) were designed to optimize the system. Strategy 1 used the lysis genes from P22 (SD121) and λ (SD122), respectively, to test the lysing abilities of different lysis enzymes. It was observed that SD122 failed to lyse on Ni²⁺-containing BG-11 agar plates, and its lysis rate in BG-11 liquid culture after Ni²⁺ induction was significantly slower than that of SD121 (Table 1). These observations led us to use P22 lysis genes for further optimization of the regulatable lysis system.

Fig. 3.

The strains and strategies used in this study. Km^R, kanamycin-resistance cassette; nrsBACD, nickel-resistance genes; nrsRS, nickel-sensing and -responding genes; P_psbAII, promoter of Synechocystis gene psbAII; P_nrsB, the nickel inducible promoter; S, R, and Rz, coliphage λ genes: holin (S), endolysin (R), and auxiliary lysis enzyme (Rz); sacB, sacB gene, which is lethal for cyanobacteria in the presence of sucrose; TT, transcriptional terminator from cyanophage Pf-WMP4; 13, 19, and 15, Salmonella phage P22 genes: holin (13), endolysin (19), and auxiliary lysis enzyme (15).

Table 1.

Comparison of different lysis strategies

Strain SD No.	Lysis strategies and descriptions	Doubling time* (hour)	Mutation rate* (10−9/generation)		Death rate* (%/hour)
			7 μM Ni2+	20 μM Ni2+	7 μM Ni2+	20 μM Ni2+	50 μM Ni2+
SD100	Wild-type Synechocystis	8.1 ± 0.7	—	—	—	—	—
SD103	Only control phage P22 holin gene 13	9.9 ± 0.8	48.2 ± 5.7	17.8 ± 2.4	29.5 ± 2.4	37.6 ± 1.0	43.8 ± 0.5
SD121	Strategy 1, using P22 lysis cassette (13 19 15)	11.1 ± 1.2	15.0 ± 1.2	9.4 ± 1.1	45.4 ± 1.8	48.7 ± 2.1	53.3 ± 0.8
SD122	Strategy 1, using phage λ lysis cassette (S R Rz)	15.1 ± 1.4	—	—	7.5 ± 3.2	11.5 ± 3.2	14.1 ± 2.8
SD123	Strategy 2, control P22 holin gene (13), which constitutively express endolysin genes (19 15)	14.1 ± 0.8	3.1 ± 0.02	2.5 ± 0.05	54.5 ± 0.7	57.4 ± 0.1	60.5 ± 0.1
SD127	Strategy 3, combination of P22 and λ lysis genes	17.9 ± 0.7	1.3 ± 0.01	0.8 ± 0.01	57.5 ± 0.03	60.3 ± 0.10	62.2 ± 0.2

*The growth and experimental conditions for doubling time, mutation rate, and lysis rate are defined in Materials and Methods.

Strategy 2 was designed to overexpress the endolysin and the auxiliary lysis factor genes (P22 19 15) under a constitutive promoter P_psbAII while regulating the holin genes (P22 13) using P_nrsB. We assumed that, before induced expression of the holin gene, the lysis enzymes would accumulate in the cytosol. Once the holin gene is expressed, the holins should produce holes in the cytoplasmic membrane and allow the accumulated lysis enzymes to degrade the peptidoglycan layer. The P_psbAII 19 15 cassette with a transcriptional terminator TP4 from cyanophage Pf-WMP4 (24) was inserted in different transcription orientations in SD123 and SD124 (Fig. 3, Table S1). SD128 was made by removing the auxiliary lysis gene 15 from SD123. Strategy 3 was to incorporate the lysis genes from λ with P22 lysis genes, with the assumption that different lysozymes attacking different bonds in the cell envelope would result in a faster lysis rate. Because the constitutively expressing cassette P_psbAII R Rz is lethal for Escherichia coli on cloning vectors, this cassette was transformed with an intermediate strain, SD126, as an overlapping PCR fragment (25) to result ultimately in SD127, a strain with lysis genes for both λ and P22.

Two-step double-crossover homologous exchanges using a K_m^R-sacB cassette (26) were applied for introducing lysis genes into the Synechocystis 6803 chromosome without leaving residual drug markers (Fig. S1). Because rapidly growing cyanobacteria are polyploid, and only 1 chromosome is involved in the initial recombination event, segregation without applying selection pressure is necessary. The phenotypic and segregation lags for recessive sucrose survival (5 days) are longer than those for dominant kanamycin resistance (1 day), because sucrose survival occurs after all chromosomes have the sacB gene fully removed, whereas kanamycin resistance occurs after a sufficient number of chromosomes have the resistance gene expressed. Essentially, the selected colonies are genotypic mixtures of cells, so isolating and testing colonies derived from a single cell after full segregation is necessary to obtain a genetically homogenous recombinant strain.

Growth of the Recombinant Strains and Genetic Stability.

The growth curves of the recombinant strains showed that the SD strains exhibited exponential growth in the cell density range of ≈10⁶-10⁸ cells/mL (Fig. S2). The doubling times calculated are given in Table 1. The growth rates for all the strains with lysis genes are significantly slower than the growth rate for the wild-type parental strain. The doubling time for SD128 (with the P22 gene 19 constitutively expressed) was 12.7 ± 0.7 h, faster than that of SD123 (14.1 ± 0.8 h, with both the P22 genes 19 and 15 constitutively expressed).

The stability of lysis genes was tested over a 75-generation period of continuous growth in liquid cultures. The presence of insertions was identified by PCR (Fig. S3). DNA sequencing data showed that all the sequences of the insertions were correct and also proved that the lysis genes were genetically stable in the Synechocystis 6803 genome over a period of 75 cell divisions (Fig. S4).

Over this period of 75 cell divisions, the frequency of Ni²⁺-resistance was evaluated by the survival ratio of the culture samples on BG-11 agar plates containing 7.0 μM and 20 μM Ni²⁺, respectively. This experiment was not applicable to SD122, because SD122 cells with the λ cassette cannot be induced to lyse on BG-11 agar plates containing Ni²⁺. As shown in Fig. S5, the frequency of resistance increased from an initial level of 10⁻⁷. At every time point, the colonies of resistant cells occurring on BG-11 agar plates containing 7.0 μM Ni²⁺ were larger than those on agar plates containing 20 μM Ni²⁺. The mutation rates for Ni²⁺ resistance (Table 1) were calculated for the first 45 generations starting with the inoculation of cultures with cells from a single colony.

Responses of Mutant Strains to the Addition of Ni²⁺.

After addition of NiSO₄, a death response was induced in the recombinant cells, which usually was accompanied by foaming of the culture. Induced cell death was measured by determining the decrease of viable cell titers as cfu/mL. The death responses of SD strains with addition of 7.0, 20, and 50 μM Ni²⁺ show that the death rates of different strains increased and became closer to each other at the higher Ni²⁺ concentrations (Fig. S6), up to a saturated level of about 60% cell death per hour (Table 1).

The cell permeability created by the synthesis of lysis enzymes on the cell envelope was indicated by penetration of the SYTOX Green nucleic acid stain. The stain easily penetrates the compromised cell envelope but does not cross the membranes of living cells (27). After brief incubation with the SYTOX Green stain, the nucleic acids of permeable cells fluoresce bright green when excited with 450- to 490-nm spectral sources; the green fluorescence masks the red fluorescence emitted by the photosynthetic pigments (Fig. S7). The penetration rates were slower than death rates in lysing cultures after the addition of Ni²⁺ (Fig. 4). Significant leakage of the cellular proteins, DNA, and phycocyanin into the culture was detected after Ni²⁺ induction (Fig. 4). The leakage rate of pigment phycocyanin was faster than that of DNA and other proteins.

Fig. 4.

The lysis response of SD123 after addition of 7 μM Ni²⁺. The value of viable cell density (cfa) and protein, DNA, and phycocyanin concentrations in the supernatant were normalized into percentage by their maximum values (i.e., 2.25 × 10⁷ cells/mL for the cell density, 0.20 mg/mL for proteins, 13.4 mg/mL for DNA, and 22.5 emission intensity at 650 nm for phycocyanin). The percentage of permeable cells of the total cells was not normalized.

Destruction of Cell Wall After Addition of Ni²⁺.

The transmission electron microscopy images of SD121 show that the expression of lysis genes causes the cell wall (peptidoglycan layers) to decrease in thickness 6 and 12 h after 7.0 μM Ni²⁺ induction and that the cell was degraded into a ghost 24 h after Ni²⁺ induction (Fig. 5).

Fig. 5.

Transmission electron microscopy images of the SD121 cells before and after the addition of 7 μM Ni²⁺. (A) SD121 cells before Ni²⁺ addition; (B) 6 h after Ni²⁺ addition; (C) 12 h after Ni²⁺ addition; (D) 24 h after Ni²⁺ addition.

Discussion

Programmed autolysis for cyanobacteria is a technique that avoids the energy requirements or environmental hazards involved in biofuel recovery from cyanobacterial biomass. This basic idea was verified by the first construction (SD101), which demonstrated the feasibility of Ni²⁺-inducible cell death by controlling synthesis of bacteriophage holins and endolysins in cyanobacteria. We plan further tests of the lysis ability of lysozymes from other phages, especially from cyanophages, and expect to find a more efficient lysis system for cyanobacteria.

The holin-endolysin system is essential for host lysis by most double-stranded DNA phages. In this system, holins are the protein clocks controlling the length of the infective bacteriophage cycle to achieve lysis at an optimal time to generate progeny phage and so are subject to intense evolutionary pressure (18). The idea for our Strategy 2 originated from this natural mechanism of precise and immediate lysis. As expected, the Strategy 2 strain SD123 exhibited a significantly faster lysis rate than the Strategy 1 strain SD121 (Table 1), suggesting that the accumulated endolysins caused a collapse in the cell wall after crossing the holin-mediated permeabilizing lesions.

It was observed that all the strains recombinant with phage lysis genes downstream of P_nrsB exhibited slower growth rates than the wild-type parent, perhaps caused by the background transcription of the lysis genes driven by basal P_nrsB activity before Ni²⁺ induction (22). It also was observed that the accumulation of intracellular lysis enzymes (P22 19 15 and λ R Rz) resulted in even slower growth rates (Table 1) and that growing SD123 and SD127 cells adhered to each other and to the vessel walls. These phenomena suggested that the cell walls of these strains were compromised before induction, perhaps by leakage of the endolysins or damage to membrane linkages by the auxiliary lysis enzymes (15 or Rz).

The death rate was faster than the rate at which cells become penetrable by the fluorescent dye (Fig. 4). The death rates, which correspond to the cells that have been induced to synthesize the lysis enzymes to a lethal level, were calculated from viable cell (cfu) counting. The cells stop forming cfus long before they are actually permeable to the dye, because it takes time for the lysis enzymes to damage the cytoplasmic membrane to enable entry of the fluorescent DNA-binding dye. Essential macromolecules and metabolites leaked out through the compromised cell envelopes (Fig. 4), contributing to the destruction of the whole cell (Fig. 5).

Missense mutations in the regulator gene nrsRS, in the promoter P_nrsB, in the binding site for the phosphorylated NrsR, or in lysis genes could cause Ni²⁺ resistance. There were fewer resistant colonies on BG-11 agar plates containing 20 μM Ni²⁺ than on BG-11 agar plates containing 7 μM Ni²⁺ (Fig. S5). This observation suggested that the resistance mutations were recessive, indicating that the phenotype of Ni²⁺ resistance occurred after the resistance mutations were present on all chromosomes. SD127 has the lowest rate of mutation to Ni²⁺ resistance (Table 1). It is possible that the slower-growing strains (e.g., SD127) take a longer time for segregation of the resistance mutation. In addition, more lysis gene backups, such as the 6 lysis genes in SD127, also resulted in a lower mutation rate to Ni²⁺ resistance.

Our study showed that the lysis genes from bacteriophages are functional and controllable in cyanobacteria. The nickel-sensing/responding system served as the trial regulator because the inducer Ni²⁺ is easy to control. Improvements in the lysis system also are expected, because the lysis genes can be controlled by other cyanobacterial inducible promoters responding to other signals such as nutrient starvation (28), growth phase (29), environmental stress (30), or other metal ions (21). The cyanobacteria programmed lysis system has other uses in addition to biofuel extraction. For example, it provides an alternative method for gently disrupting the cell wall and releasing cell contents. This system also could be developed as an inducible expression system for other genes in Synechocystis 6803.

Materials and Methods

Strains and Culture Conditions.

Synechocystis 6803 wild-type and mutant strains were grown at 30 °C in modified BG-11 medium buffered with 10 mM N-[tris(hydroxymethyl)methyl]-2-aminoethanesulfonic acid (TES)-NaOH (pH 8.0) (31) and bubbled with a continuous stream of filtered air under continuous illumination (50 μmol photons m⁻² s⁻¹). For growth on plates, 1.5% (wt/vol) agar and 0.3% (wt/vol) sodium thiosulfate were added to BG-11 agar. BG-11 medium was supplemented with kanamycin (50 μg/mL) for K_m^R strains. E. coli strain DH5α was grown at 37 °C on 1.5% (wt/vol) LB agar. When E. coli cells were used to replicate the plasmids harboring the lysis genes, cells were grown at 20 °C in LB broth and agitated by slow rotation (30 rpm) on a tube rotator to avoid lysis.

DNA Manipulations.

Unless indicated otherwise, standard DNA methods (32) were used. In some plasmids, DNA sequences were spliced by the PCR overlap extension method of Warrens et al. (25). Plasmids and constructions used in this study are listed in Table S1. The primers used in the constructions are listed in Table S2. The flanking sequences for double-crossover recombination were cloned into pUC18. Using PCR, the lysis gene cassettes were amplified from a Salmonella phage P22 lysate and an E. coli phage λ lysate. All the plasmid constructions were confirmed by DNA sequence analysis performed in the DNA Laboratory, School of Life Sciences, Arizona State University.

Introduction of Lysis Genes into Synechocystis 6803.

General conditions for transformation of Synechocystis 6803 have been optimized (33). As shown in Fig. S1, an intermediate strain first was constructed by introducing a K_m^R-sacB cassette (26) downstream of P_nrsB. Transformants were re-streaked on BG-11 agar plates containing 100 μg/mL kanamycin and grown for 5 days. To introduce lysis genes and remove the drug markers, the suicide vectors or DNA fragments (final concentration of 0.05–0.1 μg/mL) were incubated with the intermediate strain (at an OD_{730 nm} of 0.2–0.4). The transformation mixture was inoculated into 1 mL of BG-11 medium, grown for 5 days for segregation, and then plated onto BG-11 agar plates containing 4.5% sucrose (wt/vol). With the replacement of K_m^R-sacB cassette, the recombinant cells are able to grow in the presence of sucrose. To obtain a genetically homogenous strain, the cell suspensions were inoculated into BG-11 medium, grown for 2 days, and then the culture was diluted and plated on BG-11 agar plates. Cells from the individual colonies were verified by PCR for presence of the inserted regions and absence of the replaced regions. The inserted sequences were confirmed by DNA sequence analysis.

Growth Rate and Lysis Response Measurements.

For the growth-rate measurements, 200-mL recombinant strain cultures were grown in triplicate 250-mL flasks with aeration tubes under the previously described conditions. At 24-h intervals, cultures were sampled, and individual cells were counted in a hemocytometer. The inducible cell death responses of recombinant strains were tested by adding NiSO₄ to cultures (OD_{730 nm} ≈0.5). Cell death rates to Ni²⁺ were determined by the reduction in cfus. The protein and DNA concentrations in the lysing supernatant were measured by UV absorbance at 280 nm and 230 nm, respectively. The phycocyanin in the supernatant was measured by fluorescence emission at 650 nm excited at 620 nm.

Genetic Stability and Resistance Mutation.

For genetic stability tests, cultures were grown for 75 generations. When the culture OD_{730 nm} reached 1.2, the culture was subcultured by a 1:1,000 dilution in prewarmed medium. The segregation status and insertion sequences were verified from different subcultures using PCR. Ni²⁺ resistance frequencies were evaluated by the survival rates of the culture sampled on BG-11 agar plates containing Ni²⁺.

Microscopy.

The lysing cells were stained with 5 μM SYTOX Green nucleic acid stain (Invitrogen Molecular Probes, Inc.) (27) for 5 min and observed under an Axioskop40 fluorescence microscope (Zeiss). At least 400 cells were counted on the pictures taken of different samples and at time points before and after 7.0 μM Ni²⁺ addition. The degradation process of the cyanobacterial cell walls was imaged by using a CM12S transmission electron microscope (Philips Electronic Instruments) (34).

Statistical Analysis.

Most data were expressed as means ± SD. The means were evaluated with 1-way ANOVA for multiple comparisons among groups. Student's t test was used for pairwise comparisons. P < 0.05 was considered statistically significant.