Four Novel Genes Required for Optimal Photoautotrophic Growth of the Cyanobacterium Synechocystis sp. Strain PCC 6803 Identified by In Vitro Transposon Mutagenesis

Abstract

Four novel Synechocystis sp. strain PCC 6803 genes (sll1495, sll0804, slr1306, and slr1125) which encode hypothetical proteins were determined by transposon mutagenesis to be required for optimal photoautotrophic growth. Mutations were also recovered in ccmK4, a carboxysome coat protein homologue, and me, the decarboxylating NADP⁺-dependent malic enzyme. This is the first report that these known genes are required for optimal photoautotrophy.

Photosynthesis is one of the most important biological processes and occurs in a very diverse set of organisms ranging from prokaryotes to eukaryotes. Recently, much effort has been directed towards understanding the structure and function of proteins involved in photosynthesis (photosystem I, photosystem II, cytochrome b₆/f complex, Calvin-Benson cycle enzymes, etc.). While much progress has been made in the understanding of the functional organization of these proteins, relatively little is known concerning the organization of other protein components which must be involved in the regulation, assembly, and turnover of the proteins involved in photosynthesis. Cyanobacteria are photoautotrophic gram-negative eubacteria capable of performing oxygenic photosynthesis in a manner quite similar to that in eukaryotic algae and higher plants. Synechocystis sp. strain PCC 6803 is a naturally competent unicellular cyanobacterium and has proved to be one of the best model organisms for studying the mechanism and regulation of oxygenic photosynthesis (15). We are interested in identifying the genes required for oxygenic photosynthesis. In this study, we used a hyperactive Tn5-based in vitro transposition system to introduce random insertional mutations into Synechocystis and have identified a number of mutants which are incapable of undergoing optimal photoautotrophic growth. Here we describe the production, identification, and characterization of a number of these mutants. The structure and possible function of the affected genes in these mutants will also be discussed.

A glucose-tolerant strain of Synechocystis sp., PCC 6803 (15), was used as a parental control and as the DNA recipient strain in the present study. Cells of both the control strain and the derivative photosynthetic mutants were maintained under photoheterotrophic growth conditions at 30°C with a light intensity of 20 μmol of photons m⁻² s⁻¹ (fluorescent light) in liquid BG-11 growth medium (ATCC medium 616) supplemented with 10 mM TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid]-KOH (pH 8.2), 5 mM glucose, and 10 μM DCMU [N-(3,4-dichlorophenyl)-N′-dimethylurea]. Liquid cultures were bubbled continuously with air. For autotrophic cell culture, the glucose and DCMU were omitted. For cultures grown on plates, the BG-11 medium was supplemented with 1.5% agar and 0.3% sodium thiosulfate. When appropriate, kanamycin was included in the media at a final concentration of 10 μg/ml.

A Synechocystis genomic library which had been subjected to in vitro transposon mutagenesis and then amplified in Escherichia coli (1, 2) was kindly provided by D. Bhaya at the Carnegie Institute. This DNA was used to transform parental Synechocystis cells which are naturally competent and which exhibit high rates of homologous recombination (15). After selection on kanamycin-containing photoheterotrophic growth medium, followed by multiple transfers to allow sorting out, ca. 10,000 individual kanamycin-resistant cell lines were screened for the ability to grow photoautotrophically on agar plates. Those cell lines that repeatedly failed to grow photoautotrophically but which did grow photoheterotrophically (17 cell lines) were identified as putative photosynthetic mutants and selected for further analysis.

Southern blot analysis indicated that all 17 cell lines exhibited single transposon insertions (data not shown). The location of the transposon insertion was identified either by direct genomic sequencing (1) or by inverse PCR (8). Since the Synechocystis genome has been sequenced (11), the identification of the genes disrupted by the transposon mutagenesis is facile. Growth rates in liquid culture under both photoheterotrophic and photoautotrophic conditions at 30°C with a light intensity of 20 μmol of photons m⁻² s⁻¹ were monitored for 10 days. The cell sizes of the control strain and the mutants were similar (±10%) as determined by differential interference contrast microscopy. O₂ evolution activity during mid-log-phase growth was assayed by oxygen polarography with 1 mM bicarbonate as an electron acceptor.

We have determined the transposon insertion sites for 14 out of 17 mutants which we have isolated (Table 1). For three of the mutants, both direct genomic sequencing and inverse PCR failed repeatedly. Among the 14 identified mutants, 13 mutants each had a single insertion within the protein-encoding region, resulting in a premature truncation of the encoded protein. The 4YE2 mutant, unlike the other 13 mutants, had the single insertion in the 3′ noncoding region of the psbB gene. It should also be noted that for three genes, menB, menE, and ccmK4, multiple independent transposon insertions were observed (Table 1).

TABLE 1.

Photosynthetic mutants identified in this study

Mutant	Gene designationa	Name	Annotation	Function
4YE2	slr0906	psbB	PSIIb light-harvesting core protein CP47	PSII structural protein
RB12	sll0851	psbC	PSII light-harvesting core protein CP43	PSII structural protein
3DH6	sll1127	menB	Naphthoate synthasec	Menaquinone synthesis
5AH9	sll1127	menB	Naphthoate synthasec	Menaquinone synthesis
3DC8	sll1127	menB	Naphthoate synthasec	Menaquinone synthesis
AH4	slr0492	menE	O-Succinylbenzoic acid-CoA ligasec	Menaquinone synthesis
DDE5	slr0492	menE	O-Succinylbenzoic acid-CoA ligasec	Menaquinone synthesis
AAE6	slr1839	ccmK4	CO2-concentrating mechanism protein K homologuec	CO2 fixation, required for optimal photoautotrophyd
5ND9	slr1839	ccmK4	CO2-concentrating mechanism protein K homologuec	CO2 fixation, required for optimal photoautotrophyd
3WEZ	slr0721	me	Malic enzyme	Pyruvate metabolism, required for optimal photoautotrophyd
4YD9	slr1125	NAe	Probable glucosyltransferasef	Required for optimal photoautotrophyd
3ZA12	slr1306	NA	Hypothetical proteinf	Required for optimal photoautotrophyd
4BA2	sll0804	NA	Hypothetical proteinf	Required for optimal photoautotrophyd
CG4	sll1495	NA	Hypothetical proteinf	Required for optimal photoautotrophyd

^aSynechocystis sp. strain PCC 6803 open reading frame designation.

^bPSII, photosystem II.

^cPlease note that the different mutations within the same gene are independent and bear transposon insertions at different locations.

^dThis study.

^eNA, not applicable with no gene name or function yet determined.

^fCyanobase annotation (http://www.kazusa.or.jp).

A total of 10 genes were affected by the insertion of the transposon in the 14 identified mutants (Table 1). These 10 genes include six which had been previously identified and four hypothetical genes. The majority of the known genes which have been disrupted encode components known to be involved in photosynthesis. These include psbB (slr0906), which encodes the photosystem II core protein CP47, and psbC (sll0851), which encodes the photosystem II protein CP43. These are known essential structural components of photosystem II (3, 4). The ccmK4 gene (slr1839) encodes the carbon dioxide-concentrating mechanism protein homologue 4. This is one of the four carboxysome coat protein homologues present in Synechocystis. The carboxysome has been hypothesized to be involved in concentrating carbon dioxide for photosynthetic carbon fixation (5, 14). This is the first report indicating that this particular ccmK homologue is required for optimal photoautotrophy. The menB gene (sll1127) encodes the enzyme 1,4-dihydroxy-2-naphthoate synthase, and menE (slr0492) encodes the enzyme O-succinylbenzoic acid coenzyme A (CoA) ligase. These enzymes function in the menaquinone biosynthetic pathway leading to the biosynthesis of phylloquinone, the secondary electron acceptor of photosystem I (9, 10). The recovery of mutations in these known genes indicated that the screen employed in this study allows the identification of genes required for photoautotrophy.

In addition to these components, insertional inactivation of the me gene (slr0721), which encodes the malic enzyme (decarboxylating malate oxidoreductase, EC 1.1.1.39), leads to a loss of optimal photoautotrophy. This enzyme catalyzes the oxidative decarboxylation of malate into pyruvate. It is unclear, at this time, why inactivation of this gene would cause the loss of optimal photoautotrophic growth. At least two hypotheses could explain this result. First, the malic enzyme could participate in the carbon-concentrating mechanism of Synechocystis. Recently Yang et al. (16) have hypothesized that a C₄-like pathway may operate in Synechocystis involving phosphoenolpyruvate carboxylase and the malic enzyme. Their hypothesis was based on results obtained from metabolic flux analysis of the metabolism of ¹³C-labeled glucose under photoheterotrophic and mixotrophic conditions. Second, it is possible that the malic enzyme is involved in photorespiratory glycolate metabolism (6). Further studies testing these (and other) possibilities are ongoing.

The growth characteristics of the mutants 4YD9, 3ZA12, 4BA2, and CG4, which bear transposon insertions in the hypothetical genes slr1125, slr1306, sll0804, and sll1495, respectively, were further characterized. Figure 1 shows the growth curves which were obtained for these mutants in liquid culture under either photoheterotrophic (Fig. 1A) or photoautotrophic (Fig. 1B) conditions. Under photoheterotrophic conditions all of the mutants grew at rates very similar to that observed for the control strain. Under photoautotrophic conditions, however, all of the mutants exhibited significantly retarded growth rates compared to the control. The control strain grew about 4 times faster than the mutants 4BA2 and 3ZA12 and 12 times faster than the mutant 4YD9.

FIG. 1.

Growth of the mutants CG4, 4YD9, 4BA2, and 3ZA12 in comparison to the parental control Synechocystis strain. (A) Photoheterotrophic growth. (B) Photoautotrophic growth. Symbols: ▪, control strain; •, CG4; ♦, 4YD9; ▴, 4BA2; ▾, 3ZA12. These data are the averages of three experiments; standard deviations are represented by vertical bars. Please note that in some instances the error is smaller than the symbols used.

The mutant CG4 exhibited an unusual biphasic growth curve which was fully reproducible in each of three independent experiments that we performed. We have observed that this mutant exhibits a high rate of spontaneous reversion, recovering its ability to grow photoautotrophically. Sequencing of several of these revertants, however, indicated that the transposon insertion in sll1495 had been maintained (data not shown). Apparently, secondary mutations at distant unidentified sites complement the loss of sll1495 function. The accumulation of such second-site suppressor mutations which lead to growth at control rates during the photoautotrophic growth experiment may account for the biphasic growth curve observed.

We have also measured the whole-chain electron transport rates (H₂O to CO₂) in these mutants and the control strain. For most of the mutants, only modest decreases in electron transport were observed (70 to 80% of wild-type rates). In all cases, the observed rates were somewhat lower for the photoautotrophically grown cells than for those grown photoheterotrophically. Apparently, the lesions which engender the loss of optimal photoautotrophy in these mutants do not dramatically affect the ability to carry out whole-chain electron transport. Other systems such as cofactor assembly-stability, membrane assembly, photosynthate utilization, etc., must be affected in these mutants, leading to the loss of optimal photoautotrophic growth. The 4YD9 mutant, however, did exhibit significant loss of whole-chain electron transport activity, evolving oxygen at about 35% of the control rate when grown under photoautotrophic conditions. Additionally, this mutant exhibited an olive-green color which was significantly different from the brilliant blue-green coloration of the control strain and other mutant strains. Whole-cell absorption spectra for this mutant and the control strain at constant cell numbers are shown in Fig. 2. The mutant cells contain less than one-half of the chlorophyll a and no detectable phycobiloproteins. We hypothesize that this mutant may possess a defect in thylakoid membrane assembly-stability. A more complete description of the defects present in this mutant will appear elsewhere.

FIG. 2.

Absorption spectrum of the control Synechocystis strain (solid line) and the mutant 4YD9 (dashed line). The control strain has been offset for clarity.

What are the characteristics of these hypothetical genes which are required for optimal photoautotrophy? The 4YD9 mutant exhibits a transposon insertion in the slr1125 gene (Fig. 3A), which encodes a hypothetical protein of 402 amino acids and is predicted (PSORT-B, http://www.psort.org/ [7]) to be cytoplasmically localized. The protein is similar (up to 35% identity) to a few other known proteins including zeaxanthin glucosyltransferase (crtX) from Erwina uredovora, glycosyltransferase from Streptomyces coelicolor, and UDP glucuronosyltransferase from Mus musculus. Although the gene is annotated as encoding zeaxanthin glucosyltransferase in GenBank, we feel that this assignment is premature since the strongest region of similarity is in the C terminus of the protein, which is predicted to encode the glucosyltransferase domain. Similarity to other regions of the Erwinia zeaxanthin glucosyltransferase is at a significantly lower level.

FIG. 3.

Structural features of the four hypothetical genes slr1125 (A), slr1306 (B), slr0804 (C), and slr1495 (D) disrupted in the mutants 4YD9, 3ZA12, 4BA2, and CG4, respectively. Open reading frames in the vicinity of the targeted gene are shown, and the broken line below a gene indicates the direction of transcription. Putative functional domains are indicated by solid arrows. Tn, transposon; DAO, d-amino acid oxidase domain; FAD, flavin adenine dinucleotide binding domain; ATP/GTP, nucleotide-binding domain. Domains and/or motif assignments were obtained at http://www.pedant.gsf.de or, after BLAST searches, at http://www.ncbi.nlm.nih.gov.

The mutant 3ZA12 contains a transposon insertion in the slr1306 gene (Fig. 3B), which encodes a hypothetical protein of 485 amino acids. The protein contains an ATP-GTP-binding-site motif A (P loop) and has some similarities (up to 30% amino acid identities) to a number of other cyanobacterial hypothetical proteins including Slr1462 from Synechocystis and Alr4172 from Nostoc sp. Recently, this gene has been identified as one of the 181 cyanobacterial signature genes (13). These are genes which are present in the eight available cyanobacterial genomes but which are not present in other eubacteria.

The mutant 4BA2 contains a transposon insertion in the sll0804 gene (Fig. 3C), which encodes a hypothetical protein of 453 amino acids. The protein contains an ATP-GTP-binding-site motif A (P loop) and has some similarities (up to 49% identical amino acids) to several other hypothetical cyanobacterial proteins including Alr0904 from Nostoc and Sll0503 from Synechocystis. This protein is predicted to be localized to the inner membrane by PSORT-B and is predicted to contain two transmembrane α-helices in the C-terminal third of the protein. Additionally, it contains two strongly predicted coiled-coil motifs (http://www.ch.embnet.org [12]).

The mutant CG4 bears a transposon insertion in the sll1495 gene (Fig. 3D), which encodes a hypothetical protein of 397 amino acids. The Sll1495 protein appears to be a dehydrogenase and contains both a d-amino acid oxidase (DAO)-binding motif and a flavin adenine dinucleotide-binding domain. The protein has a high level of similarity (70% identical amino acids) to Alr2826 from Nostoc sp. Additionally, it shares strong similarity (50% identity) to a number of proteins which have been identified as possible transcriptional regulators, including a putative γ-aminobutyrate DTP gene cluster repressor from Salmonella sp. and the putative transcriptional regulator VCA0147 from Vibrio cholerae. This raises the interesting possibility that the Sll1495 protein may be a redox-active transcriptional regulator. No direct evidence supporting this hypothesis, however, is available at this time.

The sll1495 gene is flanked by sll1496 (mannose-1-phosphate guanyltransferase) and a valyl tRNA. These three genes could be organized in a possible operon. The insertion of the transposon in the sll1495 gene could give rise to polar effects with respect to the sll1496 gene, yielding the observed phenotype. To test this possibility, insertional mutagenesis was used to introduce a kanamycin resistance cassette in the noncoding DNA located between the sll1495 and sll1496 genes. The phenotype of this insertional mutant was identical to that of the control strain (data not shown). This indicated that the transposon insertion into the sll1495 gene was responsible for the observed phenotype of the CG4 mutant and that polar effects with respect to sll1496 were not occurring.

Transposon mutagenesis has proved to be a useful tool in the identification of genes required for photoautotrophy. In most genomes which have been sequenced, 30 to 40% of the identified genes have no known function. The identification of a substantive phenotype for such hypothetical genes is a required first step toward the elucidation of their role in cellular metabolism. While the specific functions of the hypothetical genes which we have identified in this study have not, at this time, been determined, it is clear that they are required for optimal photoautotrophic, but not photoheterotrophic, growth. We hypothesize that these genes may function in the regulation, assembly, and/or turnover of the proteins involved in the photosynthetic process.