Homologs of Circadian Clock Proteins Impact the Metabolic Switch Between Light and Dark Growth in the Cyanobacterium Synechocystis sp. PCC 6803

Abstract

The putative circadian clock system of the facultative heterotrophic cyanobacterial strain Synechocystis sp. PCC 6803 comprises the following three Kai-based systems: a KaiABC-based potential oscillator that is linked to the SasA-RpaA two-component output pathway and two additional KaiBC systems without a cognate KaiA component. Mutants lacking the genes encoding the KaiAB1C1 components or the response regulator RpaA show reduced growth in light/dark cycles and do not show heterotrophic growth in the dark. In the present study, the effect of these mutations on central metabolism was analyzed by targeted and non-targeted metabolite profiling. The strongest metabolic changes were observed in the dark in ΔrpaA and, to a lesser extent, in the ΔkaiAB1C1 mutant. These observations included the overaccumulation of 2-phosphoglycolate, which correlated with the overaccumulation of the RbcL subunit in the mutants, and taken together, these data suggest enhanced RubisCO activity in the dark. The imbalanced carbon metabolism in the ΔrpaA mutant extended to the pyruvate family of amino acids, which showed increased accumulation in the dark. Hence, the deletion of the response regulator rpaA had a more pronounced effect on metabolism than the deletion of the kai genes. The larger impact of the rpaA mutation is in agreement with previous transcriptomic analyses and likely relates to a KaiAB1C1-independent function as a transcription factor. Collectively, our data demonstrate an important role of homologs of clock proteins in Synechocystis for balanced carbon and nitrogen metabolism during light-to-dark transitions.

Keywords: RpaA; Synechocystis; carbon metabolism; circadian clock; diurnal metabolic switch.

Figure 1

Model of the Synechocystis Kai system. In Synechocystis, the KaiAB1C1 system is the closest homolog to the well-known circadian clock system from Synechococcus elongatus PCC 7942. The output signaling response of KaiC1 relies on the SasA-RpaA two-component system. The transcription factor RpaA controls gene expression in L and D conditions. RpaA can interact with other regulators (e.g., PixG and RpaB) and partially targets the same genes as the transcription factor RpaB. The KaiAB1C1 complex is intertwined with the non-standard KaiB3C3 system in Synechocystis, but the direction of signaling between the two systems remains unclear. The red circles represent phosphorylation sites.

Figure 2

Growth of Synechocystis mutants. Wild type (WT), ΔrpaA, and ΔkaiAB1C1 deletion mutants were grown in a 12-h L/12-h D cycle with ambient air (A) and 1% CO₂ (B). The light intensity during cultivation was 75 μmol photons m⁻² s⁻¹, which was increased to 150 μmol photons m⁻² s⁻¹ in (B) after the cultures reached an OD₇₅₀ of ~1. Each point represents the mean of three biological replicates (± standard deviation).

Figure 3

The ΔrpaA and ΔkaiAB1C1 strains show lower PHB accumulation at high CO₂ concentrations. PHB was quantified in cells cultivated in a 12-h L/12-h D cycle and 1% CO₂. The samples used for the PHB measurements were collected in the stationary phase at an OD₇₅₀ of ~7 in the L and D phases of the 7th L/D cycle 1 h before switching. Each bar displays the mean of three technical replicates (± standard deviation).

Figure 4

Global analysis of the metabolic L/D switch in the ΔkaiAB1C1 and ΔrpaA mutants. (A) Design of the two-factorial entrainment experiment. Wild-type and mutant cells, i.e., the first experimental factor (genotype), were precultivated under continuous light, divided into independent cultures of OD₇₅₀ = 0.6 in ambient air, and sampled in the middle of the first 12-h D phase and first 12- h L phase after entrainment, i.e., the second experimental factor (illumination). Three independent replicate cultures in each phase were analyzed by a quantitative targeted LC-MS analysis and non-targeted GC-MS metabolome profiling of the primary metabolome. (B) Hierarchical cluster analysis (HCA) of mutant and wild-type metabolomes. Annotated LC- and GC-MS metabolite data separately maximum scaled and then combined, log-transformed, autoscaled, and clustered using Pearson's correlation distance metric and average linkage. HCA includes a scale of the node height and bootstrap-support analysis by 1,000 iterations. The values represent the “approximately unbiased” bootstrap probability, and in brackets, the conventional bootstrap probability obtained using the “pvclust” R package is shown. Values of 100 represent the highest node probability. (C) Principal component analysis (PCA) of the same data. PC1 and PC3 demonstrate the variance contribution of illumination (PC1) and the ΔrpaA samples collected in the D phase (PC3) to the metabolomics data set. Yellow arrows represent the score biplot of the samples. The gray underlay indicates the loading contributions of the metabolites. A corresponding supplement contains a biplot of PC1 and PC2 and a scree plot of the variance represented by the top 10 PCs (Supplementary Figure 3).

Figure 5

Relative abundances and fold changes (FC) of metabolite concentrations in the L phase compared to the D phase. The figure includes identified metabolites that exhibited significant changes in L metabolism compared to D metabolism (Tukey test P < 0.05). Fold changes are color-coded as follows: significant increases, red (>2-FC), light red (<2-FC), significant decreases, blue (<0.5-FC), and light blue (>0.5-FC); the complete numerical data are provided in Supplementary Table 1. The grouping of the metabolites into the L/D response groups (L/D groups) was performed according to the changes in the wild type (L/D group I-II) and considered inverted (L/D group III-IV) changes in the mutants or emergent changes (L/D group V). Heat map: light yellow, low abundance; dark yellow, high abundance; gray, undetectable.

Figure 6

Summary of the metabolic changes in the mutants of Synechocystis. Venn diagram analysis of all significant (Tukey test, P < 0.05) metabolic changes among the set of identified and yet unidentified metabolites. Intersections show significant changes in the metabolite pools shared between the ΔkaiAB1C1 (brown) and ΔrpaA (blue) mutants. Zoom-ins show the contribution of the two diurnal phases to the respective sets, D phase (gray) and L phase (yellow). Note the large overlap between the mutants and the more frequent specific responses of the ΔrpaA cells.

Figure 7

Photorespiration. Box plots showing the maximum normalized relative changes in the metabolite concentrations and Tukey test results (P < 0.05, n = 3). Metabolite levels that significantly differ are indicated by lowercase letters and color coding. Note the moderate and strong increases in the RubisCO products 3PGA and 2PG in the ΔkaiAB1C1 mutant and the ΔrpaA mutant, respectively. High 2PG accumulation in the ΔrpaA mutant continues in the D phase and is associated with glycolate accumulation in the L phase. Serine and glycine respond inversely to RpaA deficiency. Glycine fails to accumulate in the D, while serine accumulates in the L. Dashed arrows indicate multiple reaction steps. Relevant enzymes and complexes include (gray) glycolate-2P phosphatase (Pgp), glycolate dehydrogenase (Glydh), aminotransferase (At), glycine decarboxylase complex (Gdc), serine hydroxymethyltransferase (Shmt), tetrahydrofolate (THF), hydroxypyruvate reductase (Hpr), and glycerate kinase (Gk).

Figure 8

Increased RubisCO abundance in mutants compared to wild type (WT). Comparison of the RubisCO levels via immunoblot analysis using anti-RbcL. Ten or fifteen microgram of crude cell extract were used for SDS-PAGE. The samples were collected 5.5 h after the transition to D or L as shown in Figure 4A.

Figure 9

Lower glycolysis, glutamate, and the pyruvate family of amino acids. Box plots showing the maximum normalized relative changes in the metabolite concentrations and Tukey test results (P < 0.05, n = 3). Metabolite levels that significantly differ are indicated by lowercase letters and color coding. Note the subtle changes in the ΔkaiAB1C1 mutant, while the ΔrpaA mutant exhibited significantly increased pyruvate family amino acids, 3PGA, PEP, and glutamate in D. These changes are accompanied by pyruvate depletion in the D phase relative to the wild type. In the L, 3PGA accumulates in the ΔkaiAB1C1 mutant relative to the wild type. In the ΔrpaA mutant, 3PGA accumulates to even higher levels, and the 2PGA and PEP increases become significant. Dashed arrows indicate multiple reaction steps. Relevant enzymes and complexes include (gray) ribulose-1,5-bisphosphate carboxylase-oxygenase (RubisCO), phosphoglycerate mutase (Pgam), enolase (Eno), phosphoenolpyruvate synthase (Peps), pyruvate kinase (Pk), pyruvate dehydrogenase (Pdh), alanine dehydrogenase (Aladh), and alanine transaminase (Alat).

Figure 10

Tricarboxylic acid pathways. Box plots showing the maximum normalized relative changes in the metabolite concentrations and Tukey test results (P < 0.05, n = 3). Metabolite levels that significantly differ are indicated by lowercase letters and color coding. Note the inverse accumulation of metabolites from the oxidative and reductive branches of the TCA pathways. The concentrations of the intermediates of the TCA reactions were largely unchanged in the ΔkaiAB1C1 and ΔrpaA cells. Dashed arrows indicate multiple reaction steps. Relevant enzymes and complexes included (gray) phosphoenolpyruvate synthase (PepS), pyruvate kinase (Pk), pyruvate dehydrogenase (Pdh), phosphoenolpyruvate carboxylase (Pepc), citrate synthase (Cs), aconitase (Acn), isocitrate dehydrogenase (Idh), succinate semialdehyde dehydrogenase (Ssadh), succinate dehydrogenase (Sdh), fumarase (Fh), malate dehydrogenase (Mdh), and malic enzyme (Me). *Citrate and Isocitrate.

Figure 11

Nitrogen assimilation and associated amino acid biosynthesis. Box plots showing the maximum normalized relative changes in the metabolite concentrations and Tukey test results (P < 0.05, n = 3). Metabolite levels that significantly differ are indicated by lowercase letters and color coding. Note the increase in aspartate and the D accumulation of glutamate in the ΔrpaA mutant. Relevant enzymes and complexes include (gray) glutamine synthetase (Gs), glutamine 2-oxoglutarate aminotransferase (Gogat), aspartate aminotransferase (Aspat), asparagine synthetase (As), aspartate decarboxylase (Aspdc), and glutamate N-acetyltransferase (Glunact).