Role of Type 2 NAD(P)H Dehydrogenase NdbC in Redox Regulation of Carbon Allocation in Synechocystis^,

NdbC has a significant function in regulating carbon allocation between storage and the biosynthesis pathways but also affects cell division.

Abstract

NAD(P)H dehydrogenases comprise type 1 (NDH-1) and type 2 (NDH-2s) enzymes. Even though the NDH-1 complex is a well-characterized protein complex in the thylakoid membrane of Synechocystis sp. PCC 6803 (hereafter Synechocystis), the exact roles of different NDH-2s remain poorly understood. To elucidate this question, we studied the function of NdbC, one of the three NDH-2s in Synechocystis, by constructing a deletion mutant (ΔndbC) for a corresponding protein and submitting the mutant to physiological and biochemical characterization as well as to comprehensive proteomics analysis. We demonstrate that the deletion of NdbC, localized to the plasma membrane, affects several metabolic pathways in Synechocystis in autotrophic growth conditions without prominent effects on photosynthesis. Foremost, the deletion of NdbC leads, directly or indirectly, to compromised sugar catabolism, to glycogen accumulation, and to distorted cell division. Deficiencies in several sugar catabolic routes were supported by severe retardation of growth of the ΔndbC mutant under light-activated heterotrophic growth conditions but not under mixotrophy. Thus, NdbC has a significant function in regulating carbon allocation between storage and the biosynthesis pathways. In addition, the deletion of NdbC increases the amount of cyclic electron transfer, possibly via the NDH-1₂ complex, and decreases the expression of several transporters in ambient CO₂ growth conditions.

In cyanobacteria, both photosynthesis and respiration occur in the same cellular compartment. The cyanobacterial thylakoid membrane hosts both the photosynthetic and respiratory electron transport chains (ETCs), while some respiratory electron transport also has been localized to the plasma membrane (PM; Lea-Smith et al., 2013; Mullineaux, 2014; Ermakova et al., 2016; for review, Cooley and Vermaas, 2001). These two major bioenergetic pathways share several redox-active components in thylakoids, including the plastoquinone (PQ) pool, the cytochrome (Cyt) b₆f complex, and soluble electron carriers in thylakoid lumen.

In respiration, a significant part of electrons derived from organic molecules is transferred by bacterial pyridine nucleotide dehydrogenases into the PQ pool. These enzymes fall into two classes called type 1 and type 2 NAD(P)H dehydrogenases (NDH-1 and NDH-2, respectively; for review, see Melo et al., 2004). The NDH-1 enzymes are proton-pumping multisubunit complexes resembling mitochondrial complex I; in cyanobacteria, they reside mostly in the thylakoid membrane and function in cyclic electron transfer (CET), respiration, and C uptake (Battchikova et al., 2011; for review, see Peltier et al., 2016). In contrast, NDH-2s are composed of single polypeptides with molecular masses around 50 kD, and they do not contribute to the generation of the proton gradient across the membrane (Yagi, 1991). The primary structure of NDH-2 commonly includes two GXGXXG motifs within the β-sheet-α-helix-β-sheet domains (Rossman fold), one for binding NAD(P)H and another for FAD or FMN (Wierenga et al., 1986). Although the crystal structures of NDH-2 from Saccharomyces cerevisiae (Ndi1; Feng et al., 2012) and the bacterial type NDH-2 from Caldalkalibacillus thermarum (Heikal et al., 2014) as well as from Staphylococcus aureus (Sena et al., 2015) have been solved, the actual mechanism of NADH:quinone oxidoreduction still remains unclear despite recently presented models (Marreiros et al., 2017). It is generally agreed that, in organisms where NADH oxidation is carried out solely by NDH-2s, the main function of these enzymes is to perform the respiratory chain-linked NADH turnover and the following donation of electrons to the protein complexes that produce the H⁺ gradient powering ATP production (for review, see Melo et al., 2004).

Besides heterotrophic organisms, where NDH-2s are usually located either in mitochondria (eukaryotes) or in the PM (prokaryotes; for review, see Melo et al., 2004), there are several photosynthetic organisms with genes encoding NDH-2s. Plant-type NDH-2s fall into three distinct groups: NDA and NDB relating to fungal sequences and NDC of cyanobacterial origin (Michalecka et al., 2003). Most of these NDH-2s are targeted to mitochondria or peroxisomes, but some also are targeted to plastids. Of the six NDH-2s in the unicellular green alga Chlamydomonas reinhardtii, the thylakoid-localized NDA2 has been produced as a recombinant protein and characterized for its enzymological properties (Desplats et al., 2009). NDA2 was shown to be able to reduce PQ by oxidizing both NADH and NADPH. However, NDA2 showed higher affinity to NADH than NADPH and had FMN as a cofactor instead of the more common FAD. Another NDH-2 enzyme, NDA3, has been localized to C. reinhardtii chloroplasts (Terashima et al., 2010). Physcomitrella patens has six NDH-2 genes, and the products of three of them are targeted to chloroplasts (Xu et al., 2013). The genome of Arabidopsis (Arabidopsis thaliana) contains seven open reading frames encoding NDH-2 homologs, of which at least five have a mitochondrial localization (Michalecka et al., 2003). NDC1, the only NDH-2 of cyanobacterial origin (Michalecka et al., 2003), has dual targeting to both mitochondria and chloroplasts (Carrie et al., 2008), being involved in prenylquinone and vitamin K₁ metabolism in the chloroplast (Piller et al., 2011).

All sequenced cyanobacterial species encode at least one NDH-2 enzyme in their genomes (Marreiros et al., 2016). The thylakoid localization has been shown for NDH-2 in Anabaena variabilis (Alpes et al., 1989) and Anabaena sp. PCC 7120 (Howitt et al., 1993). Synechocystis sp. PCC 6803 (hereafter Synechocystis) has three genes encoding NDH-2s: ndbA (slr0851), ndbB (slr1743), and ndbC (sll1484; Kaneko et al., 1996; Howitt et al., 1999). They all have conserved motifs for binding a cofactor FAD and substrate NAD(P)H, and the sequence of the binding motif suggests that NADH is strongly preferred over NADPH (Howitt et al., 1999). It was shown that NdbB of Synechocystis (Howitt et al., 1999) and NDA2 of C. reinhardtii (Desplats et al., 2009) could functionally complement an Escherichia coli mutant lacking both NDH-1 and NDH-2s. It has been suggested that the NDH-2s function as redox sensors in Synechocystis, possibly responding to the redox state of the PQ pool, rather than having any significant catalytic role in respiration (Howitt et al., 1999). Regulation by NDH-2s also may be targeted to maintain the redox balance of NAD⁺/NADH, deduced from the fact that when all three Synechocystis NDH-2s were deleted, reduced NADH was strongly prevailing (Cooley and Vermaas, 2001). In addition, it was shown recently that NdbB is involved in vitamin K₁ biosynthesis (Fatihi et al., 2015).

In order to get more detailed insights into the function of NdbC in Synechocystis, the NdbC deletion mutant was constructed and thoroughly characterized, including proteomics approaches. It is demonstrated here that, upon autotrophic growth, NdbC has a significant function in regulating carbon allocation between storage and the biosynthesis pathways, with distinct effects on cell morphology and physiology. Intriguingly, the function of NdbC becomes absolutely essential under light-activated heterotrophic growth (LAHG) conditions.

RESULTS

Construction of the ndbC Deletion Mutant (ΔndbC)

The interrupted ndbC gene from the ΔndbC mutant constructed previously (Howitt et al., 1999) was amplified using PCR, and the resulting product was used in transformation (for details, see “Materials and Methods”). Segregation was confirmed by PCR analysis (Supplemental Fig. S1). The absence of the NdbC protein in the constructed ΔndbC mutant was verified by western-blot analysis using extracted total proteins and specific antibody raised against NdbC (Agrisera; Fig. 1). To verify that the sll1485 and sll1486 genes, which are located immediately downstream of ndbC in the Synechocystis genome, were not affected, their expression was studied with reverse transcription-quantitative PCR (Supplemental Fig. S2). No decline in the expression of either sll1485 or sll1486 was observed in ΔndbC compared with the wild type.

Figure 1.

Western-blot analysis with NdbC-specific antibody. Proteins were extracted from wild-type cells and ΔndbC mutant cells grown at ambient CO₂ concentration. Samples were loaded based on total protein amount: 12.5 µg for wild-type (WT) 50%, 25 µg for wild-type 100%, and 25 µg for ΔndbC.

Growth, Morphology, and Intracellular Glycogen Content of the ΔndbC Mutant under Ambient and High-CO₂ Conditions

The growth of ΔndbC was monitored under ambient CO₂ conditions (LC) and in the presence of 3% CO₂ (HC) for several days. Based on optical density (OD₇₅₀), ΔndbC grew slightly faster in LC (Fig. 2A) but clearly faster in HC conditions compared with the respective wild type under the same conditions (Fig. 2B). Concomitant inspection of wild-type and ΔndbC mutant cells with a light microscope (Fig. 3A) and a transmission electron microscope (Fig. 3B) revealed an increased cell size of ΔndbC compared with the wild type in both LC and HC. In addition, the cell number in ΔndbC cultures, calculated from equal OD₇₅₀ of wild-type and mutant cultures, was almost 53% lower when cells were grown in LC and 59% lower in HC cultures as compared with the wild type (Table I). Since the cell size significantly affects light scattering in addition to OD, the growth was followed by cell counting. Expressing the growth as an increase in cell number reinforced the slow growth of the ΔndbC mutant in both LC and HC (Fig. 2, C and D). Based on cell number, the growth of ΔndbC was about 50% slower compared with the wild type in both LC and HC after 100 h of inoculation. In accordance with the bigger cell size, the dry weight per cell of ΔndbC was about 52% higher in LC and 142% higher in HC compared with the wild type (Table I). In line with this, the chlorophyll a (Chl) amount per cell was about 109% bigger in the ΔndbC mutant in LC and 127% bigger in HC compared with the wild type. Similarly the amount of total proteins per cell in the ΔndbC mutant was 94% higher in LC and about 115% higher in HC compared with the wild type.

Figure 2.

Growth of the wild type (WT) and the ΔndbC mutant in LC and HC conditions under continuous illumination of 50 µmol photons m⁻² s⁻¹. Growth was monitored by OD₇₅₀ (A and B) and by cell number (C and D). In the beginning of the experiments, cells were inoculated at OD₇₅₀ = 0.1. Values are means ± sd; n = 3. Photographs in A and B were taken at OD₇₅₀ = 1.

Figure 3.

Phenotypes and glycogen accumulation of wild-type (WT) and ΔndbC cells in LC and HC growth conditions. A, Light microscope images of wild-type and ΔndbC cells. B, Transmission electron microscope images of wild-type and ΔndbC cells in LC. The arrow indicates putative glycogen granules (G). C and D, Intracellular glycogen content in the wild type and the ΔndbC mutant based on the Chl amount (C) and on the cell number (D). Values are means ± sd; n = 4. Asterisks indicate statistically significant differences between the wild type and the ΔndbC mutant (P ≤ 0.05).

Table I. Comparison of cell number, dry weight, Chl, and protein content between wild-type and ΔndbC mutant cells.

The number of wild-type and ΔndbC cells in 1 mL at the OD of 1, dry weight (mg per OD or mg per 10⁷ cells), Chl amount (µg Chl per OD or µg Chl per 10⁷ cells), and total protein amount (mg per OD or mg per 10⁷ cells) are shown. Values per OD are means ± sd; n = 3 to 4, and values per cell number are calculated based on these. Asterisks indicate statistically significant differences between the wild type and the ΔndbC mutant (P < 0.05).

Conditions	Cells 107/mL	Dry Weight	Chl	Protein	Dry Weight	Chl	Protein
	OD750 = 1	mg per OD	µg per OD	mg per OD	mg per 107 cells	µg per 107 cells	mg 107 cells
Wild type LC	18.4 ± 1.21	5.4 ± 0.25	4.2 ± 0.08	0.30 ± 0.00	0.29	0.23	0.016
ΔndbC LC	8.5 ± 1.01*	3.7 ± 0.41*	4.1 ± 0.11	0.27 ± 0.01*	0.44	0.48	0.031
Wild type HC	17.6 ± 1.30	4.6 ± 0.30	3.9 ± 0.03	0.23 ± 0.02	0.26	0.22	0.013
ΔndbC HC	7.2 ± 1.19*	4.5 ± 0.16	3.6 ± 0.04*	0.20 ± 0.01	0.63	0.50	0.028

Investigation of the cell ultrastructure with the transmission electron microscope revealed an accumulation of dark granules, possibly glycogen, in the ΔndbC mutant cells (Fig. 3B). Determination of the intracellular glycogen content from the ΔndbC and wild-type cells revealed an accumulation of 50% more glycogen in ΔndbC in LC and 18% more in HC when glycogen amount was normalized to Chl concentration (Fig. 3C). Nevertheless, when the glycogen content was normalized to the cell number, one ΔndbC cell contained about 320% more glycogen in LC and 260% more in HC compared with the wild type (Fig. 3D).

Photosynthetic Characterization of the ΔndbC Mutant

The wild-type and ΔndbC cells were first subjected to membrane inlet mass spectrometry (MIMS) with ¹⁸O-enriched oxygen for gas-exchange investigations. The advantage of MIMS compared with traditional oxygen electrode is the possibility to differentiate between the gross oxygen produced by PSII and the oxygen uptake under illumination. In addition, the oxygen and CO₂ exchanges can be monitored concomitantly.

There was no statistically significant difference in dark respiration between the wild type and ΔndbC in LC conditions (Table II). In ΔndbC, the light-induced oxygen uptake rate, which is defined as the difference between the light and dark oxygen uptake rates, was about 17% higher, whereas the gross and net oxygen evolution rates were 8.5% and 11% lower, respectively, compared with the wild type. The amplitude of the fast phase of CO₂ uptake, which is related to the function of carbon-concentrating mechanisms, was almost 19% smaller in ΔndbC compared with the wild type, but no significant difference in the steady-state CO₂ uptake during illumination was detected.

Table II. Gas-exchange rates of wild-type and ΔndbC cells grown in LC.

The oxygen exchange rates are shown as µmol oxygen mg⁻¹ Chl h⁻¹ and CO₂ uptake rates as µmol CO₂ mg⁻¹ Chl h⁻¹. Values are means ± sd; n = 3. Asterisks indicate statistically significant differences between the wild type and the ΔndbC mutant (P < 0.05).

Parameter	Component	Wild Type	ΔndbC
Oxygen uptake	Dark	2.8 ± 0.2	2.3 ± 0.2
	Light induced	17.7 ± 0.5	20.6 ± 0.5*
	Total light	20.4 ± 0.4	22.9 ± 0.3*
Oxygen production	Gross	176.5 ± 5.4	161.5 ± 2.5*
	Net	156.1 ± 4.9	138.6 ± 2.2*
CO2 uptake	Fast phase	754.7 ± 34.5	613.7 ± 40.9*
	Steady state	171.6 ± 4.9	166.3 ± 6.0

Next, the Chl fluorescence and P700 oxidoreduction measurements were performed with the ΔndbC cells using the Dual-PAM fluorometer. The PSII yield, measured from dark-acclimated cells (F_v^D/F_m^D), after illumination with a short far-red (FR) light (F_v^FR/F_m^FR), and the effective PSII yield [Y(II)], measured after illumination with actinic light for 5 min, were each slightly lower in ΔndbC compared with the wild type both in LC and HC (Fig. 4A; Supplemental Table S1). Likewise, no difference was observed in the maximum quantum yield of PSII (F_v/F_m), measured with 20 µM 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU), between the wild-type and ΔndbC cells grown at LC or HC (Supplemental Table S1). When the reoxidation of Q_A⁻ in the dark was recorded after a single-turnover saturating flash, the ΔndbC mutant demonstrated slightly slower decay of fluorescence compared with the wild type, indicating modified electron transfer at the PSII acceptor side (Supplemental Fig. S3A). In the presence of DCMU, which blocks electron transfer at the Q_B site, forcing Q_A⁻ reoxidation to occur via charge recombination with the donor side components (Vass et al., 1999), no difference in the decay of fluorescence was observed between ΔndbC and the wild type (Supplemental Fig. S2B). There was no significant difference in the effective yield of PSI [Y(I)], in the maximal amount of oxidizable P700 (P_m), in the maximal fluorescence of dark-adapted cells (F_m^D), or in the maximal fluorescence after FR illumination (F_m^FR) between the wild type and the ΔndbC mutant in LC or HC conditions (Fig. 4B; Supplemental Table S1). These results along with the MIMS data indicate that the deletion of NdbC does not structurally alter PSII or PSI, yet the electron transfer downstream from PSII was slightly slower in the ΔndbC mutant compared with the wild type.

Figure 4.

Photosynthetic characterization of the wild type (WT) and the ΔndbC mutant. A and B, Y(II) (A) and Y(I) (B) of the wild type (black squares in LC, blue triangles in HC) and the ΔndbC mutant (red circles in LC, magenta triangles in HC) under illumination (50 μmol photons m⁻² s⁻¹ actinic light). Cells were dark acclimated for 10 min before each measurement, and the Chl concentration was adjusted to 15 μg mL⁻¹. First points were measured after dark acclimation, and second points were measured after 10 s of FR illumination. Values are means ± sd; n = 3. C, P700 oxidoreduction in the wild type (black line) and in ΔndbC (red line) in LC. P700 oxidation was monitored during illumination with strong FR light for 5 s (red bar), and rereduction was monitored in darkness (black bar) after illumination. D, F₀ rise measured from the wild type (black line) and the ΔndbC mutant (red line) in LC. After dark acclimation, cells were illuminated for 1 min with red actinic light (50 μmol photons m⁻² s⁻¹; yellow bar [only part is shown]), and the F₀ rise was monitored (black bar) after turning off the actinic light. Curves were normalized to the same amplitude to facilitate comparison of the kinetics. Mean values from three independent experiments are shown. r.u., Relative units.

The fact that the slower electron transfer from PSII might be due to a more reduced electron transfer chain made us next monitor the redox state of P700 during dark-light-dark transitions by the application of strong FR light (Fig. 4C). The kinetics of P700 oxidation was clearly slower and rereduction was faster in the ΔndbC mutant, suggesting elevated CET in the mutant. To verify this, the transient increase in fluorescence after illumination (F₀ rise) was monitored as a postillumination rise in Chl fluorescence, which reflects the efficiency of the NDH-1-related CET (Mi et al., 1992a; Holland et al., 2015). In the ΔndbC mutant, the F₀ rise was significantly higher and reached the maximum clearly faster compared with the wild type (Fig. 4D).

Characterization of the ΔndbC Mutant at the Protein Level

Since the characterization of the ΔndbC mutant by the methods described above did not provide any clear functional role for the NdbC protein, we next took a proteomics approach and asked whether the NdbC deletion induces any major changes in the proteome. First, we performed a global label-free tandem mass spectrometry (MS/MS) protein identification and quantitation based on a data-dependent analysis (DDA). We quantified (with P ≤ 0.05) 543 proteins in HC (Supplemental Tables S2 and S4) and 699 proteins in LC (Supplemental Tables S3 and S5) with at least two peptides. The fold change (FC) threshold of the practical significance was set to −1.3 ≤ FC and FC ≥ 1.3.

The deletion of NdbC resulted in the differential regulation of a significant amount of proteins. In the HC-grown ΔndbC cells, 128 proteins were up-regulated while 287 proteins were down-regulated in comparison with the wild type (Supplemental Table S4); in the LC-grown ΔndbC cells, the numbers were 211 and 293, respectively (Supplemental Table S5). Proteins were quantified by DDA from cells grown both in HC and LC, while selected reaction monitoring (SRM) was applied only for LC-grown cells.

Differential Protein Expression in the Wild Type and ΔndbC Revealed by a Global Proteomic Approach

In both LC and HC conditions, three enzymes, phosphofructokinase (PfkA), pyruvate kinase (Pyk1), and glyceraldehyde-3-phosphate dehydrogenase (Gap1), which are involved specifically in glycolysis, were down-regulated in the ΔndbC mutant (Table III). A similar response was observed for phosphoglycerate kinase (Pgk) functioning in glycolysis and the Calvin-Benson cycle and for two key enzymes of the oxidative pentose phosphate (OPP) pathway, Glc-6-P dehydrogenase and 6-phosphogluconolactonase, encoded by zwf and pgl, respectively. It is important to note that the deletion of NdbC did not have a significant effect on the expression of proteins functioning in glycogen biosynthesis (GlgA,C,D) or in glycogen degradation (GlgX,P; Supplemental Tables S4 and S5). Furthermore, proteins essential to cell division, including FtsZ, ZipN, MinC, MinD, and MinE, were down-regulated in both growth conditions, as were the components of the urea transporter (UrtA,E; Table III). In addition, PII and PipX, participating in the regulation of nitrogen metabolism, were down-regulated (Table III), while the global nitrogen regulator NtcA was not affected (Supplemental Tables S4 and S5). The cyanobacteria-specific AbrB-like transcriptional regulator (cyAbrB) Sll0359 and also the DNA repair protein RecA were down-regulated in LC and HC (Table III). In addition, the deletion of NdbC decreased, both in LC and HC conditions, the expression of proteins encoded by genes located in extrachromosomal plasmids. They included, among others, Sll0653 to Sll0655 and Slr6012 to Slr6016 from pSYSX and several proteins from pSYSA that have been annotated as CRISP-related proteins (Scholz et al., 2013). The latter comprised Csm6 and Cas7 from the CRISP2 group as well as Cmr6, Cmr4, and Cmr2 from CRISP3 (Table III). In contrast, the flavodiiron proteins Flv1 and Flv3 involved in the Mehler-like reaction and ChlD involved in Chl biosynthesis were up-regulated in the ΔndbC mutant in both growth conditions. In addition, the circadian clock proteins KaiB1 and KaiC1 were up-regulated in a coordinated manner in both growth conditions, but proteins involved in polysaccharide (RfbB) and amino acid (Sll0664) transport, together with a putative Mg²⁺ (MgtE) transporter, were down-regulated.

Table III. Differential protein expression in the ΔndbC mutant versus the wild type.

Proteins were quantified with DDA from cells grown in HC and LC and with SRM from LC-grown cells. The practical significance FC is in italic type when up-regulated (FC ≥ 1.3) and boldface type when down-regulated (FC ≤ −1.3; P ≤ 0.05) in the ΔndbC mutant relative to the wild type. NA, Not analyzed; ND, not detected.

Parameter	Protein	Protein Symbol	HC DDA FC	P	LC DDA FC	P	LC SRM FC	P
Cell division	Sll1633	FtsZ	−1.47	3.61E-04	−1.43	4.65E-03	−1.54	6.36E-05
	Sll0169	ZipN	−1.01	8.58E-01	−1.40	1.02E-02	−1.58	7.78E-05
	Sll0288	MinC	−1.54	4.37E-02	−1.33	1.29E-02	−1.45	4.99E-02
	Sll0289	MinD	−1.42	1.99E-02	−1.61	2.98E-04	−1.60	3.39E-05
	Ssl0546	MinE	−1.56	1.01E-02	−1.39	3.46E-04	−1.57	8.99E-05
Circadian clock	Slr0757	KaiB1	−1.57	1.24E-03	−1.69	2.04E-04	NA	NA
	Slr0758	KaiC1	−1.58	9.53E-04	−1.47	2.71E-03	NA	NA
	Slr0115	RpaA	−1.08	5.97E-01	−1.58	3.86E-03	NA	NA
DNA repair	Sll0569	RecA	−1.50	3.63E-03	−1.66	2.35E-04	NA	NA
	Slr6097		−1.75	2.03E-01	−5.16	1.41E-03	NA	NA
	Slr9101	RepA	1.50	3.50E-01	−2.79	3.45E-03	NA	NA
	Sll1165	MutS	−1.19	9.65E-01	−1.75	1.61E-02	NA	NA
Transport and binding proteins
Inorganic phosphate uptake	Sll0681	PstC1	1.22	3.46E-01	−18.67	6.27E-05	−12.21	2.00E-05
	Sll0683	PstB1	−1.00	6.20E-01	−6.85	3.06E-04	−11.22	3.96E-04
	Sll0684	PstB1'	1.01	6.12E-01	−8.10	4.64E-03	−8.26	2.19E-05
	Sll0680	PstS1	−1.12	9.51E-01	−2.78	1.61E-05	−4.02	7.20E-03
	Sll0679	SphX	1.26	2.61E-01	−1.60	3.41E-02	−1.63	1.65E-03
	Slr1247	PstS2	1.42	1.82E-01	−1.34	3.85E-03	−1.45	5.12E-03
	Sll0540	PstS	−1.35	6.69E-01	−3.16	4.21E-04	−3.09	8.97E-06
Nitrogen uptake	Sll1451	NrtB	1.07	5.13E-01	ND	ND	−1.62	1.37E-03
	Sll1452	NrtC	−1.23	2.49E-01	−2.24	1.96E-02	−1.58	6.15E-04
	Sll1453	NrtD	−1.05	7.89E-01	ND	ND	−1.92	3.05E-03
	Sll0108	Amt1	−1.12	5.38E-01	−1.20	2.57E-02	−1.30	2.93E-02
	Slr0447	UrtA	−1.67	1.02E-05	−1.91	4.61E-04	NA	NA
	Sll0764	UrtD	ND	ND	−1.71	6.98E-04	NA	NA
	Sll0374	UrtE	−1.49	2.79E-02	−1.34	1.45E-02	NA	NA
HCO3− uptake	Slr0040	CmpA	−1.37	4.19E-02	1.62	1.75E-03	1.54	5.19E-03
	Slr0041	CmpB	1.08	4.69E-01	2.44	4.26E-05	2.20	1.70E-04
	Slr0043	CmpC	1.24	2.57E-01	2.28	3.25E-04	2.13	3.15E-04
	Slr0044	CmpD	−1.03	8.89E-01	2.36	4.82E-05	1.93	2.07E-04
	Slr1512	SbtA	−1.45	9.34E-03	1.00	9.60E-01	−1.03	6.94E-01
	Slr1513	SbtB	−2.19	3.28E-07	−1.02	7.85E-01	NA	NA
Others	Slr0982	RfbB	−1.95	4.63E-03	−3.09	9.96E-06	NA	NA
	Sll0064		−1.61	1.88E-02	−2.44	1.68E-02	NA	NA
	Slr1216	MgtE	−1.99	2.25E-03	−2.40	2.35E-04	NA	NA
	Slr0559	NatB	−1.21	1.37E-01	−1.80	1.41E-03	NA	NA
	Slr1740	AppA	−1.14	4.18E-02	−1.70	4.64E-04	NA	NA
	Slr0251	Ycf85	−1.08	6.79E-01	−1.68	1.73E-02	NA	NA
	Sll1927	AppF	1.47	1.30E-01	−1.64	2.89E-04	NA	NA
	Slr0747	GgtA	ND	ND	−1.56	3.45E-03	NA	NA
	Sll1762		−1.50	1.50E-04	−1.53	1.61E-02	NA	NA
	Sll0415		1.13	4.00E-01	−1.50	3.41E-02	NA	NA
	Slr2131	AcrF	−1.49	5.94E-02	−1.37	3.76E-02	NA	NA
	Slr0369	EnvD	−1.15	1.72E-01	−1.36	1.87E-03	NA	NA
	Sll0224		1.30	1.20E-01	−1.31	9.42E-03	NA	NA
	Sll0575	RfbB	−1.18	6.57E-01	−1.31	4.34E-02	NA	NA
	Sll1270	BgtB	1.02	9.55E-01	−1.39	6.99E-03	NA	NA
	Sll0672	PacL	1.18	7.37E-02	1.53	4.05E-03	NA	NA
	Slr1295	SufA	1.05	4.86E-01	1.58	6.72E-05	NA	NA
	Sll1598	MntC	ND	ND	1.65	5.79E-04	NA	NA
	Sll1481		1.82	6.22E-03	1.80	2.05E-04	NA	NA
	Slr1729	KdpB	1.07	8.41E-01	2.34	2.02E-03	NA	NA
NDH-1	Slr1291	NdhD2	ND	ND	ND	ND	4.65	2.72E-04
	Sll1733	NdhD3	2.37	2.68E-02	1.19	1.36E-01	1.09	3.34E-01
	Sll1735	CupS	−1.41	4.12E-04	ND	ND	1.07	1.13E-01
	Sll0026	NdhF4	−1.36	4.78E-02	ND	ND	−1.57	5.11E-03
	Slr1302	CupB	−1.09	2.08E-02	−1.56	4.37E-02	−1.53	4.16E-03
Central metabolism/regulation
	Sll0745	PfkA	−1.80	7.16E-02	−1.47	5.05E-02	NA	NA
	Slr0884	Gap1	−1.75	3.36E-06	−1.49	1.82E-02	−1.43	1.85E-02
	Slr0394	Pgk	−1.46	1.22E-03	−1.52	2.58E-03	−1.44	2.09E-04
	Sll0587	Pyk1	−1.47	2.26E-02	−1.52	3.02E-03	−1.42	3.90E-01
	Slr1843	Zwf	−1.35	9.93E-05	−1.40	1.47E-03	NA	NA
	Slr0452	IlvD	1.10	3.18E-01	1.46	1.43E-03	NA	NA
	Sll1479	Pgl	−1.27	2.85E-02	−1.46	8.12E-04	NA	NA
	Sll0158	GlgB	−1.52	6.74E-06	−1.29	1.45E-02	NA	NA
	Sll1502	GltB	ND	ND	2.32	5.82E-02	NA	NA
	Sll1499	GltS	−1.08	2.24E-01	2.91	6.35E-02	NA	NA
	Slr0710	GdhA	1.33	2.82E-04	1.48	3.41E-02	NA	NA
	Sll1689	SigE	ND	ND	ND	ND	−1.67	9.22E-03
	Sll1594	CcmR	1.17	2.75E-01	1.38	2.81E-02	1.21	7.95E-02
	Ssl0707	PII	−1.82	1.22E-04	−1.68	4.15E-03	−1.70	3.13E-05
	Ssl0105	PipX	−1.54	1.83E-03	−1.42	2.15E-03	−1.40	5.72E-05
	Sll0822		1.15	7.33E-01	−1.49	9.66E-04	−1.35	9.59E-02
	Sll0359		−1.33	6.60E-03	−1.35	9.62E-04	−1.34	3.16E-03
Photosynthesis and respiration
	Sll1521	Flv1	1.88	1.55E-05	1.60	2.05E-04	1.47	4.94E-04
	Sll0550	Flv3	1.44	3.37E-03	1.62	3.87E-03	1.44	3.53E-04
	Sll1317	PetA	1.09	2.40E-01	1.51	8.92E-03	1.33	7.53E-04
	Slr0342	PetB	1.15	6.40E-02	1.66	1.50E-02	1.41	6.61E-03
	Sll1316	PetC	1.08	3.23E-01	1.56	1.84E-04	NA	NA
	Slr0343	PetD	1.14	2.26E-01	1.48	1.03E-02	1.40	1.00E-04
	Slr1834	PsaA	1.07	3.96E-01	1.35	1.30E-02	1.22	6.46E-03
	Slr1835	PsaB	1.09	3.89E-01	1.31	3.12E-02	1.20	3.13E-03
	Slr1055	ChlH	1.41	6.17E-03	ND	ND	2.08	9.75E-03
	Slr1777	ChlD	1.39	1.76E-02	1.48	6.57E-03	1.26	1.11E-03
	Sll0247	IsiA	ND	ND	3.80	1.56E-02	18.70	7.04E-06
	sll0813	CtaCII	1.55	2.05E-01	−1.36	1.82E-03	NA	NA
	Sll1484	NdbC	−10.51	1.05E-07	−5.07	1.70E-03	−195.60	4.41E-09
Plasmid encoded	Sll6053		−8.17	4.50E-05	−128.52	6.57E-06	NA	NA
	Sll6054		−996.22	1.74E-06	−288.89	2.12E-06	NA	NA
	Sll6055		−10.57	1.18E-07	−21.64	1.36E-06	NA	NA
	Slr6012		−1.31	1.54E-03	−8.50	3.25E-06	NA	NA
	Slr6013		−5.42	1.38E-07	−13.27	2.44E-06	NA	NA
	Slr6014		−4.49	4.61E-04	−7.55	1.65E-05	NA	NA
	Slr6015		−4.05	3.77E-04	−7.10	6.50E-05	NA	NA
	Slr6016		−4.30	6.29E-06	−8.84	6.05E-05	NA	NA
	Sll7062	Csm6	−2.78	1.40E-05	−5.13	1.17E-05	NA	NA
	Sll7066	Cas7	−19.76	6.19E-07	−81.87	3.15E-03	NA	NA
	Sll7085	Cmr6	−2.91	1.97E-04	−4.70	2.75E-04	NA	NA
	Sll7087	Cmr4	−3.57	6.95E-08	−35.98	2.23E-04	NA	NA
	Sll7090	Cmr2	−3.08	7.95E-06	−9.43	2.18E-05	NA	NA

Growth conditions, however, also had an impact on the expression of several proteins in the ΔndbC mutant. In HC, the deletion of NdbC specifically included the down-regulation of GlgB participating in glycogen biosynthesis and the bicarbonate transporter SbtA/B as well as the differential regulation of some subunits belonging to NDH-1 (Table III). In particular, the expression of NdhD3 increased in the ΔndbC mutant, while CupS and NdhF4 were down-regulated in HC conditions. Contrary to HC, more prominent differences between the ΔndbC mutant and the wild type were detected at the protein level when cells were grown in LC conditions. Subunits of the ATP-binding cassette (ABC)-type bicarbonate transporter, CmpA,B,C,D, and the carbon-concentrating mechanism regulator CcmR were up-regulated in a coordinated manner (Table III). In contrast, NrtC belonging to the nitrate/nitrate transporter (NRT) was down-regulated. Considerable down-regulation was observed for both phosphate transporters, the low-affinity phosphate transporter system Pst1 (PstC1, PstB1, PstB1′, PstS1, and SphX) and the high-affinity transporter Pst2 (PstS2). The IlvD protein participating in Glc metabolism (the Entner-Doudoroff pathway) was up-regulated. The expression of several proteins involved in photosynthesis was up-regulated in the ∆ndbC mutant, including the major subunits of the Cyt b₆f complex (PetA,B,C,D) that were increased cooperatively, the PSI core subunits PsaB and PsaA, and the IsiA protein (Table III). In addition, the AbrB-like transcriptional regulator Sll0822, one of the subunits of the alternate respiratory terminal oxidase CtaCII, and the circadian clock-related response regulator RpaA were down-regulated in LC conditions in the ΔndbC mutant as compared with the wild type. In LC, several proteins involved in the transport of amino acids, peptides, glucosylclyserol, and lipopolysaccharides were down-regulated in the ΔndbC mutant, in contrast to up-regulated Fe, Mn, and K transporters. Furthermore, several proteins involved in DNA repair (Slr6097, RepA, and MutS) were down-regulated in the ΔndbC mutant compared with the wild type only in LC.

Targeted Proteomic Approach Strengthened and Expanded the Scope of Proteins Affected in the ΔndbC Mutant

We also applied to LC-grown wild-type and ΔndbC cells a more precise and sensitive targeted quantification method called SRM (Anderson and Hunter, 2006; Ludwig et al., 2012; Vuorijoki et al., 2016). SRM was used in order to verify the results of DDA and, importantly, to expand the scope of quantified membrane-embedded photosynthetic and respiratory proteins. For the SRM experiments, we selected 108 proteins, including those showing differential expression in DDA, plus several other targets that could provide complementary information on proteomic changes of important metabolic routes in the ΔndbC mutant. The complete list of SRM targets is shown in Supplemental Table S6. Overall, SRM quantitation confirmed the DDA results; in the case of PsaA and PsaB proteins, however, their up-regulation was found to be lower than the FC threshold (Table III). The SRM results obtained for NdbC validated the deletion of the corresponding protein in the ΔndbC mutant. On the other hand, SRM revealed several novel differentially regulated proteins that were not detected in the global approach or were detected and quantified but with a high P value. Importantly, we discovered that various NDH-1 complexes responded differently to the NdbC mutation. SRM results showed that NDH-1₁, with the specific NdhD1 subunit, and NDH-1₃, with NdhD3, NdhF3, CupA, and CupS, were not affected (Supplemental Table S6). In contrast, NdhD2, the specific subunit of the previously elusive NDH-1₂ complex, was up-regulated substantially (Table III), while NdhF4 and CupB, belonging to the NDH-1₄ complex, were distinctly down-regulated. Common NDH-1 subunits remained unchanged (Supplemental Table S6). Furthermore, we discovered the down-regulation of SigE, the group 2 σ-factor, a decrease of NrtB and NrtD subunits of the NRT transporter, coordinately with NrtC, and an up-regulation of ChlH involved in Chl biosynthesis (Table III).

Western-Blot Analysis of Selected Proteins Provided Further Insights into the Consequences of ndbC Deletion

To verify the proteomic results described above, western-blot analysis of total proteins isolated from LC-grown cells was performed for several targets (Fig. 5A). The results obtained with protein-specific antibodies corroborated conclusions made based on mass spectrometry (MS) methods. In particular, it was confirmed that the PsaB amount was slightly up-regulated in ΔndbC compared with the wild type (Fig. 5A). Moreover, the amount of the small form of ferredoxin-NADP⁺ oxidoreductase (FNR_s) was shown to be higher in ΔndbC compared with the wild type.

Figure 5.

Western-blot analysis of the wild type (WT), ΔndbC, and different NDH-1 mutants. A, Immunoblotting with protein-specific antibodies. B, Different Synechocystis NDH-1 mutants immunoblotted with α-NdbC. C, The thylakoid membrane, PM, and soluble fractions of the wild type immunoblotted with α-NdbC, α-PsaB, and α-allophycocyanin (APC). All the cells for protein isolation were grown in LC, and samples were loaded based on total protein amount. Vertical lines indicate that the samples, although run on the same gel, are not from adjacent wells.

Since the proteomic results showed an interesting effect of the NdbC deletion on NDH-1 complexes, a possible interplay between NdbC and the NDH-1 variants was addressed. To this end, several NDH-1 mutants were grown in LC conditions, and total proteins were isolated and immunoblotted with α-NdbC (Fig. 5B). Intriguingly, NdbC was not expressed in detectable amounts if the functional NDH-1 complex (ΔndhB/M55 mutant) was missing (Fig. 5B). In line with this observation, the amount of NdbC was lower compared with the wild type when NdhD1 or NdhD2 was deleted, and the double deletion of these NDH-1 subunits decreased the NdbC amount even more. The deletion of NdhD3 or NdhD4 did not affect the NdbC amount compared with the wild type.

Furthermore, western blotting was used to reveal the localization of NdbC in the cell. Protein fractions representing the thylakoid membrane, soluble compartment, and the PM were analyzed with the NdbC-specific antibody. NdbC was clearly located only in the PM, while no signal was detected from the thylakoid membrane fraction or from the soluble proteins (Fig. 5C), in line with results obtained by proteomic studies (Pisareva et al., 2011; Liberton et al., 2016).

Response of ΔndbC to Light-Activated Heterotrophy and Photomixotrophic Conditions

Proteomics analysis of the ΔndbC mutant suggested a putative role for the NdbC protein in general carbon metabolism, which prompted us to test the behavior of the wild type and the ΔndbC mutant in the presence of an external carbon source. To this end, the wild-type and mutant cells were first subjected to LAHG conditions, where the growth medium was supported with 10 mm Glc and illumination was limited to 10 min (50 µmol photons m⁻² s⁻¹) every 24 h. Under these conditions, the ΔndbC mutant practically did not grow (Fig. 6A). Next, the length of the light period was increased to 4 h per day (4 h of 50 µmol photons m⁻² s⁻¹/20 h of dark) to change the growth conditions toward mixotrophy, which restored the growth of the ΔndbC mutant (Fig. 6B). During days 2 to 4, the growth of the ΔndbC mutant was faster compared with the wild type. Under mixotrophic conditions (50 µmol photons m⁻² s⁻¹ continuous illumination + 10 mm Glc), there was no difference in the growth based on OD between the wild type and ΔndbC (Fig. 6C). Nevertheless, it is important to note that, also in the presence of Glc, the ΔndbC cells were considerably larger compared with the wild type (Supplemental Fig. S4), which increases OD and is in line with the phenotype observed under autotrophic growth conditions (Fig. 3A).

Figure 6.

Growth of the wild type (WT; black squares) and the ΔndbC mutant (red circles) in the presence of Glc. A, Under LAHG conditions (the cells grown in darkness were exposed to 50 µmol photons m⁻² s⁻¹ light for 10 min once in 24 h). B, Under a 4-h photoperiod of 50 µmol photons m⁻² s⁻¹/20 h of darkness. C, Under mixotrophic conditions (continuous illumination of 50 µmol photons m⁻² s⁻¹). In all cases, 10 mm Glc was added to the growth medium. Values are means ± sd; n = 3.

Redox State of the Electron Transport Chain in Darkness in the ΔndbC Mutant

To monitor more precisely the redox state of the ETC in ΔndbC during incubation in darkness, the cells were grown first in autotrophic conditions and then transferred to the dark for 24 h. Q_A⁻ reoxidation kinetics and the fluorescence induction dynamics measured from the 24-h dark-adapted ΔndbC cells did not differ strikingly from those in the wild type (Fig. 7, A and B; Supplemental Fig S3, A–C). To address the possible effect of Glc on ETC in darkness, 24-h dark-incubated cells were supplemented with 10 mm Glc for 15 min before the Chl fluorescence measurements. A short incubation with Glc markedly slowed down Q_A⁻ reoxidation kinetics in both the wild-type and ΔndbC cells, yet the kinetics became much slower in ΔndbC (Fig. 7C), implying a modification of the PSII acceptor side of the mutant, presumably due to the reduced PQ pool. When the flash-induced fluorescence yield was monitored in the presence of DCMU, reflecting more the donor side function of PSII, no clear difference was observed between the wild type and ΔndbC (Supplemental Fig. S3D). Next, the dynamic changes in the redox status of the PQ pool in dark-incubated cells supplemented with Glc were studied by Chl fluorescence measurements using a saturating pulse method (Fig. 7D). A sudden exposure of the dark-incubated wild-type cells to steady-state actinic light induced a transient increase in the F_s level with a concomitant decrease of ΔF, reflecting a reduced PQ pool during the first 3 min of illumination. The ΔndbC cells demonstrated a higher F_s level and low F_m^D and ΔF values compared with the wild type (Fig. 7D). These results clearly showed that the addition of Glc in darkness triggers much stronger reduction of the PQ pool in ΔndbC than in the wild type and that the subsequent exposure to moderate light is capable of oxidizing the ETC.

Figure 7.

Fluorescence relaxation and induction curves monitored from wild-type (WT) and ΔndbC mutant cells. A and B, Cells were grown for 48 h under 50 µmol photons m⁻² s⁻¹ in LC and then shifted to darkness for 24 h. C and D, Cells were grown for 48 h in 50 µmol photons m⁻² s⁻¹ in LC and shifted to darkness for 24. For measurements, the cell suspension was supplemented with 10 mm Glc for 15 min in the dark. In the flash-induced fluorescence decay experiment, the values are means ± sd; n = 3, and the fluorescence induction curves are representative of three biological repetitions. r.u., Relative units.

DISCUSSION

In cyanobacteria, the physiological role of the NDH-2 enzymes remains largely unknown, particularly in comparison with the much better characterized NDH-1 complexes playing important roles in respiration, CET, and CO₂ uptake (for review, see Peltier et al., 2016). It was hypothesized previously that, in Synechocystis, the NDH-2s do not have a metabolically relevant role in respiratory energy production and, instead, were postulated to be involved in regulatory functions, yet with no knowledge of specific targets so far (Howitt et al., 1999). Here, we addressed the physiological role of NdbC, one of the three NDH-2s encoded in the genome of Synechocystis. The deletion of NdbC (Fig. 1) revealed several significant rearrangements in the physiology, metabolism, and morphology under autotrophic growth conditions without affecting respiration (Table II) and, in addition, disclosed an indispensable function for NdbC under LAHG conditions. It is demonstrated that NdbC, localized in the PM (Fig. 5C), has a significant role in carbon allocation between the storage and various biosynthetic pathways (for LC conditions, this is presented schematically in Fig. 8).

Figure 8.

Schematic model representing the effects of NdbC deletion on protein expression in Synechocystis cells in LC conditions compared with the wild type. The up-regulated or accumulated protein complexes, metabolic routes, or compounds in the ΔndbC mutant are marked in red, and the down-regulated ones are marked in blue. Solid arrows represent metabolic routes or uptake, dashed arrows represent linear electron transfer, and dotted arrows represent CET. TM, Thylakoid membrane.

Deletion of NdbC Distorts Cell Division

When the ΔndbC mutant was grown under autotrophic conditions, in HC or LC, visible morphological changes were observed. The cell size became considerably bigger (Fig. 3A), with increase in the content of intracellular total protein, Chl, and dry weight compared with the wild type (Table I). Based on the cell number, ΔndbC demonstrated substantially slower growth compared with the wild type in both LC and HC (Fig. 2, C and D). This was corroborated by the proteomic examination of ΔndbC. The expression of specific cell division-related proteins (Mazouni et al., 2004) was clearly altered in ΔndbC, including down-regulation of FtsZ, ZipN, MinC, MinD, and MinE proteins in HC and LC (Table III). It is possible, however, that the influence of the NdbC deletion on cell division may be a secondary effect, as some global regulators of gene expression, such as the group 2 σ-factor SigE (detected only in LC), the AbrB-like transcription factor Sll0822 (in LC), and components of the circadian clock (KaiB and KaiC), were likewise down-regulated in the ΔndbC mutant. The influence of these proteins on cell division was demonstrated previously by Osanai et al. (2013) for SigE, by Yamauchi et al. (2011) for Sll0822, and in a review by Cohen and Golden (2015) for Kai proteins. In addition, several proteins encoded by pSYSX and pSYSA were markedly down-regulated in ΔndbC compared with the wild type in both HC and LC conditions (Table III). Among them are CRISPR2 and CRISPR3 (Scholz et al., 2013), which belong to the CRISPR/Cas systems responsible for resistance to horizontal gene transfer, including phage transduction, transformation, and conjugation (Marraffini and Sontheimer, 2008; for review, see Sorek et al., 2013). However, the influence of NdbC deletion on these proteins remains unclear, since the mutant could have experienced problems with plasmid replication connected to the hindrance in cell division.

Carbon Catabolism Is Down-Regulated in the Absence of NdbC

Measurements of the intracellular glycogen showed that this important reserve of carbon was significantly higher in the ΔndbC mutant compared with the wild type, both in LC and HC (Fig. 3D). The increase of intracellular glycogen per cell in the mutant was markedly more profound than the increase of the Chl amount or of the total protein content (Table I). In line with this, the electron microscopy of ΔndbC cells in LC revealed the occurrence of multiple dark spots that, probably, are glycogen granules (Fig. 3B). The accumulation of glycogen could be due to a misbalance between synthesis and degradation. It is important to note, however, that amounts of enzymes performing glycogen synthesis (GlgA,B,C) or glycogen degradation (GlgX,P) were not coordinately altered in the ΔndbC mutant compared with the wild type either in LC or HC (Table III). Thus, the shift of the balance has to be caused by other reasons. Indeed, the proteomic analysis revealed alterations in protein contents of metabolic pathways involved in intracellular carbon allocation in LC and in HC; Gap1, Pgk, and Pyk1, the three enzymes catalyzing unidirectional reactions in glycolysis, were down-regulated in the ΔndbC mutant (Table III).

Considering the first steps of the glycolytic route, the OPP pathway and the Embden-Meyerhof-Parnas (EMP) pathway were down-regulated in both studied conditions. The amounts of two key enzymes, Glc-6-P dehydrogenase (Zwf) and 6-phosphogluconolactonase (Gnd), specific for OPP as well as PfkA involved in EMP were likewise reduced in the ΔndbC mutant (Table III). These results are in agreement with earlier studies showing that the mRNA amount of ndbC follows circadian rhythm peaking in the dark together with genes involved in sugar catabolism (Kucho et al., 2005; Layana and Diambra, 2011) as well as in the beginning of heterotrophy (Lee et al., 2007). Conversely, 6-PG hydratase (IlvD), which belongs to the recently discovered Entner-Doudoroff pathway in cyanobacteria (Chen et al., 2016), was up-regulated in ΔndbC in LC (Table III) and might partially compensate for the down-regulation of EMP and OPP.

Overall, the observed decrease in the expression levels of many glycolytic enzymes suggests that the NdbC deletion caused the slowdown of sugar catabolism, probably resulting in the reduction of pyruvate and other metabolites at the consequent steps of carbon assimilation. Since NADH is utilized in glycolysis, in particular by Gap1, which is an essential component of this process, the direct effect of the NdbC deletion cannot be excluded. However, other factors could contribute to the slowdown of sugar catabolism. For example, since SigE positively regulates the expression of genes functioning in sugar catabolism (Osanai et al., 2005), its down-regulation, mentioned above, together with the up-regulation of its anti-σ-factor ChlH (Osanai et al., 2009; Table III), could result in reduced amounts of glycolytic enzymes and, thus, promote glycogen accumulation. In addition, the RpaA protein, the master regulator of circadian control, was down-regulated in ΔndbC in LC (Table III). Its involvement in adjusting the expression levels of enzymes participating in glycolysis and OPP has been shown in Synechococcus elongatus PCC 7942 (Markson et al., 2013). Concerning metabolic reactions downstream from glycolysis, our proteomic analyses did not detect coordinated changes concerning enzymes in the tricarboxylic acid cycle. In contrast, GltB and GltS, two Glu synthetases of the GS-GOGAT pathway, were up-regulated (in LC) at the protein level together with the Gln dehydrogenase GdhA (Table III).

The Deletion of NdbC Alters the Expression of Multiple Transport and Binding Proteins Especially in LC

Even though similar changes in the proteome related to cell division and sugar catabolism were detected between the wild type and ΔndbC both in HC and LC, the deletion of NdbC had an evidently more profound effect on the expression of a large group of transporters in LC compared with HC. Most of these transporters were reported previously to localize in the PM (Pisareva et al., 2011), in accordance with the location of the NdbC protein (Fig. 5C). The transport of various nutrients, like potassium and metal ions, amino acids, and polysaccharides (Table III), demonstrated marked down-regulation in the ΔndbC mutant except for the HCO₃⁻ transporter BCT1 (Omata et al., 1999), which was up-regulated in the ΔndbC mutant compared with the wild type. The strongest down-regulation was observed for both ABC-type inorganic phosphate transporters (Table III), the constitutive low-affinity phosphate transporter Pst1 and the inducible, high-affinity phosphate transporter Pst2 (Pitt et al., 2010; Burut-Archanai et al., 2011). Furthermore, in the ΔndbC mutant, the levels of enzymes involved in nitrogen uptake, NRT transporters responsible for nitrate and nitrite uptake, as well as urea and ammonium transporters were down-regulated (Table III). Such coordinated down-regulation of different transporters in LC, when there is a shortage of terminal acceptors, unlike in HC, can be a protective response of the ΔndbC cells aimed to regulate a redox homeostasis of the cells by ion influx and efflux.

In addition to transporter proteins, the expression of several regulators of transporter functions is altered in ΔndbC. While CmpR inducing BCT1 expression (Nishimura et al., 2008; Daley et al., 2012) was not affected (Supplemental Table S5), the Sll0822 protein, which functions as a supplementary inducer of the Cmp operon (Orf et al., 2016), was down-regulated in ΔndbC, as was already mentioned above (Table III). Thus, the up-regulation of the Cmp operon in ΔndbC was not caused by known regulating mechanisms. The regulators of the nitrogen assimilation genes PII and PipX (Llácer et al., 2010; Espinosa et al., 2014; for review, see Forchhammer, 2004) were reduced in the ΔndbC mutant (Table III), while the global nitrogen regulator NtcA remained unchanged (Supplemental Table S5). Decreased amounts of PII and PipX, which promote the deactivation of NtcA (Luque et al., 2004), as well as down-regulation of Sll0822, also an inducer of nitrogen transport (Ishii and Hihara, 2008; Yamauchi et al., 2011), could cause the low accumulation of nitrogen transporters induced by the NdbC deletion, particularly in LC. Thus, in LC conditions, the deletion of NdbC disrupts the normal behavior of multiple transport systems, mainly of ABC-type transporters located in the PM.

Cross Talk between NdbC and the NDH-1 Complex: Low ATP Production from Sugar Catabolism in ΔndbC Is Compensated by CET

Cyanobacteria can incorporate inorganic carbon (C_i) in the form of CO₂ utilizing the low-CO₂-inducible high-affinity NDH-1₃ complex and the constitutively expressed low-affinity NDH-1₄ complex (for review, see Battchikova et al., 2011). MIMS measurements revealed that the fast phase of CO₂ uptake was decreased in the ΔndbC mutant in LC (Table II). The proteomic results did not indicate changes in the expression of specific subunits of NDH-1₃ in the ΔndbC mutant in LC (Table III). In contrast, NdhF4 and CupB, specific subunits of NDH-1₄, were coordinately down-regulated in ΔndbC compared with the wild type in LC (Table III), being in line with MIMS results. To get insights into the connection between the NdbC deletion and the expression of NDH-1₄, we performed a western-blot analysis of several mutants lacking specific subunits of distinct NDH-1 variants using the NdbC-specific antibody (Fig. 5B). Remarkably, NdbC was severely down-regulated in the M55 strain, indicating some cross talk between this protein and NDH-1. However, it is important to note that the amount of NdbC in the ΔndhD4 mutant (and in ΔndhD3/D4) was comparable with that in the wild type. Thus, NdbC is involved in the accumulation of NDH-1₄, but not the other way around. Most probably, NdbC affects NDH-1₄ indirectly, since they are located in different membrane compartments (Liberton et al., 2016).

In the ΔndbC mutant, CET around PSI was clearly elevated, as deduced from the high amplitude of F₀ rise (Fig. 4D; Mi et al., 1992a; Holland et al., 2015) and from slower oxidation and faster rereduction kinetics of P700 (Fig. 4C; Mi et al., 1992b; Fan et al., 2007). Proteomic analysis revealed that two variants of the NDH-1 complex, NDH-1₁ and NDH-1₂, which are involved in CET (Mi et al., 1992a), demonstrate different responses to NdbC deletion. NDH-1₁ was not affected, according to the levels of NdhD1 specific for this complex (Supplemental Table S6). In contrast, marked up-regulation was observed for NdhD2 (Table III), the specific subunit of NDH-1₂. The latter complex is present in Synechocystis cells in much smaller amounts than NDH-1₁ (Herranen et al., 2004), but, despite a low abundance, NDH-1₂ that accumulated in the ΔndbC mutant contributed to the significant increase in CET (Fig. 4, C and D). Western-blot analysis demonstrated that, in the ΔndhD2 mutant (and in ΔndhD1/D2), the NdbC amount was clearly lower compared with the wild type (Fig. 5B). Thus, NdbC and NDH-1₂ affect each other, in contrast to NDH-1₄.

Furthermore, according to western-blot analysis, FNR_s was up-regulated in ΔndbC (Fig. 5A). This form of FNR oxidizes NADPH and accumulates in heterotrophic conditions, possibly enhancing CET by transferring electrons via ferredoxin to NDH-1 (Thomas et al., 2006; Korn, 2010). The elevated CET in the mutant was accompanied by an increase of the Cyt b₆f complex (PetA,B,C,D; Table III). Enhanced CET and the corresponding increase in ATP production can compensate for lower ATP yield from glycolytic reactions in the mutant. In addition, the up-regulation of oxygen photoreduction (Table II), operated by the increased Flv1 and Flv3 proteins (Table III), possibly contributes to ATP production via a functional water-water cycle and increases the photoprotection of photosynthetic protein complexes, especially PSI (Allahverdiyeva et al., 2015).

Even though the carbon catabolic reactions were distorted due to the deletion of NdbC, photosynthesis in general was not compromised dramatically either in LC or in HC; no significant difference was detected in F_v/F_m, in P_m, or in Y(I) between the wild type and ΔndbC (Fig. 4B; Supplemental Table S1). However, both the gross and net oxygen evolution (Table II), as well as Y(II) (Fig. 4A), of ΔndbC were somewhat lower compared with the wild type, demonstrating excess electron pressure downstream of PSII, which is in agreement with elevated CET. Cooley and Vermaas (2011) showed that the PQ pool of Synechocystis deficient of all three NDH-2s (ΔndbABC) was more oxidized during illumination, which is contrary to the redox status of the PQ pool in ΔndbC. Thus, we cannot exclude the possibility that the redox state of the NADH pool (by NDH-2) and the NADPH pool (via NDH-1) can influence each other, at least in the case of ΔndbC.

NdbC Is Indispensable in LAHG Conditions

The consequences of the blockage of the glycolytic route, evident even in autotrophic conditions, should become more profound when photosynthesis is limited and all the energy for cellular needs must be derived from the degradation of sugars. This was experimentally proven by the inability of the ΔndbC mutant to grow under LAHG conditions (Fig. 6A). In heterotrophic growth conditions, 90% of supplemented Glc is directed to OPP and only about 5% enters glycolysis (Yang et al., 2002). In ΔndbC, OPP is strongly down-regulated even in autotrophic conditions, and this is expected to have serious consequences on nucleotide biosynthesis upon LAHG conditions, due to the fact that OPP is the only source for required intermediates in darkness (Kruger and von Schaewen, 2003). Similarly, ΔsigE, which is deficient in sugar catabolism, shows retarded growth in LAHG conditions (Osanai et al., 2005). Fluorescence measurements further confirmed the problems of ΔndbC. The ETC became severely and rapidly overreduced in the dark and during the dark-light transition upon supplementation of the culture with Glc (Fig. 7). The reason for Glc toxicity could be a reprogramming of carbon metabolism and a high production of reactive oxygen species triggering oxidative stress and preventing cell growth, eventually causing cell death (Latifi et al., 2009; Narainsamy et al., 2013). In mixotrophic conditions, when the active Calvin-Benson cycle functions together with the OPP pathway (Yang et al., 2002), the ΔndbC mutant was able to grow (Fig. 6C). The up-regulated Entner-Doudoroff pathway in ΔndbC (in LC), which is physiologically significant under mixotrophy (Chen et al., 2016), also could contribute to carbon flux distribution in the cells. Nevertheless, similar problems were present in cell division, as demonstrated in autotrophic conditions (Supplemental Fig. S4). Thus, the maintenance of ΔndbC growth in the medium supplemented with Glc is dependent on the regular interruption of darkness with light periods (Fig. 6B) to efficiently oxidize the ETC (Fig. 7D).

CONCLUSION

NADH and NADPH are fundamental reducing equivalents for the maintenance of cellular metabolism. In photosynthetic organisms, the NADH/NAD⁺ status is known to regulate the coordination of nitrogen assimilation and carbon metabolism by various mechanisms (Dutilleul et al., 2005; Kämäräinen et al., 2017; Pétriacq et al., 2017). By applying a number of physiological measurements and the proteomics approach, we showed that the deletion of NdbC has a distinct effect on the metabolism of Synechocystis. We postulate that NdbC located in the PM is an important component in the regulation of the cytosolic NADH/NAD⁺ balance. In photoautotrophic growth conditions with active photosynthetic carbon assimilation, Synechocystis cells do survive in the absence of NdbC. Yet, importantly, glycolytic enzymes are down-regulated in the ΔndbC mutant under these growth conditions (Fig. 8), apparently due to the elevated NADH/NAD⁺ ratio. Consequently, cells need to compensate for the lack of the enzyme by the modulation of several metabolic routes. This provides evidence that NdbC is important in fine-tuning the balance between various metabolic pathways when the cytosolic redox balance is influenced mainly by light reactions in the thylakoid membrane. However, the function of NdbC becomes crucial for cell viability in heterotrophic growth conditions. This indicates an essential role of NdbC in the regulation of the cytosolic redox environment when glycolysis is the sole pathway to assimilate carbon for metabolic needs.

MATERIALS AND METHODS

Strains and Culture Conditions

A Glc-tolerant Synechocystis sp. PCC 6803 (wild type) was used as the reference strain. The ΔndbC mutant was obtained by disruption of the ndbC gene by Zeocin resistance cassette (for details, see “Mutagenesis” below). Preexperimental cultures were always grown in BG11 medium buffered with 20 mm HEPES-NaOH (pH 7.5) under continuous white light of 50 µmol photons m⁻² s⁻¹ at 30°C, under air enriched with 3% CO₂ (HC) with agitation of 150 rpm. All experimental cultures were grown in BG11 medium buffered with 20 mm HEPES-NaOH (pH 7.5) under continuous illumination of 50 μmol photons m⁻² s⁻¹, unless mentioned otherwise, and in growth chambers with cool-white light-emitting diodes (AlgaeTron AG130 by PSI Instruments). In some experiments, cells were grown in different dark-light cycles where illumination was 50 μmol photons m⁻² s⁻¹ during the light period. During experimental cultivations, cells were grown at 30°C with agitation of 150 rpm without antibiotics, and OD₇₅₀ was measured using the Lambda 25 UV/VIS spectrometer (PerkinElmer). Glc (10 mm) was provided when mentioned. For physiological measurements and protein extraction, cells were harvested from preexperimental cultures at the midlogarithmic phase, inoculated in fresh BG11 medium at OD₇₅₀ = 0.1, and grown at HC or inoculated to OD₇₅₀ = 0.5 and shifted to ambient air (LC) for 3 d. For activity measurements, cells were harvested and resuspended in fresh BG11 medium at the desired Chl concentration and acclimated under the respective growth conditions before the measurements.

Mutagenesis

To ensure the correct wild-type background, the interrupted ndbC gene (sll1484) with Zeocin resistance cassette and its surroundings were amplified by PCR with forward primer 5′-GCTGGTACGGTGGAAAGTCT-3′ (in sll1483) and reverse primer 5′-CTACCGATTCATACCGGGGCTA-3′ (in the end of sll1484) from the ΔndbC mutant constructed by Howitt et al. (1999). The resulting PCR product was transformed into the wild-type cells using a standard established protocol (Eaton-Rye, 2011). Transformants were segregated by gradually raising the concentration of Zeocin (from 4 to 20 μg mL^{− 1}). Complete segregation of the mutation was confirmed by PCR (Supplemental Fig. S1).

Light Microscopy and Cell Counting

Cell suspensions in BG11 were examined using the Leitz Orthoplan Large Field Research Microscope and photographed with a digital microscope camera (Leica DFC420C). Digital slide photograph brightness and contrast were optimized with Leica Application Suite version 4.1. and CorelDRAW X7. A Bürker counting chamber (Marienfeld-Superior) was applied to quantify the number of cells in a culture (normalized to OD₇₅₀ = 1).

Dry Weight and Chl Determination

Twenty milliliters of cell culture was passed through a prewashed and weighed glass-fiber filter (Millipore). Filters were dried at 60°C for 24 h, kept in a desiccator for 24 h, weighed, and normalized to premeasured OD₇₅₀. Chl was extracted from cells with 100% methanol, and Chl concentrations were determined by measuring OD₆₆₅ and multiplying it with extinction coefficient factor 12.7 (Meeks and Castenholz, 1971).

Transmission Electron Microscopy

Prefixation of samples was done with 5% glutaraldehyde in 0.16 m s-collidin buffer, pH 7.4. The rest of the sample preparation and microscopy work using the JEM-1400 Plus Transmission Electron Microscope was done in the Laboratory of Electron Microscopy, University of Turku.

Glycogen Determination

Ten milliliters of cultures at OD₇₅₀ = 1 was collected, and cells were lysed by sonication (BIORUPTOR; CosmoBio). Glycogen determination was performed with the Total Starch Assay Kit (Megazyme) according to the manufacturer’s instructions.

MIMS

MIMS measurements were performed as described by Ermakova et al. (2016). Gas-exchange kinetics and rates were determined according to Beckmann et al. (2009).

Fluorescence Measurements

The Chl fluorescence from intact cells was recorded with a pulse amplitude-modulated fluorometer (Dual-PAM-100; Walz). Before measurements, cell suspensions at a Chl concentration of 15 µg mL⁻¹ were dark adapted for 10 min. Saturating pulses of 5,000 μmol photons m⁻² s⁻¹ (300 ms) and strong FR light (720 nm, 75 W m⁻²) were applied to samples when required. Y(II) was calculated as (F_m′ − F_s)/F_m′. The F_m^D was recorded after the first saturating pulse without actinic light, the F_m^FR was recorded subsequent to 8 s of strong FR illumination combined with saturating pulse, and the maximal fluorescence during illumination (F_m′) was recorded upon saturating pulse during illumination. The maximum quantum yield of PSII was calculated as F_v/F_m = (F_m − F₀)/F_m and measured in the presence of 20 μM DCMU from dark-adapted cells upon the application of red actinic light of 200 μmol photons m⁻² s⁻¹ for 1 min.

The F₀ rise was measured in darkness after the termination of actinic light. Cell suspensions at a Chl concentration of 15 µg mL⁻¹ were dark adapted for 10 min before the measurements and exposed to 1 min of red actinic light intensity of 50 μmol photons m⁻² s⁻¹.

The kinetics of the Chl fluorescence decay after a single-turnover saturating flash was monitored using a fluorometer (FL 3500; PSI Instruments) according to Vass et al. (1999). Cells were adjusted to a Chl concentration of 7.5 µg mL⁻¹ and dark adapted for 5 min before measurements. When indicated, measurements were performed in the presence of 20 μM DCMU.

The Maximum and Effective Yield of PSI and Oxidoreduction of P700

The P700 signal was recorded with a pulse amplitude-modulated fluorometer (Dual-PAM-100; Walz). Before measurements, cell suspensions at a Chl concentration of 15 µg mL⁻¹ were dark adapted for 10 min. The maximal change of P700 upon transformation of P700 from the fully reduced to the fully oxidized state, P_m, was achieved by the application of a saturation pulse after preillumination with FR light. The Y(I) was calculated as Y(I) = (P_m′ – P)/P_m.

Oxidation and rereduction of P700 were monitored from the cell suspensions at a Chl concentration of 20 μg mL⁻¹. The cells were dark adapted for 2 min before measurements. For P700 oxidation, cells were illuminated with strong FR light (720 nm, 75 W m⁻²) for 5 s, and the subsequent rereduction was recorded in darkness.

Proteomics

Western Blotting: Protein Isolation, Electrophoresis, and Immunodetection

Total protein extracts of Synechocystis cells were isolated as described by Zhang et al. (2009). Proteins were separated by 12% (w/v) SDS-PAGE containing 6 m urea, transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore), and analyzed with protein-specific antibodies. The membrane fractions to confirm the localization of NdbC were prepared as described by Zhang et al., (2004).

Sample Preparation for MS

Total proteins were isolated from cell cultures as described by Zhang et al. (2009). An equivalent of 30 µg of total protein was mixed with 2× Laemmli buffer (with 8 m urea) in a 1:1 ratio and loaded on a 12% (50% acrylamide and 1.3% bis-acrylamide) stacking gel (0.5 m Tris-HCl, pH 6.8) containing 6 m urea and no SDS. Gel bands containing all proteins were reduced, alkylated, and digested according to the protocol described by Shevchenko et al. (1996, 2006). Extracted peptides were vacuum dried and further solubilized in 2% acetonitrile (ACN) and 0.1% formic acid (FA) just before MS analysis. Samples were prepared in four biological replicates from the wild type and the ΔndbC mutant.

MS Analyses

DDA

The peptide mixtures from the wild type and the ΔndbC mutant were analyzed on a liquid chromatography-MS/MS system with the DDA method in three biological replicates. An equivalent of 200 ng of peptides was separated on a nano-liquid chromatography system (EasyNanoLC 1000; Thermo Fisher Scientific) coupled to QExactive (Thermo Fisher Scientific). The samples were first loaded on a trapping column and separated in line on a 15-cm C18 column (75 μm × 15 cm, Magic 5-μm 200 Å C₁₈; Michrom BioResources). The mobile phase consisted of water:ACN (98:2, v/v) with 0.2% FA (solvent A) or ACN:water (95:5, v/v) with 0.2% FA (solvent B). To separate the peptide mixture, a 110-min gradient was used that started with 2% B, increased to 20% B in 70 min and further to 40% B in 30 min, and then reached 100% in 5 min. MS data were acquired automatically in positive ionization mode with 2.3-kV ionization potential using Thermo Xcalibur software (Thermo Fisher Scientific). The MS method combined an MS survey scan of mass range 300 to 2,000 mass-to-charge ratio and MS/MS scans for up to 10 most intense peaks from the full scan, with charge +2 or +3. The spectra were registered with a resolution of 70,000 and 17,500 (at mass-to-charge ratio 200) for full scan and for fragment ions, respectively. Fragmentation of ions occurred in high collision dissociation mode with normalized collision energy of 27%. The automatic gain control was set to maximum fill time of 100 and 250 ms and to obtain the maximum number of 3e6 and 2e4 ions for MS and MS/MS scans, respectively.

Protein Identification

The raw files were searched against the Synechocystis protein database retrieved from Cyanobase (Kaneko et al., 1996) using an in-house Mascot (version 2.4) search engine (Perkins et al., 1999) and further analyzed using Proteome Discoverer (version 1.4) software (Thermo Scientific). The Mascot search parameters were set to trypsin as an enzyme with two miscleavages allowed, Met oxidation as variable modification, and carbamidomethylation as the fixed one. Precursor mass tolerance was restricted to monoisotopic mass of ±4 ppm and fragment ion to ±0.02 D. For validation of the spectrum identifications, the Percolator algorithm was used with a relaxed false discovery rate of 0.05.

DDA Quantification

Global label-free quantification for DDA data was performed using Progenesis QI for proteomics, LC-MS 4.0 (Nonlinear Dynamics). Peptide ions were identified with identification results from the DDA experiment. Proteins were identified and quantified with at least two peptides. Protein abundances were estimated based on volumes of the peaks representing its peptides detected in three biological replicates. Relative protein quantification was performed using all identified with ±10 ppm mass error for precursor mass, with nonconflicting peptides per protein. The statistical significance threshold in ANOVA was set to P < 0.05, and the practical significance threshold for FC was set to ±1.3.

Targeted Quantification

For targeted quantification with SRM, a spectral library was created in Skyline (MacLean et al., 2010) based on the DDA results. SRM analysis was performed on wild-type and ΔndbC mutant samples injected with four biological replicates according to the protocol described by Vuorijoki et al. (2016), with several modifications. The samples were injected with two transition lists, comprising assays for 109 proteins (106 targets and three housekeeping proteins), to ensure dwell time of at least 20 ms with cycle time of 2.5 s. The mobile phases used in chromatographic separation were 0.1% FA (A) and 80% ACN/0.1% FA (B). The 55-min gradient started with 8% B, which increased to 26% in 35 min and then to 43% in 15 min. The SRM assays for Synechocystis proteins were loaded from the Panorama Public (Sharma et al., 2014) data repository, and the assays for the missing targets were developed according to the published protocol. Data analysis was performed in RStudio with MSstats algorithm version 3.5.1 (Choi et al., 2014). Data were normalized with global standard normalization with the following housekeeping peptides: sll1818, FSLEPLDR and SYTDQPQIGR; slr0638, GVIAVTER; and sll0145, AISLSDLGLTPNNDGK.

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. PCR analysis to verify full segregation of the ΔndbC mutant.

• Supplemental Figure S2. Reverse transcription-quantitative PCR analysis of sll1485 and sll1486 expression in the wild type and the ΔndbC mutant.

• Supplemental Figure S3. Relaxation of the flash-induced fluorescence yield in darkness from the wild type and the ΔndbC mutant in different conditions.

• Supplemental Figure S4. Light microscope images from wild-type and ΔndbC cells grown under 4 h of 50 µmol photons m⁻² s⁻¹/20 h of dark + 10 mm Glc and under mixotrophic conditions.

• Supplemental Table S1. Photosynthetic parameters of the wild type and the ΔndbC mutant grown in LC and HC conditions.

• Supplemental Table S2. Identification of proteins in the wild type and the ΔndbC mutant grown under HC conditions with DDA.

• Supplemental Table S3. Identification of proteins in the wild type and the ΔndbC mutant grown under LC conditions with DDA.

• Supplemental Table S4. Results of global protein quantification for the ΔndbC mutant versus the control (wild type) grown under HC conditions (P ≤ 0.05).

• Supplemental Table S5. Results of global protein quantification for the ΔndbC mutant versus the control (wild type) grown under LC conditions (P ≤ 0.05).

• Supplemental Table S6. Results of SRM protein quantification for the ΔndbC mutant versus the control (wild type) grown under LC conditions.

Glossary

PM

plasma membrane

PQ

plastoquinone

Cyt

cytochrome

CET

cyclic electron transfer

LAHG

light-activated heterotrophic growth

Chl

chlorophyll a

LC

ambient CO₂

HC

3% CO₂

MIMS

membrane inlet mass spectrometry

DCMU

3-(3,4-dichlorophenyl)-1,1-dimethylurea

Y(II)

effective yield of PSII

Y(I)

effective yield of PSI

DDA

data-dependent analysis

FC

fold change

MS/MS

tandem mass spectrometry

SRM

selected reaction monitoring

OPP

oxidative pentose phosphate

MS

mass spectrometry

ETC

electron transport chain

EMP

Embden-Meyerhof-Parnas

ABC

ATP-binding cassette

FR

far-red

ACN

acetonitrile

FA

formic acid