Effects of global transcription factor NtcA on photosynthetic production of ethylene in recombinant Synechocystis sp. PCC 6803

Background Cyanobacteria are considered potential photosynthetic microbial cell factories for biofuel and biochemical production. Ethylene, one of the most important organic chemicals, has been successfully synthesized in cyanobacteria by introducing an exogenous ethylene-forming enzyme (Efe). However, it remains challenging to significantly improve the biosynthetic efficiency of cyanobacterial ethylene. Genetic modification of transcription factors is a powerful strategy for reprogramming cellular metabolism toward target products. In cyanobacteria, nitrogen control A (NtcA), an important global transcription regulator of primary carbon/nitrogen metabolism, is expected to play a crucial role in ethylene biosynthesis. Results The partial deletion of ntcA (MH021) enhanced ethylene production by 23%, while ntcA overexpression (MH023) in a single-copy efe recombinant Synechocystis (XX76) reduced ethylene production by 26%. Compared to XX76, the Efe protein content increased 1.5-fold in MH021. This result may be due to the release of the negative regulation of NtcA on promoter PcpcB, which controls efe expression. Glycogen content showed a 23% reduction in MH021, and the ratio of intracellular succinate to 2-oxoglutarate (2-OG) increased 4.8-fold. In a four-copy efe recombinant strain with partially deleted ntcA and a modified tricarboxylic acid (TCA) cycle (MH043), a peak specific ethylene production rate of 2463 ± 219 μL L−1 h−1 OD730−1 was achieved, which is higher than previously reported. Conclusions The effects of global transcription factor NtcA on ethylene synthesis in genetically engineered Synechocystis sp. PCC 6803 were evaluated, and the partial deletion of ntcA enhanced ethylene production in both single-copy and multi-copy efe recombinant Synechocystis strains. Increased Efe expression, accelerated TCA cycling, and redirected carbon flux from glycogen probably account for this improvement. The results show great potential for improving ethylene synthetic efficiency in cyanobacteria by modulating global regulation factors. Electronic supplementary material The online version of this article (doi:10.1186/s13068-017-0832-y) contains supplementary material, which is available to authorized users.


Background
An increased consumption of fossil resources has accelerated the development of alternative routes for producing renewable fuels and chemicals. Photoautotrophic cyanobacteria are promising solar biocatalysts for the production of various target products due to their genetic tractability, fast growth, and high photosynthetic efficiency [1,2]. Ethylene, a widely used raw material in the chemical industry and in consumer markets, has already been synthesized in model strains Synechococcus elongatus PCC 7942 (hereafter called S. elongatus PCC 7942) [3] and Synechocystis sp. PCC 6803 (hereafter called S. PCC 6803) [4][5][6][7][8] by introducing an extrinsic ethylene-forming enzyme (Efe) pathway [9,10].
Many efforts have been devoted to constructing stable and efficient ethylene-producing strains in recent years, including (1) optimizing host-preferred codons for efe expression [5,7], (2) [5,6], (3) increasing copy numbers of efe [5,7], (4) modifying ribosome binding sites upstream of efe [8], and (5) modifying related metabolic pathways [5]. In addition to genetic manipulations, a cultivation process for ethylene production was also optimized, considering factors such as light intensity, medium components, and CO 2 supply [5,7]. A volumetric ethylene production rate of 9.7 mL L −1 h −1 was achieved in our previous study by introducing three copies of efe in S. PCC 6803 with blockage of the 2-oxoglutarate decarboxylase (OGDC) and γ-aminobutyrate (GABA) shunts of the tricarboxylic acid (TCA) cycle and expression of 2-OG permease from E. coli; this is the highest cyanobacterial ethylene productivity to our knowledge [5]. Recently, it was found that S. PCC 6803 harbors a complete TCA cycle with a low flux (13% of total fixed carbon) [8,11] and that an exogenous Efe pathway can turn the bifurcated architecture into a cyclic pathway with an enhanced carbon flux (37% of total fixed carbon) [8]. With an in-depth understanding of the plasticity of cyanobacterial metabolism, a global rebalance of cellular carbon and nitrogen metabolism is important to develop more powerful ethylene photosynthetic cell factories. Global transcription machinery engineering (gTME) [12] has already been successfully applied to regulate primary metabolism toward target products in both prokaryotic and eukaryotic systems [13]. An improvement in cyanobacterial production of polyhydroxybutyrate (PHB, product of phaAB operon) [14] and hydrogen [15] has been reported by overexpression of rre37 (encoding response regulator 37) or sigE (encoding RNA polymerase sigma factor E), suggesting the huge potential of using the gTME method to enhance target metabolite production in cyanobacteria.
As a key precursor of ethylene biosynthesis, 2-OG is one of the most important indicators of carbon/nitrogen metabolic balance. The 2-OG pool is regulated by global transcription factor nitrogen control A (NtcA), which implies that NtcA should play a crucial role in ethylene production in cyanobacteria. The glutamine synthetase (GS, encoded by glnA)/glutamine oxoglutarate aminotransferase (GOGAT, encoded by glsF or gltB/gltD) [16,17] cycle is a pathway that represents a connecting step between carbon and nitrogen metabolism in S. PCC 6803 (Fig. 1). GS is a key enzyme in nitrogen assimilation, and it is regulated in multiple ways by NtcA, e.g., NtcA can positively modulate the transcription of glnA [18], negatively regulate the transcription of gifA and gifB (encoding GS inactivating factors IF7 and IF17, respectively) [18,19], and positively modulate the transcription of nitrogen stress-induced RNA 4 (NsiR4, negative regulator of IF7) [20]. In addition to genes related to nitrogen metabolism, NtcA regulates genes in a variety of other cellular processes (such as carbon metabolism and photosynthesis) as well as several sigma factors [20,21]. Although it was reported that ntcA can only be partially deleted in S. PCC 6803 [19,22], a non-completely segregated mutant exhibited abolished activations of ntcA on positive regulons (e.g., glnA, glnB, and glnN promoters) [22] and repressions of ntcA on negative regulons (e.g., gifA and gifB promoters) [19]. On the other hand, overexpression of ntcA leads to wide alterations in primary metabolism and a close to 90% loss of the intracellular 2-OG pool [23]. In addition, NtcA directly acts on sugar catabolism, which is indispensable to metabolism in cyanobacteria, by activating the transcription of sigE [24] and rre37 [25]. These results revealed that genetic modification of ntcA severely disturbed cellular metabolism in S. PCC 6803.
In this study, ntcA was chosen as a genetic engineering target to evaluate its effects on ethylene production in S. PCC 6803. We constructed ntcA partial deletion mutants and ntcA overexpression mutants using S. PCC 6803 wild-type (WT) and ethylene-producing Synechocystis recombinants as the parent strains. In addition, we analyzed the Efe protein level, glycogen content, levels of the substrate and the accompanying product of the Efe-catalyzed reaction (2-OG and succinate of TCA cycle, respectively), and ethylene productivity in these strains.

Chemicals and reagents
The kits used for molecular cloning were from TransGen (China) or Takara (Japan). Taq DNA polymerase and restriction enzymes were obtained from Fermentas (Canada) or Takara (Japan). The antibodies and chemicals used for western blotting were from Sangon Biotech (China). Pure ethylene gas standard was purchased from Heli Gas Co., Ltd (China). The α-ketoglutarate assay kit was ordered from Sigma-Aldrich (USA). Unless otherwise specified, all other chemical reagents were purchased from Sigma-Aldrich (USA).

Strain and plasmid construction
All plasmids, strains, and primers used in this study are listed in Table 1, and the details are described in this section. Clones of the PCR products were confirmed by sequencing.

Plasmids used to knock out ntcA (sll1423) by insertional inactivation
The open reading frame (ORF) of ntcA and its flanking regions were amplified by PCR from S. PCC 6803 genomic DNA using primers ntcA-d1/d4. The amplified PCR fragment was cloned into pMD19-T simple (Takara), resulting in pHM001. Plasmid pHM001 was digested completely with BamHI and BglII to remove the major coding regions of ntcA (465 bp) and was blunted with T4 DNA polymerase. Then, the plasmid was ligated with the kanamycin resistance (Km r ) cassette excised from pRL446 with PvuII, resulting in pHM002. Plasmid pHM002, which harbors the ntcA upstream flanking region (ntcAU) with 125 bp of the 5′ end of ntcA, Km r cassette, and ntcA downstream-flanking region (ntcAD) with 88 bp of the 3′ end of ntcA, was used to knock out ntcA by insertional inactivation.

Plasmids used to knock out ntcA (sll1423) by deletion
The flanking regions of ntcA (ntcAU 883 bp and ntcAD 905 bp) were amplified by PCR from S. PCC 6803 genomic DNA using primers ntcA-d1/d2 and ntcA-d3/d4, respectively. The Km r cassette was amplified by PCR from pRL446 using primers kan-1/2. The ntcAU-Km r -ntcAD cassette was generated by fusion PCR using primers ntcA-d1/d4 and was inserted into pMD19-T simple, resulting in pHM003. Plasmid pHM003, which harbors the ntcA upstream flanking region, Km r cassette and ntcA downstream-flanking region, was used to knock out ntcA by deletion.

Plasmids used to overexpress ntcA (sll1423) at the phaAB loci
The coding region of ntcA was amplified by PCR from S. PCC 6803 genomic DNA using primers ntcA-x1/x2 and was inserted into the Xcm I-digested pFL-XS BioBrick "T" vector that we previously constructed [5], resulting in pHM004. The DNA fragment containing the promoter region of the cpcB gene (gentamycin resistance (Gm r )-P cpcB cassette) was excised from pXX62 with EcoRI and SpeI and was inserted into the EcoRI and XbaI sites of pHM004 to drive the expression of ntcA, resulting in pHM005. The Gm r -P cpcB -ntcA expression cassette was excised from pHM005 with EcoRI and HindIII, blunted with T4 DNA polymerase, and inserted into the NcoI site (blunted) of pKC104 [26] with the 5′ and 3′ end flanking regions of the phaAB loci, resulting in pHM006. Plasmid pHM006 was used to simultaneously overexpress ntcA and block polyhydroxybutyrate biosynthesis.

Plasmids used to knock out ntcA and express efe at the ntcA locus in multi-copy efe recombinants
The Gm r cassette was excised from pXX58 with EcoRI and HindIII, blunted with T4 DNA polymerase, and inserted into EcoRI-digested and blunted ntcA deletion plasmid pHM003, resulting in pHM008, which was used to disrupt ntcA in strain XX109 (three-copy efe). The P cpcB -efe cassette was excised from pXX55 with EcoRI and XbaI and was inserted into the EcoRI and SpeI sites of pXX58, resulting in pHM009. The Gm r -P cpcB -efe expression cassette was excised from pHM009 with EcoRI and HindIII, blunted with T4 DNA polymerase, and inserted into EcoRIdigested, T4 DNA polymerase-blunted ntcA deletion plasmid pHM003, resulting in pHM010. Plasmid pHM010 was used to introduce another efe copy at the ntcA site.

Generation of cyanobacterial transformants
Transformation of WT and mutant S. PCC 6803 was performed according to the established procedures [27]. Plasmids pHM002 and pHM003 were transformed to S. PCC 6803 to generate the ntcA disruption strains MH013 and MH015, respectively, and pHM006 was transformed to generate the ntcA overexpression strain MH017. Plasmids pHM003 and pHM006 were transformed to the one-copy efe recombinant strain XX76 to generate MH021 and MH023, respectively. Plasmid pHM008 was transformed to the three-copy efe recombinant strain XX109 to generate MH039. Plasmid pHM010 was transformed to XX109 to generate the four-copy efe Synechocystis recombinant MH043. All the Synechocystis mutants constructed here are listed in Table 1, and their genotypes were confirmed by PCR using the corresponding primers listed in Table 1.

Quantification PCR (qPCR) of ntcA copy numbers
The genomic DNA was purified with a Genomic DNA Purification Kit (Sangon Biotech, China). Amplification was performed using SYBR ® Premix Ex Taq ™ II (Takara) with gene-specific primers (Table 1) and genomic DNA of triplicate technical replicates from duplicate biological cultures. The rnpB (encoding the RNA subunit of ribonuclease P) and rrn16Sa (encoding the 16S ribosomal RNA) genes were used as reference genes. The primers used in qPCR are listed in Table 1: rnpB-f1/f2, with a product size of 246 bp; rrn16Sa-f1/f2, with a product size of 266 bp; and ntcA-f1/f2, with a product size of 242 bp. The final PCR mixture contained primers at a final concentration of 625 nM in a total volume of 20 μL. The two-step cycling conditions were 10 min at 95 °C and 45 cycles at 95 °C for 20 s and 60 °C for 1 min. The relative genomic DNA copy numbers were determined with LightCycler 480 software according to the 2 −ΔΔCT method [28]. To correctly calculate the relative genomic DNA copy number, PCR efficiency was determined by a dilution series of the genomic DNA.

Ethylene production assay
Ethylene production was determined as described in our previous report [5]. The quantitative analysis of ethylene production was performed with the help of an ethylene calibration curve (Additional file 1: Figure S1).

Spectrophotometric determination of pigment contents
The Synechocystis strains were inoculated at OD 730 ~ 0.1 and were grown photoautotrophically. An appropriate aliquot of culture was centrifuged at the indicated times (the 2nd, 3rd, and 4th days, respectively) and was resuspended in fresh BG11 to OD 730 ~ 0.5. Then the wholecell absorbance spectra (550-750 nm) were recorded on a Beckman Coulter DU-800 UV/VIS spectrophotometer. The pigment contents were calculated based on the absorbance maxima as follows: 630 nm for phycocyanin (PC) and 680 nm for chlorophyll a (Chl a).

Protein extraction and western blotting analysis
Fifty milliliters of cell culture grown for 4 days (OD 730 ~ 1.0) was centrifuged, resuspended in 500 μL of 20 mM phosphate buffer (Na 2 HPO 4 and NaH 2 PO 4 , pH 7.4), and disrupted with glass beads (0.25-0.30 mm) by vigorous vortexing (45 min in cycles of mixing for 40 s and cooling for 20 s). The supernatant was collected after centrifugation for western blotting analysis, which was performed as described previously [5].

Determination of intracellular 2-OG
Intracellular 2-OG was extracted as previously described [18], with minor revisions. Fifty milliliters of cell culture grown for 4 days (OD 730 ~ 1.0) was harvested by centrifuging at 10,000g for 5 min at 4 °C. The pellets were resuspended in 500 μL of cold 0.3 M HClO 4 and were incubated for 15 min on ice. The resulting lysates were centrifuged at 15,000g for 15 min at 4 °C. Then the supernatant was neutralized by the addition of 37.5 μL of cold 2 M K 2 CO 3 . The KClO 4 precipitate was removed by centrifugation, and the supernatant containing 2-OG was stored at −80 °C for analysis. The intracellular 2-OG concentration was determined using the α-ketoglutarate assay kit (Sigma-Aldrich, USA).

Glycogen measurement
Glycogen content was measured as described previously [29], with minor modification. To isolate glycogen, a certain amount of cells, equivalent to 10 OD 730 units of Synechocystis culture, was collected and rinsed twice with 1 mL of deionized water. After centrifugation, the cells were suspended in 30% (W/V) KOH and were incubated at 95 °C for 2 h. Glycogen was then precipitated by the addition of ice-cold ethanol to a final concentration of 70-75% (V/V) and was subsequently incubated at −20 °C overnight. After centrifugation, the glycogen pellet was washed twice with 70 and 100% (V/V) ethanol and was dried by vacuum centrifugation at 65 °C. The isolated glycogen was resuspended in 100 mM sodium acetate (pH 4.5) and was digested to glucose by glucosidase at 65 °C for 2 h. Soluble glucose from the digests was determined using an SBA-40c biosensor analyzer (Shandong Academy of Sciences, China) equipped with a glucose oxidaseimmobilized membrane. To determine dry cell weight (DCW), cells were filtered through sterile nitrocellulose membranes (0.45 μm, Jinteng, China), washed twice with 1 mL of deionized water, and dried at 105 °C for 24 h. Then the content of glycogen was calculated based on the weight of the dried cells equivalent to those were used for glycogen determination.

Inactivation and overexpression of ntcA in S. PCC 6803 and ethylene-producing recombinants
With the aim of evaluating the effect of NtcA on ethylene production, we first inactivated the transcription factor in S. PCC 6803. Two strategies, insertional inactivation and deletion, were employed, and the resultant mutants were denoted as MH013 (Additional file 1: Figure S2) and MH015 (Fig. 2a), respectively. Unlike ntcA of S. elongatus PCC 7942 and Anabaena sp. PCC 7120, which can be completely inactivated (using ammonium instead of NaNO 3 in BG11 medium, hereafter referred to as BG11 A ) [30,31], we found that both MH013 (ntcA insertional inactivation) (Additional file 1: Figure S2) and MH015 (ntcA deletion) (Fig. 2a) were only partially inactivated. These findings are consistent with previous reports that NtcA is essential for S. PCC 6803 and cannot be fully knocked out when cultured in BG11 or BG11 A medium [19,22]. Next, we constructed MH021 (Additional file 1: Figure S3a, b) by deleting ntcA in XX76 (single-copy efe). For segregationlevel analysis, qPCR was performed to measure the relative gene copy number of ntcA in MH015 (ntcA deletion) and MH021 (ntcA deletion in single-copy efe). Our results indicated that MH015 and MH021 exhibited approximately 51 and 34% loss of nctA WT copies relative to their starting strains, S. PCC 6803 and XX76, respectively (Fig. 2c).
We also overexpressed ntcA under the control of the P cpcB promoter on the phaAB loci in S. PCC 6803 and XX76, denoted as MH017 (ntcA overexpression) and MH023 (ntcA overexpression in single-copy efe), respectively (Fig. 2b, Additional file 1: Figure S3a, b). Successful integration of the Gm r -P cpcB -ntcA cassette was achieved, and qPCR revealed that the relative gene copy number of ntcA increased by more than twofold in both MH017 and MH023 compared to their respective starting strains (Fig. 2c).
Moreover, we found that the deletion and overexpression mutants were both genetically stable with a similar segregation level after several rounds of subculture.

Ethylene production was enhanced by ntcA inactivation and was reduced by ntcA overexpression
After the generation of ntcA deletion and overexpression mutants, specific growth rates were evaluated in standard BG11 medium. As shown in Fig. 3a, the two starting strains, S. PCC 6803 and XX76 (single-copy efe), showed similar specific growth rates. However, ntcA partial deletion mutant MH015 and ntcA overexpression strain MH017 both showed restrained growth (Fig. 3b). Intriguingly, the specific growth rates of the ethylene producer (MH021) with ntcA partial deletion and the ntcA overexpression recombinant (MH023) recovered in comparison to the growth rates of MH015 and MH017, respectively. Strain MH021 even exhibited a similar specific growth rate to those of S. PCC 6803 and XX76 (Fig. 3a).
Ethylene production was also monitored, and the results indicated that all ethylene producers reached their peak production in terms of volumetric ethylene overexpression producer MH023, both the Efe level and the peak specific ethylene production were similar to those of XX76.
As the efe gene was under the control of P cpcB in this study, the increased level of Efe indicated a possible direct or indirect transcriptional regulation of NtcA on P cpcB . Hence, we set up a sequence analysis of P cpcB and found that P cpcB might harbor a putative NtcA binding site (Additional file 1: Figure S4; Additional file 2: Table S1). The 34-bp region containing the putative NtcA binding site is directly upstream of the P cpc560 identified by Zhou et al. [32]. In addition, a broader investigation of P cpcB from other cyanobacterial species indicated a widespread existence of putative NtcA production in the first 4 days (Fig. 3c). Ethylene production in the ntcA overexpression recombinant (MH023) was maximally reduced by 26% compared to that of the baseline strain XX76 (Fig. 3c). In contrast, volumetric ethylene production in the ntcA partial deletion mutant (MH021) was maximally increased by 23% compared to that of XX76 (Fig. 3c). Regarding specific ethylene production, MH021 reached a peak value of 1572 ± 77 μL L −1 h −1 OD −1 730 , which was 1.7-fold higher than that of XX76. However, MH023 (ntcA overexpression in single-copy efe) generally showed slightly reduced or unchanged specific ethylene production, despite its 7% lower specific growth rate (Fig. 3a, d). Thus, our results indicate that the deficiency of ntcA is an effective strategy for enhancing ethylene production in S. PCC 6803.

Deficiency of ntcA enhanced the expression of the P cpcB -controlled exogenous efe gene
To explore why the inactivation of ntcA positively affected ethylene production, we first detected the level of Efe in recombinant Synechocystis. From western blotting and subsequent quantitative analysis, we found that strain MH021 (ntcA deletion in single-copy efe) produced 1.5-fold more Efe protein than the starting strain, XX76 (single-copy efe) (Fig. 4a-c). This alteration was in accordance with the 1.7-fold increase in peak specific ethylene production of MH021 (Fig. 4d). In the ntcA The open reading frame (ORF) of ntcA (sll1423) was replaced by a kanamycin resistance (Km r ) cassette through homologous recombination with plasmid pHM003. DNA fragments were amplified by PCR and were analyzed by agarose gel electrophoresis, showing the partial segregation of ntcA in the mutant strains. PCR products from S. PCC 6803 were generated using the indicated primer pairs. Primer sequences are listed in Table 1. Lanes were loaded with PCR products that were generated with genomic DNA from the indicated strain as a template. The sizes of the PCR products are indicated on the right. b Schematic representation of the overexpression of ntcA. The Gm r -P cpcB -ntcA expression cassette was inserted into the phaAB loci through homologous recombination with plasmid pHM006. DNA fragments were amplified by PCR and were analyzed by agarose gel electrophoresis, which showed the complete replacement of phaAB by the Gm r -P cpcB -ntcA expression cassette. PCR products from S. PCC 6803 were generated using the indicated primer pairs. Primer sequences are listed in Table 1. Lanes were loaded with the PCR products that were generated with genomic DNA from the indicated strains as templates. The sizes of the PCR products are indicated on the right. c Quantitative PCR results of ntcA deletion and overexpression mutants. The copy numbers of ntcA were measured through qPCR. The reference genes for S. PCC 6803 were rnpB and the 16S rRNA gene. The relative ratios of gene copy numbers of ntcA were quantified in the WT strain, ethylene producer XX76, and ntcA mutant strains (MH015, MH021, MH017, and MH023). Data represent the means ± standard deviations from three independent experiments ▸ recognition sites (Additional file 2: Table S1). Furthermore, we measured the PC content and PC/Chl a ratio of the Synechocystis strains with whole-cell absorbance spectra. MH015 (ntcA deletion) and MH021 (ntcA deletion in single-copy efe) showed increases at most of the investigated culture stages, while MH017 (ntcA overexpression) and MH023 (ntcA overexpression in single-copy efe) showed decreases in PC content and PC/Chl a ratio relative to S. PCC 6803 and XX76, respectively (Additional file 1: Figure  S5a, b). Thus, improved cellular Efe levels and ethylene production are partially related to the release of negative regulation of NtcA on the P cpcB -efe cassette.

NtcA modification mobilized cellular glycogen for ethylene production
Since NtcA was found to regulate sugar metabolism, the glycogen content of S. PCC 6803 and the recombinant strains were determined to explore the relationship between cellular glycogen and ethylene production. We found that the glycogen content of S. PCC 6803 ranged from 6 to 7.4% of the DCW in the period corresponding to robust ethylene synthesis in the recombinant strains (Fig. 5). Compared to S. PCC 6803, the single-copy efe strain XX76 showed a similar pattern of glycogen accumulation. The deletion of ntcA (MH015) caused a 20% loss of glycogen on the 2nd day but rapid glycogen accumulation relative to S. PCC 6803, and 20% more glycogen content was observed on the 4th day (Fig. 5). As expected, MH021 (ntcA deletion in single-copy efe) showed a 14-23% reduction in glycogen content compared to XX76 (Fig. 5). In terms of the overexpression of ntcA, both MH017 (ntcA overexpression) and MH023 (ntcA overexpression in single-copy efe) showed significantly lower glycogen contents relative to their parent strains, S. PCC 6803 and XX76, respectively. Interestingly, the glycogen content of MH023 was much higher than that of MH017 by the 2nd day, and the glycogen accumulation trends in MH017 and MH023 were reversed; the glycogen level in MH017 was lower at the beginning and increased later, while the glycogen level in MH023 was higher at the beginning and decreased later (Fig. 5). The results of glycogen content analysis revealed that the introduction of the Efe pathway did not change c Volumetric ethylene production. d Specific ethylene productivity. Data represent the means ± standard deviations from three independent experiments the cellular glycogen level. However, the glycogen level was changed by simultaneously introducing the Efe pathway and modifying NtcA. NtcA deficiency might mobilize cellular glycogen for ethylene synthesis considering the enhanced ethylene production in MH021.

NtcA inactivation affected the intracellular pool of succinate and 2-OG
Succinate and 2-OG, two intermediates of the TCA cycle, are directly related to the Efe reaction. To investigate the metabolic effect of ntcA deletion on the unusual TCA cycle, the intracellular levels of succinate and 2-OG in the Synechocystis strains were monitored on the 4th day. We found that the pool of 2-OG (1.62 ± 0.26 nmol mL −1 OD −1 730 ) was over 6 times larger than that of succinate (0.22 ± 0.05 nmol mL −1 OD −1 730 ) under standard culture conditions in S. PCC 6803 WT (Fig. 6). These results confirmed the reported metabolic gap between these two metabolites, even with functional OGDC and GABA shunts [11]. In ethylene producer XX76 (single-copy efe), the levels of succinate and 2-OG both increased, and the ratio of succinate to 2-OG (Ratio Suc/2-OG ) increased by 73% compared to that of S. PCC 6803 (Fig. 6), indicating enhanced TCA cycling, as previously noted [8]. In MH015 (ntcA deletion), the level of intracellular 2-OG decreased while succinate increased, and the Ratio Suc/2-OG increased threefold compared to that of S. PCC 6803. Deleting ntcA in XX76 (MH021) caused a dramatic enlargement of the intracellular pool of succinate and a slight increase in 2-OG, with the Ratio Suc/2-OG being enlarged by 9-, 4.8-, and 1.5-fold compared to that of S. PCC 6803, XX76, and  MH015, respectively. These results indicated that the flux from 2-OG to succinate dramatically increased by the introduction of Efe when coupled with the partial deletion of ntcA and that this genetic modification reprogrammed the metabolic network of the TCA cycle.

NtcA inactivation in combination with TCA cycle modification enhanced cellular ethylene production in recombinant Synechocystis
In our previous study, we constructed Synechocystis recombinant strain XX109 with three efe copies. In this recombinant, the ogdc (sll1981) and ssadh (slr0370) genes were inactivated by inserting the Km r -P cpcB -efe and Cm r -P cpcB -efe cassettes, respectively [5]. Theoretically, this strain should get an Efe-dominated TCA cycle with primarily cyclic architecture because the OGDC and GABA pathways were both blocked.
To further improve ethylene production, we constructed strains MH039 and MH043 by replacing ntcA with a gentamycin cassette (Gm r ) or a Gm r -P cpcB -efe cassette, respectively, based on the three-copy efe strain XX109 (Additional file 1: Figure S3c, d). Quantitative PCR analysis revealed that MH039 (ntcA deletion in three-copy efe) lost approximately 56% and that MH043 (ntcA deletion in four-copy efe) lost 30% of the nctA WT copies (Additional file 1: Figure S6). Compared to XX109, MH039 showed a similar specific growth rate, and MH043 showed a slightly improved specific growth rate (Fig. 7a). In addition, specific ethylene production was enhanced in these strains (Fig. 7b). (single-copy efe), MH021 (ntcA deletion in single-copy efe), and MH023 (ntcA overexpression in single-copy efe). Data represent the means ± standard deviations from at least two independent experiments Fig. 7 Effect of the deletion of ntcA on ethylene productivity and Efe level in multi-copy efe recombinant Synechocystis strains (XX109, MH039, and MH043). a Specific growth rate. b Maximal specific productivity. c Western blotting detection of Efe. d Quantitative analysis of Efe level from the results of two repeats. Data in a and b represent the means ± standard deviations from three independent experiments A maximal specific ethylene production rate of 2463 ± 219 μL L −1 h −1 OD −1 730 was achieved in MH043 (Fig. 7b), which was higher than the previously reported rate.
The Efe level (Fig. 7c, d) and PC content (PC/Chl a ratio as well) (Additional file 1: Figure S5c, d) both increased in MH039 (ntcA deletion in three-copy efe) and MH043 (ntcA deletion in four-copy efe) compared to that in XX109. The results confirmed the possible negative regulation of NtcA on the P cpcB -efe cassette, even when the OGDC and GABA shunts of the TCA cycle were blocked. However, despite harboring four efe copies, MH043 exhibited a similar Efe level to that of MH039.

Discussion
Here, we enhanced ethylene production by the partial deletion of ntcA and analyzed the corresponding physiological and metabolic effects in Synechocystis strains. NtcA was the target for genetic modification based on its essential role in integrating the signals of nitrogen and carbon balance (Fig. 1). We found that ntcA deletion and overexpression mutants both showed a stable genotype ( Fig. 2; Additional file 1: Figures S3, S6). Compared to previous studies, the lower segregation level of ntcA observed here might be due to different nitrogen resources. Furthermore, cell growth of the non-segregated ntcA mutant was not seriously restrained. Recently, it has been found that ethylene can modulate several important aspects of Synechocystis biology by the functional ethylene receptor (SynEtr1, encoded by slr1212) [33]. NtcA has also been shown to control transcription at a genome-wide level and widely affect the primary metabolism of S. PCC 6803. Consequently, the recovered growth of MH021 (ntcA deletion in single-copy efe) and MH023 (ntcA overexpression in single-copy efe) might relate to the combined effect of ethylene modulation and ntcA modification. On the other hand, the stable genotype, in combination with the recovered physiological feature, demonstrated the feasibility of improving ethylene production in S. PCC 6803.
Previous studies have revealed that the modification of transcription regulators can cause alterations in primary metabolism and can enhance the production of native biological products in S. PCC 6803 [25,34,35]. Thus, it is reasonable that ethylene production was affected by the deletion or overexpression of transcription factor NtcA. We first suspected that the enhanced ethylene production in MH021 (ntcA deletion in single-copy efe) (Fig. 3c,  d) was related to only the cellular metabolic reprogramming imposed by ntcA deletion. However, Efe level detection ( Fig. 4 and 7) and PC content analysis (Additional file 1: Figure S5) implied the possible native regulation of NtcA on P cpcB . Moreover, we found a putative NtcA binding site between 560 bp and 594 bp upstream of the initiation codon of the cpcB gene (Additional file 1: Figure S4; Additional file 2: Table S1). In fact, it has been reported that the transcription factor binding sites (TFBSs) on P cpcB are crucial for its promoter strength in S. PCC 6803; however, the NtcA site was not included (only 560 bp upstream of the initiation codon was analyzed) [32]. NtcA belongs to the CRP (cAMP receptor protein) family and binds as a homodimer to the conserved palindromic sequence GTAN 8 TAC [20,21]. As NtcA promoters were thought to serve as regulatory points to coordinate photosynthesis and nitrogen uptake [21], the genes involved in various stages of photosynthesis (such as allophycocyanin apc, PC cpc) were supposed to be new members of the NtcA regulons [21]. The possible regulation of NtcA on the P cpcB -efe cassette found in this study is a good example of this hypothesis.
Glycogen is the main carbohydrate storage compound that can divert a significant portion of fixed carbon in S. PCC 6803 [36,37]. A rapid accumulation of glycogen usually occurs when cyanobacteria face stress conditions [37]. In this study, we observed rapid glycogen accumulation in the ntcA mutant MH015, but a lower level of glycogen (maximal loss of 23%) and a slow glycogen accumulation rate were observed in MH021 (ntcA deletion in single-copy efe) under standard culture conditions (Fig. 5). Thus, it is plausible that the Efe reaction can draw many carbons from glycogen when the ntcA transcription factor is deleted. Lower glycogen content results from enhanced catabolism and/or attenuated anabolism. It was reported that the key genes involved in glycogen catabolic pathways, such as glycolysis and the oxidative pentose phosphate (OPP) pathway, were regulated by several transcription regulators and sigma factors, including NtcA [22], Rre37 [25,35,38], and SigE [25,34,35]. The coding genes of such regulators (including NtcA itself ) were among the regulons of NtcA [21,39,40]. Thus, under the regulation of NtcA, the intermediates involved in glycogen catabolic pathways can be converted to acetyl-CoA and malate, which are inputted into the TCA cycle. In agreement, a sharply increased Ratio Suc/2-OG was found in MH021 (Fig. 6), which indicated a remodeled metabolic network for the TCA cycle. The results were also supported by identification of the predominantly cyclic architecture of the TCA cycle in recombinant Synechocystis with an efficient Efe pathway. In addition, the positive regulation of NtcA on the GS-GOGAT cycle of nitrogen metabolism might also bring carbon flux to the Efe reaction.
As previously found, ethylene production in cyanobacteria is a complex process involving multiple genetic and environmental factors [4][5][6][7][8]. The highest specific ethylene production in the multi-copy efe strain MH043