Author Correction: The transcriptional regulator NtrC controls glucose-6-phosphate dehydrogenase expression and polyhydroxybutyrate synthesis through NADPH availability in Herbaspirillum seropedicae

An amendment to this paper has been published and can be accessed via a link at the top of the paper.


Results
The ntrC mutant of H. seropedicae SmR1 produces more PHB than the parental strain. The C/N ratio is one of the key factors controlling PHB accumulation in bacteria 26,27 . Accordingly, we anticipated that mutant strains of the NTR system would be a useful tool to investigate its involvement in PHB synthesis. We measured the content of PHB in glnB, glnK, glnD, amtB and ntrC knock-out mutants as well as in the parental strain H. seropedicae SmR1. The strains were grown in low C/N ratio (37 mM DL-malate and 20 mM NH 4 Cl) and high C/N ratio (37 mM DL-malate and 5 mM NH 4 Cl). The cultures were sampled in four stages of growth: early-exponential phase (OD 600 0.4), mid-exponential phase (OD 600 0.8), early stationary phase (OD 600 1.2) and Figure 1. The NTR system and its regulation. In nitrogen excess, high glutamine concentration activates the deuridylylation of GlnB by GlnD. The unmodified form of GlnB stimulates the adenylyl transfer activity of GlnE, resulting in glutamine synthetase adenylation (GS-AMP) and therefore its inactivation. GlnB also interacts with the NtrB stimulating its phosphatase activity, resulting the dephosphorylation of the transcriptionally active form of NtrC. On the other hand, when the bacterium faces a nitrogen-limiting condition, the low glutamine concentration leads to the uridylylation of GlnB (GlnB-UMP) by GlnD. GlnB-UMP stimulates the deadenylylation of GS-AMP, turning GS active. The 2-oxoglutarate (2-OG) is an effector of GlnB, and its concentration is high under nitrogen-limiting conditions. late stationary phase (OD 600 1.6). The PHB contents of the parental, glnK and glnD strains were very similar since their maximum PHB content was around 25%/cdw (cell dry weight) in high C/N (Fig. 2). The PHB was reduced to 16%/cdw when the strains grew in low C/N. Interestingly, the glnB and amtB mutants presented a significant reduction of PHB accumulation compared to the parental strain (Fig. 2). This effect was remarkable in low C/N where glnB and amtB mutants produced only half of the PHB level of the parental strain. In contrast, PHB production in the ntrC mutant was unregulated and presented the highest content of PHB amongst all strains analysed, reaching 32% of PHB regardless the initial C/N ratio (Fig. 2). This higher PHB production regardless the nitrogen level in the medium suggests that ntrC knock-out decouples the synthesis of PHB of the C/N ratio.
The ntrC mutant also produces more PHB with monosaccharides as carbon source. As malate is not a conventional substrate for biotechnological applications due to its high price compared to other sources, we measured PHB accumulation in the ntrC mutant on D-glucose and D-xylose, which are monosaccharides highly abundant in cheap feedstocks and agro-industrial residues. The C/N ratios were the same. When grown in the presence of D-glucose, the maximum PHB production of the ntrC mutant was 1.8-and 1.7-fold higher than the parental production, in low and high C/N ratio, respectively ( Fig. 3A and D). On D-fructose, the maximum PHB production of the ntrC mutant was 2.0-and 2.15-fold higher than the parental production, in low and high C/N ratio respectively ( Fig. 3B and E). Interestingly, when D-xylose was used as the carbon source, in low C/N ratio there was no difference of PHB between both strains, but in high C/N ratio, the ntrC mutant produced 1.5-fold more PHB (Fig. 3C and F). Table 1 shows the maximum content of PHB and the productivities obtained for both strains in the different conditions assayed. The profile of sugar consumption and the data of PHB concentration (g/L) and yield of g PHB/g of substrate for the SmR1 and ntrC strains are shown in the supplemental information ( Figure S1 and Tables S1 and S2). Complementation of the ntrC mutant restores PHB production to the parental level. The ntrC gene is clustered in an operon downstream from glnA and ntrB 18,28 . The ntrC mutant was complemented through pKRT1 conjugation. The pKRT1 is a pLAFR3-derivative containing a glnAntrBC operon copy from H. seropedicae SmR1 23,29 . PHB accumulated in strains harbouring pKRT1 was reduced when grown in malate or glucose, regardless the C/N ratio applied (Fig. 4). Therefore, the complementation of the ntrC mutant demonstrates that the NtrC is directly involved in the higher PHB production observed.
The derepression of the zwf (glucose-6-phosphate dehydrogenase) gene results in high NADPH in the ntrC mutant. Previous works have shown that NtrC is involved in the regulation of the expression of glucose-6-phosphate dehydrogenase (G6PDH) and glutamate dehydrogenase (GDH) 30,31 . Specifically, the G6PDH activity is widely implicated in the generation and maintenance of the NADPH/NADP + ratio 32,33 . The activity of the NADP + -dependent malic enzyme (ME) is also correlated with NADPH/NADP + balance 34 . This observation led us to determine the activity of these enzymes in the parental strain SmR1 and ntrC mutant. The G6PDH and NADP + -dependent ME activities were 2.3-and 1.6-fold higher in the ntrC mutant when grown in NFb-glucose with 20 mM NH 4 Cl ( Fig. 5A and C). GDH activity did not differ statistically between both strains (Fig. 5B). The specific activities of G6PDH, ME and GDH enzymes for the SmR1 and ntrC strains are shown in the Supplemental Figure S2. The repression of zwf transcription by NtrC has been previously reported for   the bacterium Pseudomonas putida 3,30 . The zwf of H. seropedicae is located downstream from pgi (Hsero1099, encoding a phosphoglucoisomerase) and upstream from a gene encoding a transcriptional regulator of the HexR family (Hsero1097) and talB (Hsero1096, encoding a transaldolase). Since the 76 bp region between pgi and zwf seems not to contain a promoter region, the upstream region of pgi was cloned with GFP in the pEKGFP01 to determine the expression profile of the operon carrying zwf in H. seropedicae SmR1 and ntrC mutant. The activity of the Ppgi-gfp fusion in the ntrC mutant was higher than that of the parental strain during all growth phases, achieving a maximum difference of 2.8-fold (Fig. 5D). The transcription of zwf and other genes involved in the Entner-Doudoroff pathway and in the PHB metabolism were compared between the SmR1 and ntrC strains through RNA-seq analysis. The data corroborate the higher expression of zwf in the ntrC mutant (Table S3). Also, the transcription of genes involved in the PHB metabolism was lower in the ntrC mutant (Table S4), indicating that the higher PHB production measured is a consequence of a metabolic factor. This prompted us to measure the NAD(P)H/NAD(P) + ratio in both strains. The NADH/NAD + ratio had no statistically significant difference between the parental strain SmR1 complemented or not with an additional copy of the operon glnAntrBntrC cloned into pLAFR3 (pKRT1) (Fig. 5E). The same was observed for the NADH/NAD + ratio in the ntrC mutant (Fig. 5F). However, the NADPH/NADP + ratio was statistically higher for both strains when they were not complemented with pKRT1. For the ntrC mutant, the NADPH/NADP + ratio was 2.1-fold greater than in the parental strain ( Fig. 5E and F, comparing the red bars). Taken together these results indicate that the knock-out in ntrC increases the expression of G6PDH and the regeneration of NADP + in NADPH.
The ntrC mutant of H. seropedicae SmR1 is more resistant to oxidative stress caused by hydrogen peroxide. A high NADPH/NADP + ratio stimulates PHB production 35,36 . High [NADPH] increases the flux towards PhaB (NADPH-dependent acetoacetyl-CoA reductase), yielding PHB not only as a carbon stock but also a redox sink 2,36 . Furthermore, NADPH is crucial to anti-oxidative defences in most organisms, ensuring a reductive cellular environment to mitigate the deleterious effects of oxidative species as hydrogen peroxide, hydroxyl radical and superoxide 37 . Therefore we determine the resistance to oxidative stress of both strains. Taken these results together, we conclude that the ntrC mutant has a more efficient defence against the oxidative stress, probably due to the NADPH accumulation which can be applied to mitigate the deleterious effects of the oxidative insult. The high NADPH generation also stimulated the PHB synthesis observed in the mutant.

Discussion
The nitrogen level in the growth medium is a key factor interfering with PHB production in bacteria [38][39][40] . Therefore, the interruption of nitrogen regulatory systems can be a useful strategy to reduce the C/N ratio effect on PHB production and ultimately to improve PHB production. Since the NTR system is the master regulator of the nitrogen metabolism in several bacterial species 41 , wherein we investigate the PHB accumulation using as a model the bacterium H. seropedicae SmR1 and a set of mutants defective in the expression of regulatory proteins of the NTR system. Among all strains evaluated only the ntrC mutant produced higher PHB contents than the parental strain. The ntrC mutant produced around the double of PHB than the parental strain (in % PHB/mg of cdw), in all tested conditions. Similar results were also reported to ntrB and ntrC mutants of A. brasilense Sp7 24 , suggesting that possibly the effect of NtrC on PHB synthesis is widespread among prokaryotes. Based on these findings, we anticipated some possible explanations for our results. The synthesis of PHB is largely dependent on high levels of acetyl-CoA and NADPH. Normally, both metabolites are in high intracellular concentrations when the bacterium faces a condition of carbon excess and limitation in another nutrient, such as nitrogen 1 . This condition is permissive for PHB synthesis due to elevating acetyl-CoA level generated by carbon overflow and high NADPH/NADP + ratio as consequence of nitrogen limitation 7,9,35 .
Considering that H. seropedicae ntrC mutant produced more PHB than the parental strain even under unfavourable conditions (low C/N ratio) suggests that NtrC somehow is modulating acetyl-CoA and/or NADPH concentrations. A role of NtrC on directly activating transcription of genes involved in PHB synthesis is unlikely since PHB production is higher in the ntrC mutant even under an excess of ammonium, a condition where NtrC is mainly dephosphorylated and therefore inactive 22 ( Fig. 2A and B). A transcriptomic analysis of the ∆ntrC mutant of the PHA-producing bacterium Pseudomonas putida KT2442 brought important elements to understand the link between NtrC and PHA metabolism 3 . In the ∆ntrC strain, the transcription of zwf-1 (encoding glucose-6-phosphate dehydrogenase -locus tag PP1022) and gap-1 (glyceraldehyde-3-phosphate dehydrogenase -PP1009) was upregulated 5.7-and 2.6-fold, indicating that NtrC represses their expression. Hervás et al. suggested that repression of zwf shows that NtrC controls hexose catabolism in bacteria, likely to prevent a carbon overflow under nitrogen-limiting conditions for growth 3 . In another work, the ntrC deletion in P. putida KT2440 also rendered cells more resistant to oxidative stress 30 . Furthermore, zwf-1 (PP1022) was also up-regulated in the ∆ntrC mutant of P. putida KT2440, as determined by transcriptomic analysis. H. seropedicae SmR1 and P. putida have the same incomplete glycolytic pathway since it lacks 6-phosphofructokinase (PFK-1). As a consequence, the Entner-Doudoroff (ED) pathway converts glucose and fructose into pyruvate and glyceraldehyde-3-phosphate. Particularly, in mutants defective in ntrC expression, it is expected a higher metabolic flux through glucose-6-phosphate dehydrogenase, generating more acetyl-CoA and NADPH, which in turn increases PHB accumulation.
These findings suggest that the up-expression of zwf could be a major factor leading ntrC defective mutants to accumulate more PHB under permissive conditions. The observation that the H. seropedicae ntrC mutant produced more PHB even in unfavourable conditions (low C/N ratio) suggests a higher pool of reducing power as compared to the parental strain. Therefore is likely that the bacterium switches the synthesis of PHB as an electron sink to avoid the deleterious effect of a high reductive environment.
In fact, we have already reported that mutant strains defective in PHB synthesis or accumulation presented a severe growth penalty on glucose 17 . Probably, as a consequence of the redox imbalance caused by the inability of those mutants to divert electrons towards PHB synthesis.
NtrC is widely recognised as a transcriptional regulator restricted to regulate genes involved in nitrogen metabolism. Indeed, a significant portion of genes corresponding to 2% of the E. coli genome was determined to be under control of the NtrC transcriptional activity, most of them involved in amino acids and ammonium transport, nitrogen assimilation and nitrogen recycling 42 . Our results demonstrate that NtrC interferes with PHB synthesis, bring to the discussion if NtrC could be one of the connection points between nitrogen and carbon metabolism. Alternatively, the NtrC control of the NADPH recycling may help cells dealing with a stressful condition such as nitrogen limitation. In fact, the glutamate synthase (GOGAT) and the glutamate dehydrogenase (GDH), which are important reactions for ammonium assimilation require NADPH as an electron donor to catalyse the amination of 2-oxoglutarate 43 . In E. coli, NADPH concentration was homeostatic after ammonium upshift 44 . Since NADPH consumption increases when ammonium assimilation is higher, it is likely that other pathways are generating the NADPH demanding 44 . Whether or not the phosphorylation state of NtrC interferes in NADPH production through zwf expression is uncertain. The repression of gdhA encoding the glutamate dehydrogenase in P. putida KT2442 by NtrC was shown to be independent of the phosphorylation state, since both wild-type and the mutant NtrC D55E,S161F (mimicking the phosphorylated protein) can bind to the gdhA promoter and repress transcription 31 . The effect of phosphorylation could be further studied employing a ntrB mutant of H. seropedicae and complementation of the ntrC mutant with NtrC variants unable to be phosphorylated. Such experiments will determine if the reduction in PHB production observed in the glnB and amtB mutants derived from the phosphorylation state of NtrC.
It would be interesting to investigate further the role of NtrC on PHB synthesis in other PHB-producing models, such as Ralstonia eutropha, Azotobacter vinelandii and P. putida. It would point if the involvement of NtrC on PHB synthesis is conserved among other classes of bacteria, serving as a strategy for metabolic engineering aiming to improve PHB production.

Methods
Bacterial Strains, Plasmids, and Growth Conditions. Strains and plasmids used are listed in Table 2.
Escherichia coli strain Top10 (Thermo Fisher Scientific Inc., Waltham, MA, USA) and E. coli S17-1 45 were used for cloning and conjugational procedures, respectively. E. coli strains grew in LB medium at 37 °C and shaken at 160 rpm. H. seropedicae parental strain SmR1 46 and mutant strains were cultured in NFbHP media with 37 mM DL-malate and the indicated concentration of NH 4 Cl at 30 °C and shaken at 120 rpm 49 . D-glucose at 25 mM, D-fructose at 25 mM and D-xylose at 30 mM were applied as alternative carbon sources for PHB production measurement as a replacement for malate. Two regimes of carbon-to-nitrogen (C/N) ratio were used at the start of cultivation: high C/N ratio with 5 mM of NH 4 Cl and low C/N with 20 mM of NH 4 Cl.
Quantification of PHB. PHB was quantified by methanolysis, followed by GC-FID (gas chromatography coupled to a flame ionisation detector) analyses as described previously 50 with 5 to 10 mg of lyophilised bacteria. Amounts of PHB in each sample were normalised to cell dry weight (cdw; the weight of the lyophilised bacterial pellet) and expressed as % of PHB cell dry weight −1 .

Construction of Ppgi-gfp transcriptional fusion. The intergenic region of the pgi (locus-tag Hsero1099)
was amplified using the primers Fw_Ppgi_Hs 5′TATCTCGAGTGTCGGGTTCCTGTTAGCGT 3′, containing a XhoI site (underlined) and Rev_Ppgi_Hs 5′TATACTAGTCATATGGGTCTGGTGTCGGTTGGCGG 3′, containing a SpeI site (underlined) as previously described 51 . The amplified product was cloned into the sites XhoI and SpeI of the pBBR1MCS-3 52 , generating the pEK07. The reporter gene gfp containing upstream the rbs site B0034 and downstream the double terminator B0015 was extracted from the plasmid BBa_I13504 (Registry of Standard Biological Parts, partsregistry.org) digested with the EcoRI and SpeI enzymes and cloned into pBlueScript II KS(+) digested with EcoRI and XbaI. Then, the gfp cassette was removed by digestion with the XbaI and SacI enzymes and cloned into the pEK07 digested with the same enzymes, generating the pEKGFP01. The pEKGFP01 was transformed in E. coli S17-1 and conjugated to H. seropedicae by bi-parental mating.  Complementation of ntrC mutant. The pKRT1 cosmid was conjugated by bi-parental mating between E. coli S17-1 and H. seropedicae strains. The transconjugant colonies were selected in NFbHP-malate with 20 mM NH 4 Cl agar with 10 µg/mL of tetracycline. The complemented strains were cultivated in NFbHP-malate and NFbHP-glucose at high and low-C/N ratios.
The reactions were monitored for 5 min at 585 nm in a Shimadzu ™ spectrophotometer. The activity was expressed as µmol of formazan/min of reaction per OD 600 of the culture.
Determination of NAD(P)H/NAD(P) + ratio. Intracellular levels of NAD + , NADP + , NADH and NADPH were determined by the improved cyclic assay using either ADH (Sigma #A3263) or G6PDH (Sigma #G6378), respectively 53 . The dinucleotides were extracted using cell pellets from 1 mL of culture, cultivated either until the mid-log (OD 600 of 0.4-0.5) or late-log (OD 600 of 1.0-1.2) phases. Reduced and oxidised nicotinamide adenine dinucleotides were differentially extracted by treatment with alkali or acid, respectively, followed by extract neutralisation. The assays were performed in 200 μL in a water bath for 30 minutes at 37 °C containing the following components: 0.1 M Tricine-NaOH buffer (pH 8.0); 4.2 mM MTT; 40 mM EDTA (disodium salt); 16.6 mM PES; 5 M ethanol as substrate for alcohol dehydrogenase to determine NADH/NAD + or 25 mM glucose 6-phosphate (dipotassium salt) as substrate for G6PDH to determine NADPH/NADP + . To determine NADPH/NADP + and NADH/NAD + , 10 µL of a baker's yeast G6PDH solution (14 units/mL) or 10 µL of a baker's yeast alcohol dehydrogenase solution (100 units/mL) were added per reaction, respectively. Reactions were stopped by adding 100 μL of 5 M NaCl followed by 5 minutes of ice incubation. The precipitated formazan was centrifuged for 5 min at 14,000 x g and solubilised in 500 μL of 96% ethanol. The formazan was quantified as a function of the absorbance at 550 nm of 200 µL of sample in 96-well plates in a Biotek ELX-800 microplate reader. The standard calibration curve was run in triplicate using up to 30 pmol/assay of either NAD(P)H or NAD(P) + standards.
Analysis of intracellular ROS levels using flow cytometry. Cells from 1 mL of culture were collected by centrifugation at 14,000 × g for 1 min and then resuspended in 500 μL of PBS buffer supplemented with 1 mM EDTA, 0.01% Tween 20 and 0.1% Triton X-100. Cells were subsequently incubated with 50 μM 2′-7′-dichlorofluorescein diacetate (H 2 -DCFDA) for 30 min at 30 °C in the dark. Control experiments without H2-DCFDA addition were also set up under the same conditions. Treatment with H 2 O 2 was performed by pre-incubation of cells with increasing concentration of H 2 O 2 for 30 min at 120 rpm and 30 °C, before addition of H 2 -DCFDA. The samples were analysed by flow cytometry using a BD AccuriTM C5 flow cytometer equipped with a 488 nm argon laser and a 533/30 nm bandpass filter (FL1-H). The median fluorescence intensity was used to determine the intracellular ROS levels.
Data availability. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.