Prevention of PHB formation and its effect on amorphadiene biosynthesis
Culturing R. sphaeroides under nitrogen-limited conditions could theoretically result in growth-independent isoprenoid synthesis via the native MEP and the heterologous MVA pathways. However, upon consumption of the limited available nitrogen, R. sphaeroides stores excess carbon intracellularly as PHB, a nitrogen-free carbon and energy storage compound (29). Aiming to increase isoprenoid production, we reasoned that deletion of one of the PHB synthesis genes would block PHB production under nitrogen-limited conditions and could therefore increase the flux through the MVA pathway. Deletion of the phaC1 and phaC2 genes, that code for the PHB polymerase, is the established approach for eliminating PHB biosynthesis in R. sphaeroides (32–35). Nonetheless, this does not prevent activity of the NADPH-dependent acetoacetyl-CoA reductase PhaB, which could result in the undesired accumulation of 3-hydroxybutyryl-CoA or excretion of 3-hydroxybutyrate (Fig. 1). In a recent study, we demonstrated that deletion of the phaB gene prevents PHB biosynthesis (31). Here, we confirmed that the deletion of either the phaB gene (Rs265_∆phaB strain) or the combined deletion of the phaC1 and phaC2 genes (Rs265_∆phaC1∆phaC2 strain) prevents PHB formation both under nitrogen excess and nitrogen-limited conditions (Fig. 2A). As observed before (29), the wild type (Rs265) strain produced substantial amounts of PHB, especially under nitrogen-limiting conditions.
The pBBR-ads plasmid harbouring the heterologous amorphadiene synthase gene was transferred to the various R. sphaeroides strains by conjugation. The resulting strains were cultured under both nitrogen excess and nitrogen-limited conditions. At glucose depletion, we determined (Table 3): amorphadiene titers (CP), yield of amorphadiene on glucose (YP/S), active biomass concentration (CX) and the amorphadiene on biomass ration (YP/X). We observed that, for the Rs265_∆phaB:pBBR-ads and Rs265_∆phaC1∆phaC2:pBBR-ads strains that use only the endogenous MEP pathway (MEP-only strains), elimination of PHB synthesis does not result in higher amorphadiene/biomass ratios compared to the Rs265 wt strain (Fig. 2B, Table 3). In fact, these ratios remained unaffected when moving from nitrogen excess to nitrogen-limited conditions (Table 3) at a value of 2.9 ± 0.2 mg ∙ g of biomass-1. Interestingly, the Rs265_∆phaC1∆phaC2:pBBR-ads strain showed an even lower amorphadiene/biomass ratio compared to the Rs265:pBBR-ads and the Rs265_∆phaB:pBBR-ads strains (Table 3). In summary, when only the endogenous MEP pathway was active, isoprenoid biosynthesis appeared to be strictly growth-coupled. Moreover, the MEP flux was insensitive or negatively affected by the impaired PHB synthesis.
We subsequently transformed the available Rs265, Rs265_∆phaC1∆phaC2 and Rs265_∆phaB strains with the orthogonal MVA pathway, cloned in the pBBR-MVA-ads plasmid. The amorphadiene/biomass ratio increased 10- to 20-fold for all the strains tested (Fig. 2C, Table 3). The highest increase was observed for the ∆phaB strain (Rs265_∆phaB:pBBR-MVA-ads), reaching a ratio of 63.7 ± 4.0 mg ∙ g of biomass-1 (Fig. 2C, Table 3). This value was significantly higher than the ratio reached by the strain with a functional PHB synthesis (Rs265:pBBR-MVA-ads), which was 35.9 ± 3.6 mg ∙ g of biomass-1 (Fig. 2C, Table 3).
Increase of the MVA pathway flux, as consequence of the phaB deletion, was confirmed also for the Rs265-MVA_∆dxr strain, for which the MEP pathway is inactivated via the deletion of the 1-deoxy-D-xylulose 5-phosphate reductoisomerase (dxr) gene, after genomic integration of the MVA pathway (Fig. 1). This strain relies exclusively on the non-native isoprenoid route (MVA-only) (16). Also here, a substantial increase in the amorphadiene/biomass ratio was observed during both nitrogen excess and nitrogen-limited conditions (Fig. 2D, Table 3). The highest value observed was for the Rs265-MVA_∆dxr∆phaB:pBBR-MVA-ads strain, with 36.9 ± 1.6 mg ∙ g of biomass-1 during nitrogen limitation.
In summary, although both the ∆phaC1∆phaC2 and ∆phaB knockouts were equally effective in reducing PHB synthesis, the ∆phaB knockout strain produced more amorphadiene, both volumetrically and per biomass unit (Table 3).
Organic acids secretion as consequence of PHB deletion
We reasoned that comparing the secretion profiles between ∆phaB and ∆phaC1∆phaC2 could provide additional insights on the beneficial effect of ∆phaB on isoprenoid synthesis. Therefore, we quantified by HLPC analysis the organic acids in the spent medium of Rs265, Rs265_∆phaC1∆phaC2 and Rs265_∆phaB harbouring either pBBR-ads or pBBR-MVA-ads plasmids (Fig. 2E-H). For both ∆phaB and ∆phaC1∆phaC2 strains, 2-oxoglutarate (80 to 340 mg ∙ g of biomass-1, Fig. 2F) and pyruvate (150 to 500 mg ∙ g of biomass-1, Fig. 2G) were the main by-products. Both compounds require coenzyme A (CoA) for proceeding further in the metabolism via oxidative decarboxylation (Fig. 1). Excretion of these compounds suggests that free CoA is limiting when PHB biosynthesis is prevented.
The ∆phaC1∆phaC2 strains secreted an additional unknown compound, which was identified as 3-hydroxybutyrate (3HB) by NMR (Additional file 1: Fig. S1). The spent medium for the Rs265_∆phaC1∆phaC2:pBBR-ads strain showed a value of 0.002 ± 0.001 mg 3HB ∙ g of biomass-1 In contrast, a value of 0.004 ± 0.001 mg 3HB ∙ g of biomass-1 was observed for the Rs265_∆phaC1∆phaC2:pBBR-MVA-ads strain, which expressed the heterologous MVA pathway (Fig. 2H). Therefore, under nitrogen limitation and upon expression of the MVA pathway, the amount of 3HB secreted increased significantly compared to when only the MEP pathway was active.
13C metabolic flux ratio analysis of isoprenoid biosynthesis under different growth conditions
The Rs265_∆phaB:pBBR-MVA-ads strain, that overexpresses the MVA pathway and still has an active MEP pathway, showed the highest amorphadiene/biomass ratio under nitrogen-limited conditions (Fig. 2C, Table 3). Additionally, we observed that, independent from the cultivation conditions and the presence of an active PHB synthesis pathway, the dual-pathway (co-expressing MEP and MVA pathways) strains largely outperformed the single-pathway strains (Fig. 2B-D, Table 3). We therefore decided to further investigate the separate and combined contribution of the isoprenoid pathways to the amorphadiene production by 13C flux ratio analysis.
MEP and MVA pathways are known to exert a reciprocal stimulation (30,36). To better understand their mode of interaction under the conditions tested, we determined their contribution via 13C metabolic flux ratio analysis of the Rs265:pBBR-MVA-ads and Rs265_∆phaB:pBBR-MVA-ads strains. We therefore compared the resulting amorphadiene/biomass ratios for each pathway with the ones determined for i) the Rs265:pBBR-ads and Rs265_∆phaB:pBBR-ads (MEP-only) strains, and ii) the Rs265-MVA_∆dxr:pBBR-MVA-ads and Rs265-MVA_∆dxr∆phaB:pBBR-MVA-ads (MVA-only) strains. Previously, this 13C-method provided important insights on the flux ratios upon co-expression of the MEP and MVA pathways (30). Under nitrogen excess condition, we observed that the dual-pathway strains with active MEP and MVA pathways showed a higher amorphadiene/biomass ratio for each isoprenoid route compared to when these were active individually (Fig. 3A). We therefore confirmed that, during nitrogen excess conditions, co-expression of the two isoprenoid pathways resulted in enhancement of their capacities in the Rs265 and Rs265_∆phaB strains harbouring the pBBR-MVA-ads plasmid (Fig. 3A, Table 4). Moreover, the capacity of the MVA pathway in the Rs265_∆phaB and Rs265-MVA_∆dxr∆phaB strains was even further enhanced by the phaB deletion, as made obvious by comparison to strains that still contain the phaB gene (Fig. 3A, Table 4). In contrast, the flux through the native MEP pathway remained unaffected by the phaB deletion. Thus, under nitrogen excess conditions, deletion of the phaB gene results in an increase of the isoprenoid flux exclusively via the MVA pathway (Table 4).
We further studied the isoprenoid flux ratio under nitrogen-limited conditions (Fig. 3B). The Rs265:pBBR-ads and Rs265_∆phaB:pBBR-ads strains, which rely exclusively on the MEP pathway for isoprenoid production did not show any increase in the amorphadiene/biomass ratio when compared to nitrogen excess conditions (Fig. 3B). In contrast, the amorphadiene/biomass ratio for the strains that express only the MVA pathway was increased 3-fold when compared to this value under nitrogen excess conditions (Rs265-MVA_∆dxr and Rs265-MVA_∆dxr∆phaB strains harbouring pBBR-MVA-ads plasmid, Fig. 3). Interestingly, for the strains that express both pathways (Rs265:pBBR-MVA-ads and Rs265_∆phaB:pBBR-MVA-ads strains, Fig. 3B), there was only a minor increase of the MVA pathway capacity in the Rs265_∆phaB strain when compared to nitrogen excess conditions (Table 4). On the other hand, the MEP pathway capacity increased by 80% for the Rs265 strain (from 9.4 ± 0.3 to 16.9 ± 0.4 mg amorphadiene ∙ g of biomass-1) and by 300% for the Rs265_∆phaB strain (from 9.2 ± 1.3 to 28.7 ± 1.3 mg amorphadiene ∙ g of biomass-1). Thus, for the dual-pathway strain, the increase of the amorphadiene/biomass ratio under nitrogen limitation conditions is attributed to the endogenous MEP pathway (Table 4).
Amorphadiene biosynthesis during resting cells conditions
Under nitrogen-limited conditions a short exponential growth phase occurred, and therefore a short growth-associated amorphadiene production phase could not be avoided. This resulted in non-linear growth and production kinetics, making it difficult to assess yields (mg amorphadiene ∙ glucose-1) and productivities (mg amorphadiene ∙ L-1 ∙ h-1). In order to focus exclusively on growth-uncoupled production, and to obtain linear kinetics, we decided to assess amorphadiene production during resting cell conditions in nitrogen-free medium. This cultivation setup simulates the production phase of a two-stage fermentation setup where growth and production are separated.
Since deletion of phaB and expression of the MVA pathway increased production during nitrogen limitation, we reasoned to assess the amorphadiene production levels in the presence of also an active MEP pathway. Therefore, we further cultivated the strains Rs265, Rs265_∆phaB, Rs265-MVA_∆dxr and Rs265-MVA_∆dxr∆phaB under resting cells condition. All these strains contained the pBBR-MVA-ads plasmid (Fig. 4A).
A linear increase was observed in the OD600 of the strains with a functional PHB biosynthetic pathway (Rs265 and Rs265-MVA_∆dxr, Fig. 4B). This trend is known to be associated with the accumulation of this storage compound (29), and it is associated with cell expansion rather than cell division. Accordingly, the corresponding ∆phaB strains (Rs265_∆phaB and Rs265-MVA_∆dxr∆phaB) did not show any increase in OD600. Glucose consumption (Fig. 4C, Additional file 1: Fig. S2), pH and amorphadiene concentrations (Fig. 4D, E) were followed over time. A decrease in the pH of the Rs265_∆phaB and Rs265-MVA_∆dxr∆phaB strains was observed (Fig. 4D), which can be explained by the secretion of organic acids upon prevention of PHB accumulation, mainly of pyruvate and 2-oxoglutarate (Fig. 4F).
Amorphadiene samples were collected over time from all the cultures (Fig. 4E), and yields and productivities were calculated for the first 24 h (Fig. 4G-I). The corresponding values for the Rs265 strain were the lowest among the tested strains. Deletion of phaB (Rs265_∆phaB) resulted in a 2-fold increase of the amorphadiene/biomass ratio (Fig. 4G, Additional File 1: Table S2), the volumetric productivity (Fig. 4H) and the yield on glucose (Fig. 4I) compared to the Rs265 strain. Also, inactivation of the endogenous MEP pathway in the Rs265-MVA_∆dxr strain resulted in an increase of those values compared to Rs265 strain (Fig. 4H, I). Hence, inactivation of either the PHB production pathway or the endogenous MEP pathway stimulates growth-independent production. Combined inactivation of the MEP and PHB production pathways (Rs265-MVA_∆dxr∆phaB strain) allowed this strain to reach the highest amorphadiene/biomass ratio, volumetric amorphadiene productivity (Fig. 4H) and yield on glucose (Fig. 4I). All these values were 2,5-fold higher in the Rs265-MVA_∆dxr∆phaB strain, compared to the Rs265 control strain. Hence, deletion of the endogenous MEP and PHB biosynthetic pathways resulted in the best metabolic setup for exploiting non-growing conditions for amorphadiene production.