Inactivation of two genes encoding glycolate dehydrogenase in Synechocystis resulted in glycolate production
In Synechocystis, glycolate is converted to glyoxylate by two glycolate dehydrogenases (GlcD1 and GlcD2), and subsequently metabolized by three branched routes (Eisenhut M, Kahlon S et al. 2006, Eisenhut M, Ruth W et al. 2008). To completely block the glycolate metabolism, both GlcD1 and GlcD2 encoded by glcD1 and glcD2, respectively, were inactivated (Fig. 1). The resulting mutant was designated as WT-ΔglcD (Table 1). Complete segregation and correct gene insertions at both glcD1 and glcD2 sites were verified by PCR and sequencing (Fig. S1).
Table 1
The Synechocystis strains used in this study
Strain
|
Genetic background
|
Source of Rubisco
|
Wild type
|
Synechocystis sp. PCC 6803
|
-
|
WT-ΔglcD
|
WT ΔglcD1::emr; ΔglcD2::specr
|
-
|
RPE-ΔglcD
|
ΔglcD Δpta::Pcpc560-rpe-Trbcs-cmr
|
Riftia pachyptila endosymbiont
|
4Pm-ΔglcD
|
ΔglcD Δpta::Pcpc560-4pm-Trbcs-cmr
|
Phaeospirillum molischianum
|
5St-ΔglcD
|
ΔglcD Δpta::Pcpc560-5st-Trbcs-cmr
|
Sedimenticola thiotaurini
|
6RBC-ΔglcD
|
ΔglcD Δpta::Pcpc560-6rbcL-6rbcS-Trbcs-cmr
|
Synechocystis sp. PCC 6803
|
As glycolate metabolism was completely blocked, we next investigated glycolate accumulation in strain WT-ΔglcD. Both the intracellular and extracellular glycolate concentration of WT-ΔglcD were analyzed and compared with that of the WT strain. Samples were taken after three days cultivation supplemented with or without 50 mM NaHCO3. The intracellular glycolate concentration of the WT strain was 0.004 µmol L− 1OD730 − 1 and 0.02 µmol L− 1OD730 − 1 respectively, when supplemented with or without 50 mM NaHCO3 (Fig. S2). Moreover, the extracellular glycolate concentration was undetectable in the WT strain under both conditions (data not shown). It is evident that glycolate could be rapidly metabolized in the WT strain. On the contrary, strain WT-ΔglcD accumulated glycolate intracellularly and extracellularly under both conditions (Fig. 2 and Fig. S2). The intracellular glycolate concentration of strain WT-ΔglcD was 0.51 µmol L− 1OD730 − 1 when supplied with 50 mM NaHCO3, and increased to 1.75 µmol L− 1OD730 − 1 without the supply of NaHCO3 (Fig. S2). Furthermore, the glycolate concentration in the medium of strain WT-ΔglcD reached 86.47 µmol L− 1OD730 − 1 (mass concentration of 0.02 g/L) and 317.77 µmol L− 1OD730 − 1 (mass concentration of 0.06 g/L) after 3 days cultivation respectively, with or without 50 mM NaHCO3 (Fig. 2). Apparently, the majority of glycolate was excreted to the culture by strain WT-ΔglcD, and the intercellular glycolate accumulation could be negligible. We further monitored the glycolate concentration in the medium every three days and found that strain WT-ΔglcD produced 0.19 g/L and 0.34 g/L of glycolate after 18 days cultivation respectively with or without the supply of 50 mM NaHCO3 (Fig. 2). In other words, glycolate can be produced from CO2 and secreted extracellularly upon inactivation of the two glycolate dehydrogenases in Synechocystis. Moreover, strain WT-ΔglcD produces higher concentration of glycolate when no additional NaHCO3 was supplemented, suggesting ambient level CO2 is sufficient for glycolate production to occur.
Overexpression of the native carboxysome-located Rubisco does not contribute to glycolate production
Given the multiple industrial applications of glycolate, we were encouraged to further increase glycolate production. Glycolate synthetic pathway comprises two reactions (Fig. 1). RuBP reacts with O2 to generate one molecule of 2PG and one molecule of 3-Phosphoglycerate (3PGA) (Eisenhut M, Ruth W et al. 2008, Fernie AR and Bauwe H 2020). 2PG is then dephosphorylated to glycolate and 3PGA enters the CBB cycle to regenerate RuBP (Eisenhut M, Ruth W et al. 2008, Fernie AR and Bauwe H 2020). In order to identify the bottleneck of glycolate production, the intercellular 2PG concentration in the WT strain and strain WT-ΔglcD were measured. Samples were taken after three days cultivation with or without the supply of 50 mM NaHCO3. With the intact glycolate metabolism, the intracellular 2PG concentration in the WT strain were below 0.03 µmol L− 1OD730 − 1 under both growth conditions (Fig. S2). The intracellular 2PG level in strain WT-ΔglcD was at the same level as compared to the WT strain. However, as mentioned above, the intracellular glycolate concentration in strain WT-ΔglcD became about 100-fold higher than that of the WT strain irrespective of the supply of 50 mM NaHCO3 (Fig. S2). This indicated that the conversion from 2PG to glycolate in strain WT-ΔglcD was efficient and that the oxygenation of RuBP catalyzed by Rubisco was the rate-limiting step of glycolate production.
Thus, to increase glycolate production, the native Rubisco of Synechocystis was overexpressed in strain WT-ΔglcD. The resulting mutant was designated as strain 6RBC-ΔglcD (Table 1) and its capacity for glycolate production was determined with the same growth conditions as mentioned above. After 18 days of cultivation, strain 6RBC-ΔglcD produced 0.16 g/L and 0.35 g/L of glycolate when supplied with or without 50 mM NaHCO3, respectively. Neither titer is significantly higher than that of strain WT-ΔglcD under the same condition (Fig. 3a and b). In addition, no significant difference was observed in the growth rate of strains 6RBC-ΔglcD and WT-ΔglcD under both conditions (Fig. 3c and 3d). Moreover, the SDS PAGE and native PAGE results suggested that 6RBC was successfully overexpressed and assembled under both conditions (Fig. 3e and f). These results together suggested that overexpression of 6RBC Rubisco did not contribute to increase glycolate production. The reason behind is likely that the native 6RBC Rubisco is encapsulated in a microcompartment found in all cyanobacteria, termed as the carboxysome. It reduces the oxygenase activity of Rubisco by inhibiting the entrance of O2 and increasing CO2 concentration around Rubisco (Espie GS and Kimber MS 2011). Thus, to increase glycolate production, the selected Rubisco is expected to be located outside the carboxysome so as its oxygenase activity can play a role.
Overexpression of Form II Rubiscos enhanced glycolate production
It was previously reported that replacing the native Rubisco of cyanobacteria with Form II Rubisco could not support the biogenesis of carboxysome, indicating the Form II Rubisco resides outside the carboxysome (Baker SH et al. 1998, Durao P et al. 2015). If the Rubisco is located in the cytosol, it is accessible to molecule oxygen and a reduced CO2 level due to the absence of carbonic anhydrase in the cytosol (Price GD et al. 2008, Price GD 2011). Thus, we hypothesized that Form II Rubiscos might be promising candidates to increase glycolate production. To this end, three form II Rubiscos from Riftia pachyptila endosymbiont (RPE Rubisco), Phaeospirillum molischianum (4Pm Rubisco) and Sedimenticola thiotaurini (5St Rubisco) were selected and individually overexpressed by using the strong promoter Pcpc560 in strain WT-ΔglcD (Table 1), resulting in strains RPE-ΔglcD, 4Pm-ΔglcD and 5St-ΔglcD, respectively (Fig. S1).
Subsequently, glycolate production of these three strains were determined without additional NaHCO3, which seemed to be more favorable for strain WT-ΔglcD to produce glycolate. After 18 days of cultivation, strain 5St-ΔglcD produced 0.3 g/L glycolate, which is not significantly higher than that of strain WT-ΔglcD (Fig. 3a). Moreover, no significant difference on growth were observed between them (Fig. 3c). This incapacity for increasing glycolate production could be attributed to the undetectable expression and assembly of 5St Rubisco (Fig. 3e). In contrast, glycolate production was dramatically enhanced in strains RPE-ΔglcD and 4Pm-ΔglcD (Fig. 3a). After 18 days cultivation, strain 4Pm-ΔglcD produced 0.66 g/L of glycolate, about twofold of strain WT-ΔglcD, while strain RPE-ΔglcD produced 0.87 g/L of glycolate, 2.6-fold of strain WT-ΔglcD (Fig. 3a). However, the growth of strains RPE-ΔglcD and 4Pm-ΔglcD were significantly impaired (Fig. 3c). The expression and assembly of RPE Rubisco and 4Pm Rubisco were also detected (Fig. 3e). RPE Rubisco was copiously overexpressed and well assembled. By contrast, 4Pm Rubisco was successfully overexpressed but not assembled well. This explained their different capacity on enhancement of glycolate production. Taken together, these results showed that overexpression of Form II Rubisco indeed increased glycolate production.
Supply of NaHCO3 increased glycolate production by strains RPE-ΔglcD and 4Pm-ΔglcD
As mentioned above, glycolate production by strain WT-ΔglcD decreased when supplied with 50 mM NaHCO3 (Fig. 2). Thus, we further investigated whether glycolate production of strains RPE-ΔglcD and 4Pm-ΔglcD would also be repressed when supplied with 50 mM NaHCO3.
Surprisingly, glycolate production by strains RPE-ΔglcD and 4Pm-ΔglcD was not decreased, but instead sharply when NaHCO3 was available (Fig. 3b). Strain 4Pm-ΔglcD produced 1.46 g/L of glycolate in 18 days when supplemented with 50 mM NaHCO3, which is about 7.7-fold of the titer of strain WT-ΔglcD under the same condition (Fig. 3b). This is also more than twofold of the titer produced by Strain 4Pm-ΔglcD without additional NaHCO3. Additionally, 4Pm Rubisco assembled better in strain 4Pm-ΔglcD upon addition of 50 mM NaHCO3, which could contribute to the increased glycolate production (Fig. 3f). Among these three strains, strain RPE-ΔglcD was inarguably the best glycolate producer, generating 2.82 g/L after 18 days of cultivation, about 15-fold of the titer of strain WT-ΔglcD under the same growth condition (Fig. 3b). Moreover, the expression and assembly of RPE did not differ upon addition of NaHCO3 (Fig. 3f), suggesting that the increased glycolate production was not related to the assembly of RPE Rubisco. However, the growth of strains RPE-ΔglcD and 4Pm-ΔglcD were also significantly impaired under this condition (Fig. 3d)
Thus, we further investigated glycolate production of strain RPE-ΔglcD when supplied with different concentration of NaHCO3. Glycolate production of strain RPE-ΔglcD increased along with increasing the concentration of NaHCO3, and approached a plateau of 2.84 g/L when supplied with 30 mM NaHCO3 (Fig. 4a). Notably, the growth of strain RPE-ΔglcD gradually reduced along with the increased glycolate production (Fig. 4b). The intracellular glycolate concentration in RPE-ΔglcD was also increased, from 5.6 μmol L-1OD730-1 in the absence of NaHCO3, to 10.4 μmol L-1OD730-1 when adding 50 mM NaHCO3 (Fig. S2). It was previously reported that intracellular accumulation of glycolate is toxic to the cell (Eisenhut M, Ruth W et al. 2008). The retarded growth of strain RPC-ΔglcD upon adding increased concentration of NaHCO3 was probably related to the elevated intracellular glycolate concentration in strain RPE-ΔglcD.
Supply of CO2 decreased glycolate production by strain RPE-ΔglcD.
Cyanobacteria can use both HCO3- and CO2 as external inorganic carbon source (Price GD, Badger MR et al. 2008, Price GD 2011). As supply of HCO3- increased glycolate production of strains RPE-ΔglcD and 4Pm-ΔglcD, we then wondered what would be the effect if supplying CO2. Since strain RPE-ΔglcD produces much higher glycolate concentration than that of strain 4Pm-ΔglcD, we chose strain RPE-ΔglcD to study the effect of CO2.
To this end, the external organic carbon supplied was changed from NaHCO3 to CO2. The glycolate production and growth of strain RPE-ΔglcD were evaluated under 1% or 3% CO2 (Fig. 4c and 4d). After 12 days of cultivation, strain RPE-ΔglcD produced 0.87 g/L glycolate under 1% CO2, and the glycolate titer decreased to 0.47 g/L under 3% CO2 (Fig.4c). Additionally, the growth of strain RPE-ΔglcD increased positively with increasing the CO2 level (Fig. 4d). The increased growth and reduced glycolate production of RPE-ΔglcD together indicated that supply of CO2 enhanced the carboxylation reaction of RPE and consequently inhibited the oxygenation reaction.
RPE Rubisco is located in the cytosol
The enhanced glycolate production indicated an active oxygenation reaction catalyzed by RPE Rubisco and 4Pm Rubisco. This suggested that they are probably located in the cytosol rather than in the carboxysome as the O2 concentration in cytosol is much higher. To provide direct evidence, we visualized their location in vivo by fluorescent labelling. We first tried to carry out the co-localization analysis by labelling RPE Rubisco with cyan fluorescent protein (CFP) and 6RBC with yellow florescent protein (YFP). RPE Rubisco was labelled with CFP at its C-terminal (termed as RPE-CFP). YFP was fused to the C-terminal of the large subunit of 6RBC (termed as 6RBCL-YFP). RPE-CFP and 6RBCL-YFP were individually expressed in the WT strain to give single fluorescent signal and co-expressed in the WT-strain to test whether these two fluorescent signals could be overlayed together. However, the fluorescent signals of RPF-CFP and 6RBCL-YFP were too week to give the location information (data not shown).
We next fused green fluorescent protein (GFP) to the C-terminal of RPE Rubisco or the large subunit of 6RBC Rubisco, termed as RPE-GFP and 6RBCL-GFP, respectively. RPE-GFP and 6RCBL-GFP were individually expressed in the WT strain to give single fluorescent signal. Meanwhile, the red fluorescence of endogenous chlorophyll-a of Synechocystis was used to indicate the shape of the whole cell (Cameron J et al. 2013). RPE-GFP gave rise to a large single fluorescent punctum at the cell polar, suggesting that RPE proteins intended to aggregate at the edge of cell (Fig. 5a). By contrast, 6RBCL-GFP intended to exhibit several fluorescent spots at a more central position within the cell, indicating the location of mature carboxysomes, which was in agreement with the previous report (Fig. 5b) (Cameron J, Wilson S et al. 2013). The different position of fluorescent signals between RPE-GFP and 6RBC-GFP indicated that RPE is not located in the carboxysome where 6RBC-GFP resides. The bacterial Form II Rubisco from Rhodospirillum rubrum was previously expressed in the Δrbc strain of Synechocystis (Durao P, Aigner H et al. 2015). The resulting mutant could not support the biogenesis of carboxysome and photoautotrophic growth at ambient CO2 concentration (Durao P, Aigner H et al. 2015). Thus, it is conceivable that the aggregate of RPE-CFP observed here is most likely in the cytosol.