Stable Platform for Mevalonate Bioproduction from CO2

Stable production of value-added products using a microbial chassis is pivotal for determining the industrial suitability of the engineered biocatalyst. Microbial cells often lose the multicopy expression plasmids during long-term cultivations. Owing to the advantages related to titers, yields, and productivities when using a multicopy expression system compared with genomic integrations, plasmid stability is essential for industrially relevant biobased processes. Cupriavidus necator H16, a facultative chemolithoautotrophic bacterium, has been successfully engineered to convert inorganic carbon obtained from CO2 fixation into value-added products. The application of this unique capability in the biotech industry has been hindered by C. necator H16 inability to stably maintain multicopy plasmids. In this study, we designed and tested plasmid addiction systems based on the complementation of essential genes. Among these, implementation of a plasmid addiction tool based on the complementation of mutants lacking RubisCO, which is essential for CO2 fixation, successfully stabilized a multicopy plasmid. Expressing the mevalonate pathway operon (MvaES) using this addiction system resulted in the production of ∼10 g/L mevalonate with carbon yields of ∼25%. The mevalonate titers and yields obtained here using CO2 are the highest achieved to date for the production of C6 compounds from C1 feedstocks.

Schematic diagram of the overall concept for the plasmid addition system and product synthesis from CO 2 in engineered C. necator H16 strains.

Figure S1 -Overview of pathways involved in mevalonate-producing strains of
C. necator H16.CO 2 is converted to pyruvate by the reactions of the Calvin cycle and glycolysis.Pyruvate acts as the starting compound for PHB, mevalonate, and pantothenate biosynthesis.The PHB biosynthetic operon can be inactivated to redirect the carbon flux towards mevalonate synthesis.To facilitate retention of the mevalonate production pathway during continuous fermentation, the genomic copy of pantothenate synthetase (panC) or the large subunits of the RuBisCO (ΔΔcbbLS) were deleted.The panC gene was then placed, with its native promoter on the backbone of the plasmid carrying the mevalonate pathway.Similarly to the panC complementation, the large subunit of the RuBisCo gene was placed under a mediumstrength promoter on the backbone of the plasmid carrying the mevalonate pathway.The addiction systems described above allowed for continuous production of mevalonate from CO 2 .Dashed lines indicated multiple reactions.Acetyl-CoA acetyltransferase/HMG-CoA reductase (MvaE), HMG-CoA synthase (MvaS), pantothenate synthetase (PanC).Products from top to bottom: artemisinin (antimalarial compound), β-methyl-Δ-valerolactone ((βMΔVL) precursor to biodegradable polymer), and β-carotene (precursor to compounds relevant to the pharmaceutical, cosmetics, and food industries).

Molecular cloning strategies to construct the plasmids used in this study
The panC deletion plasmid pMTL70621::ΔpanC was constructed by amplifying the 750 bp DNA regions located just upstream and downstream of the panC gene on C. necator H16 chromosome 1, using primer pairs panC_LHA_FW + panC_LHA_REV and panC_RHA_FW + panC_RHA_REV.The resulting fragments were assembled into a deletion cassette via size overlap extension PCR.Plasmid pMTL70621-SacB was digested with NotI and NdeI restriction enzymes and the deletion cassette ligated into the plasmid.
Plasmids pMTL70621::ΔcbbLS2 and pMTL70621::ΔcbbLSp, respectively used to delete the cbbLS2 and cbbLSp genes, were constructed by amplifying the circa 1000 bp DNA regions located just upstream and downstream of the cbbLS2 operon on C. necator H16 chromosome 2 and of the cbbLSp operon on the pHG1 megaplasmid.In the case of ΔcbbLS2, the upstream and downstream homology arms were amplified using primer pairs LHA_cbbLSch_FW + LHA_cbbLSch_RV and RHA_cbbLSch_FW + RHA_cbbLSch_RV, respectively.Primer pairs LHA_cbbLSm_FW + LHA_cbbLSm_RV and RHA_cbbLSm_FW + RHA_cbbLSm_RV were instead used to obtain the left and right homology arms for cbbLSp.The deletion cassettes were then cloned in the XbaI/NcoI-digested pMTL70621-SacB vector, using the NEBuilder ® HiFi DNA assembly method, to obtain pMTL70621::ΔcbbLS2 and pMTL70621::ΔcbbLSp.
The panC complementation plasmid pMTL71301::panC was obtained by cloning the panC gene from C. necator H16, including its native promoter and Rho-independent terminator of transcription, in the AscI/FseI-digested modular shuttle vector pMTL71301.The P panC -panC-T panC DNA construct was amplified from C. necator H16 chromosome 1, using primers panC_FW_HiFi and panC_REV_HiFi.
The cbbLS complementation plasmid pMTL71301::T500-P phaC -cbbLS2 was constructed by amplifying the promoter of the phaCAB operon from C. necator H16 chromosome 1, using primers T_PphaC_fwd (which contains the T500 synthetic terminator sequence) and PphaC_rev, and the cbbLS2 operon from chromosome 2, using the primers cbbLS_fwd and cbbLS_rev.The T500-P phaC and cbbLS2 DNA parts were then cloned in the XbaI/NheI-digested pMTL71301 vector, using the NEBuilder ® HiFi DNA assembly method.
To construct plasmid pMTL70621::KI_A3739::araC-P BAD -mvaES::A3740, the circa 780bp DNA sequences encompassing the intergenic regions found between the H16_A3739 and H16_A3740 genes were amplified using primer pairs 3739INT3740_UP_FWD + 3739INT3740_UP_REV and 3739INT3740_DOWN_FWD + 3739INT3740_DOWN_REV.The UP and DOWN homology arms were then cloned in the NotI/XhoI-digested pMTL70621-SacB vector, using the NEBuilder ® HiFi DNA assembly method, resulting in plasmid pMTL70621::KI_A3739::INT::A3740.To allow cloning of the araC-P BAD -mvaES operon in between the homology arms, a cutting site for XbaI (provided as a spacer in the 3739INT3740_UP_REV primer sequence) was introduced downstream of the UP sequence.Primer pairs IR_ara_FWD + A3743_ara_REV, A3743_mvaE_FWD + A3743_mvaE_REV and A3743_mvaS_T500_FWD + IR_mvaS_T500_REV were used to amplify the araC-P BAD promoter cassette, the mvaE and mvaS genes, respectively.Plasmid pMTL71301::araC-P BAD -mvaES was used as a template for these PCRs.The synthetic Rho-independent terminator T500 was introduced downstream of mvaS by including it in the IR_mvaS_T500_REV primer sequence.These DNA parts were then cloned together in the XbaI-digested pMTL70621::KI_A3739::INT::A3740, using NEBuilder ® HiFi, to obtain plasmid pMTL70621::KI_A3739::araC-P BAD -mvaES::A3740, which was used to integrate the MVA pathway into C. necator chromosome 1.
Table S1.Strains and plasmids used in this study.

MVA production in C. necator H16 under heterotrophic conditions
MVA is obtained from the condensation of three Acetyl-CoA molecules by the action of the acetyl-CoA acetyltransferase/HMG-CoA reductase (MvaE) and HMG-CoA synthase (MvaS) enzymes (Katsuki and Bloch, 1967).A schematic representation of the upper part of the MVA biosynthetic pathway is shown in Figure 1A.In this study, the MVA biosynthetic genes mvaE and mvaS from Enterococcus faecalis (Ef-mvaE and Ef-mvaS) were selected since these have been successfully employed to produce MVA and terpenoids in different bacterial species, including Escherichia coli and Methylobacterium extorquens (Tabata and Hashimoto, 2004;Wang et al., 2010;Yoon et al., 2009;Zhu et al., 2016).In particular, the E. coli codon-optimised versions of the Ef-mvaE and Ef-mvaS genes were cloned in the pMTL71301 modular shuttle vector (Ehsaan et al., 2021), under the control of the araC/P BAD L-arabinose-inducible promoter cassette, to obtain plasmid pMTL71301::araC-P BAD -mvaES.This plasmid was transferred to C. necator H16 to obtain the C. necator H16/pMTL71301::araC-P BAD -mvaES strain, which was then cultivated in nitrogen-limited MM, supplemented with 2.5% fructose and 10 µg/mL tetracycline.Ability to produce MVA under heterotrophic conditions was assessed, using the method described in the following paragraph.A C. necator H16 derivative carrying the empty pMTL71301 plasmid was used as the negative control strain in these experiments.To test whether MvaE and MvaS were being expressed correctly, samples were collected from the C. necator H16/pMTL71301 and C. necator H16/pMTL71301::araC-P BAD -mvaES cultures right before providing the inducer L-arabinose into the media and after 24 h of induction.
Whole-cell protein extracts were prepared from these culture samples and analysed by SDS-PAGE, as shown in Figure S1A.Significant amounts of the MvaE and MvaS proteins were detected in the H16/pMTL71301::araC-P BAD -mvaES strain after 24h of induction with L-arabinose (Figure S1A).MvaE expression levels were significantly higher with respect to MvaS.This is likely because this protein is encoded by the first gene of the MVA biosynthetic operon mvaES.No expression of these proteins could be observed before adding the inducer L-arabinose into the culture media, suggesting that transcription of the mvaES operon is tightly regulated by the araC/P BAD promoter cassette, despite the presence of multiple copies of this synthetic DNA construct per cell.In line with the protein expression data, the C. necator H16/pMTL71301::araC-P BAD -mvaES strain also produced MVA under heterotrophic conditions, with a maximum MVA titre of around 2.4 g/L that was measured at the 48 h post-induction time point (Figure S1B).The concentration of MVA in the culture media did not vary significantly following this time point, suggesting that MVA production by C. necator H16/pMTL71301::araC-P BAD -mvaES might have stopped after 48 h of induction.The concentrations of MVA produced over time by C. necator H16/pMTL71301 and C.
necator H16/pMTL71301::araC-P BAD -mvaES strains, under these experimental conditions, were quantified by HPLC analysis (Figure S1).As expected, neither expression of MvaE and MvaS or MVA production could be observed in the negative control strain carrying the empty pMTL71301 vector.

Shake flask cultivation experiments
Single colonies of C. necator strains H16/pMTL71301 and H16/pMTL71301::araC-P BAD -mvaES were used to inoculate 5 ml of LB liquid medium supplemented with 15 µg/mL tetracycline (in 50 mL Falcon tubes) and grown overnight at 30°C with shaking (200 rpm).After overnight incubation, the optical density (OD 600 ) of the cultures was measured and normalised to OD 600 =0.1 in 100 mL of LB medium supplemented with 15 µg/mL tetracycline (in 500 mL flasks).The bacterial cultures were then incubated at 30 °C with shaking (200 rpm) until they reached an OD 600 of around 0.6 -0.7.At this point, 1 mL samples were collected from each culture, centrifuged at 14000 rpm for 1 minute and the cell pellets were stored at -20°C to be used for the purification of pre-induction protein samples for SDS-PAGE.In addition, 1 mL of 20% (w/v) Larabinose solution was added to the cultures, thus obtaining a final L-arabinose concentration of 0.2%.Following induction with L-arabinose the bacterial cultures were incubated overnight at 30°C, with shaking.The following morning, the OD 600 of the cultures was measured and 1 mL samples were removed from each of them to obtain the post-induction protein samples for SDS-PAGE.At this point, cells were harvested by centrifuging the cultures at 8000 rpm for 5 minutes.The pellets were then resuspended in suitable volumes of 2.5% (w/v) Fructose Nitrogen-limited MM, supplemented with 15 μg/mL tetracycline, to obtain cultures with an OD 600 of around 15.These were then transferred to 250 mL flasks and incubated at 30°C, with shaking.

Preparation of protein samples for SDS-PAGE analysis
The pre-and post-induction cell pellets prepared as described in paragraph 2.

Correlation between levels of "social cheaters" and MVA productivity in PAN
A clear correlation was observed between the levels of social cheaters present in the PAN population and MVA production rates.To estimate the amounts of social cheaters, we firstly quantified the effects of pantothenate supplementation on the viable counts of the control strain (CTRL), by calculating the average increase in the number of cfu/mL observed on Pan, with respect to LB.This amount (15.87%) was then subtracted from the percentage of tetracycline sensitive (100 -% Tet-resistant cells) PAN cells at each time point to estimate the percentage of social cheaters (Figure S4).The partial MVA production rates, associated with each of the postinduction 24 h time intervals in between culture samples withdrawals, were determined by dividing the increments in MVA concentrations observed between two consecutive time intervals (ΔMVA) by 24 h (Figure S4).
As can be observed in Figure S4, the percentage of social cheaters in the PAN population increased significantly from the initial value determined at the time of inoculation (around 11%) to over 70% at the time of induction with 0.2% L-arabinose.
This was the highest level of cheaters observed throughout the fermentation experiment, suggesting that higher percentage of cheaters may lead to population collapse.Indeed, under these conditions, the concentration of pantothenate present in the media is likely to be too low to sustain growth of such a large population of cheaters.As a result of this, during the first day of induction the average amounts of cheaters decreased by ~24% and the cooperators, which now represented just over half of the total PAN population, started producing MVA, with a productivity of around 0.030 g/L/h.MVA productivity (0.031 g/L/h) remained more or less constant during the second day of induction, at the end of which a further reduction (-11.61%) of the cheaters levels was observed.It is important to note that the highest MVA productivity (0.044 g/L/h) for this fermentation process was observed between the 48 h and 72 h post-induction time points, precisely the time interval when the portion of cheaters reached the lowest level observed throughout the experiment (~19%).It should also be noted that cooperators, besides producing MVA, would also synthesize pantothenate.Since these constituted around 80% of the total population at the 72 h post-induction time point, it could be speculated that extracellular pantothenate concentrations would be high at this point.Therefore, an increasing number of cells could have exploited this freely available pantothenate in order to survive without the need to keep the pMTL71301::araC-PBAD-mvaES::panC plasmid.Consistently with these predictions, the levels of cheaters increased sharply (by 31.73%) between the 72 h and 96 h post-induction time points, while MVA productivity decreased from 0.044 to 0.028 g/L/h (-36.36%),during the same time interval.Even though a reduction in the percentage of cheaters (-11.59%) was observed in the following time interval, between the 96 h and 120 h post-induction time points, MVA productivity (0.024 g/L/h) also decreased slightly during this time.This is likely to reflect the significant decrease in the total number of viable cells observed during the same time interval (Figure 2B), which may indicate the presence of nutrient limitations in the culture media.Indeed, MVA productivity kept decreasing up to 0.011 g/L/h between the 120 h and 144 h postinduction time points, even though the percentage of cheaters did not change significantly during this time.It is possible that MVA production may have stopped completely around the 144 h time point, when the highest MVA titre was recorded.
Indeed, MVA concentration decreased slightly during the last 24 h time interval.
Overall, our observations suggest that the implementation of the panC-based plasmid addiction system in C. necator H16 provided significant advantages in terms of MVA production, when applied to batch fermentation processes carried out under chemolithoautotrophic conditions.Approximately 4 g/L MVA were produced by the PAN strain, while no MVA synthesis was observed in the CTRL strain (no plasmid addiction system).However, the data obtained also show that the overproduction of pantothenate, resulting from the presence of multiple copies of the panC gene in each cell, led to the generation of significant levels of plasmid-free social cheaters.Although it would have been interesting to measure the variations in pantothenate levels over time to identify the concentration range allowing for high MVA productivity, these were not analysed since we considered this investigation to be beyond the scope of the present study.
araC-P BAD -phaA-mvaES::cbbLS2 pMTL71301 expressing panC from C. necator H16, mvaE and mvaS from Enterococcus faecalis pMTL71301 expressing cbbLS2 from C.necator H16 pMTL71301 expressing cbbLS2 from C.necator H16, mvaE and mvaS from Enterococcus faecalis pMTL71301 expressing cbbLS2 and phaA from C.necator H16, mvaE and mvaS amplify the DNA sequence upstream of the panC gene in the C. necator H16 genome to obtain the left homology arm for constructing the panC in-frame deletion panC_RHA_FW TGAGCGGCAGCGCCGCACCAGGC panC_RHA_REV TGTTGCTGACCACGCTGGTGGGCGTG Primers used to amplify the DNA sequence downstream of the panC gene in the C. necator H16 genome to obtain the right homology arm for constructing the panC in-frame deletion LHA_cbbLSch_FW gctcggtacccggggatcctctagaGTGCGCAGCTCCAGTCC CAGCGTG LHA_cbbLSch_RV tcgagcatccGCTTGTCTCCTTGCGTGGTTGAGCG Primers used to amplify the DNA sequence upstream of the cbbLS2 operon on C.
6 were lysed using the BugBuster ® protein extraction reagent(Merck-Millipore).The protein extraction mix was prepared by diluting BugBuster ® (1:10); Protease Inhibitor Cocktail Set VII (1:50); rLysozyme TM (1:1000) and benzonase (1:10000) in appropriate volumes of PBS (all these reagents were purchased from Merck-Millipore).The total protein concentration in each sample was normalised using the formula: [(OD 600 / 0.2) * 45] / 2. This was used to calculate the volume (in µL) of protein extraction mix to be added to each cell pellet.Cell pellets were then incubated at room temperature for 20 min, with gentle shaking (50 rpm).Resuspended cell pellets were then centrifuged at 14000 rpm for 10 min.10 µL of the resulting supernatants (soluble protein fraction) were then mixed with 10 µL NuPAGE TM LDS sample buffer (Thermo Fisher Scientific) and boiled at 100ºC for 5 min.SDS-PAGE of protein samples was performed using NuPAGE TM Bis-Tris 4-12 % precast polyacrylamide gels (Thermo Fisher Scientific).Protein bands were visualised using GelCode TM Blue Safe Protein Stain (Thermo Fisher Scientific).

Figure
Figure S1.MvaE and MvaS expression and MVA production from fructose

Table S2 .
Oligonucleotide primers used in this study.