The strains used in this study originate from the CLiP library (Li et al., 2019): the reference strain (CC-4533) and ΔPRK (LMJ.RY0402.119555), both obtained from the Chlamydomonas Resource Center. Chlamydomonas cells were grown on agar plates or liquid medium, using Tris-acetate-phosphate (TAP) medium (Gorman and Levine, 1965) or High salt medium (HSM) (Sueoka, 1960) at 25°C, under continuous light (40-60 µmol photons m^-2 s^-1) or dark (in particular for the ΔPRK strain), and shaking for liquid cultures (130 rpm). Antibiotics used were hygromycin B (10 µg/mL) and paromomycin (15 µg/mL). Growth analyses were performed in the Algem^® labscale double photobioreactor systems (Algenuity, Stewartby, UK) for large volume cultures (400 mL) or in the Algem^® HT-24 (Algenuity) photobioreactor for small volume cultures (25 mL). All chemicals were obtained from Sigma-Aldrich unless otherwise specified.

Protein and nucleic acid designs were performed in silico using Serial Cloner 2.6.1 software. All recipient plasmids are derived from the MoClo original toolkit (Weber et al., 2011). Phosphoribulokinase PRK cDNA sequence (Cre12.g554800; Uniprot accession P19824) was obtained by PCR on reverse translated mRNA extracts from a D66 strain (Schnell and Lefebvre, 1993) using primers TTGAAGACTTAATGGCTTTCACTATGCGCGC and TTGAAGACAACGAACCCACGGGCACAACGTCC. The resulting PRK coding sequence was designed for the position B3-B4 of the Chlamydomonas MoClo toolkit (Crozet et al., 2018), and cloned into the plasmid pAGM1287 (Weber et al., 2011). Two other constructs were obtained by PCR on genomic DNA of CC-4533 strain using the following primers:

TTGAAGACTTCTCAGGAGGCCCTGGGCTTTAGCCCC and AAGAAGACAACTCGAGTACATGATGCATGTAACAGCAGCAATGAT for the PRK promoter, TTGAAGACTTCTCATACTGCGTCTTGGGTCGGTGCGCT and CAGAAGACAACTCGCATTGGTTGCTAACAGCTCGACGC for the PRK 5’UTR, and TTGAAGACTTCTCAAATGGCTTTCACTATGCGCGC and TTGAAGACTTCTCAAATGGCTTTCACTATGCGCGC for the PRK CDS with introns.

Each part was cloned into the plasmid pAGM9121 (Patron et al., 2015). The level 1 plasmids were built in pICH47742 with the endogenous promoter and 5’UTR of PRK and the 3’UTR/Terminator of PSAD controlling the expression of the PRK CDS with or without its endogenous introns. Level M plasmids were built in pAGM8031 combining p1-013 (hygromycin resistance gene) (de Carpentier et al., 2020) with the PRK transcriptional unit resulting in pCMM-24 (P_PSAD-PRK_CDS), pCMM-25 (P_PRK-PRK_CDS) and pCMM-26 (P_PRK-PRK_(i)).

Transformations were performed as previously described (Crozet et al., 2018), using 55 fmol of purified cassette after BbsI-HF (for photoautotrophy screening with p1 plasmids) or BsaI-HF (for antibiotic screening with pM plasmids) digestion (New England Biolabs) of the corresponding plasmid. The transformation leads to random insertion of the transgene in the nuclear genome. Transformants were selected on HSM-agar medium or TAP-agar containing hygromycin B (20 mg/L), Plates and transformants were analyzed after 5 to 7 days of growth in medium light (50 µmol photons m^-2 s^-1).

Cells were grown in TAP medium up to 4-5 x 10⁶ cells mL⁻¹, harvested by centrifugation at 2500 g for 5 min at room temperature (RT), and lysed in 400 μL of extraction buffer (0.2 M Tris-HCl pH 7.5, 200 mM NaCl; 25 mM EDTA; 0.5% SDS) for 10 min at 37°C under agitation (1400 rpm). After centrifugation at 17000 g for 3 min at RT, the genomic DNA contained in the supernatant was precipitated with one volume of isopropanol for 10 min at room temperature and collected by centrifugation at 17000 g for 10 min at RT. The pellet DNA was then washed with 70% ethanol, spinned (17000 g for 3 min at RT) and the pellet was air-dried prior to resuspension in water. PCR was performed using the Quick-LoadR^©Taq2×Master Mix (New England Biolabs) according to the manufacturer’s instructions. Primers used were Plex5: AAGGACGCTGACATG, Plm1: CCTGATGGATGGTTC, PLPSAD: TTGAAGACAATCATCTCAATGGGTGTG and Plen3U: AGGTGCCAAAGCAAC.

Cells were grown in TAP medium up to 4-5 x 10⁶ cells mL⁻¹, harvested by centrifugation at 5000 g for 10 min at 4°C, resuspended in 500 µL of Buffer B (30 mM Tris-HCl pH 7.9, 0.5 mM EDTA, antiprotease complete tablets (Roche)), and lysed twice by using glass beads and vortexing (30 sec vortex, 1 min on ice). The total extract was then clarified by centrifugation (2x10 min at 21000 g) and the concentration of the total soluble protein was determined by BCA Protein Assay using bovine serum albumin (BSA) as standard.

Total soluble proteins were analyzed by western blotting with a custom rabbit polyclonal primary antibody raised against Chlamydomonas PRK (Covalab, Bron, France) subsequently detected by secondary anti-rabbit antibody coupled to horseradish peroxidase (Sigma-Aldrich reference A9169, Saint Louis, USA). Detection was done with commercial ECL peroxidase assay (GE Healthcare, Chicago IL USA) with a Chemidoc (Bio-Rad, Hercules CA USA).

Cells were grown until exponential phase (2-6 x 10⁶ cells mL⁻¹) and serial dilutions in TAP or HSM media were made. The dilution was spotted (10 µL) onto TAP and HSM agar plates at different light intensities. The plates were then scanned after 7 days using a Perfection V800 Photo scanner (Epson). This analysis was automatized using an Opentrons OT-2.

Growth analyses were performed using the Algem^® labscale double photobioreactor system (Algenuity, Stewartby, United Kingdom) under continuous light (100 µmol photons m^-2 s^-1) and 120 rpm agitation in TAP or high salt medium (HSM). The strains were pre-acclimated for at least 24h in the same medium prior to growth analysis. The absorbance at 740 nm was recorded every 10 min using the built-in sensor. The maximal growth rate was determined as the maximal slope of the growth curve (ΔAbs/Δtime). Growth curves obtained with the AlgemHT24 photobioreactor were performed in autotrophic conditions and with three different light intensities, i.e. low light at 38 μmol.m^-².s^-1, medium light at 100 μmol.m^-².s^-1, and high light at 300 μmol.m^-².s^-1. The resulting growth curves were analyzed using GraphPad Prism software, and in particular a fit was made with the Gompertz growth equation (Y = YM*(Y0/YM)^(exp(-K*X))) to obtain an accurate value of lag phase and μmax, the maximum specific growth rate.

To eliminate metabolites that could interfere with enzymatic activity measurements, the crude extract was desalted in Sephadex G-25 Columns (GE Healthcare) and reduced with 20 mM DTT in 50 mM Tris-HCl (pH 7.5) for 30 min at 30°C. The PRK activity was measured as previously described (Gurrieri et al., 2019). Briefly, the reaction mixture contained 50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 40 mM KCl, 10 mM MgCl_2, 5 U/mL Pyruvate kinase, 6 U/mL Lactate dehydrogenase, 2.5 mM phosphoenolpyruvate, 2 mM ATP, and 0.2 mM NADH. The desalted crude extract was added and the background was recorded for 1-2 min. PRK activity was then measured by adding 0.5 mM ribulose-5-phosphate and monitoring the oxidation of NADH at 340 nm.

Phosphoribulokinase is an enzyme unique to the CBBC encoded by a single gene (Cre12.g554800_4532) in Chlamydomonas reinhardtii. To comprehensively investigate the function of the PRK-encoding gene (PRK1), we used the mutant LMJ.RY0402.119555 (hereafter named ΔPRK) from the Clip Library, generated by random insertion of the paromomycin resistance gene (Li et al., 2019). This mutant strain harbors an insertion (named CIB1) in the exon 7 of the PRK1 gene ( Figure 1A ) as identified by junction sequencing with a 95% confidence (Li et al., 2019). In the mutant strain, the position of the CIB insertion in exon 7 was verified by PCR analysis with primer couples allowing specific detection of a single band related to either the WT PRK1 gene or the disrupted PRK1 gene containing the CIB1 insertion ( Figure 1B ). To determine the impact of the CIB1 insertion on PRK protein expression, total protein extracts from the mutant strain and the reference strain were analyzed by western blotting. The PRK signal was very strong in the reference strain, while no signal could be detected in the mutant strain ( Figure 1C ). To evaluate the detection threshold of our custom made anti-PRK polyclonal antibody, a gradient of total protein from the reference strain CC-4533 was utilized and a PRK signal could be still be detected using 5% of the total protein extract ( Figure 1C ). This finding indicates that the mutant strain has a PRK content lower than 5% of that in CC-4533, strongly suggesting that the presence of the CIB insertion disrupts PRK protein expression. Since PRK activity is unique in Chlamydomonas, the absence of the PRK protein should correlate with a loss of the corresponding enzymatic activity. As PRK is activated by reduction (Marri et al., 2008), we treated the total soluble protein extract from each strain with the strong reducing agent DTT to obtain maximal PRK activity. Consistently, no PRK activity was detected in the mutant strain compared to the reference strain ( Figure 1D ). Consequently, the LMJ.RY0402.119555 strain is a knock-out strain with undetectable levels of both PRK protein and activity.

To examine the functional consequences of PRK deficiency in Chlamydomonas we analyzed the growth phenotype of the ΔPRK mutant under various conditions. Since PRK is central to the CBBC, we tested the ability of this mutant to grow in the presence of light and acetate as a carbon source (TAP light, mixotrophic conditions) or in the presence of light in a minimal medium without acetate (HSM light, strictly photoautotrophic conditions). In solid media, the ΔPRK mutant could not grow photoautotrophically but was able to grow in the presence of light and acetate, albeit more slowly than the reference strain ( Figure 1E ). To further characterize the growth phenotype of the ΔPRK strain in mixotrophic conditions, we monitored its growth over time in a photobioreactor in TAP liquid medium, compared with the CC-4533 reference strain ( Figure 1F ). The growth profiles revealed a pronounced difference between the two strains. Indeed, the mutant grew much slower than CC-4533, explaining the difference observed on solid media. This growth defect of the PRK mutant is characterized by an extended lag phase ( Figure 1G ) and a reduced maximum growth rate (µ_max) compared to the reference strain ( Figure 1G ). These data indicate that the ΔPRK strain is non-photoautotrophic and strongly imply that the PRK1 gene is essential for photosynthetic carbon fixation, and that the absence of the CBBC limits the growth of the knock-out strain. To demonstrate that this phenotype is due to the absence of PRK, we sought to functionally complement the ΔPRK strain.

To verify whether the loss of photoautotrophic growth is due to the absence of PRK in the mutant, we functionally complemented the ΔPRK strain with a synthetic construct designed to restore PRK expression. This construct was built using the intron-less coding sequence (CDS) of the PRK1 gene, obtained from cDNA, fused to a triple HA tag under the control of the promoter, 5’UTR and 3’UTR/Terminator of PSAD gene. This transcriptional unit was coupled to a hygromycin resistance gene ( Figure 2A ). The PSAD genetic elements were previously shown to allow strong constitutive expression of a reporter gene (Crozet et al., 2018). The synthetic construct containing the two transcriptional units was introduced into Chlamydomonas nuclear genome by transformation. Multiple clones from the transformation were selected in the light on TAP solid medium supplemented with hygromycin and named C_x, for complemented strain X. The C_x strains were genotyped by PCR using primer couples allowing specific detection of either the CIB insertion characteristic of the ΔPRK strain background or of the synthetic transgene ( Figure 2B ). All five chosen C_x and the mutant were found to contain the CIB1 insertion confirming the ΔPRK background. All selected clones also displayed the presence of the synthetic PRK transgene that was absent in the ΔPRK strain. We performed anti-PRK western blots to confirm the expression of the transgenic PRK. As shown in Figure 2C , 3xHA tagged PRK, which has a higher molecular weight than endogenous PRK, is present in all C_x complemented strains. In comparison with the gradient of CC-4533 total protein extract, we observed that the expression of the transgenic PRK-3HA is much lower than the endogenous PRK and shows some variability between C_x clones. This may be due to the design of the synthetic PRK construct and/or to position effects due to random insertion of the transgene in distinct genomic sites.

If the loss of PRK in the mutant is responsible for its growth phenotype, a functional complementation restoring photoautotrophic growth in minimal medium should be observed in the C_x strains. The growth phenotype of the different strains was assessed using a spot test analysis on solid media ( Figure 2D ).

In TAP medium in the dark, Chlamydomonas grows heterotrophically and all C_x strains grew like the reference strain. In TAP medium in the light, the complemented strains C_x grew similarly to CC-4533 although C₁₂ appeared to grow significantly more slowly. By contrast the ΔPRK strain grew much more slowly than all other strains, as shown previously ( Figure 1E ). Nevertheless, after 10 days the growth the ΔPRK strain was clearly visible ( Figure 2D ). In HSM medium in the light, a restoration of the photoautotrophic growth was observed in the C_x strains. This demonstrates that PRK is essential for photoautotrophy in Chlamydomonas. However, the growth of C_x strains appeared slower compared to CC-4533 suggesting a partial functional complementation. To assess whether this phenotype was caused by a lower level of PRK protein, the relative quantity of PRK expressed in the complemented strains was determined. All the C_x strains showed a relatively low expression of the PRK enzyme, with a maximum of 40% of the PRK content compared to the reference strain ( Figure 2E ). The C₁₂ strain that has a barely detectable level of PRK still partially restored photoautotrophy. PRK activity measurements correlated with the PRK levels estimated by western blot ( Figure 2F ). To quantitatively assess the growth limitation imposed by decreased PRK contents, we performed growth kinetics in a controlled photobioreactor. The C_x strains grew in TAP liquid medium in accordance with their PRK content ( Figure 2G ), as observed in the spot test assays ( Figure 2D ). Importantly, the strain with 40% of PRK content (C₁₁) behaved as the reference strain in TAP medium but not in minimal medium ( Figure 2D ). This suggests that 40% of PRK content compared to the reference strain is limiting the CBBC in Chlamydomonas. In tobacco, the content of PRK leading to a phenotype was below 15% of the reference (Paul et al., 1995; Banks et al., 1999). Therefore, the excess of PRK may be more limited in Chlamydomonas than in tobacco. In order to determine more precisely the PRK content limiting the CBBC and to test whether overexpression of PRK may increase the CBBC turnover, a higher level of expression of the transgene needs to be achieved.

Although position effect may account for the variations observed between transformants, the limitation is more likely due to the design of the synthetic transgene ( Figure 2A ). Indeed, we screened hundreds of transformants and never found any clone with a PRK level above 40% of the level of the endogenous PRK, as exemplified in Figure 2 . This suggested that the genetic elements controlling PRK expression may not be appropriate to ensure an expression level comparable to the reference strain.

The limitation to about 40% of the protein level found in the reference strain when expressing PRK from the synthetic construct may stem from various factors. PRK, like numerous other CBBC enzymes, is highly abundant and is estimated to represent 0.25% of total cellular proteins in Chlamydomonas (Hammel et al., 2020). Although the strong constitutive PSAD promoter and 5’UTR were employed to drive PRK expression in the synthetic pCMM-24 construct, it is possible that the endogenous PRK promoter is significantly stronger and would ensure a higher expression level. The absence of introns in the synthetic gene may also be partly responsible for the low expression. Indeed, it is now well-established that the presence of introns in the synthetic construct is frequently required to ensure high-level expression of a transgene in Chlamydomonas (Lumbreras and Purton, 1998; Fuhrmann et al., 1999; Baier et al., 2018). The presence of introns could boost gene expression due to the presence of a transcriptional enhancer, through a process called Intron-mediated enhancement that stimulates directly transcription or through interaction with the spliceosome (Schroda, 2019; Baier et al., 2020). An online tool has been developed to design Chlamydomonas transgenes with artificial introns (Jaeger et al., 2019). The fusion of the recombinant PRK with a triple HA-tag may also impact expression, potentially by destabilizing the protein or affecting its activity.

To investigate the significance of the promoter strength and of the presence of introns we have designed new synthetic constructs that incorporate the endogenous PRK1 promoter and its 5’UTR to drive either expression of the PRK1 CDS devoid of introns and also lacking triple HA-tag or the native PRK1 coding sequence containing natural introns ( Figure 3A ). The PRK1 promoter and the native PRK1 gene coding sequences were amplified by PCR from Chlamydomonas CC-4533 genomic DNA and cloned as level 0 MoClo parts. These parts were then used to build the two transcriptional units later on coupled to the hygromycin resistance gene to generate pCMM-25 and pCMM-26 ( Figure 3A ). Each of these synthetic constructs were introduced in the nuclear genome of the ΔPRK strain. The original synthetic gene (pCMM-24) containing the PSAD promoter and PRK1 CDS was also transformed again in the ΔPRK strain as a control. The transformants were selected on photoautotrophic growth restoration in minimal medium. For each transformation, roughly 50 colonies were selected for further analysis.

As we previously observed with the first synthetic construct that in the ΔPRK background cell growth is directly correlated with PRK expression level ( Figure 2 ), we selected the fastest growers to increase the probability to select strains with higher PRK levels through spot test assays on HSM medium after growth to saturation in TAP medium ( Figure 3B ). For several representative fast growers, we analyzed the PRK content by immunoblot ( Figure 3C ) and the growth properties in liquid HSM medium in a photobioreactor ( Figures 3D, E ) or in solid HSM medium ( Figure 3F ). As previously shown ( Figure 2 ), the pCMM-24 construct (α strains) only allowed partial complementation of the ΔPRK phenotype. The level of PRK was limited to 21% in the best case and photoautotrophic growth was only partially restored compared to CC-4533. Comparable results were obtained with the pCMM-25 construct as exemplified by the best growing strain D5 which PRK level was limited to 25% of the reference strain CC-4533 and consistently showed partial complementation of photoautotrophic growth. This indicates that neither the expression without the triple-HA tag nor the use of the PRK1 promoter to drive PRK1 CDS expression resulted in increased PRK expression level. Conversely, the pCMM-26 construct expressing the intron-containing PRK endogenous coding sequence allowed to fully complement the phenotype of the ΔPRK strain ( Figure 3 ) and even overexpress the protein ( Figure 4 ). This indicates that the presence of introns in the PRK1 coding sequence is crucial for high level expression of the transgene. This result is consistent with previous studies (Lumbreras and Purton, 1998; Fuhrmann et al., 1999; Baier et al., 2018; Lauersen et al., 2018; Schroda, 2019; Baier et al., 2020). It would be interesting in future studies to determine whether the high-level expression is linked to a specific PRK1 intron. Even with the pCMM-26 intron containing construct, a range of PRK expression levels (34%-250% relative to CC-4533) was observed in the transformants. This variability is most likely attributable to random insertion of the construct in regions of the genome with a variable ability to drive gene expression. This position effect is classically observed in Chlamydomonas and can be exploited to generate strains with diverse levels of expression of a given transgene. In our case, immunoblot and growth assays revealed that a PRK content of 86% is sufficient to fully restore photoautotrophic growth with kinetics comparable to the reference strain CC-4533 ( Figure 3 ). This result suggests that PRK may be present in excess in Chlamydomonas. Nevertheless, this excess appears much more limited than previously reported in tobacco where PRK limitation was only observed below 15% of the WT level (Paul et al., 1995; Banks et al., 1999). To confirm that PRK endogenous level does not limit the CBBC we analyzed strains overexpressing PRK.

To analyze the impact of PRK overexpression in Chlamydomonas, we examined strains transformed with the pCMM-26 construct either in the ΔPRK or the CC-4533 backgrounds ( Figure 4A ). The growth of strains overexpressing PRK 1.5-fold (A2 strain) or 2.5-fold (β3 strain) ( Figure 4B ) was analyzed in a photobioreactor in liquid HSM medium under three different light intensities ( Figures 4C–E ). Strains expressing 100% (CC-4533) or 25% (D5) of PRK were used as controls. The growth of the two overexpressor strains was comparable to the CC-4533 strain. At high light intensity (300 μmol m^-2 s^-1) growth of the overexpressor was similar to CC-4533 while at lower light intensities the 1.5-fold overexpressor showed a marginally slower growth kinetics compared to the two other strains. The cause of this slower growth in low light of the A2 strain is not known but could be related to the site of insertion of the transgene. Nonetheless, these results clearly show that overexpression of PRK does not lead to an increased growth. This suggests that the endogenous level of PRK in Chlamydomonas is not limiting the Calvin-Benson-Bassham cycle under optimal growth conditions. It would be interesting to determine if PRK may become limiting in conditions of increased CBBC flux, such as overexpression of a limiting enzyme like SBPase.