Chlamydomonas reinhardtii strain UVM4 was used as the parental strain for transformations. This strain was derived from several rounds of mutation in the parental CC-4350 by Dr. Juliane Neupert in the lab of Prof. Dr. Ralph Bock (Neupert et al., 2009) and contains a mutation in Sir2-type histone deacetylase (SRTA), which enables improved transgene expression rates from the algal nuclear genome (Neupert et al., 2020). UVM4 is not able to use nitrate as a nitrogen source due to nit1/nit2 locus mutations (Freudenberg et al., 2021). Microalgal cultures were routinely maintained in a Tris acetate phosphate (TAP) medium (Gorman and Levine, 1965) with updated trace element solution (Kropat et al., 2011) and maintained under 150 μmol m⁻² s⁻¹ mixed cold and warm LED lights with 120–190 rpm agitation in shake flasks or microtiter plates. Light intensities and spectra were measured with a handheld spectrometer (Spectromaster C-7000, Sekonic). Ammonium in TAP salts solution was replaced with equimolar NaNO₃ to make TAP-NO₃. Replacements of phosphate with phosphite to make TAPhi and TAPhi-NO₃ media are described in Supplementary Material 1.

A high-density 6xP medium was prepared as described in Freudenberg et al. (2021), and buffered phosphite solutions to match molar concentrations of phosphorous to make a 6xPhi medium are as described in Supplementary Material 1. All phosphite solutions were filter sterilized and added to media after autoclaving. Cultivation in CellDeg HD100 cultivators (CellDeg GmbH, Germany) was performed with the indicated light and CO₂ regimes by the growth-control unit using either 6xP or 6xPhi media. Precultures were conducted as follows: 20 ml of the late-exponential phase (~1.5 × 10⁷ cells mL⁻¹) C. reinhardtii pre-cultured in TAPhi-NO₃ was centrifuged and resuspended in either 6xP or 6xPhi media. The cells were diluted 1:2 with a fresh medium, and 5 mL was added to 95 ml in the CellDeg reactor at T0. Illumination was delivered by a Valoya broad spectrum LED board supplied by CellDeg GmbH (Germany, the spectrum is presented in Supplementary Material 2).

Growth of algae and contaminants was analyzed by flow cytometry using an Invitrogen Attune NxT flow cytometer (Thermo Fisher Scientific, UK) equipped with a 488 nm blue laser for forward-scatter and side-scatter measurements, and a 695/40 nm filter to detect chlorophyll and non-fluorescent particles, respectively. All culture samples were diluted 1/100 with 0.9% NaCl solution and measured in technical triplicates using previously described settings (Overmans and Lauersen, 2022).

Plasmids used in this study are listed in Supplementary Table 1. All cloning and plasmid linearization was performed with Thermofisher FastDigest restriction enzymes, New England Biolabs Quick Ligase, and Q5 polymerase following the manufacturer's protocols. Plasmids were maintained in chemically competent Escherichia coli DH5a transformed by heat shock. Glass bead transformation of C. reinhardtii was performed as previously described for both chloroplast and nuclear-targeted genetic constructs (Kindle, 1990; Kindle et al., 1991). Chloroplast transformation of the pPO3 plasmid [(Changko et al., 2020) graciously provided by Prof. Saul Purton] was performed following the protocol described in Economou et al. (2014) with the following modifications: 10 μg circular DNA and 0.1 mm diameter glass beads rather than 0.424–0.600 mm as commonly used for nuclear transformation (Economou et al., 2014). Recovery was performed in 45 mL TAPhi liquid for 5 days with 150 μE PAR prior to plating. Selection was achieved by plating on TAPhi agar plates incubated at 200 μE for 2–3 weeks. Transformation and selection resulted in only 3–10 colonies per event, and multiple cycles of transformations were used to collect ~20 colonies capable of growth in liquid TAPhi. Colonies were then grown in TAPhi liquid in microtiter plates until green for 1 week, but were not checked for homoplasmy before next transformations.

One transformant with clear growth in liquid TAPhi, hereafter named UVM4-Phi, was transformed for complementation of nitrate metabolic capacity by co-transformation of linearized pMN24 (Fernández et al., 1989) and pMN68 (Schnell and Lefebvre, 1993; Chlamydomonas Resource Center, https://www.chlamycollection.org) by glass beads as previously described (Freudenberg et al., 2021) with overnight recovery and subsequent selection on TAPhi-NO₃ agar plates. Transformant colonies recovered on TAPhi-NO₃ plates were recovered after 2 weeks. Resultant colonies were then compared in liquid media in 24-well microtiter plates with standard lighting conditions at 180 rpm. Homoplasmy of the pPO3 integration into the chloroplast genome was determined only in the colonies, which showed growth in this medium, as Phi and NO₃ metabolic capacity was used as a main selection criterion. Homoplasmy was determined in final strains by PCR using primers Fw: AATTGTATGGGCTCACAACAAACTTAAAGT and Rv: TAAAATTGTGAGACCATGAGTAATGTTCCTCC. The resulting transformants were also screened by an iodine vapor assay as previously described (Wichmann et al., 2018) to determine if random integration had caused starch synthesis modifications. Modified UVM4 transformants, which grew with phosphite and nitrate, are referred to as UVM4-phosphite-nitrate (UPN) strains.

Efficiency of nuclear transgene expression of intermediate strains was investigated by glass bead transformation of the pOpt2_mVenus_Paro plasmid (Wichmann et al., 2018) followed by selection on each respective modified medium with 10 mg L⁻¹ paromomycin and fluorescent reporter expression analysis. Fluorescent mVenus expression intensities were analyzed by picking primary transformant colonies using a PIXL robot (Singer Instruments, UK) to 384 colonies/plate layout on manufacturer-supplied rectangular Petri dishes. After 1 week, colonies were replicated using the Singer Instruments ROTOR to generate imaging-ready colonies. White-light pictures of algae colony plates were taken in the built-in PIXL camera. Chlorophyll and mVenus fluorescence signals were captured in an Analytik Jena Chemstudio Plus gel doc with an eLite halogen light source and excitation filters. Chlorophyll fluorescence was captured by 475/20 nm excitation with orange DNA gel emission filter with 1 s exposure, while an mVenus signal was captured with 510/10 nm excitation and 530/10 nm emission filter with 30 s exposure.

Plasmids for algal nuclear genome-based expression of the Pogostemon cablin patchoulol synthase (PcPS, UniProtQ49SP3) were adapted from (Lauersen et al., 2016). The gene expression cassette for PcPS expression was modified from the pOpt2 vector concept of (Wichmann et al., 2018) to contain transgene designs presented in Baier et al. (2020), Einhaus et al. (2021), and Freudenberg et al. (2021). Briefly, PcPS expression here was driven by the hybrid heatshock 70A beta tubulin promoter described by Einhaus et al. (2021), and the mVenus cassette was modified to contain two copies of the C. reinhardtii ribulose-1,5-bisphosphate carboxylase/oxygenase small subunit (RBCS2) intron 1. The RBCS2 intron 2 was moved into the C-terminal strep-II tag of the gene-of-interest expression cassette in the pOpt2_mVenus_Paro plasmid to match that recently described (Baier et al., 2020). This plasmid confers paromomycin resistance in C. reinhardtii (Wichmann et al., 2018). PcPS was subcloned into BamHI-BglII, and 2X, 3X, and 4X PcPS expression cassettes were built by ScaI-BglII inserts from the previous plasmid subcloned into ScaI-BamHI of the progenitor plasmid described in Supplementary Figure 1. All constructs contain the C-terminal mVenus (YFP) fusion, which enabled plate-level fluorescence detection in UPN colonies picked by the PIXL robot. C. reinhardtii squalene synthase (UniProt A8IE29) knockdown was achieved by secondary transformation using the previously described pOpt2_cCA-gLuc_i3-SQS_Spect plasmid (Wichmann et al., 2018). UPN PcPS-YFP + SQS k.d. double transformants were selected on TAPhi-NO₃ agar media containing 10 mg L⁻¹ paromomycin and 200 mg L⁻¹ spectinomycin as previously described (Wichmann et al., 2018). YFP and luciferase signals of UPN colonies were captured in the ChemstudioPLUS with previously described buffers and reagents for Gaussia princeps luciferase bioluminescence analysis (Lauersen et al., 2013a). Full-length target recombinant protein was determined by SDS PAGE and in-gel fluorescence of whole cell pellets in the Chemstudio PLUS with YFP filters described for colony screening. All plasmid sequence files used in this work are given in Supplementary Material 3.

To test the ability of engineered C. reinhardtii UPN lines to withstand contamination in non-sterile conditions using media modifications presented in this work, cultivation was performed with intentional yeast contamination. TAP-NO₃, TAPhi-NO₃, 6xP, and 6xPhi media were used to cultivate an engineered SQS k.d. + 2XPcPs expressing UPN transformant. Saccharomyces cerevisiae (yeast) cells were cultured in a yeast extract-peptone-dextrose (YPD) medium (Cold Spring Harbor Protocols) overnight at 28°C. The following day, pelleted cells were resuspended in 300 ml 6xP or 6xPhi media. The 2XPcPS-SQS k.d. strain was cultivated in TAPhi-NO₃ until a mid-late exponential phase (1.7 × 10⁷ cells mL⁻¹); 50 mL was spun down and resuspended with either 6xP or 6xPhi media. The cells were washed two times with the target test medium, and then diluted into 300 mL of the same media. About 4 mL of these dilute cultures was added to microtiter plates for each condition as described. Triplicate wells in 6-well microtiter plates containing 4 mL dilute UPN patchoulol culture in 6xP or 6xPhi media were inoculated with either 500 μl of yeast solutions (above) or a clean medium as controls. About 500 μl of an n-dodecane overlay was also added to each well. Approximately 3 ml of a concentrated potassium bicarbonate buffer was added between the wells to provide a dilute CO₂ atmosphere as previously described (Dienst et al., 2020). Cultures in TAP-media were grown for 6 days and 6xP/Phi for 9 days on laboratory shakers at 120 rpm with a 12-h:12-h light:dark cycle (150 μE). Cell densities and patchoulol productivities were analyzed as described above.

The C. reinhardtii mutants UVM4 and UVM11 (Neupert et al., 2009) exhibit reduced transgene silencing due to a mutation in the in Sir2-type histone deacetylase (SRTA; Neupert et al., 2020). UVM4 has served as a powerful parent strain for many recent examples of metabolic engineering in this green microalga (Lauersen et al., 2016, 2018; Wichmann et al., 2018; Einhaus et al., 2021; Freudenberg et al., 2021). Despite its value for past experiments, the alga is grown at neutral pH and contains mutations in its nitrate metabolism, which prevent use of this nitrogen source. These two features manifest in high risk of microbial contamination and the inability to use industrially relevant culture media (Changko et al., 2020; Freudenberg et al., 2021). To prepare UVM4 strains for broader applications, we set to complement it with the capacity to use phosphite as a P source and nitrate as an N source.

We transformed UVM4 with plasmid pPO3 (Changko et al., 2020) to express the P. stutzeri WM88 phosphite NAD⁺ oxidoreductase ptxD (López-Arredondo and Herrera-Estrella, 2012) that converts inorganic phosphite into organic phosphate from the algal chloroplast genome (Figure 1A). We found it was possible to transform UVM4 with this chloroplast genome-integrating plasmid by glass bead transformation and select colonies on a TAPhi medium with no additional selection pressure. A resulting UVM4-Phi transformant was then subsequently transformed with pMN24 and pMN68 plasmids, which contain genomic copies of the nit1 and nit2 loci, respectively, to complement nitrate metabolism capacity (Fernández et al., 1989; Schnell and Lefebvre, 1993; Figure 1A).

Colonies were recovered by selection on TAPhi-NO₃ plates with no antibiotic (Figure 1A). Complementation with nit1/2 can sometimes lead to colonies that survive on the agar plate, but do not perform well in a liquid medium. Therefore, we also benchmarked performance of 24 UVM4-phosphite-nitrate (UPN) colonies derived from these transformations in TAPhi-NO₃ and photoautotrophic cultivation with CO₂ as a carbon source (Figure 1A, lower right). UPN strains grown in a liquid medium with nitrate exhibited variable performance, especially in photoautotrophic conditions (Figures 1A,B). In the TAP medium, with ammonium and phosphate, UVM4-Phi and UPN strains were found to reach lower cell densities than UVM4. These strains exhibited lower portions of small cell debris and attained reasonable cell densities (Supplementary Figure 2). Most colonies maintained normal starch accumulation, which was qualitatively assessed by iodine vapor; however, Colonies 19 and 23 showed reduced iodine staining (Figure 1C). Homoplasmy of pPO3 integration was determined in UVM4-Phi and nitrate-complemented strains (Supplementary Figure 3).

To confirm that the three plasmid integrations did not modify the performance of the parent UVM4 capacities for nuclear transgene expression, several UPN strains with acceptable growth in liquid phosphite-nitrate media were transformed with a YFP reporter plasmid (Wichmann et al., 2018). High-throughput robotics-assisted colony picking allowed analysis of between 943 and 1,079 colonies for each strain, and plate-level fluorescence imaging was used to quantify reporter expression across the populations. YFP reporter expression efficiencies in the final UPN strains were comparable to parent UVM4 (Figure 1D).

Using mono- and di-basic forms of phosphite, we generated a buffered phosphite solution to emulate the phosphate buffer solution of a recently published 6xP medium (Freudenberg et al., 2021; Supplementary Material 1). In order to compare whether our high-density phosphite medium (6xPhi) could be used in a comparable fashion to 6xP, we benchmarked growth of a UPN strain in a high-density cultivation concept using high-light and membrane-delivered CO₂ in CellDeg HD100 cultivators (Figure 2A). We did not observe differences in performance for the UPN strain cultivated in 6xP or 6xPhi, which reached comparable cell densities throughout cultivation (Figure 2B).

We then chose to combine two proven engineering strategies for sesquiterpenoid production from a UPN strain grown only in the TAPhi-NO₃ medium. The C. reinhardtii codon-optimized P. cablin patchoulol synthase (PcPS) was expressed in 1, 2, 3, and 4X copy fusion protein constructs with C-terminal YFP from the nuclear genome of this alga (Figure 3). Robotics-assisted colony picking and YFP screening allowed selection of six transformants per plasmid with confirmed PcPS expression, which were benchmarked for patchoulol productivity as previously described (Lauersen et al., 2018). The best-performing transformants were subsequently transformed with a secreted luciferase-artificial-micro-RNA expression construct, targeting the C. reinhardtii squalene synthase (SQS; Wichmann et al., 2018). After colony recovery and robotics picking, plate-level imaging was used to isolate colonies with a YFP fluorescence signal (PcPS) and luciferase activity (SQS k.d.; Figure 3). Patchoulol productivity analysis indicated striking improvements in patchoulol productivity for SQS k.d. strains compared to parentals with the best-performing strains, generating ~145 fg patchoulol cell⁻¹ (Figure 3). Full-length fusion protein expression could be confirmed only for 1-3XPcPS-YFP constructs by in gel fluorescence (Supplementary Figure 4).

As contamination of cultures can be an issue with neutral pH cultivation, we wanted to determine if phosphite and nitrate could permit algal growth in the presence of contamination. We intentionally contaminated the best-performing UPN PcPS SQS k.d. strain in mixotrophic (acetic acid) and photoautotrophic cultures in media with nitrate as a nitrogen source and either phosphate or phosphite as a phosphorous source. Yeast cells were added to cultures directly in higher cellular abundances than algal cells (Figure 4). In all media conditions, yeast cells did not proliferate, regardless of the presence of organic carbon but also did not reduce in number. When acetic acid (TAP-derived) media were used, algal growth was reduced compared to cultivations without yeast, also with phosphite (Figure 4). No difference in performance was noted in photoautotrophic cultures. In all conditions, the presence of high concentrations of yeasts in cultivations did not inhibit heterologous patchoulol production (Figure 4).

We then benchmarked patchoulol productivity in a 200 mL culture in a membrane gas delivery bioreactor, containing 10% dodecane overlay to capture a heterologous sesquiterpenoid product. Culture volume was adjusted to 200 mL to avoid contact of dodecane with the hydrophobic gas delivery membrane during shaking, and the culture was operated in non-sterile conditions. The culture accumulated up to 6.5 x 10⁷ ± 1.9 x 10⁶ cells mL⁻¹ and generated 6.2 mg L⁻¹ patchoulol in 6 days using this system.

We chose to introduce phosphite metabolism by transformation of plasmid pPO3 as this also contains extra future chloroplast genome engineering potential through the addition of the W_TGA tRNA for tryptophan as previously described (Changko et al., 2020). When filter sterilization was used, transformation and selection on phosphite solutions were greatly improved, and appearance of background algal growth at the plate level was reduced (data not shown). After confirmed growth of pPO3 transformants in liquid phosphite, nitrate metabolism was complemented by transformation of both pMN24 (NIT1) and pMN68 (NIT2) plasmids in a UVM4-Phi strain. This double transformation is relatively inefficient; nevertheless, we could generate several dozen colonies per transformation, which recovered on nitrate plates. Colonies, which recovered on nitrate plates did, however, not all perform well in liquid culture growth in nitrate-containing liquid media. We chose to move forward with only those colonies that appeared to grow to a dark-green stationary phase (Figure 1A). Colonies were then checked for homoplasmy integration of the pPO3 phosphite metabolism-conferring plasmid (Supplementary Figure 3), and two transformants were benchmarked for their growth in phosphite- and nitrate-containing media compared to their parental strains (Figure 1B).

To determine if our strategy for UVM4 augmentation would allow future engineering to benefit from these metabolic enhancements, two questions remained: 1) was nuclear transformation expression efficiency disturbed during these events in UVM4 derivatives? 2) Can inorganic phosphite be used in a similar way to organic phosphate for buffered media solutions? We benchmarked two fully complemented UPN transformants, their UVM4-Phi parents, and the UVM4 starting strain for YFP efficiency expression from the nuclear genome. Using high-throughput colony picking, we were able to analyze ~1,000 colonies per transformation event and compare YFP expression efficiencies across the populations (Figure 1C). Although some variance, little difference could be seen in the total ratio of high and mid-range YFP expressing colonies, suggesting these three plasmid integrations had not modified nuclear transgene expression capacity from the UVM4 background.

We then set out to make a buffered phosphite solution, which could replace buffered phosphate solutions in culture media (Supplementary Material 1). In direct comparison growth tests, the UPN strain did not show performance differences in phosphite compared to phosphate in photoautotrophic high-density cultivations (Figure 2). Our results indicate the use of inorganic buffered phosphite solutions as media components is not different than phosphate for the augmented strains. As phosphate is a globally dwindling resource important to agriculture, bio-conversion of phosphite into phosphate may also enable the use of this waste mineral to yield bio-fertilizers through engineered algal cultivation.

UPN strains were maintained exclusively on the TAPhi-NO₃ medium for all routine lab work. A further aim was to determine if it was possible to conduct additional metabolic engineering in these strains for heterologous isoprenoid production using Phi-NO₃ media. Plasmids were constructed based on previous designs to express the patchoulol synthase (PcPS) and localize it in the cytoplasm of the alga where this enzyme is known to convert freely available farnesyl pyrophosphate (FPP) into patchouli alcohol (patchoulol; Lauersen et al., 2016). We chose to combine recently published modifications in promoter (Einhaus et al., 2021) and intron use (Baier et al., 2020; Freudenberg et al., 2021; Supplementary Figure 1) in order to assist recombinant protein accumulation in an effort to enhance product yields. Previous work on PcPS indicated cellular patchoulol productivities could be enhanced when the protein was fused to itself in a repetitive fashion to yield more active sites per translated protein (Lauersen et al., 2016). Here, we copied this gene design strategy and combined it with an artificial micro-RNA (amiRNA) knockdown of the squalene synthase (SQS; Figure 3). It was previously found that SQS k.d. improved (E)-α-biabolene titers from the cytoplasm of C. reinhardtii as this is the direct competitor for FPP precursor (Wichmann et al., 2018). Additive PcPS units were found here to increase cellular patchoulol yields as previously observed (Lauersen et al., 2016). However, increasing repetitions past 3XPcPS copies were found to be unstable and did not generate reliable patterns in patchoulol production, despite some production being observed (Figure 3, Supplementary Figure 3). As expected, based on past work with bisabolene (Wichmann et al., 2018), addition of the SQS k.d. to the best performing PcPS variant of each plasmid led to transformants with drastic improvements in patchoulol productivity. Previous engineering of patchoulol production for C. reinhardtii led to a maximal volumetric productivity of ~350 μg patchoulol L⁻¹ in mixotrophic 400 mL bioreactor conditions from a 3XPcPS-YFP transformant (Lauersen et al., 2016). Here, a 2XPcPS transformant subsequently transformed with the SQS k.d.-generated lines, producing 700–1,400 μg patchoulol L⁻¹ culture (Figure 3, Supplementary Figures 5, 6). Improvements were observed across 1–4X PcPS-YFP lines by SQS k.d. and from 1–3XPcPS-YFP; mean production increased with increasing PcPS units. Maximal cellular productivity was observed in a single 2XPcPS SQS k.d. line, with up to 143 fg patchoulol cell⁻¹ (Figure 3).

A risk to scaled cultivation of engineered C. reinhardtii in bio-production concepts is contamination and reduced productivities, which is especially true for cell-wall-deficient strains that may be more readily outcompeted by contaminants. We intentionally contaminated mixotrophic and photoautotrophic media, containing either NO₃ or Phi and NO₃ with yeast, and cultivated a UPN-patchoulol-producing strain in these sub-optimal conditions (Figure 4A). We inoculated the 2XPcPS-SQS k.d. UPN strain into media containing 6 x 10⁶ cells mL⁻¹ yeast, the same cell density as reached in the mid-log phase for the algal cells. We chose to provide CO₂ using potassium bicarbonate buffers (Dienst et al., 2020) rather than direct gas delivery to further challenge the phototrophic cultures. In all conditions, yeast cells did not proliferate, with either acetic acid as a carbon source, or in the photoautotrophic media (Figure 4A). Mixotrophic cultures exhibited lower algal cell densities in later stages of cultivation than in the absence of yeast, likely due to the yeast sequestering some of the acetic acid. However, in the establishment phase (days 0–3), growth with yeast was not markedly different than that of cultures without yeast. In photoautotrophic cultures, yeast was inoculated at higher starting cell densities (8 x 10⁶ cells mL⁻¹) as there is no organic carbon source. This density was chosen to determine if the yeast cells would competitively inhibit the inoculated algal cells. Here, the UPN strain was able to overtake the yeast cells, demonstrating linear growth in both conditions relative to carbon diffusion rates in the medium. Under all conditions, the presence of yeast contaminants did not hinder the accumulation of heterologous patchoulol in dodecane overlays, which exhibited similar productivities per algal cell (Figure 4A). Our results indicate that both nitrate and phosphite are a powerful combination to limit contaminating microbial competitors in engineered algal cultivation concepts.

To determine if we could produce patchoulol in non-sterile conditions, we cultivated this strain in a CellDEG HD100 bioreactor with dodecane overlay using the 6xPhi medium (Figure 4B). Dodecane impairs the hydrophobic gas delivery membrane of the reactors, so we used 200 mL culture volume to prevent the solvent interacting with the membrane. Previous photoautotrophic yields of this product were only 350 μg L⁻¹ in 8 days (Lauersen et al., 2016). Here, without any process optimization, ~6.2 mg patchoulol L⁻¹ was produced from CO₂ in 7 days (Figure 4). A previous study with Synechocystis sp. PCC 6803 used a similar membrane gas delivery system for 10-ml cultures to generate up to 17.3-mg patchoulol L⁻¹ in 8 days using a two-stage semi batch mode where half of the culture medium was replaced after 96 h. We did not further optimize our production experiments in the HD100, as the risk of dodecane-membrane wetting means cultivation must be performed with volumes not intended for the system. A further issue of the dodecane overlay in turbid algal cultures is the formation of emulsions with hydrophobic cellular components and the dodecane solvent (Lauersen, 2019). The surface interaction of culture and dodecane causes significant emulsion formation in this volume ratio (Figure 4B), which is less pronounced in smaller volume cultivation units. Indeed, better isoprenoid extraction methods are needed for performance benchmarking and production concepts, which do not rely on hard-to-handle solvents. Nevertheless, our results indicated the combination of nitrate and phosphite metabolic capacities enables high-density cultivations of engineered strains to be performed with reduced risks of contamination without affecting process yields. To our knowledge, this is the first demonstration of combinatorial plastid and nuclear genome engineering in a green alga, which helped facilitate growth in non-sterile conditions, coupled with nuclear transgene expression for heterologous metabolite production.