Contribution of Complex I NADH Dehydrogenase to Respiratory Energy Coupling in Glucose-Grown Cultures of Ogataea parapolymorpha

Yeast strains, culture conditions, and maintenance.The yeast strains used in this study were CBS11895 (DL-1), a wild-type Ogataea parapolymorpha strain ordered from CBS-KNAW (Westerdijk Fungal Biodiversity Institute, Utrecht, The Netherlands), and IMD010, a CBS11895-derived mutant with a disruption in the NUBM gene (nubm^G445GC). Strains were grown in an Innova shaker incubator (New Brunswick Scientific, Edison, NJ, USA) set to 30°C and 200 rpm, in 500-ml shake flasks containing 100 ml of medium. Heat-sterilized (120°C for 20 min) YPD medium (10 g liter⁻¹ Bacto yeast extract, 20 g liter⁻¹ Bacto peptone, 20 g liter⁻¹ glucose, demineralized water) was used for strain construction and maintenance. Solid medium was prepared by addition of 2% (wt/vol) agar to YPD medium. Frozen stock cultures were prepared from exponentially growing shake flask cultures by addition of glycerol to a final concentration of 30% (vol/vol) and aseptically stored in 1-ml aliquots at −80°C.

Molecular biology techniques and strain construction.Escherichia coli strain DH5α was used for plasmid transformation, amplification, and storage. Plasmids were isolated from E. coli using a GenElute Plasmid Miniprep kit (Sigma-Aldrich, St. Louis, MO, USA). Genomic DNA of yeast colonies used as the template for diagnostic PCR was isolated using the lithium acetate (LiAc)-sodium dodecyl sulfate method (66). Diagnostic PCR was performed using DreamTaq polymerase (Thermo Fisher Scientific, Waltham, MA, USA) and desalted primers (Sigma-Aldrich). DNA fragments obtained by PCR were separated by gel electrophoresis, and PCR purification was performed with a GenElute PCR Clean-Up kit (Sigma-Aldrich).

For the construction of strain IMD010 (O. parapolymorpha nubm^G445GC), the O. parapolymorpha open reading frame (ORF) encoding the complex I NUBM 51-kDa subunit (OpNUBM, locus tag HPODL_04625; GenBank accession number XM_014080963.1) was identified. OpNUBM was found via a homology search (blastn [https://blast.ncbi.nlm.nih.gov]) in the O. parapolymorpha CBS11895 (DL-1) RefSeq assembly (NCBI accession number GCF_000187245.1) (67) using the O. polymorpha (Pichia angusta) partial sequence identified as coding for the complex I Nubm subunit (GenBank accession no. AL434382) (21, 68) as input. Other O. parapolymorpha complex I subunits were assigned based on protein sequence homology (tblastn [https://blast.ncbi.nlm.nih.gov]) with known P. pastoris subunits (67, 69, 70). OpNUBM was disrupted using the pUDP CRISPR/Cas9 system described previously (71). The guide RNA (gRNA) donor plasmid pUD676 was de novo synthesized by GeneArt (Thermo Fisher Scientific) and contained the synthetic 233-bp double-stranded DNA (dsDNA) gRNA construct with a gRNA spacer sequence (5′-CCTGATGTAAATATACGCTG-3′) targeting OpNUBM after bp 445 out of 1,467 bp. pUD676 was then integrated into pUDP002 via BsaI-mediated assembly as described previously (71), yielding OpNUBM-targeting plasmid pUDP084. For disruption of OpNUBM, wild-type O. parapolymorpha was transformed with pUDP084 via electroporation and subjected to a prolonged liquid incubation protocol, as described previously for deletion of OpADE2 and OpKU80 (71). Primers 12200 and 12201 (5′-CCCAGCTACGATCTCAAGAC-3′ and 5′-AACTTGGTGCCCGAGTTAC-3′, respectively) were then used for PCR amplification of the OpNUBM locus of seven randomly picked single colonies, and subsequent Sanger sequencing (Baseclear, Leiden, The Netherlands) revealed that three out of seven tested colonies contained an indel at the gRNA target site. A single colony of one of the mutants, containing a single cysteine nucleotide insertion between position 445 and 446 of the OpNUBM ORF, was restreaked three times subsequently on nonselective YPD medium to remove pUDP084 and renamed IMD010.

Bioreactor cultivation.Bioreactor cultivation was performed using synthetic medium (SM) with the addition of 0.15 g liter⁻¹ of Pluronic 6100 PE antifoaming agent (BASF, Ludwigshafen, Germany). SM was prepared according to C. Verduyn et al. (72) and autoclaved at 120°C for 20 min. Glucose and vitamins (72) were prepared separately and filter sterilized (vitamins) or heat sterilized at 110°C for 20 min (glucose). Bioreactors were inoculated with exponentially growing cells from independent shake flask cultures (grown as described above) with SM and 20 g liter⁻¹ of glucose. All cultures were performed in 2-liter benchtop bioreactors (Applikon, Delft, The Netherlands) with initial volumes of 1.4 liters (batch) and working volumes of 1.0 or 1.4 liters (chemostat) or 1.4 liters (retentostat). Cultures were sparged with dried, compressed air (0.5 vvm [volume of gas per volume of liquid per minute]) and stirred at 800 rpm. Temperature was maintained at 30°C, and pH was controlled at 5.0 by automatic addition of a 2 M KOH (batch and chemostat) or 10% (wt/vol) NH₄OH (retentostat) solution by an ADI 1030 Bio Controller system (Applikon) or by an ez-Control bioreactor controller (Applikon). In chemostats and retentostats, the working volume was maintained by an electrical level sensor that controlled the effluent pump. Culture exhaust gas from bioreactors was cooled with a condenser (2°C) and dried with a Perma Pure dryer (Inacom Instruments, Veenendaal, The Netherlands) prior to online analysis of carbon dioxide and oxygen with a Rosemount NGA 2000 analyzer (Emerson, St. Louis, MO, USA). Batch cultures were performed with 7.5 g liter⁻¹ of glucose as the sole carbon source and an initial optical density at 660 nm (OD₆₆₀) of 0.3 (approximately 0.05 g liter⁻¹ of biomass dry weight). In chemostat cultures, 7.5 g liter⁻¹ of glucose was used as a sole carbon source, and the dilution rate (D) was set by maintaining a constant inflow rate. Cultures were assumed to have reached a steady state when, after a minimum of 5 volume changes, the oxygen consumption rate, carbon dioxide production rate, and biomass concentration changed by less than 3% over two consecutive volume changes. Retentostat cultivation was performed essentially as described by C. Rebnegger et al. (29). To predict the accumulation of biomass in retentostat cultures of strains CBS11985 and IMD010, the predictive biomass accumulation script of C. Rebnegger et al. (29) was used with m_S and Y_X/S^max values as estimated from chemostat cultures with D values of 0.1 and 0.025 h⁻¹. Based on the assumption that maintenance energy requirements are growth rate independent, a feeding regime was selected for O. parapolymorpha in which 10 g liter⁻¹ of glucose was used for the preceding chemostat phase, and 5 g liter⁻¹ was used for the retentostat phase in combination with a 1.2-liter mixing vessel, which was identical to the setup used for the earlier work on retentostat cultivation of P. pastoris (29). The dilution rate was determined by maintaining a constant inflow rate of medium from the mixing vessel, a 3-liter benchtop bioreactor (Applikon) with a working volume of 1.2 liters, and stirred at 500 rpm. The volume in the mixing vessel was kept constant by an electrical level sensor that controlled the feed pump of the mixing vessel. The medium was SM as described above but contained an additional 0.5 ml liter⁻¹ of the concentrated trace element solution (1.5× final concentration) and an additional 1 ml liter⁻¹ of the vitamin stock solution (2× final concentration) (72). Retentostat cultivation was preceded by a chemostat cultivation (D of 0.025 h⁻¹) using the same conditions as described for the subsequent retentostat cultivation. During the chemostat phase, the medium flowing into the mixing vessel contained 10 g liter⁻¹ of glucose. Once a steady state was achieved, the retentostat phase was initiated by two changes: (i) the medium flowing into the mixing vessel was changed to be drawn from a medium vessel with identical medium composition but with 5 g liter⁻¹ of glucose as the limiting compound, and (ii) the culture effluent was redirected through a filtered effluent port, equipped with a hollow stainless steel filter support with an autoclavable hydrophobic polypropylene filter with 0.22-μm pore size (Trace Analytics, Braunschweig, Germany). Prior to heat sterilization, the filter was soaked overnight in a 96% ethanol solution and subsequently rinsed with 1× phosphate-buffered saline (Sigma-Aldrich).

Biomass measurements.Optical density was measured at 660 nm on a Jenway 7200 spectrophotometer (Jenway, Staffordshire, UK). For biomass dry weight determination (typically performed in triplicate), exactly 10 ml of culture broth was filtered over predried and preweighed membrane filters (0.45-μm pore size; Pall Corporation, Ann Arbor, MI, USA), which were washed with demineralized water, dried in a microwave oven at 350 W for 20 min, and weighed immediately (73). Samples from chemostat and retentostat cultures were diluted with demineralized water prior to filtration to obtain a biomass dry weight concentration of approximately 2 g liter⁻¹. The exact dilution was calculated by weighing the amount of sample and diluent and assuming a density of 1 g ml⁻¹ for both fractions. For samples from late-stage retentostat cultures of strain IMD010 (after approximately 14 days and onwards), membrane filters were placed in glass bowls and covered with plastic funnels for microwave drying as under normal drying conditions biomass flakes formed that detached from the membrane filters, preventing accurate determination. Membrane filters were routinely redried and reweighed to ensure complete drying. Biomass protein content was determined using dried bovine serum albumin (BSA) (fatty acid free; Sigma-Aldrich) as described previously (74), with the modifications that NaOH was used instead of KOH and absorbance was measured at 510 instead of at 550 nm. Culture samples were diluted with demineralized water to biomass dry weight concentrations between 2.2 and 3.8 g liter⁻¹ prior to protein content analysis.

Metabolite analysis.For the determination of extracellular metabolite concentrations during batch fermentations, 1-ml aliquots of culture sample were centrifuged for 3 min at 20,000 × g, and the supernatant was used for analysis. Samples from chemostat and retentostat cultures were rapidly quenched with the cold steel beads method (75). Metabolite concentrations were analyzed by high-performance liquid chromatography (HPLC) on an Agilent 1100 HPLC (Agilent Technologies, Santa Clara, CA, USA) with an Aminex HPX-87H ion exchange column (Bio-Rad, Veenendaal, The Netherlands) operated at 60°C with 5 mM H₂SO₄ as the mobile phase at a flow rate of 0.6 ml min⁻¹.

Viability assays.For cell viability determination based on membrane integrity (via propidium iodide [PI]), approximately 0.5 ml of culture broth was sampled into 15 ml of ice-cold 10 mM Na-HEPES buffer (pH 7.2) containing 2% (wt/vol) glucose and kept on ice. Cell concentrations were determined using a Z2 Coulter counter (Beckman Coulter, Fullerton, CA, USA) set to a detection interval of 1.5 to 5.8 μm. The buffered sample was then diluted in isotone II diluent (Beckman Coulter, Woerden, The Netherlands) to a suspension containing 10⁷ cells ml⁻¹ and stained with PI (Sigma-Aldrich) as described previously (76). The stained samples were analyzed on an Accuri C6 flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA), equipped with a 488-nm laser, and detected by the FL-3 channel (620-nm band pass filter) for PI staining. Per sample, 30,000 events (cells) were analyzed. The viability was determined using Flowing software, version 2.5.1 (Perrtu Terho, Turku Centre for Biotechnology, University of Turku, Finland), by subtracting the percentage of PI-stained cells from a starting value of 100%. For determination of cell viability based on CFU counts, cultures were sampled into Na-HEPES buffer as described above and analyzed on a BD FACSAria II SORP cell sorter (BD Biosciences, Franklin Lakes, NJ), equipped with a 70-μm nozzle and operated with filtered FACSFlow sheath fluid (BD Biosciences). Evaluation of cytometer performance, analysis of cell morphology, and cell sorting were essentially performed as described previously (77). Gating of cell populations for CFU count determination by plating was performed so that typically more than 90% of all detected events (cells) would be sorted. Viability was determined as the average percentage of sorted cells able to form a colony after 3 days of incubation at 30°C on quintuplicate YPD plates (96 cells sorted per plate).

Calculation of growth rate dependency of maintenance energy requirements.For chemostat cultures, specific growth rate (μ) and biomass-specific glucose uptake rate (q_S) were calculated by solving biomass and substrate mass balances assuming steady-state conditions, and least squares linear regression was used to estimate maintenance energy substrate requirements (m_S; intercept with y-axis) and theoretical maximum biomass yield (Y_X/S^max; reciprocal of slope) coefficients (26) from q_S/μ relationships. Calculations for retentostat cultures were performed essentially as described by C. Rebnegger et al. (29): pairs of μ and q_S values were calculated from biomass accumulation between adjacent sampling points, and m_S values were estimated via least squares linear regression from moving windows of continuous pairs of calculated μ and q_S values (including from chemostat cultivations), with the exception that moving windows of 5 q_S-μ pairs were used for m_S estimation.

Whole-genome sequencing and stability of NUBM disruption.Genomic DNA of CBS11895 and IMD010 was isolated using a Qiagen 100/G kit (Qiagen, Hilden, Germany) from a shake flask culture grown in YPD medium to stationary phase, according to the manufacturer’s instructions. DNA concentrations were quantified using a Qubit fluorometer, version 2.0 (Thermo Fisher Scientific). CBS11895 was sequenced by Novogene Bioinformatics Technology Co., Ltd. (Yuen Long, Hong Kong) on a HiSeq 2500 instrument (Illumina, San Diego, CA, USA) with 150-bp paired-end reads using a True-seq PCR-free library preparation (Illumina). IMD010 was sequenced on a MiSeq instrument (Illumina) using a TruSeq DNA PCR-free library preparation as described previously (78).

In order to verify the genetic stability of the nubm^G445GC disruption in strain IMD010 during the prolonged glucose-limited cultivations, a minimum of four single-colony isolates from each individual chemostat and retentostat cultivation with IMD010 was tested for the presence of the mutation. To this end, the cultures were plated for single colonies on solid YPD medium at the last sampling point of each fermentation, their genomic DNA was isolated, and primers 12200 and 12201 were used to PCR amplify and Sanger sequence the site containing the nubm^G445GC disruption, as described above. The nubm^G445GC genotype was still present in all investigated colonies, and no additional mutations were detected within any of the sequenced 688-bp amplicons.

RNA extraction, RNA sequencing, and transcriptome data analysis.Sampling for transcriptome analysis was performed by quenching culture broth directly into liquid nitrogen to immediately stop mRNA turnover (79), followed by storage at –80°C. In the case of batch cultures, sampling for transcriptome analysis was done in mid-exponential phase at a biomass dry weight concentration of approximately 0.9 g liter⁻¹ with >75% of the initial glucose concentration remaining in the reactor. Processing of samples for long-term storage using AE buffer, acid-phenol-chloroform-isoamyl alcohol (125:24:1, pH 4.5; Thermo Fisher Scientific), and sodium dodecyl sulfate, as well as total RNA isolation was performed as described previously (77). The quality of the total extracted RNA was evaluated with an Agilent 2200 Tapestation (Agilent Technologies, Santa Clara, CA), and the RNA concentration was determined using a Qubit 2.0 fluorometer (Thermo Fisher Scientific) combined with a Qubit RNA BR (broad-range) assay kit (Thermo Fisher Scientific). Library preparation and RNA sequencing were performed by Novogene Bioinformatics Technology Co., Ltd. (Yuen Long, Hong Kong). Sequencing was done with an Illumina paired-end 150-bp sequencing read system (PE150) using a 250- to 300-bp insert strand-specific library which was prepared by Novogene. For the library preparation, mRNA enrichment was done using oligo(dT) beads. After random fragmentation of the mRNA, cDNA was synthetized from the mRNA using random hexamer primers. Afterwards, second-strand synthesis was done by addition of a custom second-strand synthesis buffer (Illumina), deoxynucleoside triphosphates (dNTPs), RNase H, and DNA polymerase I. Finally, after terminal repair, A ligation, and adaptor ligation, the double-stranded cDNA library was finalized by size selection and PCR enrichment.

The sequencing data for the samples obtained by Novogene had an average read depth of 21 million reads per sample. For each sample, reads were aligned to the genome of CBS11895 (DL-1) RefSeq assembly (NCBI accession number GCF_000187245.1) (67) with the two-pass STAR procedure (80). In the first pass, a splice junction database was assembled which was used to inform the second round of alignments. Introns were allowed to be between 15 and 4,000 bp, and soft clipping was disabled to prevent low-quality reads from being spuriously aligned. Ambiguously mapped reads were removed. Expression was quantified per transcript using HTSeq count in union intersection mode (81). To exclude from the analysis genes expressed at low levels, genes with an average fragments per kilobase per million (FPKM) value below 10 in all samples were removed. Counts were normalized by TMM normalization using the edgeR package (82), and subsequently differentially expressed genes were determined with an absolute log₂ fold change of >2 and a false-discovery rate of < 0.001. Mean normalization of transcript data was performed per gene and separately for strains CBS11895 and IMD010 using the TMM-normalized FPKM values and was done either including (for comparative expression analysis) or excluding (for analysis of growth rate-correlated gene expression) data from the batch fermentations. For identification of growth rate-correlated gene clusters, analysis of the mean-normalized transcript values versus specific growth rate was performed using the maSigPro R package (83, 84). Genes with a trend significantly different from the mean were selected with a Benjamini-Hochberg corrected P value of <0.1, and subsequently the regression parameters for two clusters of genes were identified with a significance cutoff of 0.05 and an R² of >0.8. For determination of enriched Gene Ontology (GO) terms in growth rate-correlated gene clusters and sets of differentially expressed genes, the online generic GO Term Finder tool (http://go.princeton.edu/cgi-bin/GOTermFinder) and Saccharomyces Genome Database annotation were used. A cutoff of 0.01 was used for the corrected P value (Bonferroni correction), and a background list was provided containing all O. parapolymorpha CBS11895 protein-coding genes for which S. cerevisiae S288C orthologs could be identified (3,094 out of 5,325; obtained using the Orthologous Matrix Database [85]). The number of GO terms was reduced using REVIGO with an allowed similarity setting of 0.5 (86) (see Data Set S1 in the supplemental material for all identified GO terms).

Proteome processing and data analysis.For proteome sampling, a culture sample equivalent to 2 to 4 mg of biomass dry weight was sampled into precooled microcentrifuge tubes, pelleted by centrifugation at 4°C at 4,700 × g for 5 min, and washed with 1.5 ml of ice-cold 1× phosphate-buffered saline (Sigma-Aldrich). After an additional centrifugation step under identical conditions and subsequent removal of the phosphate-buffered saline, cell pellets were stored at –80°C until further processing. In the case of batch cultures, sampling for proteome analysis was done in mid-exponential phase at a biomass dry weight concentration of approximately 0.9 g liter⁻¹ with >75% of the initial glucose concentration remaining in the reactor. To process samples for analysis, cell mass was normalized to a dry weight of 1.6 mg and mechanically lysed using 0.5-mm zirconium beads and a PreCellys homogenizer. Proteins were isolated using Bligh and Dyer extraction (87), followed by reduction, alkylation, and digestion using trypsin. Samples were analyzed in technical triplicates by liquid chromatography tandem mass spectrometry (LC-MS/MS) using a Vanquish UHPLC coupled to a Q Exactive Plus Orbitrap MS (Thermo Fisher Scientific). Peptides were separated using reverse-phase chromatography using a gradient of water with 0.1% formic acid (solvent A) and acetonitrile with 0.1% formic acid (solvent B) from 2% B to 45% B in 50 min. Data-dependent acquisition (DDA) was performed with a resolution setting at 70,000 within the 400- to 1,600-m/z range and a maximum injection time of 75 ms, followed by high-energy collision-induced dissociation-activated (HCD) MS/MS on the top 15 most abundant precursors using a resolution setting of 17,500 and a 200- to 2,000-m/z range with a maximum injection time of 50 ms. The minimum intensity threshold for MS/MS was 1,000 counts, and peptide species with 1 and >8 charges were excluded. MS/MS spectra were analyzed with the SEQUEST HT search engine and Proteome Discoverer, version 2.3, against the proteins of the CBS11895 (DL-1) RefSeq assembly (NCBI accession no. GCF_000187245.1) (67). Label-free quantification was performed using the top three unique peptides measured for each protein. Retention time alignment was performed on the most abundant signals obtained from nonmodified peptides measured in all samples, and results were corrected for the total ion intensities measured for each sample. For subsequent analysis, only proteins were taken along that achieved a combined detection confidence with an FDR of <1% and additionally were individually detected with an FDR of <1% in at least 5 out of the total 48 LC-MS analyses (6 per condition). For proteins that passed these requirements, protein abundance was set to 0 for individual analyses that did not exhibit a detection confidence with an FDR of <1%, and the average abundance of all analyses per condition was used for further calculation. Proteins were considered not detected for a specific condition if they were not measured at least once with a detection confidence of an FDR of <1% for that condition. Mean normalization of the protein data was performed per gene and separately for strains CBS11895 and IMD010 using the total ion intensity-normalized protein abundances.

Data availability.Transcript abundances, lists of differentially expressed genes, sets of growth rate-correlated genes, identified S. cerevisiae orthologs of O. parapolymorpha protein-coding genes, complete lists of enriched GO terms, and total ion intensity-normalized protein abundances are available in Data Set S1 in the supplemental material.

Genome sequencing data of CBS11895 and IMD010 are available at NCBI (https://www.ncbi.nlm.nih.gov/) under BioProject accession number PRJNA588376. RNA-seq data are available at NCBI (https://www.ncbi.nlm.nih.gov/) under Gene Expression Omnibus (GEO) accession number GSE140480. Raw proteomics data are available on figshare (https://doi.org/10.6084/m9.figshare.11398773) (88).