Transcriptomic studies on the product stress response revealed that YCF1 is a beneficial factor for progesterone production in Yarrowia lipolytica.

Ying Wang: Writing – review & editing, Supervision, Project administration, Investigation, Funding acquisition, Conceptualization. Ruosi Zhang: Formal analysis, Data curation. Mingdong Yao: Writing – review & editing, Supervision. Wenhai Xiao: Writing – review & editing, Supervision, Project administration, Investigation, Funding acquisition, Conceptualization. Ying Wang: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Ying-Jin Yuan: Writing – review & editing, Supervision.

The biomass of Y. lipolytica plateaus with the accumulation of progesterone, and its environmental resistance also increases accordingly. This led us to consider whether steroids might induce a product stress response. To examine the mechanisms of the progesterone-induced stress response in Y. lipolytica , omics techniques were used to identify genes whose transcription levels changed after progesterone exposure. As shown in Fig. 2 A, the number and intensity of gene responses peaked 30 min after induction, and the majority of differentially expressed genes were restored to their baseline levels 90 min later. Similar to the transcriptome results of the progesterone response of Y. lipolytica in this study, that of C . albicans and S. cerevisiae under stimulation with progesterone at supraphysiological concentrations was extensive, rapid, transient, and reversible and was most intense 30 min after stimulation. This finding indicates that the time-dimension response changes of the three yeasts under progesterone stimulation are relatively conservative [ 34 ]. However, there were 389 genes whose expression differed during the stable induction period (90 min), indicating that progesterone stimulation may have a long-term effect on cell physiology. Fig. 2 The response of Y. lipolytica to progesterone stress is reflected mainly in amino acid metabolism, lipid metabolism, and transport. A. Profile of differential gene expression under progesterone stimulation. B. Functional categories of KEGG (left)- and GO (right)-annotated genes that were differentially expressed in response to 15 min of progesterone treatment. C. The highly enriched functional categories of the KEGG-annotated genes found to be induced and repressed after 15 min of progesterone treatment. Fig. 2 The response of Y. lipolytica to progesterone stress is reflected mainly in amino acid metabolism, lipid metabolism, and transport. A. Profile of differential gene expression under progesterone stimulation. B. Functional categories of KEGG (left)- and GO (right)-annotated genes that were differentially expressed in response to 15 min of progesterone treatment. C. The highly enriched functional categories of the KEGG-annotated genes found to be induced and repressed after 15 min of progesterone treatment. In the microbial stress response, the signal transduction system senses stress signals and activates self-defense mechanisms. Cells preferentially synthesize specific proteins to repair damage, ensure their own activity, and then regulate the contents of stress-responsive metabolites to cope with stress. Moreover, this process is rapid and transient. Therefore, in this study, 1782 DEGs at the early stage (15 min) of P4 induction, to which the strain was sensitive, were selected for annotation and classification. According to the KEGG pathway and GO annotation ( Fig. 2 B) analyses, the DEGs were enriched mainly in processes such as transport and catabolism. Analysis of the number of induced and repressed differentially expressed genes in the top 7 KEGG categories with the highest gene enrichment ( Fig. 2 C) indicated that within the first 30 min after progesterone exposure, all selected KEGG categories presented both upregulated and downregulated genes. Among these genes, genes in more than five categories were predominantly upregulated, whereas the majority of DEGs in the transcription and translation categories were downregulated. Ninety minutes after induction, the number of responsive genes in the 7 classification groups markedly decreased, with the differentially expressed genes primarily being upregulated. Progesterone stimulation triggered widespread decreases in cellular transcription and translation levels and upregulation of genes related to signal transduction, transport and catabolism, as well as lipid metabolism under stress, and during the induced stable phase, the expression of these genes returned to levels close to the baseline levels. Compared with the progesterone stress responses of S . cerevisiae and C . albicans [ 35 , 36 ], the downregulation of transcription and translation metabolism is partially conserved in Y. lipolytica ; the overall upregulation of genes related to signal transduction, transport and catabolism, and lipid metabolism in response to progesterone stimulation in yeast is relatively conserved; and the overall upregulation of carbohydrate metabolism and amino acid metabolism serves as a complement to the steroid hormone stimulation response behavior observed in yeast. From the perspective of differential gene expression, the results of this study suggest that Y. lipolytica exhibits stress responses after progesterone exposure. Accumulating evidence indicates that progesterone has dual effects on microbial phenotypes. On the one hand, progesterone temporarily inhibits cell growth; on the other hand, progesterone stress activates a series of protective mechanisms that help cells overcome growth inhibition and increase resistance to various environmental stresses, with different modes of action [ 37 , 38 ]. Considering that fungi need to balance cellular stress resistance and growth rates under stressful conditions, the extensive gene response that occurs in Y. lipolytica under simple progesterone stress may be accompanied by inhibited cell growth. Transcriptome analysis revealed that the metabolism and cellular processes of yeast underwent reorganization in response to progesterone accumulation. Manipulating the restructured metabolic pathways or cellular processes to improve the biosynthesis of progesterone may lead to unexpected results. Irrespectively of the treatment's duration, comparative transcriptomic analysis revealed that the expression of 246 genes was upregulated or downregulated ( Fig. 3 A). GO enrichment analysis of these 246 DEGs revealed that they were related primarily to metabolic processes and catabolic processes ( Fig. 3 B), which is consistent with the global gene response classification. Eighteen genes whose expression was upregulated at each time point (log2(fold change) ≥ 1.0, Table S5 ) were screened as candidate genes, and most candidate genes were involved in multiple metabolic processes. These candidates were classified into five groups on the basis of their main functions ( Table 2 ). Fig. 3 C shows that under progesterone stress, differentially expressed genes related to sterol esterification, transport proteins, fatty acid synthesis-degradation pathways, and the TCA cycle pathway are generally upregulated. Fig. 3 Identification and overexpression of candidate genes. A. A Venn diagram. The circles represent the genes whose expression was upregulated or downregulated during the response period, and the numbers represent the number of DEGs. The overlapping regions represent differentially expressed genes that were present during two or three response periods. B. Combination of a Sankey diagram and a bubble chart. Functional enrichment of genes induced after 15 min of progesterone treatment and cellular processes associated with specific candidate genes. C. Transcriptional profiles of sterol synthesis, amino acid metabolism, lipid metabolism and transport under progesterone treatment. D. Effects of the overexpression of candidate genes on biomass accumulation and progesterone production. E. Effects of the overexpression of candidate genes on precursor production. Fig. 3 Table 2 Classification, ID, name and function of candidate genes. Table 2 Classification Gene ID Gene Name Function Transporter activity 2908366 ATP-binding cassette transporter SNQ2 ABC drug resistance transporter, involved in a variety of azole drug reactions 2912709 YALI0E25179g Transmembrane transporter activity 2912325 Glycerol uptake facilitator protein (glpF) The diffusion of glycerol across the cytoplasmic membrane is mediated by a pore-type mechanism, with high permeability to glycerol Lipid metabolism 2909298 Acyl-CoA dehydrogenase (APDG) Involved in the process of β-oxidation of fatty acids 2911816 Oxalate---CoA ligase (AAE3) Catalyzes oxalate degradation to produce CO 2 , which protects cells from oxalate damage 2910372 Lipid-transfer protein May play a role in regulating steroidogenesis 2912831 Long-chain fatty acid-CoA ligases(ACSLs) Convert free long-chain FA into acyl-CoAs, which are key intermediates in both anabolic and catabolic pathways 2911520 Acetyl-CoA acetyltransferase (atoB) Catalyzes the conversion of acetyl-CoA to acetylacetyl-CoA, which is involved in the β oxidation of fatty acids Amino acid metabolism 2912696 Lactoylglutathione lyase (gloA) Involved in methylglyoxal detoxification 2912887 Citrate synthase (Cs) Citrate synthase, which catalyzes the production of citric acid from oxaloacetate and acetyl-CoA, is involved in the TCA cycle 2911154 Malate synthase (glcB) Malic acid synthase, involved in the TCA cycle 2908832 Methylglutaconyl-CoA hydratase(AUH) Involved in amino acid degradation pathways 2912486 Tryptophan aminotransferase (Tam1) Has scavenger receptor activity, binds to any modified low-density lipoprotein (LDL) or other polyanionic ligand, and delivers the ligand into the cell via endocytosis Glucose metabolism 2909987 Carotenoid dioxygenase(cao-2) Carotenoid dioxygenase, involved in the synthesis of xanthin in Neurospora 2908079 Sterol O-acyltransferase 2(SOAT2) Catalyzes the formation of sterol esters from long-chain fatty acyl-CoA with sterols Others 2912943 Mucin-19 (MUC19) A mucin that helps in the formation of a mucus barrier and has a protective effect against infection 2907331 nuclear GTP-binding protein (NUG1) Involved in purine nucleotide metabolism; GTPase required for 60S ribosomal subunit export to the cytoplasm 2911552 Protein STE50 It participates in growth arrest during the coupling process and plays an important role in signaling interactions and signal transmission for multiple MAPK-mediated communication pathways Identification and overexpression of candidate genes. A. A Venn diagram. The circles represent the genes whose expression was upregulated or downregulated during the response period, and the numbers represent the number of DEGs. The overlapping regions represent differentially expressed genes that were present during two or three response periods. B. Combination of a Sankey diagram and a bubble chart. Functional enrichment of genes induced after 15 min of progesterone treatment and cellular processes associated with specific candidate genes. C. Transcriptional profiles of sterol synthesis, amino acid metabolism, lipid metabolism and transport under progesterone treatment. D. Effects of the overexpression of candidate genes on biomass accumulation and progesterone production. E. Effects of the overexpression of candidate genes on precursor production. Classification, ID, name and function of candidate genes. In terms of specific defense mechanisms against progesterone pressure, we hypothesized that the upregulation of steroid esterification genes to some extent helps alleviate the excessive concentration of membrane steroid analogs caused by the progesterone that has entered the membrane system [ 39 ]. A significant increase in fatty acid synthesis and catabolism can accelerate membrane damage repair and the renewal of membrane components [ 40 , 41 ], allowing exogenous progesterone in the membrane system to be converted into less toxic compounds via intracellular processes. The significant upregulation of factors related to the TCA cycle indicates that progesterone stress may trigger the activation of metabolic pathways related to acetyl-CoA. In particular, the upregulation of factors related to the TCA cycle provides energy for ATP-dependent transport proteins involved in drug efflux and metabolite transport within the cell membrane system, maintaining the physiological balance of the cell [ 42 ]. Under stressful conditions, microorganisms tend to prioritize allocating resources for stress resistance over cell growth to maintain viability [ 43 ]. Considering the important role of the upregulation of factors related to the TCA cycle and fatty acid metabolism in defense responses, we hypothesized that, under progesterone stress, when cellular defense mechanisms are highly upregulated, the overall upregulation of the TCA cycle and lipid metabolic processes provides raw materials and energy for product synthesis [ 44 ]. Moreover, the physiological activities involved in the upregulation of these processes may focus primarily on stress resistance, with a relatively limited ability to promote cell growth. To investigate whether altered metabolic or cellular processes are beneficial for progesterone production, the 18 candidate genes in Table 2 were overexpressed in a progesterone-producing strain that utilizes glucose to produce campesterol, which is then transformed into progesterone via pregnenolone. The introduction of these candidate genes can help strains grow to some extent ( Fig. 3 D). Four genes have been demonstrated to be more effective at promoting progesterone production, especially two genes (Gene IDs: 2912325 and 2908366) encoding the glycerol uptake facilitator protein (glpF) and transport protein SNQ2, which improved production by 29.2 % and 51.7 %, respectively ( Fig. 3 D). Protein alignment revealed that 2908366 is homologous to SNQ2p and Aus1p, the proteins of S . cerevisiae , both of which are related to sterol transport [ 45 , 46 ]. Compared with the control strain, strain 325 (SyBE_Yl08090003, Table S1 ) accumulated more pregnenolone, while the amount of accumulated campesterol decreased. The precursor of strain 366 (SyBE_Yl08090001, Table S1 ) showed the opposite trend ( Fig. 3 E). The two enzymes that catalyze the synthesis of progesterone from campesterol in different strains may result in varied growth across environmental conditions, resulting in different catalytic capacities and, as a result, changes in precursor accumulation. With respect to biochemical production in microbial cells, cellular transporter systems can provide a direct mechanism for reducing product stress [ 47 ]. In this study, transporters were also shown to play a significant role in the progesterone stress response and progesterone excretion. ABC transporters (ATP-binding cassette transporters) are ubiquitously found in all species from prokaryotes to humans and are among the most prevalent families of transporters in living creatures. These transporters depend on ATP hydrolysis to directly stimulate the substrate to bind to the binding protein (SBP) to achieve transport [ 48 ]. The sterol transporters Pdr5p and Ycf1p, which reportedly export various toxic compounds [ 49 ] and thereby confer multidrug resistance to yeast cells, as well as Candida tropicalis ABCE transport proteins, were selected, and a BLAST protein comparison was conducted to identify endogenous transport proteins in Y. lipolytica . The comparison results are shown in Table 3 . Compared with the previously reported transporter ABC2 of Y. lipolytica [ 50 ], the investigated proteins were overexpressed in progesterone-producing strains. To characterize the function of the gene, strain 366 was renamed SNQ2. Table 3 Endogenous transporter proteins employed in this study. Table 3 Gene ID Protein name Protein identity 2908366 SNQ2p/AUS1p 93 % 2906684 Pdr5p 96 % 2912704 Ycf1p 99 % 2911636 ABC2 100 % 2906445 ABCE 99 % Endogenous transporter proteins employed in this study. Compared with the control, the overexpression of endogenous transporters did not significantly affect the final growth of the cells ( Fig. 4 A). In addition to the ABC2 transporter of Y. lipolytica , the upregulation of the other three transporters, Pdr5, Ycf1, and ABCE, increased progesterone production by 21.96 %, 37.94 %, and 30.88 %, respectively. Furthermore, the production of this transporter increased the export of the precursor campesterol into the extracellular space ( Fig. 4 B). The transporter ABC2 in yeast did not show an advantage in progesterone production, which may be due to the accumulation of the precursor pregnenolone. The conversion of pregnenolone to progesterone is regarded as the rate-limiting step [ 21 , 51 ]. Owing to the complex stress response mechanism of microorganisms, the influence of heterologous protein expression on strains is difficult to predict, and sometimes, it can cause negative effects [ 52 ]. Although the expression of endogenous proteins promoted progesterone accumulation to some extent, the effect was relatively small, and there was no further enhancement compared with the effect of the initially discovered transport protein SNQ2. Fig. 4 Effects of the overexpression of endogenous transporter genes on biomass accumulation and the production of progesterone and its precursors. A. Final OD 600 and progesterone production of strains overexpressing heterologous transporters. B. Intracellular and extracellular campesterol production in heterologous transporter-overexpressing strains. The percentage is the proportion of extracellular campesterol. (The error bars for all the data represent the standard deviations calculated from triplicate experiments. Significance levels of Student's t -test: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.) Fig. 4 Effects of the overexpression of endogenous transporter genes on biomass accumulation and the production of progesterone and its precursors. A. Final OD 600 and progesterone production of strains overexpressing heterologous transporters. B. Intracellular and extracellular campesterol production in heterologous transporter-overexpressing strains. The percentage is the proportion of extracellular campesterol. (The error bars for all the data represent the standard deviations calculated from triplicate experiments. Significance levels of Student's t -test: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.) To investigate the amplification effect of transport proteins, the advantageous targets identified in section 3.2 , the SNQ2 gene was integrated via multicopy integration, or the Ycf1 gene was introduced into the chassis strains simultaneously. However, compared with that of the control strain, the accumulation of de novo synthesized progesterone did not improve. In addition, we applied the target SNQ2 to the biocatalytic process of converting pregnenolone to progesterone, aiming to obtain more progesterone. During substrate feeding, the efficiency of the conversion of pregnenolone to progesterone even decreased ( Fig. S4 ). The overexpression of transport proteins may place an extra burden on cells, or it may disrupt other metabolic pathways, thereby affecting progesterone production [ 53 ]. There may be other limiting factors in the progesterone biosynthesis pathway. To determine whether the introduction of heterologous transport proteins can further improve the stress response and progesterone product accumulation, a phylogenetic analysis of the transporters we identified was conducted ( Fig. 5 A) to select candidates. The results of the phylogenetic tree indicate that the correlation between transporters and species evolution is relatively weak. Five proteins, Pdr5p and Ycf1p from S . cerevisiae , ABCE from Arabidopsis thaliana , and ABCG1 and ABCC1 from Homo sapiens , were chosen for their possible involvement in sterol transport ( Table 4 ). Fig. 5 Identification and overexpression of exogenous transporters. A. The general phylogenetic topology of heterologous transporters. B. Final OD 600 and progesterone production of strains overexpressing heterologous transporters. C. Intracellular (purple) and extracellular (cherry red) campesterol production in heterologous transporter-overexpressing strains. The percentage is the proportion of extracellular campesterol. D. The effect of the transporter identified in this study on different progesterone-producing strains. E. Plane diagram of the molecular docking of progesterone with transporters. Fig. 5 Table 4 Research progress related to strain tolerance based on transporter engineering. Table 4 Protein Name Source of Protein Gene ID in Yl Substrate Reference SNQ2p S. cerevisiae 2908366 Estradiol, phospholipid [ 60 ] Pdr5p S. cerevisiae 2906684 Progesterone, estradiol [ 45 , 61 ] Ycf1p S. cerevisiae 2912704 Testosterone, progesterone [ 49 , 62 ] ABC2 Y . lipolytica 2911636 Decane [ 50 ] ABCG1 H . sapiens 2910127 Cholesterol [ 63 ] ABCG2 H . sapiens , C . familiaris 2910127 Estrone-3-sulfate, Geraniol [ 64 ] ABCC1 H . sapiens 2912704 Glucocorticoid [ 65 ] ABCE A . thaliana 2906445 _ b [ 66 ] b Substrate not reported. Identification and overexpression of exogenous transporters. A. The general phylogenetic topology of heterologous transporters. B. Final OD 600 and progesterone production of strains overexpressing heterologous transporters. C. Intracellular (purple) and extracellular (cherry red) campesterol production in heterologous transporter-overexpressing strains. The percentage is the proportion of extracellular campesterol. D. The effect of the transporter identified in this study on different progesterone-producing strains. E. Plane diagram of the molecular docking of progesterone with transporters. Research progress related to strain tolerance based on transporter engineering. b Substrate not reported. These heterologous transport proteins were successfully overexpressed in the progesterone-producing strains. We attempted to demonstrate the impact of these proteins through the final cell density and progesterone accumulation in the strains. The growth impact assays demonstrated that the strains presented specific growth benefits compared with the control strain and outperformed the endogenous transporter SNQ2 ( Fig. 5 B). Moreover, these transporters increased the accumulation of progesterone in Y. lipolytica . Two of the most advantageous proteins among them are Sc _Ycf1p and At _ABCE. YCF1p was associated with a statistically significant increase in performance, resulting in a 69.6 % increase in progesterone accumulation. The expression of heterologous transport proteins considerably stimulates the production of the precursor campesterol, and the expression of Sc _Ycf1 also increases the extracellular content of this precursor ( Fig. 5 C). In addition, the intermediate pregnenolone ketone accumulated in most strains, which may suggest that the efflux of progesterone further promotes this pathway. To assess the effect of the transporter identified in this study, the Sc _Ycf1 gene were introduced into the previously developed progesterone-producing strain SyBE_Yl2092067 in our laboratory. The genotype of SyBE_Yl2092067 was listed in Table S1 . Meanwhile, pregnenolone was added at a final concentration of 50 mg/l during the logarithmic growth phase of the strain (24 h of fermentation) to verify the effect of the transporter proteins on high-yield progesterone producing strain. The results ( Fig. 5 D) indicated that the transporter proteins discovered in this study remain effective under high progesterone conditions. To examine the differences between distinct transport proteins and the accumulation of progesterone synthesis proteins, the binding ability between progesterone and transporters was predicted by simulating molecular docking with Maestro ( Fig. 5 E and Table 5 ). The increase in progesterone titer caused by the overexpression of the YCF1 protein of S . cerevisiae and the ABCE protein of A . thaliana may be due to the number of covalent bonds between the substrate and proteins. The specificity and affinity of transport proteins determine the efficiency of substrate transfer [ 54 ]. Therefore, protein engineering [ 55 ] can be applied to adjust the substrate binding specificity and affinity of transport proteins, enabling them to adapt to new substrates and improving the ability of cells to expel specific substances. Table 5 Predicted binding energy for progesterone and transporters via molecular docking. Table 5 Transporter MMGBSA ΔG Bind 2908366 −39.63 Sc _YCF1 −41.35 At _ABCE −44.10 Predicted binding energy for progesterone and transporters via molecular docking.

The Escherichia coli strain DH5α was used for plasmid construction and amplification. E. coli strains were grown in complete Luria–Bertani (LB) media (10 g/L NaCl, 10 g/L tryptone, and 5 g/L yeast extract), with 20 g/L agar added to prepare solid media. Ampicillin or kanamycin (100 mg/L) was added to LB when necessary. The E. coli strains were cultured at 37 °C and 200 rpm. All the Y. lipolytica strains used in this study were derived from ATCC201249 and purchased from the American Type Culture Collection ( Table S1 ). YPD medium was used for the cultivation of Y. lipolytica strains. YPD medium supplemented with nourseothricin sulfate was used for transformant selection. Specifically, recombinant cells were inoculated into 5 mL of 2 % YPD medium (20 g/L glucose, 20 g/L tryptone, and 10 g/L yeast extract) and grown at 28 °C and 220 rpm for 12–24 h until they reached the exponential phase. The precultures were subsequently inoculated into 5 mL of 2 % YPD with an initial OD 600 of 0.2 for an additional 18 h of incubation. The seed culture was subsequently inoculated into 50 mL of 5 % YPD medium to attain an initial OD 600 of 0.1, followed by incubation at 28 °C with shaking at 220 rpm for 192 h [ 26 ]. The primers used for plasmid construction are listed in Table S2 and were synthesized by TsingKe (Beijing, China). The vectors used for building the gene expression cassettes contained one specific promoter and one terminator with two back-to-back Bsa I sites in between. The IntF (YALI0F3161413-3162449) locus was chosen for single-copy gene integration. rDNA loci were chosen for multicopy gene integration. The genes encoding heterologous ABCE, ABCG1, and ABCC1 were optimized and synthesized by TsingKe (Beijing, China; Table S3 ). Endogenous genes were amplified from the genome of Y. lipolytica via PCR. Gene DNA fragments containing 20 bp homologous arm segments were obtained via polymerase chain reaction (PCR) via Phanta Max Super-Fidelity DNA Polymerase purchased from Vazyme (Nanjing, China). The configuration of the PCR amplification system (100 μl) was as follows: 20 μL of 5 × Phanta polymerase buffer, 2 μL of 10 mM dNTPs, 2 μL of Phanta polymerase, 4 μL of 10 μM forward primer mixture, 4 μL of 10 μM reverse primer mixture, 2 μL of DNA template mixture, and 66 μL of ultrapure water. The above components were mixed to obtain a 100 μL system, and 50 μL was aliquoted into PCR tubes. The PCR program was as follows: predenaturation at 95 °C for 2 min; 35 cycles of denaturation at 95 °C for 30 s, annealing at 55–68 °C (based on the primer Tm value) for 30 s, and extension at 72 °C for 30 s/1 kb; and extension at 72 °C for 5 min. The PCR products were inserted into the digested vector via an In-Fusion reaction ( Fig. S3 ) [ 27 ]. The restriction enzymes used in this study were purchased from New England Biolabs (Beijing, China). Gibson assembly [ 28 ] was used for plasmid construction. All the plasmids used in this study are shown in Table S4 . Genetic transformation of the plasmids into Y. lipolytica strains was conducted with the Frozen-FZ Yeast Transformation IITM Kit (Zymo Research, USA). Y. lipolytica SyBE_Yl2091025 [ 21 ] was employed as the initial strain (Con 01). The engineered strains were selected on YPD media supplemented with nourseothricin sulfate. The clones were verified via PCR with the primers listed in Table S2 . To avoid the impact of errors caused by the byproducts of the progesterone-producing strain itself, the chassis strain ATCC201249 was chosen as the starting strain for transcriptome analysis. ATCC201249 was inoculated with 50 mL of 2 % glucose YPD at an initial concentration of OD600 = 0.1 and fermented at 220 rpm at 28 °C for 12 h (to the mid-logarithmic growth stage). For both Candida albicans and S. cerevisiae , the number of genes whose expression changed significantly and reproducibly was greatest at the 10 −4  M dose. In the experimental group, the steroid substrate progesterone was added at a final concentration of 105 mg/L (3.3 × 10 −4  M; the progesterone was first ultrasonically dissolved in ethanol, filter sterilized, and then mixed with an equal volume of sterilized Tween-80 (Solarbio, Beijing, China).) [ 29 ], and in the control group, the same amount of steroid mother liquor solvent (50 % ethanol-Tween 80) was added. Progesterone stress experiments on other yeasts [ 30 ] have shown that the steroid response was fast and transient, peaking at 30 min for both C . albicans and S. cerevisiae . Therefore, in this study, 15 min, 30 min, and 90 min were selected. After incubation in a shaker three times, approximately 200 mg of cells was collected by two rounds of centrifugation and resuspended in ultrapure water. After centrifugation, the yeast was collected and quickly frozen in liquid nitrogen. The procedure used for RNA-Seq was as follows: the total RNA of the sample to be tested was extracted according to the instructions of the yeast total RNA extraction kit ( PureLink™ Pro 96 , purchased from Thermo Fisher Scientific, CA, USA), and the mRNA was enriched and purified. cDNA was obtained via reverse transcription of the fragmented RNA. The loop single-strand DNA library was obtained via double-strand complementation, amplification, denaturation and cyclization with bridge primers. The samples were sequenced, and their quality was determined via an Agilent 2100 Fragment Analyzer (Agilent, CA, USA). After the data were filtered, RSEM software and the FPKM method were used to calculate the relative gene expression. GCF_000002525.2_ASM252v1 from the NCBI was used as the reference genome of the yeast Y . lipolytica , and the data were further analyzed via an interactive reporting system from BGI, Dr. Tom ( http://report.bgi.com/ps ). Between-group differential gene analysis was performed using DEseq2 [ 31 ]/DEGSeq [ 32 ] under the conditions of a fold change ≥2 and adjusted P value ≤ 0.001 (depending on the project). PoissonDis was used for between-sample differential gene analysis under the conditions of a fold change ≥2 and an FDR ≤0.001 (depending on the project). The pheatmap function was used on the differential gene set to draw a heatmap of the differential gene clusters. According to the GO and KEGG annotation results and official classifications, the DEGs were functionally classified, the phyper function in R software was used for KEGG enrichment analysis, and the TermFinder package was used for GO enrichment analysis. With a Q value of ≤0.05 as the threshold, candidate genes that met this condition were defined as being significantly enriched. Sample extraction and steroid analysis were performed according to Zhang et al. [ 21 ] with modifications. A 1 mL culture was centrifuged at 12000 rpm for 2 min, and the supernatant was collected and suspended in 1.5 mL of acetic ether for 10 min and mixed by vortexing to extract the extracellular progesterone. The ethyl acetate phase was collected by centrifugation at 12000 rpm for 2 min and then dried. In addition, the harvested cells were washed twice with 1 mL of Milli-Q double deionized water, suspended in 1 mL of 3 M HCl and boiled for 10 min. Afterward, the cell samples were centrifuged at 12000 rpm for 1 min. The cell debris was washed away with 1 mL of Milli-Q double deionized water, and the samples were resuspended in 500 μL of 20 % (w/v) KOH–methanol solution and incubated at 60 °C for 6 h for saponification. Then, 200 μL of n-hexane was added to the tube to extract pregnenolone, which was vortexed for 10 min. After centrifugation, the n-hexane phase was collected and dried by using a centrifugal vacuum evaporator (EYELA CVE-3100, Tokyo, Japan). The derivatization of the dried product was conducted with 200 μL of N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA, SIGMA-Aldrich, USA) at 37 °C for 2 h. Before GC−MS analysis, each sample was diluted with 200 μL of hexane. The GC−MS instrument (SHIMADZU GC-2010 PLUS with a SHIMADZU GCMS-QP2020 detector) was used with a SHIMADZU SH-Rxi-5Sil MS column. The injector temperature was 290 °C. The column temperature was initially maintained at 70 °C for 1 min, then increased to 250 °C at a temperature ramp of 30 °C/min and maintained at 250 °C for 2 min, followed by an increase to 280 °C at a temperature ramp of 20 °C/min, and finally maintained at 280 °C for 20 min. The split ratio was 50:1. The steroid standards used in our study were as follows: campesterol (Solarbio, Shanghai, China), pregnenolone (Aladdin, Shanghai, China), and progesterone (Aladdin, Shanghai, China). For all the data, the error bars represent the standard deviation calculated from triplicate experiments. Significance levels were determined by Student's t -test [ 33 ]: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

In this work, we found that the accumulation of progesterone imposes stress on Y . lipolytica and that yeast responds to this stress through metabolic or cellular process reorganization. Omics studies revealed that the progesterone stress response of Y . lipolytica involves mainly the activation of transport proteins, lipid metabolism, and amino acid metabolism. Among these response targets, transport proteins present significant advantages for progesterone production. ABC transporter-mediated efflux is a common mechanism of pleiotropic drug resistance (PDR) in cells [ 56 ]. The PDR pathway is one of the most common resistance mechanisms in yeast, and it is also the pathway that is most sensitive to the progesterone stress response in S . cerevisiae and C . albicans [ 46 , 57 ]. Steroid hormones can stimulate high expression of a series of PDR genes and induce the activation of a range of membrane-localized ABC transporter family efflux pumps. These findings indicate that progesterone induction in Y. lipolytica can also trigger a conserved efflux transport mechanism similar to that of yeast. Although endogenous transport proteins cannot promote strain growth, they do increase the production of progesterone to some extent. The heterologous transport of proteins in Y. lipolytica can not only increase a strain's resistance to progesterone but also increase progesterone accumulation. Ultimately, the Ycf1 transport protein from S . cerevisiae increased progesterone production by 69.6 %. The expression of transport proteins in this study had a positive effect on only progesterone accumulation, without leading to a significant accumulation of progesterone or the fact that progesterone cannot be used on a large scale. This is largely still limited by the expression of the P450Scc system. Moreover, studies [ 13 ] have shown that the positioning of enzymes within the system directly affects the interaction between the system and the substrate, significantly increasing the difficulty of expressing this pathway. Therefore, greater efforts are needed to address the limitations in enzyme expression. Once the difficulties in pathway expression are resolved, applying the strategies described in this article may lead to unexpected improvements in progesterone accumulation. This study argues that the identification of key gene targets is crucial for enhancing the production of the desired progesterone in engineered microorganisms. Further efforts are needed to achieve multilevel, large-scale identification of beneficial gene targets involved in the progesterone production process.

Progesterone is widely distributed and plays irreplaceable physiological roles in the reproductive system and physiological processes. It is used in cases such as threatened or recurrent abortion during pregnancy, dysfunctional uterine bleeding, infertility due to luteal insufficiency, dysmenorrhea, endometriosis, secondary amenorrhea, menstrual disorders, and premenstrual syndrome and as part of menopausal hormone therapy [ [1] , [2] , [3] ]. On the other hand, progesterone plays an important role in brain function as a neurosteroid [ 4 , 5 ]. Recently, progesterone has also been identified as a potential candidate plant hormone [ 6 ]. As the application of progesterone becomes more widespread, its market value is also increasing. The number of people who need progesterone for luteal support after embryo transfer can reach hundreds of thousands or even more every year. The annual sales volume of contraceptive pills with levonorgestrel as the main ingredient is tens of millions of boxes. With changes in fertility rates, the popularization of assisted reproductive technologies and increased demand for the treatment of gynecological diseases, etc., worldwide, the market demand for progesterone is increasing. Global Market Insight noted that the market size of progesterone in 2023 is valued at $1.3 billion, and it is estimated that it will grow at a CAGR (Compound Annual Growth Rate) of 12.7 % from 2024 to 2032. Currently, the main strategies for progesterone synthesis include extraction from animal raw materials [ 7 ], chemical synthesis [ 8 ], microbiological transformation [ 9 ] and total microbial synthesis. However, extraction is limited by low progesterone concentrations, which contributes to high costs, whereas chemical synthesis is limited by low yields, expensive catalysts, and environmental pollution. Artificial microbial cell factories, as important tools in synthetic biology, provide fast, stable, and sustainable technological approaches for steroid hormone production. In recent decades, research related to the creation of microbial strains capable of producing steroid hormones has been actively developing, which opens up prospects for replacing multistage chemical syntheses with environmentally friendly biotechnologies [ 10 , 11 ]. Through modification of the P450scc oxidase system [ [12] , [13] , [14] , [15] ] and regulation of metabolic flux [ 16 ], the synthesis of pregnenolone/progesterone has been successfully achieved in Saccharomyces cerevisiae , Yarrowia lipolytica , and Mycolicibacterium neoaurum ( Table 1 ). Owing to various advantages, the oleaginous yeast Y . lipolytica is considered a promising host for steroid production, including its native MVA pathway and sufficient supply of acetyl-CoA, NADPH, and ATP [ 17 ]. Owing to the existence of lipid bodies, they can dissolve many hydrophobic or weakly polar substances and provide storage space for the accumulation of hydrophobic products [ 18 ]. Y . lipolytica is also regarded as a generally recognized as safe (GRAS) strain and is widely used in the production of pharmaceuticals and food. Owing to these advantages, Y . lipolytica has become a promising strain for the synthesis of steroids, such as 7-dehydrocholesterol [ 19 ] and campesterol [ 20 ]. Table 1 The progress of progesterone synthesis in different hosts and the engineering strategies and the corresponding yields. Table 1 Host Engineering strategies Corresponding yields Ref. S . cerevisiae Disruption of the Δ22-desaturase gene and introduction of Arabidopsis thaliana Δ7-reductase (DWF5), mature forms of CYP11A1 (P450 side chain cleaving), ADX (FDX1), ADR (FDXR), and 3β-HSD _ a [ 58 ] Y. lipolytica Disruption of the Δ22-desaturase gene and introduction of Xenopus laevis DHCR7, mature forms of Sus scrofa CYP11A1, Bos taurus ADX, ADR, and 3β-HSD ∼5 mg/L [ 29 ] M. neoaurum Mutant CYP11A1 (mCYP11A1) and adrenodoxin reductase (ADR) were connected by a flexible linker (L) and expressed the ADR-related homolog ARH1 45 mg/L [ 59 ] a Corresponding yields not reported. The progress of progesterone synthesis in different hosts and the engineering strategies and the corresponding yields. a Corresponding yields not reported. Our research group successfully obtained progesterone-producing strains [ 21 ] by introducing the P450 system (side chain cleavage cytochrome P450, including CYP11A1 from Sus scrofa , adrenodoxin (AdX) and adrenodoxin reductase (AdR) from Bos taurus ) and 3β-hydroxysteroid dehydrogenase/isomerase into Y . lipolytica with high campesterol production ( Fig. 1 ). Moreover, we successfully synthesized 4AD in Y . lipolytica by coculturing progesterone-producing strain with another strain expressing CYP17A1-CYB5. In the mixed bacterial system used in this study, the upstream module was the initial synthesis site for progesterone. Under different cultivation conditions, intense growth competition occurred among the components of the microbial community before the total biomass of the community reached a stable phase. When the microbial population entered the stable phase, the proportion of the upstream module decreased to 82.6 % (yeast extract-peptone dextrose medium, YPD) and 72.3 % (nitrogen-limited media, NLM) ( Fig. S2 ). Here, we hypothesized that during the early stage of mixed fermentation, a microenvironment with relatively high steroid levels may limit the biomass accumulation of upstream microbial communities while enhancing the resistance of strains to the environment. To date, ample research has shown that the synthesis of heterologous products typically triggers microbial stress responses [ [22] , [23] , [24] ]. This stress may have an impact on cell growth and limit the product synthesis efficiency of strains. Therefore, understanding how microorganisms respond to product pressure and developing strategies to regulate stress resistance on the basis of this understanding are crucial for improving the growth and fermentation performance of microbial cells. Transcriptomic methods involve the use of effective technologies that are able to elucidate the mechanisms underlying microbial responses to target compounds at the transcriptional level, allowing the identification of potential targets for modification to improve cell resistance and survival [ 25 ]. Fig. 1 The impact of the progesterone stress response on cellular metabolism. The genes (Δ7-reductase, CYP11A1, ADR, ADX and 3β-HSD) are heterologous genes overexpressed in the progesterone synthesis pathway, while the gene ERG5 (shown in red) needs to be knocked out for the pathway's expression. Fig. 1 The impact of the progesterone stress response on cellular metabolism. The genes (Δ7-reductase, CYP11A1, ADR, ADX and 3β-HSD) are heterologous genes overexpressed in the progesterone synthesis pathway, while the gene ERG5 (shown in red) needs to be knocked out for the pathway's expression. This work aimed to utilize omics techniques to identify beneficial factors for progesterone production. By analyzing the response mechanisms of Y. lipolytica under progesterone stress, we found that the effects of progesterone product stress on cell physiology are long-lasting and cause changes in various metabolic pathways or cellular processes, which are reflected mainly in amino acid metabolism, lipid metabolism, and protein transport ( Fig. 1 ). Through transcriptome data analysis, we identified several potential key targets that were upregulated (log2-fold change >2) during three different periods. Upon the overexpression of these key targets, transport proteins stand out. We then screened and overexpressed the isoenzymes of native and exogenous transporters in the progesterone-producing strain. The YCF1 protein from S . cerevisiae had the greatest benefit, increasing progesterone synthesis by 69.6 %. Taken together, the results of this study provide a theoretical basis for elucidating the self-regulatory mechanisms of Y. lipolytica in an environment with steroid hormones, as well as new targets for the systematic construction of progesterone-producing strains, which could serve as excellent chassis strains for the synthesis of downstream steroid substances.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.