Search
Close this search box.

Coupling thermotolerance and high production of recombinant protein by CYR1N1546K mutation via cAMP signaling cascades – Communications Biology

The thermotolerant strain FDHY23 is obtained by screening under high temperature condition

In this study, a thermotolerant K. marxianus strain, FDHY23, was obtained by subjecting the wild-type strain LHP1044 to high temperature selection at 46 °C. FDHY23 exhibited remarkably improved heat resistance, as evidenced by its higher cell viability in spot assays compared to the reference strain LHP1044 at temperatures of 46 °C, 47 °C, and 48 °C (Fig. 1a). Growth curves for strains FDHY23 and LHP1044 were analyzed at different temperatures (Fig. 1b). There was no significant difference in the maximum OD600 reached during the stationary phase between these two strains at both 30 °C and high temperatures (Fig. 1b). However, under high temperature conditions, FDHY23 exhibited a shorter lag phase (Fig. 1b) and a significantly faster maximum specific growth rate during the logarithmic phase (Fig. 1c) compared to LHP1044. These results demonstrate that FDHY23 is better adapted to high temperature environments and exhibits a faster growth.

Fig. 1: Comparison of high temperature tolerance between strains FDHY23 and LHP1044.
figure 1

a Thermotolerance spot assays. The cultures were diluted 5 times with a 5-fold dilution factor from an initial OD600 of 0.6. They were then incubated at temperatures of 30 °C, 46 °C, 47 °C, and 48 °C, respectively. b Growth curves. It was plotted based on the correlation between cell density (OD600) and culture time at 30 °C, 46 °C, and 47 °C, respectively. c Maximum specific growth rate during the logarithmic phase. For growth curves and maximum specific growth, values were calculated as average±SD from four biological replicates. The red and blue colors represent strains FDHY23 and LHP1044, respectively. **: p < 0.01.

CYR1N1546K mutation is the key determinant of heat resistance phenotype

We conducted whole-genome DNA sequencing on FDHY23 and LHP1044 strains and identified specific genetic variations in FDHY23 compared to LHP1044. The mutations on the nuclear chromosomes and mitochondria are documented in Table S1 and Table S2 of Supplementary Data 1, respectively. There were only two SNP mutations in the intergenic region on the mitochondria, while multiple mutations were identified in the coding regions of the nuclear chromosomes. The identified coding region variations were validated using PCR, confirming two genes with non-synonymous mutations (Fig. 2a). One is the leukotriene A4 hydrolase gene LAP2, with a G to C mutation at position 1054, resulting in the substitution of glycine (G) with arginine (R) at position 352. The other is the adenylyl cyclase gene CYR1, with a C to G mutation at position 4638, leading to an amino acid change from asparagine (N) to lysine (K) at position 1546.

Fig. 2: The non-synonymous mutations detected in FDHY23 and their impact on the thermotolerance phenotype.
figure 2

a Mutation positions in the LAP2 and CYR1 genes. The yellow marks represent the mutation sites. b The distinct effects of LAP2G352R and CYR1N1546K mutations on the thermotolerance phenotype. Spot assays to assess heat resistance were performed at 30 °C and 46 °C, respectively.

To assess the effect of these two mutations on the high-temperature tolerance phenotype, we introduced the respective mutations into strain LHP1044 using CRISPR/Cas9 technology. The LHP1044-LAP2G352R strain carrying the introduced LAP2 point mutation did not exhibit any obvious alteration in heat resistance (Fig. 2b). However, the LHP1044-CYR1N1546K strain carrying the introduced CYR1 point mutation displayed a marked enhancement in high-temperature tolerance, with no discernible difference in growth compared to LHP1044 at 30 °C (Fig. 2b). Furthermore, when the CYR1 gene in FDHY23 was subjected to a reversion mutation, the resulting FDHY23-CYR1 strain exhibited a noticeable decrease in growth capability at 46 °C (Fig. 2b). These results demonstrate that the CYR1 mutation is essential for the thermotolerance ability of FDHY23.

FDHY23 improved recombinant protein production at 30 °C

We further evaluated FDHY23’s capability of recombinant protein production, using recombinant soybean hemoglobin LBA as an example. The intracellular expression levels of LBA protein were compared between FDHY23 and LHP1044 in shake flasks at 30 °C. Remarkably, FDHY23 exhibited a 2.6-fold increase in LBA production compared to LHP1044 (Fig. 3a). We compared the production levels of other five recombinant proteins in FDHY23 and LHP1044 strains in shake flasks. These include the intracellularly expressed Fowlpox virus coat protein HVP2 and Porcine circovirus coat protein PCV2, as well as the secreted Est1E esterase from yak rumen, AnFaeA esterase from Aspergillus niger, and Badgla glycosidase from Talaromyces purpureogenus. As shown in Fig. 3b–f, the expression levels of HVP2, PCV2, Est1E, AnFaeA, and Badgla were 1.7-fold, 1.6-fold, 2.1-fold, 1.7-fold, and 5.5-fold higher in FDHY23 compared to LHP1044, respectively. Additionally, the intracellular retention rate of the Badgla protein was remarkably reduced in FDHY23, decreasing from 32.9% in LHP1044 to 8.9% in FDHY23. These findings demonstrate that FDHY23 is capable of producing higher levels of recombinant proteins and greatly improve the secretion process (Fig. 3g). The uncropped blot images for Fig. 3 can be found in Supplementary Fig. 1.

Fig. 3: The CYR1N1546K mutation improves recombinant protein yield in cells.
figure 3

Strains FDHY23 and LHP1044 exhibited distinct expression levels of six recombinant proteins on SDS-PAGE at 30 °C (A-F). a Soybean hemoglobin LBA. Four biological replicates. b Fowlpox virus coat protein HVP2. Four biological replicates. c Porcine circovirus coat protein PCV2. Four biological replicates. d Est1E esterase. Five biological replicates. e AnFaeA esterase. Six biological replicates. f Badgla glycosidase. Five biological replicates. g Intracellular and extracellular enzymatic activity of the glycosidase Badgla. Five biological replicates. h The essential effect of CYR1N1546K mutation on recombinant LBA protein expression. Values were calculated as average±SD from biological replicates. **: p < 0.01; ***: p < 0.001.

To determine whether the genetic mutations contribute to the enhanced recombinant protein yield in FDHY23, we expressed the recombinant LBA protein in the LHP1044-CYR1N1546K strain. It showed remarkably higher expression levels than LHP1044 and comparable yield to FDHY23 (Fig. 3h). The LHP1044-LAP2G352R strain carrying the introduced LAP2 point mutation did not exhibit any obvious alteration in LBA production compared to LHP1044 (Fig. 3h). Therefore, the CYR1N1546K mutation not only enhances thermotolerance but also essentially improves recombinant protein yield.

We further analyzed the temporal production of LBA in FDHY23 and LHP1044 during shake flask cultivation at 30 °C. There was no significant difference in the growth curves of the two strains (Fig. 4a). Both strains showed an increase in LBA production over time, with a noticeable surge at the beginning of the stationary phase (Fig. 4b). Notably, FDHY23 consistently displayed remarkably higher LBA production compared to LHP1044, corroborating the above findings (Fig. 3a).

Fig. 4: Comparison of recombinant LBA protein temporal production between FDHY23 and LHP1044 under shake flask cultivation and industrial fermentation conditions.
figure 4

a Growth curves of strains FDHY23 and LHP1044 during recombinant LBA production under shake flask cultivation at 30 °C. b Expressions of recombinant LBA protein under shake flask cultivation at 30 °C. LBA production was analyzed at 24 h, 36 h, 48 h, 60 h, and 72 h by SDS-PAGE. c Growth curves during high-density fermentation process at 32 °C. d Recombinant LBA protein yield during high-density fermentation with inorganic salts. LBA production was analyzed at 12 h, 24 h, 36 h, 48 h, 60 h, and 72 h by SDS-PAGE. For growth curves, values were calculated as average±SD from three biological replicates. The red and blue colors represent strains FDHY23 and LHP1044, respectively.

To simulate high-density fermentation in industrial settings, we investigated the expression capacity of recombinant LBA protein of FDHY23 compared to LHP1044, using a 5 L fermenter and an inorganic salt synthetic medium at 32 °C. No significant difference was observed in the growth curves between the two strains (Fig. 4c). It is worth noting that in the fermenter, both strains exhibited visible LBA bands as early as 24 h, with FDHY23 and LHP1044 reaching peak LBA concentrations of 4.79 g/L and 2.87 g/L, respectively (Fig. 4d). Therefore, the CYR1N1546K mutation has demonstrated the ability to effectively enhance recombinant protein production in an industrial high-density inorganic salt fermentation environment, indicating promising prospects for industrial applications. The corresponding SDS-PAGE analysis for absolute quantification is presented in Supplementary Fig. 2. The uncropped blot images for Fig. 4 can be found in Supplementary Fig. 3.

FDHY23 demonstrates high-level recombinant protein production even at high temperatures

We further compared the ability of strains to express recombinant LBA protein under extreme high-temperature conditions (46 °C). It was observed that at 46 °C, the expression of recombinant LBA protein was barely detectable in LHP1044 (Fig. 5a), while evident expression levels were still observed in FDHY23, peaking at 60 h (Fig. 5a). Their growth curves at 46 °C showed no significant difference (Fig. 5b). The LHP1044-CYR1N1546K strain also expressed remarkably higher LBA production at 46 °C (Fig. 5c). These results confirm that the CYR1 mutation indeed possesses the combined phenotype of thermotolerance and high productivity of recombinant proteins. The uncropped blot images for Fig. 5 are presented in Supplementary Fig. 4.

Fig. 5: Comparison of recombinant LBA protein production between FDHY23 and LHP1044 under high-temperature conditions.
figure 5

a Expressions of recombinant LBA protein at 46 °C. Protein production was analyzed at 24 h, 36 h, 48 h, and 60 h by SDS-PAGE. b Growth curves during recombinant LBA production at 46 °C. The red and blue lines represent strains FDHY23 and LHP1044, respectively. c Recombinant protein yield elevation at 46 °C for the effect of CYR1N1546K mutation. The strain LHP1044-CYR1N1546K, which was constructed by introducing the CYR1 mutation into strain LHP1044, also showed higher LBA production at 46 °C compared to LHP1044. For growth curves, values were calculated as average±SD from three biological replicates.

CYR1N1546K mutation weakens the enzymatic activity of adenylate cyclase and reduces intracellular cAMP levels

Through conservation domain analysis provided by NCBI, the CYR1N1546K mutation was identified to be located within the catalytic domain (1537–1705) of adenylate cyclase (Fig. 6a). This domain is essential for catalyzing the conversion of ATP to cAMP. The asparagine (N) at this position is conserved between K. marxianus and S. cerevisiae. When N is mutated to K, the amino acid changes from acidic to basic, which could potentially disrupt interactions with surrounding amino acids and thereby affect the catalytic efficiency of the enzyme.

Fig. 6: The CYR1N1546K mutation affects adenylate cyclase activity and intracellular cAMP levels.
figure 6

a Conserved domain containing the mutation position. It depicts the arrangement of conserved domains along the amino acid sequence of K.marxianusCYR1 gene. The mutation site at position 1546 is situated within the catalytic domain (CYCc) of the CYR1 gene. b The enzymatic activity of adenylate cyclase measured at 30 °C during the mid-log phase and stationary phase of strains FDHY23, LHP1044-CYR1N1546K, and LHP1044. Four biological replicates. c cAMP levels measured at 30 °C during mid-log phase and stationary phase of these three strains. Three biological replicates. d cAMP levels measured at 46 °C during the mid-log and stationary phases of strains FDHY23 and LHP1044. Three biological replicates. e cAMP levels measured at 30 °C during the stationary phase of strains FDHY23 and LHP1044 producing recombinant protein LBA. Four biological replicates. Values were calculated as average±SD from biological replicates. *: p < 0.05; **: p < 0.01; ***: p < 0.001.

We measured the enzymatic activity of adenylyl cyclase and the levels of cAMP in mid-log phase and stationary phase cells at 30 °C. Compared to LHP1044 (Fig. 6b), the adenylyl cyclase activity decreased by 36.3% and 30.6% in the mid-log phase and stationary phase cells of strain LHP1044-CYR1N1546K, respectively. Similarly, FDHY23 showed a decrease of 34.4% and 37.2% in the same phases, respectively. The intracellular cAMP levels also showed a decrease, with strains LHP1044-CYR1N1546K and FDHY23 experiencing a significant drop compared to LHP1044 in the mid-log and stationary phases (Fig. 6c), respectively. Therefore, the CYR1N1546K mutation weakens the enzymatic activity of adenylyl cyclase and leads to a decrease in intracellular cAMP levels at 30 °C.

We further investigated the intracellular cAMP levels of strains FDHY23 and LHP1044 under two different stress conditions: high temperature (46 °C) and production of recombinant protein LBA. In the mid-log phase at 46 °C (Fig. 6d), the cAMP level of LHP1044 was significantly lower than its cAMP level at 30 °C (p = 0.009, Cohen’s d value = −5.34). In contrast, the cAMP level of FDHY23 at 46 °C was higher than its cAMP level at 30 °C (p = 0.213, Cohen’s d value = 1.67), with average values of 4.50 and 3.83, respectively. Since the Cohen’s d value is larger than 0.8, indicating a substantial difference between the two groups, it can be concluded that the higher cAMP level in the mid-log phase of FDHY23 at 46 °C compared to 30 °C was meaningful, rather than merely occurring by chance, despite the p-value not being significant. While in the stationary phase at 46 °C (Fig. 6d), the cAMP levels of LHP1044 and FDHY23 were both significantly lower than their levels at 30 °C (p = 0.012 and p = 0.003, respectively). These results suggest that with the increase in temperature, FDHY23 may enhance the production of cAMP during mid-log phase, while LHP1044 significantly drops its cAMP production. For LBA production at the stationary phase at 30 °C (Fig. 6e), the cAMP level of FDHY23 remained significantly lower than LHP1044. Please note that the units for Fig. 6e differ from those in 6c and 6d. This is because when producing the recombinant protein LBA, the cell protein content cannot be used as the denominator. Therefore, the cAMP levels during LBA production (Fig. 6e) are not compared to those during normal growth (Fig. 6c).

FDHY23 commonly enhances mitochondrial function, respiratory chain, protein folding, antioxidant response, and vesicle transport processes under high temperature and recombinant protein production conditions

To investigate the possible molecular mechanisms underlying the enhanced high-temperature tolerance and recombinant protein production capacity caused by the CYR1N1546K mutation, we conducted transcriptomic RNA-seq analysis on strains FDHY23 and LHP1044 during two conditions: during the mid-log phase at 46 °C (Supplementary Data 2) and during the stationary phase while expressing recombinant protein LBA at 30 °C (Supplementary Data 3). The mitochondrial-encoded genes showed no significant expression changes, therefore the subsequent analysis primarily centered on nuclear-encoded genes.

At 46 °C, FDHY23 exhibited upregulation of 1570 genes (Fig. 7a) and downregulation of 497 genes (Fig. 7b) compared to LHP1044. The upregulated genes were enriched in GO pathways related to mitochondrial respiratory chain ATP generation, ribosome biogenesis, protein translation, protein folding, glycolysis, gluconeogenesis, TCA cycle, acetyl-CoA generation, lipid synthesis, vesicle transport, protein glycosylation, ROS metabolism, DNA replication, and cytokinesis. The downregulated genes were mainly associated with transmembrane transport, fatty acid beta-oxidation, meiosis, and unfolded protein response. During recombinant protein production, FDHY23 showed upregulated 325 genes (Fig. 7a) and 47 downregulated genes (Fig. 7b) compared to LHP1044. The upregulated genes were enriched in GO pathways such as mitochondrial translation, respiratory chain assembly, heme biosynthesis, protein folding, cellular response to heat, GPI anchor biosynthesis, spindle assembly, and peroxide metabolism. While the downregulated genes were enriched in processes related to glycolysis, glycogen biosynthesis, hormone-mediated signaling, and stress response. There were 192 genes commonly upregulated in high temperature and recombinant protein production conditions, enriched in GO pathways such as protein translation, protein folding, mitochondrial function, respiratory chain ATP synthase assembly, heme biosynthesis, peroxide metabolism, Golgi vesicle fusion, and GPI anchor biosynthesis. Only 9 genes were both downregulated, mainly involved in lipid degradation and vacuole function. Interestingly, in both high temperature and recombinant protein production conditions, there were remarkably more upregulated genes than downregulated genes, indicating that the upregulated cellular processes (Supplementary Data 4) are crucial in facilitating these favorable conditions. Therefore, the commonly upregulated processes of mitochondrial function, respiratory chain ATP generation, protein folding, ribosome biogenesis, ROS resistance, and vesicle transport likely predominantly contribute to the coupling of high temperature tolerance and high recombinant protein production phenotypes.

Fig. 7: GO enrichment analysis of differentially expressed genes between FDHY23 and LHP1044 under high temperature and recombinant protein production conditions.
figure 7

a GO enrichments of upregulated genes in FDHY23 compared to LHP1044. ‘LBA production’ represents the recombinant protein production condition. The enrichment results were denoted as follows: pink for all genes upregulated at high temperature, orange for all genes upregulated during LBA production, and coral for genes commonly upregulated under both high temperature and LBA production conditions. b GO enrichments of downregulated genes in FDHY23 compared to LHP1044. Light blue and light green represent the downregulated genes at high temperature and LBA production, respectively. The numbers of differentially expressed genes are labeled in the Venn diagram. c qRT-PCR analysis of 25 crucial genes involved in various cellular processes. The red and purple colors represent the high temperature and LBA production conditions, respectively. Values were calculated as average±SD from three biological replicates.

It should be noted that during recombinant protein production, there were fewer differentially expressed genes compared to high temperature conditions. This is likely because the transcriptome during LBA production captures the stationary phase rather than the logarithmic growth phase observed at 46 °C. As a result, in the stationary phase of recombinant protein production, while metabolic changes observed at high temperature may not be sustained, key functions involved in cellular apparatus such as mitochondrial assembly, respiratory chain, protein translation and folding, ER and Golgi system are retained.

We further conducted quantitative real-time PCR (qRT-PCR) analysis of 25 crucial genes (Fig. 7c) involved in various cellular processes, including the respiratory chain, ATP generation, protein folding, anti-ROS, ribosome assembly, protein translation, secretory pathway, vesicle transport, central carbon metabolism, and cAMP signaling cascade. The qRT-PCR results were mostly consistent with the RNA-seq data (Supplementary Data 2 and Data 3), thus confirming the reliability of our RNA-seq findings.

At high temperatures, FDHY23 reconfigures carbon flux to enhance the production of ATP, glycogen, acetyl-CoA, lysine, and nucleotides

As mentioned earlier, in strain FDHY23 compared to LHP1044, during the mid-log phase at high temperatures, there is an upregulation of genes involved in both glycolysis and gluconeogenesis (Fig. 7a), indicating the activation of two opposing cellular processes. This raises curiosity about how FDHY23 efficiently allocates carbon sources during rapid growth under high-temperature conditions. To address this, we conducted a systematic analysis of differentially expressed genes related to metabolic pathways at 46 °C (Fig. 8).

Fig. 8: The distribution map of differentially expressed genes involved in carbon metabolism pathways during logarithmic growth of FDHY23 at high temperatures.
figure 8

LHP1044 serves as the control. The upregulated, downregulated, and unchanged genes are represented by the colors orange, green, and blue-gray, respectively.

As shown in the top portion of Fig. 8, glucose uptake genes (HXT4, STL1, and SKS1) and glucose phosphorylation genes (GLK1 and RAG5) were significantly upregulated, resulting in increased production of glucose-6-phosphate. For glucose-6-P, it can be directed towards four downstream pathways, with glycolysis being the primary pathway leading to pyruvate production. Several steps in glycolysis are also shared with gluconeogenesis. Shared genes involved in both glycolysis and gluconeogenesis, such as FBA1 and PGK1, were significantly upregulated. Additionally, specific genes related to glycolysis (e.g., PFK1, PFK2, and PYK1) and gluconeogenesis (e.g., PYC2, PCK1, and FBP1) were also upregulated, indicating an overall enhancement of both glycolysis and gluconeogenesis processes.

Acetyl-CoA production from pyruvate was upregulated in both the mitochondria and cytoplasm pathways at high temperatures. In the mitochondrial pathway, the conversion of pyruvate to acetyl-CoA through the upregulation of the pyruvate dehydrogenase complex (PDA1, PDB1, and LAT1) was observed. Furthermore, there was an upregulation in the synthesis of the necessary coenzyme lipoamide. In the cytoplasm pathway, the conversion of pyruvate to acetaldehyde, and then to acetate and finally to acetyl-CoA, was upregulated at each step. While the degradation of acetyl-CoA to acetate was downregulated. It is worth noting that the enhanced conversion of acetaldehyde to acetate was accompanied by the generation of NADPH from NADP+, providing sufficient reducing power for many biosynthetic reactions, including lipid and DNA synthesis. Additionally, the conversion of acetaldehyde to ethanol in the cytoplasm (coupled with converting NADH to NAD+) and the generation of ethanol from acetaldehyde in the mitochondria (accompanied with NAD+ to NADH) were both upregulated. This “acetaldehyde-ethanol shuttle” provides NAD+ for the cytoplasmic oxidative breakdown processes like glycolysis, while also supplying sufficient NADH as an electron donor for the mitochondrial respiratory chain.

Acetyl-CoA can be used for lipid/sterol synthesis or enter the tricarboxylic acid (TCA) cycle (Fig. 8). We observed upregulation of genes involved in both processes (Supplementary Data 4), indicating increased acetyl-CoA production by FDHY23 at high temperatures. The increased synthesis of lipids and sterols can enhance membrane stability. The upregulation of the TCA cycle provides electrons for the respiratory chain and building blocks for biosynthesis. Additionally, lysine and heme synthesis, branching off from the TCA cycle, were also upregulated. Heme promotes the respiratory chain, while lysine facilitates the generation of acetyl-CoA from pyruvate, lipid synthesis, and protein translation activation.

For glucose-6-P, its second fate is to continue gluconeogenesis and produce four different forms of glycogen chains, including trehalose, branched glycogen, cell wall insoluble β-glucan, and N-glycan on glycoproteins (shown in the upper left part of Fig. 8). All four downstream pathways were upregulated, while glycogen degradation into glucose was downregulated. Intracellular branched glycogen and trehalose serve as thermoprotectants, stabilizing the cell membrane and preserving the conformation of intracellular proteins17. Genes BGL2 and EXG1 were upregulated, facilitating the hydrolysis of linear β-1,3-glucan on the cell wall. Conversely, the gene YBR056W, responsible for the hydrolysis of branched β-1,6-glucan, was downregulated. These findings suggest that strain FDHY23 reinforces its cell wall into a mesh-like structure to withstand high temperatures more effectively.

For glucose-6-P, the pentose phosphate pathway is its third downstream pathway (shown in the upper right part of Fig. 8). This pathway has four different directions. (1) Returning to glycolysis, the involved genes TKL1 and TAL1 were downregulated. (2) Proceeding from the intermediate product ribulose-5-P to synthesize FMN and pyridoxal-5-P, both of which were upregulated. FMN mediates mitochondrial respiratory chain, while pyridoxal-5-P is involved in amino acid deamination reactions, providing a nitrogen source for nucleotide synthesis. (3) Proceeding from ribose-5-P to generate ribose-1-P and PRPP, both of which were also upregulated, promoting nucleotide synthesis. (4) Proceeding from the intermediate product erythrose-4-P, together with the main glycolytic intermediate P-enolpyruvate, to generate aromatic amino acids and tetrahydrofolate (THF). Aromatic amino acids synthesis was downregulated, while THF generation was upregulated, promoting purine and pyrimidine synthesis. The fourth downstream pathway of glucose-6-P is inositol-3-P production, which was upregulated and may facilitate the synthesis of inositol phospholipids in the cell membrane.

As shown above, strain FDHY23 effectively reprogrammed carbon flux allocation at high temperature, by boosting the processes of lipid biogenesis, respiratory chain ATP generation, glycogen synthesis and nucleotide production, accompanied with elevated levels of intracellular NADPH and NAD+. FDHY23 not only enhanced glucose transport but also increased nitrogen uptake for material synthesis. Additionally, it enhanced vacuolar function and microautophagy to facilitate resource recycling (Supplementary Data 4). These adaptations enabled efficient resource utilization and optimized material synthesis at high temperatures. As a result, a large number of genes involved in cell cycle proliferation were upregulated in FDHY23. Additionally, some genes involved in cell mating and sporulation were upregulated, indicating a slight activation of sporulation as a response to high temperature. However, due to abundant ATP supply, rapid proliferation took priority over sporulation in FDHY23 at high temperatures.

CYR1 mutation modulates the cAMP signaling pathway to rebuild global cellular processes under high temperature and recombinant protein production pressures

The CYR1 mutation directly decreases intracellular cAMP levels. Therefore, the detailed analysis below (Fig. 9) explores the impact of altered cAMP generation and its downstream cascades on overall cellular processes under conditions of high temperature and recombinant protein production.

Fig. 9: The altered cAMP signaling cascades caused by the CYR1 mutation during high temperature and recombinant protein production conditions.
figure 9

The color-gene differential expression correspondence is shown in the top right corner of the figure.

For cAMP generation (Fig. 9), genes associated with mitochondrial function and the respiratory chain in FDHY23 were generally upregulated under high temperature and recombinant protein production conditions. This led to an increase in ATP production, which serves as the precursor substrate of cAMP. The enzyme CYR1, responsible for catalyzing cAMP generation, was also upregulated at high temperatures, while its inhibitory gene IRA2 showed downregulation. The genes PDE1 and PDE2, involved in cAMP degradation, did not exhibit significant change in expression. These findings indicate that FDHY23 upregulates CYR1 expression and increases ATP levels as a substrate for cAMP generation at high temperatures, resulting in a modest restoration of cAMP levels observed during mid-log phase of FDHY23 at high temperatures (Fig. 6d).

The protein kinase A (PKA) is the sole known downstream effector of cAMP in yeast. cAMP activates PKA by binding to its inhibitory subunit BCY1 (Fig. 9). The expression levels of PKA subunits in FDHY23 decreased under high temperature and recombinant protein production conditions (Fig. 9), consistent with the reduction in cAMP. In S. cerevisiae, repression of PKA activity promoted protein folding, stimulated gluconeogenesis and glycogen synthesis, inhibited glycolysis, enhanced mating and sporulation, and suppressed cell proliferation under heat stress18,19,20,21. In FDHY23 (Fig. 9), reducing PKA activity at high temperature improved protein folding and glycogen synthesis, similar to S. cerevisiae. However, unlike S. cerevisiae, glycolysis and cell proliferation were actually increased in FDHY23. This could be due to the heightened need for respiratory chain production, which upregulates glycolysis. Additionally, the ample supply of ATP may have fueled the upregulation of cell proliferation.

In addition, in S. cerevisiae, PKA activated the transcription factor RAP1, which in turn promoted the transcription of ribosomal genes through RAP1 binding sites their promoters22. In FDHY23, despite reduced PKA activity, most ribosomal genes were upregulated under high temperature and recombinant protein production conditions (Supplementary Data 4). Meanwhile, RAP1 expression remained unchanged, and its binding sites were only present in a few ribosomal genes. This indicates that the regulation of ribosomal gene transcription in FDHY23 may involve alternative mechanisms, rather than the PKA-RAP1 pathway. It has been reported that decreasing cAMP levels in E.coli can stimulate ribosome recovery23. Thus, the reduction in cAMP levels in FDHY23 may facilitate ribosome biogenesis through specific mechanisms, compensating for the loss caused by protein denaturation and degradation at high temperatures.

To identify new potential cAMP-regulated proteins in K. marxianus, we performed a conserved domain search on the entire genome. Based on the known cAMP-binding domains such as BCY1’s CAP_ED domain and PDE2’s GAF domain, through the “Conserved Domains” tool provided by NCBI, we identified three proteins in K. marxianus that contain known cAMP-binding domains. These include NTE1 and FBXL7, both containing the CAP_ED domain, and YKL069W, with the GAF domain. NTE1 is a phospholipase that converts phosphatidylcholine (PC) to glycerophosphocholine, playing a role in the rapid turnover of PC during temperature increase24. At high temperatures, NTE1 was upregulated in FDHY23, along with the majority of genes involved in lipid and sterol synthesis. This suggests that the decrease in cAMP at high temperatures may stimulate the upregulation of its binding protein NTE1, leading to PC turnover as the membrane lipid signal to trigger lipid and sterol production. FBXL7 is an E3 ubiquitin ligase subunit and its deficiency has been shown to enhance the transcription of mitochondrial genes25. In S. cerevisiae, a brief pulse of cAMP led to a notable but temporary rise in mitochondrial transcript levels26. In FDHY23, FBXL7 was downregulated at high temperatures (Fig. 9), while mitochondrial genes were predominantly upregulated. This indicates that at high temperature, FBXL7 may mediate the promotion of mitochondrial gene transcription by transient cAMP pulses induced by respiratory chain augmentation. Gene YKL069W’s function is unknown, but it was upregulated in FDHY23 during both high temperature and recombinant protein production. Additionally, genes associated with Golgi vesicle transport and resistance to ROS were commonly upregulated in these conditions (Supplementary Data 4). Previous researches demonstrated that the addition of cAMP can enhance the expression of genes involved in vesicle transport at high temperatures27 and boost cellular resistance to ROS28. Therefore, it is possible that YKL069W mediate the pulsatile increase in cAMP levels, leading to the upregulation of vesicle transport genes and anti-ROS genes.

In summary, the CYR1 mutation reduces cAMP levels, which in turn activates multiple cascade pathways during high temperature and recombinant protein production, including respiratory chain, lipid and sterol synthesis, vesicular transport, ROS resistance, protein folding, glycogen biosynthesis, and ribosome biogenesis. These changed cellular processes subsequently affect the allocation of metabolic resources within the cell.