The T. reesei Δ res2 mutant shows growth and germination defects

The res2 ortholog in T. reesei was identified with fungiDB. Pair-wise sequence alignment with the N. crassa RES2 showed a 60% sequence identity on the amino acid level between both sequences. To study the role of the res2 orthologue of T. reesei, the gene was deleted using CRISPR-Cas9. After genetic validation by PCR and sequencing, confirming the correct insertion of the deletion cassette, three of the validated transformants were randomly kept for further analysis. The absence of off-target integrations was also confirmed in these mutants by qPCR.

Although deletants showed no phenotype on transformation plates, we first wanted to investigate if the res2 deletion impacts basic functions such as growth and germination. To address this question, Rut-C30 and the three deletion transformants were plated on solid minimal medium containing glucose or a cellulase secretion-inducing substrate such as lactose in the presence or absence of 1 or 5 mM DTT. Figure 1A shows that growth of both Rut-C30 and the mutant was faster on glucose and slower on lactose. However, growth of the ∆res2 mutant was clearly reduced compared to that in the Rut-C30 parental strain on both substrates, but especially on lactose, Fig. 1A. It is interesting to note that the mutant tended to grow better in the presence of glucose + DTT, reducing the difference with Rut-C30 (Fig. 1B). One hypothesis to explain this observation is that the reducing activity of DTT enhances the activity of some enzymes, as it is the case for cellobiase [20], which leads to increased availability of nutrients and increased fungal growth.

Fig. 1

Growth of three ∆res2 transformants and Rut-C30 on different substrates. A Rut-C30 and Δres2 transformants were grown on minimal medium supplemented with 2% non-inducing substrate (glucose) or 2% inducing substrate (lactose), with or without addition of DTT. Photos and measurements were done on the fifth day after inoculation. For each strain, three biological replicates were assayed, but only one replicate is shown representatively. B Size of colonies of the ∆res2 mutant compared to its parental strain Rut-C30. The error bars indicate the standard deviation of three biological replicates. Significance of the difference between the Rut-C30 and the mutant strains was calculated with the two-tailed t-test with independent variables (**P < 0.01)

To investigate if growth was already affected at the germination step, we assessed the germination rate for both Rut-C30 and the Δres2 transformant 2 on three media: minimal medium with glucose as a non-inducing condition, and minimal medium containing lactose or soluble hydroxyethyl cellulose (HEC) as inducing substrates. Hyphal growth was detected after 14 h on glucose and lactose for the Rut-C30 strain (Additional file 1 and Fig. 2). Δres2 showed slower germination on the three substrates. On HEC, hyphae appeared after 21 h for both strains with a slight delay for the mutant strain (Fig. 2). However, even if the germination rate was not or not much reduced for the Δres2 after 21 h on the three substrates, it can clearly be seen on Additional file 1 that the hyphal length of the mutant was much shorter.

Fig. 2

Germination rate of Rut-C30 and Δres2 glucose, lactose and HEC. The error bars indicate the standard deviation of three biological replicates of one representative transformant of Δres2

Deletion of res2 leads to reduced protein production

Protein secretion of the two strains was quantitatively measured after growing them in shake flasks in the presence of either glucose or cellulose + lactose. As expected, higher protein secretion was observed in the cellulase-inducing condition than with glucose for both strains. However, protein secretion in the three ∆res2 mutants was lower compared to that in the Rut-C30 parental strain (Fig. 3A). The biomass of the mutants and the parental strain were very similar or even slightly higher in the mutants, leading also to a reduced protein production yield in the ∆res2 strains (Fig. 3B). These results are in agreement with those of [19] where protein secretion decreased approximately by 30% in N. crassa ∆res2 strains grown on 1% w/v crystalline cellulose supporting the hypothesis that the TF Res2 is involved the regulation of protein secretion.

Fig. 3

A Concentration of secreted proteins of ∆res2 mutant and its parental strain Rut-C30. B Biomass and yield of ∆res2 and Rut-C30 in glucose cultures. Strains were cultured for seven daysin 25 mL of BTCA medium in the presence of either glucose or cellulose and lactose. All cultures were inoculated with the same number of spores (~2x105 spores). The error bars indicate the standard deviation of three biological replicates for glucose cultures and two biological replicates for cellulose + lactose cultures. Significance of the difference between the Rut-C30 and the mutant strains was calculated with the two-tailed t-test with independent variables (*P < 0.05, **P < 0.01, ***P < 0.001)

Impact of secretion stress on gene expression in Rut-C30 and ∆res2

To determine if RES2 plays a role in the secretion pathway regulation or the secretion stress response in the hyperproducer Rut-C30, a transcriptomic study was conducted in conditions inducing secretion stress. It had previously been demonstrated that growth of Rut-C30 in lactose fed-batch culture [18] and addition of DTT [12] to T. reesei culture induce UPR. To choose appropriate stress conditions, preliminary fed-batch cultures on lactose and glucose with or without addition of DTT were conducted and the transcript levels of the UPR marker gene bip1 determined at different time points by qPCR. Fed-batch fermentation allows to obtain a higher protein production in lactose culture and to control the pH, thus reducing variables that can affect gene regulation and mimicking industrial like conditions. The results showed that induction of cellulase production by lactose feeding and exposure to 10 mM DTT led to an increase in the UPR marker gene bip1 (Fig. 4). The highest increase of bip1 transcript levels was detected at 2 h after addition of DTT. In the lactose cultures, there was only a low increase of bip1 mRNA levels compared to the glucose cultures, indicating a moderate secretion stress. Similarly to bip1, transcript levels of the UPR marker genes pdi1 and hac1 (spliced version) were also highest 2 h after DTT addition (Additional file 2). In addition, lactose induced the transcription of cellulases as expected, evidenced by an increase of the transcription levels of the cbh1 gene at all time points. However, the cbh1 gene was repressed after 4 and 6 h in the glucose culture with DTT, and much less induced in lactose cultures with DTT (Additional file 2). Taking into account these results, we decided to analyze gene expression by RNA-seq 2 h after exposure (D) or not (C) to 10 mM DTT in the reference strain Rut-C30 (R) and one representative transformant of the Δres2 mutant (∆r) in glucose (G) and lactose (L) fed-batch cultures, leading to four culture conditions : Glucose Control (GC), Glucose DTT (GD), Lactose Control (LC), Lactose DTT (LD). Transcriptome data were used for pair-wise comparison of cultures within each strain (GD/GC, LD/LC, LC/GC, or LD/GD) and between strains for the same conditions. The results of the fed-batch fermentations (protein concentration and biomass) are presented in Additional file 3.

Fig. 4

Relative gene expression of the secretion stress biomarker bip1 in Rut-C30 in fed-batch cultures. RNA samples were taken at three timepoints (2h, 4h, 6h) after the addition of DTT. The relative expression of bip1 with respect to that on glucose at 2h is shown. The error bars indicate standard deviation of two biological replicates

A total of 1163 and 1664 genes (~ 10% of the genome) were differentially expressed (DE) in at least one of the four comparisons in the Rut-C30 and the Δres2 strains, Additional files 4 and 6, respectively. Exposure to 10 mM DTT led to a much higher change in gene expression in Rut-C30: 851 and 849 Differentially Expressed Genes (DEGs) in GD/GC and LD/LC compared to 171 and 129 in lactose-induced stress conditions LC/GC and LD/GD, respectively. Similarly, 1188 and 1172 genes were DE in Δres2 in GD/GC and LD/LC compared to 318 and 259 in LC/GC and LD/GD, respectively. Remarkably, in Δres2 more genes were differentially expressed genes than in Rut-C30. Whereas approximately as many genes were up – and downregulated due to the addition of DTT, there were much more upregulated genes in the lactose compared to the glucose cultures for both strains. The only exception is the LD/GD of the Δres2 strain with more down- than upregulated genes (Fig. 5A).

Fig. 5

DEGs of Rut-C30 (R) and Δres2 (Δr) in two stress conditions. A Bar graph displaying the number of up and down-regulated genes in Rut-C30 and Δres2 in each condition using a 0.5% false discovery rate cut-off and with an absolute log2 fold change greater than 2. Venn diagrams showing both unique and common DEGs in Rut-C30 and Δres2 in B the presence vs. absence of DTT and C glucose vs. lactose cultures

Out of 1163 DEGs, only 192 (~ 16.5%) were uniquely DE in Rut-C30 while 693 (~ 41.6%) out of 1664 DEGs were uniquely DE in Δres2. 971 DEGs were common between the two strains. 934 genes (~ 56%) of the DEGs in Δres2 were common with Rut-C30 in DTT-induced stress conditions (GD/GC and LD/LC) (Fig. 5B). What is intriguing is that only few DEGs (141, ~ 8.5%) were common between the two strains in the lactose-induced stress conditions (LC/GC and LD/GD) (Fig. 5C). Most of them code for CAZymes (e.g. CEL3D), and a smaller number code for transporters (e.g. TrSTR1) or redox regulators (e.g. AOD1). Interestingly, a much higher number of genes, i.e. 336 were differentially regulated in the Δres2 strain only, but only 175 were unique to Rut-C30 in the lactose conditions.

res2 and Rut-C30 have similar clusters of DE genes

A clustering analysis of DEGs of both strains was performed. Five different expression profiles were found for Rut-C30 and Δres2, but Δres2 had a sixth cluster of unassigned genes (Fig. 6). The first five clusters display very similar profiles.

Fig. 6

Clustering of the DEGs. The average profile of each cluster is shown for Rut-C30 (on the left) and Δres2 (on the right). The histograms in the middle represent the most enriched GO terms expressed as a percentage of genes compared to the number of background genes for each cluster of the two strains based on fungiDB. Only GO terms with a p-value < 0.005 are indicated

Gene ontology enrichment analysis (Additional files 5 and 7) revealed that the first cluster was enriched in genes encoding secreted proteins such as cellulases and other lignocellulose-degrading enzymes (e.g. TRIREDRAFT_123989 cel7a/cbh1, TRIREDRAFT_74223 xyn1, TRIREDRAFT_76672 cel3a/bgl1) in both strains. Transcript levels were greatly reduced in the presence of DTT, both with lactose and glucose as carbon source, and significantly increased in the presence of lactose compared to the glucose control culture. Rut-C30 had an additional enriched GO term cell wall polysaccharide metabolic process (e.g. TRIREDRAFT_120229 xyn3 and TRIREDRAFT_121127 bxl1). Thus, the res2 deletion did not impact lactose induction of genes involved in biomass degradation.

The second cluster contained genes involved in carbohydrate metabolic process for both strains. These genes (e.g. TRIREDRAFT_104797 bgl3j and TRIREDRAFT_46816 cel3d) showed moderately lower expression levels in the presence of DTT and either no change in expression in the presence of lactose or some increase in transcript levels. In addition, small molecule metabolic process genes (e.g. TRIREDRAFT_110414 uga1 and TRIREDRAFT_123288 xki1) and transmembrane transport (e.g. TRIREDRAFT_104072 xlt1) were also enriched. Some GO terms were specifically enriched in Δres2 with 430 genes in this cluster, compared to only 118 in Rut-C30: carboxylic acid metabolic process (e.g. TRIREDRAFT_102382 glo2 and TRIREDRAFT_121449 his3) and branched chain amino acid metabolism (e.g. TRIREDRAFT_122868 hom6 and TRIREDRAFT_51499 ilv5) were found to be enriched.

The third cluster containing approximately a hundred genes in both strains was characterized by a high upregulation of genes (e.g. TRIREDRAFT_106245 cta1 and cytochrome P450-encoding genes) acting against oxidative stress in conditions where DTT was present. An increase in transcript levels of Redox active genes is unsurprising as DTT is a reductant that has a variety of stress effects on fungal cells.

The fourth cluster grouped genes which displayed upregulation in the presence of DTT, but at a lower level than genes in cluster 3. Additionally, they were also slightly upregulated on LC/GC. Genes in this cluster have functions in protein targeting, secretion and lipid metabolism. Also, UPR genes such as TRIREDRAFT_122920 bip1 and TRIREDRAFT_122415 pdi1, as well as genes of the ERAD pathway like TRIREDRAFT_50647 hrd1 and TRIREDRAFT_47330 lcl2 fall in this group. In Rut-C30, protein glycosylation related genes were also enriched, whereas the GO terms DNA repair and cellular response to stress were more specifically enriched in the mutant.

Although the fifth cluster grouped genes with similar gene expression profiles, i.e. a moderate downregulation with DTT and in LC/GC, the most enriched GO terms in each strain were different. The three most enriched GO terms in RUT-C30 were ribosome biogenesis (e.g. TRIREDRAFT_104595 snu13 and TRIREDRAFT_124149 nhp2), purine-containing compound metabolic process (e.g. TRIREDRAFT_120568 eno1 and TRIREDRAFT_47221 ynk1) and branched-chain amino acid biosynthetic process, while in Δres2, genes related to amide transport (e.g. TRIREDRAFT_59364 opt2) were enriched. For the sixth cluster of unassigned genes in Δres2, no specific enrichment of functions was found, but it contained genes coding for proteins with putative functions in pseudouridine synthesis (e.g. TRIREDRAFT_3671 gar1 and TRIREDRAFT_44449 cbf5) or carbohydrate catabolic process (e.g. TRIREDRAFT_121735 cel3b and TRIREDRAFT_55319 abf2).

Function of differentially regulated genes in the Δres2 strain

Although clustering did not reveal major differences in global the gene expression patterns between Rut-C30 and the res2 deletion strain, we looked for individual genes that were differentially regulated in both strains. To potentially identify genes related to the protein secretion pathway, we focused on the fourth cluster grouping many of these genes of which 443 and 622 (~ 5% of the genome) were DE in Rut-C30 and Δres2, respectively. However, genes involved in secretion or secretion stress response such as UPR or ERAD were not differentially expressed in the Δres2 strain compared to Rut-C30 under any condition implying that RES2 is not involved in the regulation of the secretion stress response. But several other genes related to the secretion pathway and to other metabolic functions displayed differential expression in the two strains in various clusters and are described in more detail below.

As DTT might impact a lot of cellular functions and lead to a high number of DEGs, we first concentrated on lactose-induced stress conditions (LC/GC or LD/GD). This condition more specifically highlights DEGs that are potentially involved in protein secretion rather than other kinds of biological processes. For example, some MFS (Major facilitator superfamily) and ABC transporters were found to be upregulated in the condition LC/GC in Δres2 but not regulated in Rut-C30, such as low-affinity glucose transporter TRIREDRAFT_106556 (hxt13), TRIREDRAFT_62747, and TRIREDRAFT_47897 which is involved in the oxidative stress response. One MFS transporter in this cluster is downregulated in the mutant (TRIREDRAFT_58561). On the other hand, TRIREDRAFT_61278 encoding a putative high affinity glucose transporter was upregulated specifically in Rut-C30 in the LD/LC condition (Table 1). Another gene, TRIREDRAFT_124198 coding for a putative secreted protein of unknown function, displayed a Log2 fold change (L2FC) that increased from 1.56 in Rut-C30 to 4.3 in Δres2 in the condition LC/GC.

Table 1 Significantly differentially regulated genes in only one of the strains in the indicated condition(s)

Other upregulated genes in LC/GC in Δres2 but not DE in Rut-C30 include TRIREDRAFT_46285 encoding the chaperon HSP30, the putative alcohol dehydrogenase involved in redox reactions TRIREDRAFT_77770 and pks2 (TRIREDRAFT_65891) having roles in secretion and secretion metabolism. Also, a G-protein-coupled receptor involved in cellulose sensing csg1 (TRIREDRAFT_27948) is upregulated in GD/GC and LC/GC in in Δres2 only and falls in cluster 4.

Interestingly, several lipid metabolism genes of this cluster were specifically DE in the lactose culture in the mutant strain, such as phospholipase D (pldB, TRIREDRAFT_22331) and TRIREDRAFT_45980, a gene encoding a protein putatively involved in phospholipid translocation. They are both upregulated in the in Δres2 strain in this condition (Table 1). The ple gene (TRIREDRAFT_21960) encoding phospholipase E showed a similar upregulation in LC/GC but was grouped in the unassigned genes cluster of Δres2. On the other hand, the acyltransferase CST26 gene (TRIREDRAFT_109980) was induced in Rut-C30 only in the LD/LC condition. This Candida albicans orthologue encodes a transferase involved in phospholipid synthesis and its lack of induction suggests that its expression could be regulated by RES2 (fungiDB).

An important component of fungal membranes is ergosterol. But only two genes of the ergosterol pathway were found to be differentially regulated in the Δres2 strain compared to Rut-C30: the are2 gene encoding a sterol-O-acyltransferase (TRIREDRAFT_50607) and erg3, a putative C-14 sterol reductase (TRIREDRAFT_81049). The former was specifically induced in LC/GC whereas the latter was repressed in the presence of both DTT and lactose (LD/LC) in the mutant strain. Therefore, biosynthesis of ergosterol did not seem to be significantly affected by the lack of RES2.

In yeast and filamentous fungi, UPR is linked to cell wall integrity [21] and in yeast, the two pathways are coordinately regulated. In Rut-C30 and the Δres2 mutant, UPR is induced with DTT and cellulase induction by lactose. Even if UPR related genes were not DE in the two strains, we verified if the expression of genes coding for putative cell wall modifying enzymes were differentially affected. Indeed, a chitinase (TRIREDRAFT_120953), two β-1,3 glucanases belonging to family GH55 (TRIREDRAFT_73248 and TRIREDRAFT_ 121746) and a GH71 α-1,3 glucanase were differentially regulated in the two strains in either of the two secretion stress conditions. The GH18 gene was induced in LC/GC in the Δres2 strain only whereas both GH55 genes were repressed in the same strain in the presence of either lactose or DTT. GH71 gene is downregulated in LC/GC, but only in Rut-C30 (Table 1). All four proteins are predicted to be secreted or cell wall located.

We also analyzed the effect of the res2 deletion on media containing the redox stress agent DTT in more detail. Again, genes encoding proteins involved in the translocation or intracellular transport of phospholipids were found to be upregulated uniquely in the Δres2 mutant (ept1, TRIREDRAFT_55627 and Sec14, TRIREDRAFT_8192). These results suggest that the deletion of res2 impacts the expression of genes involved in the ER membrane synthesis or homeostasis, which is particularly important in conditions of secretion stress.

Another functional group of genes for which the expression was found to be impacted by the deletion of res2 encode glycosyltransferases (GT). Among other functions, these enzymes are catalyzing several protein glycosylation steps during maturation in the ER and Golgi. Three GT (TRIREDRAFT_72788, TRIREDRAFT_77283 and TRIREDRAFT_120923) belonging to CAZy families GT2, GT31 and GT32 were upregulated in the presence of DTT, but in the Δres2 mutant only. Three other GT (TRIREDRAFT_66687, TRIREDRAFT_64925 and TRIREDRAFT_122992) were upregulated with DTT uniquely in Rut-C30. It is noteworthy that all the mentioned GT whose genes were DE are predicted, by localization prediction online tools, to be located in the ER or Golgi apparatus. Even if it is not possible at this stage to evaluate the real impact on glycosyl side chains of secreted proteins, glycosylation is probably altered in the Δres2 mutant in the presence of DTT which could interfere with normal secretion and/or activity of secreted enzymes.

These results suggest that RES2 is, probably among other functions, somehow involved in the regulation of protein synthesis and protein fate and of pathways related to secretion. It is not known if RES2 acts directly on the regulated genes or if its action involves other factors. Therefore, we analyzed the data for differentially regulated TFs. In Rut-C30, only two genes encoding putative transcriptional regulators were DE compared to the Δres2 strain, one up- and the other downregulated (Table 2). In contrast, in the latter, six genes were specifically DE in the deletion strain and four of them were downregulated in the presence of DTT. TRIREDRAFT_109538 belonging to cluster 4 was upregulated in the GD/GC condition whereas the sixth one, TRIREDRAFT_12107, a homeodomain-containing protein, was upregulated in the lactose culture in the mutant only. Interestingly, the ortholog of this gene in Penicillium oxalicum was found to be involved in cellulase and xylanase production [22, 23]. Finally, the VIB1 TF (TRIREDRAFT_54675) which belongs to cluster 2 was downregulated in both strains in the presence of DTT (GD/GC and LD/LC) but upregulated by Log2fold change factor of 1.85 in the presence of lactose in the Δres2 strain only. This points to an eventual interaction between the two transcription factors in cellulase inducing conditions.

Table 2 Genes encoding putative transcriptional regulators differentially regulated in only one strain in the indicated condition