Evolution and screening of Trichoderma reesei mutants for secreted protein production at elevated temperature

Abstract

 

The filamentous fungus Trichoderma reesei is a mesophilic ascomycete commercially used to produce industrial enzymes for a variety of applications. Strain improvement efforts over many years have resulted not only in more productive hosts, but also in undesirable traits such as the need for lower temperatures to achieve maximum protein secretion rates. Lower fermentation temperatures increase the need for cooling resulting in higher manufacturing costs. We used a droplet-based evolution strategy to increase the protein secretion temperature of a highly productive T. reesei whole cellulase strain from 25°C to 28°C by first isolating an improved mutant and subsequently tracing the causative high-temperature mutation to one gene designated gef1. An industrial host with a gef1 deletion was found to be capable of improved productivity at higher temperature under industrially relevant fermentation conditions.

One-Sentence Summary

High-temperature droplet-based evolution resulted in the identification of a mutation in Trichoderma reesei gef1 enabling high productivity at elevated temperatures.

Introduction

Trichoderma reesei (teleomorph Hypocrea jecorina) is a mesophilic ascomycete that is used as a host to manufacture a wide range of secreted proteins, including both native and heterologous enzymes, with applications in a variety of industries (reviewed by Cherry & Fidantsef, 2003; Bischof et al., 2016; Liu et al., 2023). Early strain improvement efforts, starting with the original isolate QM6a, gave rise to strains such as RUT-C30 and RL-P37 that secreted higher levels of the native mix of cellulase and hemicellulase enzymes (Montenecourt & Eveleigh, 1979; Sheir-Neiss & Montenecourt, 1984). With the application of recombinant DNA techniques, it has been possible to manipulate the mix of native proteins produced. For example, genes encoding the major secreted proteins (CBHI/Cel7A, CBHII/Cel6A, EGLI/Cel7B, and EGLII/Cel5A) have been deleted and native proteins such as individual endoglucanases, xylanases, protease, or glucoamylase have been over-produced by International Flavors and Fragrances (IFF). Among the heterologous enzymes produced in T. reesei at IFF are cellulases, hemicellulases, alpha-amylases, lipases, and phytases. Typically, the strong cbh1 (cel7a) promoter is employed to drive high level, regulated expression of the gene of interest.

Strains QM6a, RUT-C30, and RL-P37 all grow fastest at temperatures of approximately 34°C. However, the optimum temperature for secreted protein production is 25°C–28°C (Suh et al., 1986). Industrial fermentation processes for T. reesei are typically fed-batch. The initial growth phase begins with a high, repressing concentration of glucose and is run at approximately 34°C. When glucose is exhausted and growth slows, the temperature is reduced to 28°C and an inducing feed is initiated at a rate dependent on the concentration of biomass. Maximal secreted cellulase production requires de-repression, an inducer, and slow growth rate (Pakula et al., 2005). The inducer may be cellulose, lactose, or a glucose/sophorose mixture generated by the action of cellulase on concentrated glucose (England et al., 2010).

We have continued to improve the RL-P37 strain lineage to enhance secreted protein production. A strain, designated A83, was isolated through mutagenesis and screening of RL-P37 that demonstrated significantly improved total protein productivity. A single point mutation in the nik1 gene, encoding a two-component histidine kinase, was shown to be responsible (Bodie et al., 2021a; manuscript in preparation). Strain T4 was isolated from A83 showing modest improvements in fermentation characteristics, but the responsible mutation has not been identified. Unlike RL-P37, for which the optimum temperature for protein production is 28°C, the optimum temperature for A83 and T4 is 25°C.

Industrial scale fermentation requires cooling to maintain temperatures as low as 25°C–28°C. Depending on geography this may incur significant capital and operating costs. Reduction or elimination of these expenses by raising the production temperature to 28°C or above without loss of productivity would reduce the cost of manufacturing. We developed a high-temperature droplet-based evolution strategy that has resulted in the isolation of mutant T4abc E1 that has elevated protein secretion at higher temperatures. The mutation responsible for this phenotype in the genome of mutant T4abc E1 has been identified as inactivation of gef1, encoding a Rho-type guanine exchange factor.

Materials and Methods

Strains

All strains were derived from RL-P37, a high-producing cellulase strain (Sheir-Neiss & Montenecourt, 1984), available at the USDA-ARS culture collection (NRRL, Peoria, IL, USA). The T. reesei strains used or generated in this study are listed in Table 1. RL-P37 nik1 was made by transforming RL-P37 with a nik1  ^M743T gene replacement cassette made by fusing a DNA fragment containing the 5´ region upstream of the nik1 locus, a loxP-flanked hygromycin B-resistance marker cassette and two DNA fragments containing the promoter and nik1 gene that includes the T to C substitution that changes the amino acid at position 743 from methionine to threonine. For further details, see Bodie et. al. (2021a).

Table 1.

Trichoderma reesei Strains Used in This Study

Strain name	Parent strain	Modification from parent
RL-P37	RUT NG14	UV mutagenesis for cellulase over-production
RL-P37 nik1	RL-P37	Replacement of wild-type nik1 with nik1M743T
T4	A-83	UV and NTG mutagenesis for cellulase over-production; mutation of nik1 to nik1M743T
T4abc	T4	Morphological changes not disclosed as proprietary industrial host
T4abc E1	T4abc	NTG mutagenesis for high-temperature cellulase over-production. Mutation of gef1
T4abcGEF1	T4abc	Deletion of gef1
T4abcGEF1 nik1 wt.	T4abcGEF1	nik1 M743T restored to wild type

T4abc is derived from T4, a high-productivity mutant isolated based on improved cellulase activity. T4 contains the nik1 mutation inherited from A83. T4abc is a proprietary industrial strain containing morphological changes similar to those discussed in Bodie et al. (2021b). T4abc has been used in another study involving the modification of transcription factor ACE (Luo et al., 2020). T4abc E1 is an evolution mutant derived from T4abc that has improved protein secretion at higher temperatures due to a mutation in gef1 (Bodie et al., 2024a). T4abcGEF1 is derived from T4abc and contains an inactive gef1 by insertion of pyr2 into the coding region using a Cas9-based method. For further details, see supplementary resource 1. T4abcGEF1 nik1 wt. is derived from T4abcGEF1 where nik1  ^M743T has been reverted to wild type using a Cas9-based method (supplementary resource 2). All strains used in this study were spore purified so that only pure homokaryotic strains were used for further strain characterization. Media used to maintain the strains include Vogels minimal medium (MM) agar (Vogel, 1956) and Bird-E agar (Metzenberg, 2004) incubated at 28°C for 4–7 days with alternate light/dark 12-hr cycles.

NTG Mutagenesis and Evolution Screening for High-Temperature Mutations

Using evolution methods described by Bachmann et al. (2013), enriched high-temperature mutant libraries were made using serial propagation in emulsions. T4abc spores were mutated with N-methyl-N´-nitro-N-nitrosoguanidine (NTG) until a 93% kill was obtained. Mutated spores were added to MM containing 4 g/l (NH₄)₂S0₄, 4.5 g/l KH₂PO₄, 1.0 g/l MgSO₄·7H₂0, 1.0 g/l CaCl₂·2H₂0, 0.01 NaCl, and 2.5 ml/l trace elements solution containing 175 g/l citric acid, 200 g/l FeSO₄·7H₂O, 16 g/l ZnSO₄·7H₂O, 3.2 g/l CuSO₄·5H₂O, 1.4 g/l MnSO₄·H₂O, and 0.8 g/l H₃BO₃. Cellulose was used as the sole carbon source. Attempts to use insoluble celluloses, such as microcrystalline cellulose or acid swollen cellulose, interfered with the formation of the emulsion, so soluble carboxymethylcellulose (CMC) (Sigma–Aldrich C5678, St. Louis, MO, USA) was used at 0.5%. Emulsion droplets encapsulating mutated spores were made by combining 700 µl HFE7500 oil (3M Novec), 300 µl 2.0 × 10⁶ CFU/ml mutated T4abc spores in MM containing CMC described above, and 0.8% vol/vol Pico Surf1 surfactant (Sphere Fluidics, Cambridge, UK) into a 15 ml conical centrifuge capped tube (Thermo Fisher Scientific, MA, USA). The emulsion was formed by vortexing the tube at 3 000 rpm for about 5 min. Subsequently, the excess oil phase was removed from the bottom of the tube. The emulsion was incubated statically at 31°C for 72 hr and subsequently broken by the addition of 300 µL of 1H, 1H,2H,2H-perfluorooctanol (Alfa Aesar, Heysham, UK). The evolved cells were recovered by washing in MM without a carbon source. Mutants were recovered following plating on Bird-E agar medium containing 0.5% CMC as the sole carbon source at 31°C for 7 days. Spores were resuspended on the agar plate by adding 10 ml of MM without a carbon source and the spores were recovered by centrifugation at 4 000 rpm for 5 min. The spores were washed by repeating this step three times. After the final wash, the spores were suspended in MM and re-encapsulated as described above. Serial propagation was repeated for 11 rounds and spores from surviving mutants were frozen in a 15% glycerol library. To verify that protein secretion was not adversely affected during evolution, a simple cellulase agar plate assay was used to assess cellulase secretion properties of a subsample of mutants in the evolution library. The library was screened using a cellulase plate assay developed by Montenecourt and Eveleigh (1977) except Bird-E agar plates were made with 15 g/l acid swollen cellulose as the sole carbon source. The library was diluted so that individual colonies were visible on the plates. Plates containing a total of 553 mutants were incubated at 31°C for 5 days. Subsequently, 500 mutants with the highest halo to colony size ratios were selected for more quantitative evaluation of protein secretion at 31°C in slow-release microtiter plates (srMTPs) described by Chan et al., 2018. Briefly, srMTPs are 24-well microtiter plates made of a polymer incorporating culture components, such as lactose, releasable into NREL culture media over time (Ouedraaogo et al., 2015). About 1 × 10⁶ spores of T4abc mutants or the T4abc parent were inoculated into 2 ml NREL medium in individual wells. Depending on the experiment, cultures were incubated between 25°C and 31°C, at 250 rpm, and 85% humidity for 4 days. Total protein was assayed on desalted supernatants (Zeba spin plates, Thermo Fisher Scientific, Rockford, IL, USA) using the Pierce BCA Assay Kit (Thermo Fisher Scientific, Rockford, IL, USA).

The Genomic Sequencing and Mutation Identification

To identify the range of mutations that were introduced into the parent T4abc genome, we obtained and analyzed the genome sequences of RL-P37, T4abc, and T4abc E1. Briefly, the genomic DNA was isolated using the Zymo QuickDNA Fungal/Bacterial Miniprep Kit (Zymo Research) for Illumina short read sequencing according to the manufacturer’s recommended procedure. The Illumina paired-end sequencing library was prepared according to the manufacturer's recommended protocol. The library was sequenced on HiSeq2000 (Illumina) for a 2x151 bp paired-end run. Illumina short reads were first mapped to the reference QM6a genome sequence published by the Joint Genome Institute (JGI) by BWA (Li and Durbin, 2009), and the Single Nucleotide Polymorphisms (SNPs) were called by samtools (Li, 2011) using default parameters. A pairwise SNP comparison between strains T4abc and T4abc E1 revealed that out of a number of SNPs not common to both strains, only one non-synonymous SNP was found in the coding sequence of a gene named gef1 in T4abc E1.

Evaluation of Protein Production in Fed-Batch Fermentations

Trichoderma reesei T4abc parental, T4abc E1 mutant, and T4abcGEF1 engineered strain were grown under identical liquid culture conditions in 2 l DASGIP fermenters (Eppendorf, Hamburg, Germany) as previously described (England et al., 2010).

To create a seed culture for the 2 l scale fermentations, spores from each strain were added separately to 50 ml of medium in 250-ml flasks with baffles. The medium contained 5 g/l (NH₄)₂SO₄, 4.5 g/l KH₂PO₄, 1 g/l MgSO₄·7H₂O, and 14.4 g/l citric acid, adjusted to pH 5.5 with 5% NaOH. After autoclaving for 30 min, sterile 60% glucose was added to a final concentration of 27.5 g/l along with 2.5 ml/l of the trace element solution described above. The cultures were grown for 48 hr at 28°C with shaking. The contents of each flask were added separately to 2 l fermenters containing 900 ml of medium containing 4.7 g/l KH₂PO₄, 1.0 g/l MgSO₄·7H₂O, 4.3 g/l (NH₄)₂SO₄, and 2.5 ml/l of the trace element solution. These components were heat sterilized at 121°C for 30 min prior to inoculation. Post-sterile additions included 145.6 ml of 50% glucose and 0.6 g/kg of CaCl₂. Thereafter, the temperature was maintained at 34°C and pH at 3.5. After batch glucose depletion, a glucose-sophorose feed was started to induce protein production, and the temperature was dropped to 25°C and pH was increased to 4.8 unless otherwise stated.

Post-fermentation, either total protein concentrations or specific productivity was determined and compared to the parent strain. For the latter, dry cell weight measurements were determined using methods described by England et al. (2010). Ultra performance liquid chromatography (UPLC) was used to compare changes in the individual cellulase components of T4abc and T4abc E1 using methods described by Mitchinson et al. (2012). Briefly, the column used was a ZORBAX 300 SB-C3 rapid resolution high-definition 2.1 × 100 mm, 1.8 μm column (Agilent BioLC Columns (PN:858750–909). The mobile phase A was composed of 0.1% (v/v) trifluoroacetic acid in water, while mobile phase B was 99.9% (v/v) acetonitrile and 0.1% (v/v) trifluoroacetic acid.

Samples were diluted in 10 mM NaAc pH 5.0 and 8 μg/ml EndoH to 0.8 μg/ml total protein and mixed in a thermomixer at 37°C, 400 rpm for 24 hr then filtered through a 0.2 µM filter.

Retention time of purified proteins including CBH1, CBH2, EG1, and EG2 were used as standards.

Results

Effect of a nik1 Mutation on Optimal Protein Secretion Temperature

To evaluate the effect of nik1^M743T on protein secretion at different temperatures, RL P-37 and RL P-37 nik1 were evaluated at 25°C and 28°C in DASGIP fermentations and total protein secretion was compared (Fig. 1).

Fig. 1.

Effect of histidine kinase nik1^M743T on total protein secretion at 25°C and 28°C by whole cellulase strains RL-P37 and RL-P37 nik1 in DASGIP fermentation for 140 hr. The mean ± standard deviation was calculated from independent experiments (N = 2–3). Two sample t-tests assuming equal variances identified statistically significant differences between RL-P37 nik1 at 25°C and 28°C (P ≤ 0.05). RL-P37 at 25°C and 28°C had similar total protein secretion (P ≥ 0.1).

RL-P37 has similar total protein secretion at both 25°C and 28°C. RL-P37 nik1 has the highest protein secretion at 25°C, but at 28°C, the total protein secretion decreases by about 15%. In a similar experiment, proprietary strain T4abc containing nik1^M743T also shows ∼20% decrease in total protein secretion at 28°C compared to 25°C (Fig. 2). Based on these results, it is evident that the nik1^M743T mutation is responsible for lowering the optimal protein secretion temperature from 28°C to 25°C.

Fig. 2.

Comparison of total protein secretion of T4abc and mutant T4abc E1 at 25°C, 28°C, and 31°C in srMTPs. The mean ± standard deviation was calculated from at least four biological replicates. Two sample t-tests assuming equal variances identified statistically significant differences between T4abc at 25°C compared to 28°C (P ≤ 0.05). T4abc E1 at 25°C and 28°C are similar (P ≥ 0.1). At 31°C, both T4abc and T4abc E1 have reduced protein secretion compared to conditions at 25°C–28°C (P ≤ 0.05).

Screening, Isolation, and Characterization of T4abc E1 Mutant

Adaptive evolution is a versatile tool that has been used successfully to improve properties of industrial strains for decades (Lin, et. al., 2017; Lee & Kim, 2020; Mavrommati, et. al., 2022; Wang et. al., 2023). For this study, evolution methods were designed to select for mutants capable of cellulase secretion under adverse temperature conditions enabling only efficient growers to survive multiple rounds of serial propagation.

Evolution in emulsion droplets has the advantage of eliminating cross feeding and other cell-to-cell interactions. Diffusion of molecules between droplets is also prevented by the outer oil phase. The selection occurs on cellulase-induced cells that have reached a filamentous growth stage ensuring that mutants are isolated based on growth resulting from the action of induced secreted enzymes on the cellulose substrate (Samlali et al., 2022).

Prior to encapsulation in droplets, T4abc was mutated and about 1 × 10⁶ spores/ml were diluted so that approximately no more than one spore would be present in a droplet as estimated by Poisson distribution and confirmed by microscopy. To achieve this, typically at least 80% of emulsion droplets were observed to be empty or contained non-growing spores, similar to results reported by Luu et al. (2023).

The disadvantage of encapsulating spores by vortexing is that the emulsion contains polydisperse droplets that contain different amounts of nutrients and have different maximum incubation times before polar hyphal growth ruptures the droplet. Therefore, harvest time was standardized to approximately 72 hr based on the presence of filamentous growth in droplets measuring 100–120 µm in stable emulsions. The altered morphology of strain T4abc decreased hyphal extension, increased branching and enabled greater biomass accumulation prior to rupture of droplets. After 11 rounds of evolution, libraries containing enriched T4abc mutants were characterized for total protein production.

Screening and Small-Scale Protein Evaluation

Cellulase secretion properties of 553 evolved mutants were evaluated using a cellulase agar plate assay. Mutant halo to colony size ratios were compared to the T4abc parent. About 96.5% of mutants produced similar ratios to the T4abc parent, 3% produced slightly higher ratios, and 0.5% produced lower ratios. For this screen, halo to colony ratio measurements are indicative of cellulase secretion; however, ratios can be affected by differences in growth rate and changes in morphology are both likely to be present in mutant populations. Therefore, 500 mutants with halo to colony size ratios at least comparable to the T4abc parent were selected for more quantitative evaluation of protein secretion at higher temperatures in small scale srMTP fermentations.

To identify mutants capable of protein production at higher temperatures, the srMTP plates were incubated at 31°C, which is suboptimal for cellulase production and, therefore, total protein production by the parent. After srMTP fermentation, 5 mutants out of 500 evaluated produced more total protein than the T4abc parent. After further evaluation of the five mutants in DASGIP fermentations, only one mutant produced more secreted protein than the parent at 31°C. This mutant designated T4abc E1 was further characterized for protein secretion at 25°C, 28°C, and 31°C in srMTPs (Fig. 2). At 25°C, T4abc and T4abc E1 produced similar levels of total secreted protein. However, at the higher temperatures of 28°C and 31°C, T4abc E1 compared to T4abc parent produced more than 22% and 14%, respectively. There was no significant loss of total secreted protein production by T4abc E1 when cultured at 28°C versus 25°C, but reduced productivity was observed at 31°C.

UPLC was used to compare the major cellulase components, including CBH1, CBH2, EG1, and EG2, produced by T4abc and T4abc E1 after fermentation at 31°C for 133 hr. In Fig. 3, mutant T4abc E1 has a similar cellulase profile as the T4abc parent. This indicates that the high-temperature improvement in mutant T4abc E1 is due to an increase in overall secreted protein while maintaining a similar secreted protein ratio.

Fig. 3.

UPLC of T4abc and T4abc E1 from a fermentation incubated at 31°C for 133 hr, using methods described by Mitchinson et al. (2012). Cellulase components CBH1, CBH2, EG1, and EG2 were identified based on UPLC results of purified enzymes.

Identification of Causative High-Temperature E1 Mutation

Analysis of the genome sequence of strain T4abc E1 identified a mutation in a gene annotated as estExt_fgenesh5_pg.C_30459 that resides on chromosome 3 at position 1532639–1538781 in the wild-type T. reesei QM6a v2.0 genome sequence assembly (https://mycocosm.jgi.doe.gov/Trire2/Trire2.home.html). The deduced amino acid sequence (PID:120482) includes a region of sequence homology to the Rho-type guanine nucleotide exchange factor domain (RhoGEF domain; Pfam PF00621, also called Dbl-homologous or DH domain). This homology region is at amino acid residue positions 1 242–1 454 in the deduced amino acid sequence. The Rho/Rac/Cdc42 family GTPases regulate many different cellular processes and are activated through release of bound GDP and subsequent binding of GTP, which is catalyzed by guanine exchange factors in the RhoGEF family (Harris, 2011). The mutation in the variant T4abc E1 strain replaces a glutamate codon at amino acid position 1 245 with a premature stop codon, thereby truncating the protein near the beginning of the RhoGEF domain. We have named this T. reesei gene and protein gef1/GEF1.

Unlike some other RhoGEF domain proteins, no Plekstrin homology domain (PH domain, Pfam PF00169) is found in T. reesei GEF1. The PH domain can bind phosphatidylinositol in membranes and the beta/gamma subunits of heterotrimeric G proteins or protein kinase C, and thereby is associated with signal transduction pathways (Lemmon & Ferguson, 2000). There are four other deduced protein sequences (PIDs 4 751, 75 210, 80 277, and 123 030) found in the T. reesei QM6a genome that contain RhoGEF domains, of which two (PIDs 4 751 and 123 030) also contain PH domains.

The GEF1 protein also has a region of homology to the BAR or AH/BAR domain (Pfam PF03114) at residues 1 576–1 677. The BAR domain is found in a variety of proteins. In some cases, it has been shown to interact with membranes, dimerize or to bind Rho family GTPases (Peter et al., 2004). A BAR domain is not observed in the other four T. reesei RhoGEF domain proteins.

The deduced amino acid sequence of GEF1 (supplementary  resource 3) also includes a C-terminal sequence (residues 1 918–1 990) that is highly conserved in some other filamentous fungal RhoGEF domain proteins, including from Fusarium fujikuroi, Fusarium graminearum, Thielavia terrestris, Podospora anserina, Penicillium rube ns, and Aspergillus niger. However, this C-terminal sequence domain of unknown function is not present in the other four RhoGEF domain proteins identified in the T. reesei genome.

T4abc E1 and T4abcGEF1 were evaluated in srMTP at 28°C to compare total protein secretion levels. In Fig. 4, it is evident that T4abc E1 and T4abcGEF1 are making similar amounts of total secreted protein verifying that the GEF1 truncation or deletion has similar effects on secreted protein at 28°C and that the mutation in GEF1 is causative for the high-temperature protein secretion observed in mutant T4abc E1.

Fig. 4.

Comparison of mutant T4abc E1 containing the gef1 truncation and T4abcGEF1 containing a gef1 deletion at 28°C in srMTPs. The mean ± standard deviation was calculated from at least three biological replicates.

T4abcGEF1 was further characterized by measuring total protein secretion over a temperature range of 25°C, 27°C, 28°C, 29°C, and 30°C in DASGIP fermentations (Fig. 5). Total protein secretion is similar at 25°C–28°C. At 29°C–30°C, there is a sharp drop in total protein secretion of about 40% (P ≤ 0.05).

Fig. 5.

Comparison of protein secretion of T4abcGEF1 at 25°C–30°C in DASGIP fermentations at 140 hr. The mean ± standard deviation was calculated from independent experiments (N = 3 at 25°C–28°C, N = 2 at 29°C–30°C). Two sample t-tests assuming equal variances identified statistically significant differences between T4abcGEF1 at 25°C–28°C compared to 29°C and 30°C (P ≤ 0.05). Due to the proprietary nature of T4abcGEF1, total protein secretion of T4abcGEF1 at 25°C was arbitrarily set to 1 and the relative amounts of protein at each condition appropriately adjusted and plotted.

To study the effect of GEF1 and nik1^M743T interactions, the nik1 mutation was reverted to wild type and total protein secretion was measured in DASGIP fermenters at temperatures from 25°C to 30°C (Fig. 6).

Fig. 6.

Comparison of protein secretion of T4abcGEF1 nik1 wt. at 25°C–30°C in DASGIP at 140 hr. The mean ± standard deviation was calculated from independent experiments (N = 2–3). As T4abcGEF1 nik1 wt. is a proprietary strain, the total protein produced at 25°C at 140 hr was arbitrarily set at 1 and the relative amounts of protein at each condition appropriately adjusted and plotted.

As shown in Fig. 6, protein secretion by T4abcGEF nik1 wt. improves 10%–20% at increasing temperatures from 25°C to 28°C. At 30°C, protein secretion decreases by 20%.

Lastly as specific productivity is an important parameter for the evaluation of new hosts for industrial fermentations, for the key strains used in this study specific productivity was compared at 28°C. Parent T4abc, mutant T4abc E1, and T4abcGEF1 were evaluated in DASGIP fermentations at 28°C for 200 hr. As a control, fermentations of T4abc were included at 25°C, which is the optimal temperature for protein production by this strain. Results in Fig. 7 indicate that T4abc E1 and T4abcGEF1 have similar productivity at 28°C to T4abc at 25°C. For T4abc fermentations at 28°C, productivity starts to decline around 100 hr and at 200 hr is reduced by almost 30%. High protein secretion by T4abc combined with the higher temperature of 28°C may result in the induction of stress responses resulting in the strain switching from protein secretion to growth late in the fermentation.

Fig. 7.

Comparison of specific productivity of T4abc, T4abc E1, and T4abcGEF1 in DASGIP fermentations at 28°C for 200 hr. T4abc at 25°C is included as a control for maximum productivity achievable by this host. Specific productivity is similar for T4abc E1 and T4abcGEF at 28°C comparable to T4abc at 25°C. For T4abc at 28°C, productivity is 30% lower by the end of the fermentation at 200 hr. These results indicate the optimal temperature for high-productivity host T4abc has been increased from 25°C to 28°C by deletion of gef1 in the new T4abcGEF1 host. The mean ± standard deviation was calculated from independent experiments (N = 3). Relative productivity data are based on T4abc at 25°C arbitrarily being set to 1 at the end of the fermentation. The remaining data sets were adjusted accordingly.

Discussion

Droplet microfluidics is a versatile technique useful in many applications including the development of high-throughput screens to improve secretion of products by many microorganisms (Bachmann et al., 2013; Huang, et al., 2015; He et al.,2022). For T. reesei and other fungi, droplet microfluidics has been used to improve productivity by monitoring cellulase activity within droplets using fluorescent substrates (He et al., 2019; Luu et al., 2023), and by combining microfluidics with flow cytometry (Li et al., 2023).

Here, we used a simple droplet-based evolution screen selecting for growth based on secretion of cellulases under non-optimal high-temperature conditions. This evolution screen was successful at isolating mutant T4abc E1 capable of optimal protein production at higher temperatures compared to the T4abc parent strain. The temperature improvement essentially restored the optimal temperature to that of the RL-P37 ancestor prior to the addition of the high-productivity nik1 mutation into the lineage. Recent work has shown that by using similar evolution methods combined with high-throughput screens, it is possible to go to even higher temperatures outside the normal range reported for T. reesei, while still maintaining full productivity (Bodie et al., 2024b).

The evolution screen described here could be further improved using microfluidics to produce larger, more uniform droplets (Luu et. al, 2023), and subsequently screened using high-throughput methods capable of detecting increases in biomass within the droplet. This evolution screen could also be used to improve other important industrial strain characteristics including productivity and yield. The causative mutation in strain T4abc E1 that enables increased productivity at elevated temperature was shown to be a premature stop codon within a coding sequence (CDS) that encodes PID:120482. We have named this T. reesei gene and protein gef1/GEF1 due to similarity with proteins in other fungal genomes. GEF1 is a 1990 amino acid protein that includes a RhoGEF domain (Pfam PF00621). This domain is found in guanidine nucleotide exchange factors (GEFs), which activate members of the Rho family of small GTPases by stimulating release of bound GDP and subsequent binding of GTP. In turn, Rho GTPases regulate many different cellular processes (Harris, 2011).

Filamentous fungal genomes contain multiple gene sequences with homology to Rho GTPases and GEFs with the best studied being Cdc42 and Rac1 GTPase homologs and the GEF Cdc24 (Araujo-Palomares et al., 2011; Harris, 2011; Kwon et al., 2011). In these fungi, both Cdc42 and Rac1 play important roles in hyphal polarity and morphogenesis. However, the exact role of each differs between species. Neither cdc42 nor racA are essential in Aspergillus nidulans but are synthetically lethal showing that they share at least one essential function (Virag et al., 2007). In this fungus, Cdc42 is involved in polarity and the formation of lateral branches but loss of RacA has little effect on hyphal morphology and the cytoskeleton and Spitzenkörper appear unaffected. GEFs encoded by the A. nidulans cdc24 and fubA genes interact with Cdc42 and RacA, respectively. While cdc24 is an essential gene, deletion of fubA has only minor effects on morphology (Si et al., 2016). In contrast, in Neurospora crassa, loss of either CDC-42 or RAC1 causes severe morphological defects with loss of apical polarity and the Spitzenkörper. CDC-24, which is essential for viability, seems to be the primary GEF that interacts with both GTPases (Araujo-Palomares et al., 2011). In A. niger, RacA plays the more important role in hyphal growth maintaining polarity, while loss of CftA (Cdc42) has little effect on hyphal morphology although, again, the two genes are synthetically lethal (Kwon et al., 2011). Loss of RacA caused hyper-branching but resulted in a reduced number of secretory vesicles per tip leading to no change in the amount of secreted glucoamylase (Kwon et al., 2013). However, when the glucoamylase gene was overexpressed, higher levels of secreted protein could be achieved in a racA deletion strain than in the parent strain (Fiedler et al., 2018). Similarly, deletion of rac1 in T. reesei, altered strain morphology and increased secreted cellulase production under certain culture conditions (Fitz et. al., 2019). On the other hand, deletion of Cdc42 in T. reesei resulted in a growth defect and impaired cellulase production, while overexpression showed increased expression and diversity of extracellular proteins (Jiang et al., 2023).

Like other filamentous fungi, and unlike yeast, T. reesei has both CDC42 and RAC1 Rho GTPases (PID:50335 and PID:47055, respectively). We have examined sequence similarities between T. reesei GEF1 and RhoGEFs in A. nidulans, A. niger, and N. crassa. Reciprocal BLASTP searches showed that the closest homolog in A.nidulans is PID:2393 in the JGI mycocosm database (encoded by ANID_10442), named FUBA by Si et al. (2016) and shown to interact with RacA. In A. niger, NRRL3 and N. crassa reciprocal BLASTP searches showed the closest homologs are PID:1473 and PID:2497, respectively in the JGI mycocosm database, which are distinct from the Cdc24 homologs. Based on the sequence similarities, we would predict that GEF1 interacts with, and activates, RAC1. Sequence similarities would also predict that the T. reesei RhoGEF PID:4751 is an ortholog of Cdc24, which interacts with Cdc42. Loss of GEF1 activity might be expected to confer a similar phenotype to loss of RAC1, including altered hyphal morphology (Fitz et al., 2019). We have not observed an obvious alteration in morphology as a result of inactivation of gef1, but this may be because the T4abc strain used in this study incorporates other mutations that affect morphology (Bodie et. al., 2021b). Conversely, it is possible that mutation of rac1 could influence the temperature optimum for secreted protein production similar to that of the gef1 mutation, but we have not tested this.

As described above, the effect of rac1 and gef1 mutations varies between different species of filamentous fungi, and it is not possible to predict the effect for T. reesei. However, there seems to be a general theme that regulatory networks related to hyphal morphology, cell wall integrity and osmosensing can affect secreted cellulase productivity by T. reesei (Wang et al., 2013; Lv et al., 2015; Wang et. al., 2018; Bodie et al., 2021b). Carbon source and cellulase inducer sensing networks may also be inter-related.

Inactivation of the T. reesei gef1 gene was initially identified in a strain that also had a mutation in the nik1 gene encoding a two-component histidine kinase thought to be part of the HOG1 pathway and response to osmotic stress. The nik1 mutation enabled increased secreted protein productivity but was also responsible for a reduction in the optimum temperature for secreted protein production. In the presence of mutated nik1, mutation of gef1 restored the optimum temperature of production to 28°C as observed in parental strain RL-P37. However, in the absence of mutated nik1, mutation of gef1 did not raise the optimum temperature of production above that of strain RL-P37 (Fig. 6) suggesting that NIK1 and GEF1 proteins may operate in interacting pathways.

Our research demonstrated that it is possible to increase the production temperature in industrially relevant T. reesei strains without an impact on secreted protein levels. Identification of mutations that raise the optimum temperature of protein secretion still further would be valuable.

Author Contributions

E.B. initiated the study, designed experiments, preformed evolution, and strain evaluation in small scale fermentations. Z.C. preformed Bioinformatics and high-temperature causative gene identification. K.C. developed srMTP methods used in this study and preformed small scale fermentation and data analysis. C.L. preformed strain construction of nik1 wild-type strain. C.L. evaluated strains in fermenters for protein production and analyzed fermentation experiments. S.S. performed UPLC and data analysis. M.V. preformed gene cloning, strain construction, and high- temperature causative gene identification. E.B. and M.W. drafted the manuscript. All authors read and approved the final manuscript.

Funding

None declared.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary material with the exception of proprietary datasets as indicated. These datasets are not public but are available from the corresponding author upon reasonable request.