Inactivation of Cellobiose Dehydrogenases Modifies the Cellulose Degradation Mechanism of Podospora anserina

ABSTRACT

Conversion of biomass into high-value products, including biofuels, is of great interest to developing sustainable biorefineries. Fungi are an inexhaustible source of enzymes to degrade plant biomass. Cellobiose dehydrogenases (CDHs) play an important role in the breakdown through synergistic action with fungal lytic polysaccharide monooxygenases (LPMOs). The three CDH genes of the model fungus Podospora anserina were inactivated, resulting in single and multiple CDH mutants. We detected almost no difference in growth and fertility of the mutants on various lignocellulose sources, except on crystalline cellulose, on which a 2-fold decrease in fertility of the mutants lacking P. anserina CDH1 (PaCDH1) and PaCDH2 was observed. A striking difference between wild-type and mutant secretomes was observed. The secretome of the mutant lacking all CDHs contained five beta-glucosidases, whereas the wild type had only one. P. anserina seems to compensate for the lack of CDH with secretion of beta-glucosidases. The addition of P. anserina LPMO to either the wild-type or mutant secretome resulted in improvement of cellulose degradation in both cases, suggesting that other redox partners present in the mutant secretome provided electrons to LPMOs. Overall, the data showed that oxidative degradation of cellulosic biomass relies on different types of mechanisms in fungi.

IMPORTANCE Plant biomass degradation by fungi is a complex process involving dozens of enzymes. The roles of each enzyme or enzyme class are not fully understood, and utilization of a model amenable to genetic analysis should increase the comprehension of how fungi cope with highly recalcitrant biomass. Here, we report that the cellobiose dehydrogenases of the model fungus Podospora anserina enable it to consume crystalline cellulose yet seem to play a minor role on actual substrates, such as wood shavings or miscanthus. Analysis of secreted proteins suggests that Podospora anserina compensates for the lack of cellobiose dehydrogenase by increasing beta-glucosidase expression and using an alternate electron donor for LPMO.

INTRODUCTION

Cellobiose dehydrogenases (CDHs; EC 1.1.99.18) are extracellular fungal enzymes that catalyze the oxidation of cellobiose into cellobiono-1,5-lactone (1). CDHs sometimes harbor one flavin-containing dehydrogenase (FAD-DH) domain from the AA3_1 family, but in most cases both an FAD-DH domain and a cytochrome domain from the AA8 family (www.cazy.org) are present (2). FAD-DH oxidizes cellobiose at the C-1 position to cellobiolactone by reduction of FAD, followed by its reoxidation thanks to acceptors such as quinones and phenoxy radicals. CDHs are secreted by a variety of fungi, including white-rot, brown-rot, and soft-rot fungi, as well as molds (1), upon growth on cellulose and plant biomass (3, 4). The first CDHs, initially named cellobiose oxidase, were discovered as enzymes able to catalyze the cellobiose-dependent reduction of quinones (5, 6). Their enzymatic properties have been characterized in many fungi, including Phanerochaete chrysosporium (7), Trametes versicolor (8), Schizophyllum commune (9), Pycnoporus cinnabarinus (3), Pycnoporus sanguineus (10), Cerrena unicolor (11), Podospora anserina (12, 13), and Neurospora crassa (14–17). Zámocký et al. (18) divided the CDH family into three phylogenetic branches. Class I consists of CDH from basidiomycetes and does not contain a family 1 carbohydrate binding module (CBM1), although most CDHs bind to cellulose (14). Class II contains only ascomycete CDHs that do (class IIA) or do not (class IIB) harbor a CBM1 module.

CDH appears to be involved in the degradation of both lignin and cellulose by fungi (1, 8, 9, 15, 16, 19, 20). Nevertheless, their actual function is still not fully understood (1, 9, 15, 16). Recently, fungal CDHs were shown to act as a reductant for lytic polysaccharide mono-oxygenases (LPMOs), providing electrons for the redox-mediated oxidative cleavage of cellulose (14, 17, 21–23). The electrons are rapidly transferred from the flavin to the heme domain of CDH via the heme propionate group (24–26). Following heme reduction, the cytochrome domain of CDH reduces the copper in the active site of the LPMO to initiate oxidative cellulose breakdown (27).

Although large amounts of data are available on the biochemistry of these enzymes, little is known about their genetics. In basidiomycetes, a CDH gene has only been inactivated in T. versicolor (28), and the single CDH of this species was found to be important for wood invasion but not for kraft pulp delignification. Recently, Phillips et al. (14) and Zhang et al. (15) analyzed the phenotypes of the deletions of cdh1 and cdh2 genes in N. crassa and found a reduction in CDH activity. However, they did not investigate thoroughly the effect of the deletion on fungal physiology and on plant biomass degradation. To gain insight into the actual role of CDH, we performed a thorough analysis of the role of the three enzymes encoded by the genome of the filamentous ascomycete P. anserina. This filamentous fungus is easy to manipulate and is able to rapidly complete its life cycle with wood as the sole carbon source (29–31). Gene deletions and construction of multiple mutants can rapidly be achieved. Moreover, genome sequencing (30) has revealed that its genome contains a large number of genes putatively involved in lignocellulose breakdown. Comparison with other fungi revealed that P. anserina has a large number of such genes, like basidiomycetes (2, 30). Being genetically tractable and having a complement of genes implicated in lignocellulose breakdown similar to that of basidiomycetes make it a good model to decipher how fungi manage to break down lignocellulose. Here, we investigated the in vivo role of three P. anserina genes encoding CDHs by targeted gene deletion and analyzed phenotypes and secretomes of the mutants.

RESULTS AND DISCUSSION

Sequence analyses of the three CDHs of Podospora anserina.

As in previous analyses of the first version of the P. anserina genome sequence (12), analysis of a corrected version of the complete genome of P. anserina (32) revealed the presence of three sequences potentially coding for CDHs, Pa_7_2650 or P. anserina CDH1 (PaCDH1) (formerly CDHpa) (12), Pa_0_280 or PaCDH2 (formerly PaCDHB) (13), and Pa_6_11360 or PaCDH3 (Fig. 1). A domain search revealed that PaCDH2 lacked a CBM1 module and thus belonged to class IIB, while PaCDH1 contained a CBM1 module at its C terminus and thus belonged to class IIA. These two CDHs were previously characterized as able to oxidize cellobiose (12, 13) and to provide electrons to P. anserina AA9 LPMOs for efficient cleavage of cellulose (13). Intriguingly, PaCDH3 does not contain an AA8 cytochrome domain and lacks a CBM1 module. Lack of an AA8 cytochrome domain is not unprecedented, as an active CDH missing this domain thanks to proteolytic cleavage has been purified from T. versicolor (8). SignalP (33) and Predisi (33) predicted a secretion signal for PaCDH1 and PaCDH2 but not for PaCDH3. Analysis of a P. anserina expressed sequenced tag database (30), transcriptome sequencing (RNA-seq; R. Debuchy, personal communication), as well as microarray data (34) validated the annotation of the three genes and showed that they were expressed but not regulated during mycelium growth. However, expression of PaCDH1 was significantly increased in mutants of the PaNox1, PaMpk1, and PaMpk2 genes; PaCDH3 expression was significantly increased in mutants of PaMpk1 and PaMpk2 genes and PaCDH2 in mutants of the PaMpk2 genes. PaNox1, PaMpk1, and PaMpk2 belong to a signaling pathway involved in numerous aspects of the P. anserina life cycle, including regulation of biomass degradation, and microarray analysis of the transcriptomes of these mutant are available (34).

FIG 1.

Modular organization of Podospora anserina CDHs. Domain organization of the three P. anserina CDHs. AA8, cytochrome domain from the AA8 family; FAD, FAD-DH domain; CBM1, carbohydrate binding module 1.

Exploration of the genomes of other filamentous fungi showed that CDH genes were not present in the available genomes of basal fungi, i.e., Chrytridiomycota, Neocallimastigomycota, Blastocladiomycota, Kickxellomycotina, Mucoromycotina, Entomophthoromycotina, and Glomeromycota. They were also lacking in the available genomes of Taphrinomycotina, Saccharomycotina, Pucciniomycotina, and Ustilaginomycotina. Moreover, they were not present in all genomes from Pezizomycotina and Agaricomycotina, including some species known to be excellent wood degraders, such as Fomitopsis pinicola and Neolentinus lepideus. Phylogenetic analysis performed with a selected group of CDHs (Fig. 2) showed that PaCDH1 was orthologous to Neurospora crassa CDH1, also a class IIA enzyme, and PaCDH2 was orthologous to CDH2, a class IIB one. PaCDH3 has no orthologous gene in N. crassa, which only has two enzymes, and clustered with those from Magnaporthe grisea (Maggr1I113216) and Daldinia eschscholzii (DalEC12I83623). Unlike PaCDH3, Maggr1I113216 has a predicted AA8 cytochrome domain, while DalEC12I83623 does not. Note that M. grisea has four additional genes encoding putative CDH, and only one of those has the cytochrome binding domain while the three others do not. All of the M. grisea CDHs have predicted secretion signals, but DalEC12I83623 does not. Loss/gain of the AA8 domain, associated with loss or retention of the secretion signal, thus seems a frequent event during fungal evolution.

FIG 2.

Phylogenetic tree of selected fungal CDHs. The Podospora anserina CDHs are boxed.

Phenotypic analyses of CDH-deleted strains.

To explore the role of PaCDH1, PaCDH2, and PaCDH3, the three genes were inactivated by targeted gene deletion and the mutants were validated by Southern blotting (see Fig. S1 in the supplemental material). Three strains, PaCDH1^Δ, PaCDH2^Δ, and PaCDH3^Δ, inactivated for PaCDH1, PaCDH2, and PaCDH3, respectively, were obtained. To explore possible functional redundancy, the PaCDH1^Δ PaCDH2^Δ, PaCDH1^Δ PaCDH3^Δ, and PaCDH2^Δ PaCDH3^Δ double mutants, as well as the PaCDH1^Δ PaCDH2^Δ PaCDH3^Δ (CDH^Δ) triple mutant lacking all of the CDH genes, were constructed by genetic crosses thanks to the different selection markers used to inactivate each gene. A thorough phenotypic analysis then was performed and compared to the wild type and the CAT^ΔΔΔΔΔ mutant inactivated for all five catalase genes present in the P. anserina genome (35). Growth and fertility of the strains were first evaluated on M2, the minimal medium used to assess P. anserina development. Because formation of fruiting bodies requires energy, assessing not only growth but also fertility enabled us to assess whether the mutants were as proficient as the wild type at retrieving nutrients from the offered food sources. The M2 medium contains dextrin as the sole carbon source. As seen in Fig. 3A, the growth rate was not altered in any of the CDH mutants on this medium; fertility of all of the mutants, including germination efficiency of the ascospores, also was not modified in this medium (Fig. 3B).

FIG 3.

Growth rate (A) and fertility (B) of the wild type and CDH mutants on M2 containing dextrins as a carbon source. (A) Diameters of cultures 3 and 5 days after inoculation. Data are averages from three cultures. (B) On this medium, fruiting bodies are formed along a ring that centers on the inoculation point. The fruiting bodies (inset on the wild type) are the small black dots.

Growth and fertility were then assayed on media containing various carbon sources, including glucose, crystalline and fibrous cellulose, alkali-sulfonated lignin, Guibourtia demeusi wood shavings, and shredded miscanthus. On glucose medium, growth of wild-type P. anserina is slowed and fertility severely diminished. As seen in Fig. S2, no modification of growth rate was observed on glucose medium in any of the CDH mutants, and their fertility, but not the germination efficiency of the ascospores, was diminished as described for the wild type. On wood shavings and miscanthus, we did not detect any obvious phenotype (Fig. S3), suggesting that CDH is dispensable for scavenging nutrients from these carbon sources. This is different from the results of Dumonceaux et al. (28), who observed an inability to grow on wood. Unfortunately, Southern blot analysis did not provide full confirmation that inactivation of the single CDH gene was responsible for the phenotype in the Dumonceaux et al. paper, leaving the possibility that lack of growth on wood was due to an additional genetic defect. Alternatively, the difference may be due to the fact that ascomycetes and basidiomycetes rely on different levels of (compensated or uncompensated) CDH activity. Moreover, because both wood shavings and miscanthus are very heterogeneous, growth rates, which can be different from the plate assay ones, cannot be measured, and small differences in fertility are somewhat problematic to evaluate correctly. Similarly, growth rate was not modified on alkali-sulfonated lignin, unlike that of the CAT^ΔΔΔΔΔ mutant, which was decreased as described previously (Fig. S3) (35). When assayed with paper which contained fibrous cellulose, fertility of all mutants also was not affected (Fig. S4). Lack of effect of the mutation on the ability to degrade paper was confirmed by evaluating the weight loss of paper pads, as we could not detect any statistically significant differences between the wild type and the analyzed mutants (Fig. S4). On the contrary, and interestingly, when assayed with crystalline cellulose, fertility as measured by the number of perithecia was diminished by a factor of two specifically for PaCDH1^Δ PaCDH2^Δ and CDH^Δ, while germination efficiency, growth rate, and mycelium density were not altered (Fig. 4). Owing to the lack of phenotype of the single mutants, we could not easily engage complementation experiments. Thus, we could not completely rule out that this phenotype was due to an additional unintended genetic defect. Nevertheless, the phenotype was observed in three independently constructed double mutants and three independently constructed triple mutants. Unaltered growth rate and mycelium density indicated that these mutants could not retrieve more nutrients from crystalline cellulose. Thus, the lower fertility of these mutants was due to impairment in their ability to scavenge nutrients from crystalline cellulose. Overall, we could not detect such dramatic effects as those created by deleting catalases (Fig. S3) (35) or laccases (36, 37), suggesting either a minor role of CDH in degradation of complex biomass or that P. anserina has compensatory mechanisms that enable it to cope with a lack of CDH. It is known from previous work that many compounds or proteins may be used as electron donors by LPMOs (38–40). Nonetheless, CDH activity is important in P. anserina for the degradation of crystalline cellulose, as seen by the lower fertility of the PaCDH1^Δ PaCDH2^Δ and CDH^Δ mutants when grown on this food source. Crystalline cellulose is present at various levels in the different biomasses eaten by herbivores, i.e., the growth substrates of this coprophilous fungus. Therefore, small decreases in growth and fertility of strains lacking CDH, not detected here in the wood shaving/miscanthus experiments because of their heterogeneous nature or because crystalline cellulose represents only a minor fraction of the total cellulose, could in the wild and in the long term have a large impact on fitness.

FIG 4.

Growth rate (A) and fertility (B) of the wild type and CDH mutants on M4 containing crystalline cellulose as the sole carbon source. (A) Diameters of cultures 3 and 5 days after inoculation. Data are averages from three cultures. (B) On this medium, fruiting bodies are also formed along a ring that centers on the inoculation point. However, this ring is larger than that on M2. It may vary from one experiment to the other, owing to the heterogeneous nature of the medium. Nevertheless, PaCDH1^Δ PaCDH2^Δ (CDH1^Δ CDH2^Δ) and CDH^Δ (CDH1^Δ CDH2^Δ CDH3^Δ) exhibited a constant diminution of fertility, as seen by the estimated number of perithecia (indicated at the bottom of each petri plate).

Because CDH may produce peroxide as a final product, we assayed production of peroxide and superoxide by wild-type and CDH mutant thalli using the diaminobenzidine (DAB) and nitroblue tetrazolium assays (NBT), respectively (41). We could not detect any differences among the mutants in the production of both peroxide and superoxide, as previously seen for the CAT^ΔΔΔΔΔ mutant (Fig. S5) (35). This suggested that CDHs were minor sources of reactive oxygen species (ROS) in P. anserina; likewise, catalase had a minor role in removing peroxide under the conditions used for the assays, as previously hypothesized (35). We also assayed with the DAB assay the production of ROS when P. anserina encounters Penicillium chrysogenum. At the contact point, P. anserina exerts hyphal interference and kills P. chrysogenum hyphae (42). As seen in Fig. S5, this process was not modified in the CDH mutants. Overall, it thus appears that CDHs have a minor role in peroxide/superoxide production in P. anserina, as detected here by the DAB and NBT assays.

Proteomic analyses of Podospora anserina secretomes.

Analyses of wild-type P. anserina and the CDH^Δ triple mutant was performed using a secretomic analysis approach. P. anserina was grown on lignocellulosic biomass (i.e., wheat straw) to favor the secretion of lignocellulose-acting enzymes. Enzymatic assays showed that the CDH activity was equal to 0.429 nkat/ml in the wild type and null in CDH^Δ. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analyses of secretomes allowed the identification of about 50 secreted proteins (Table 1). Surprisingly, the most abundant protein in both secretomes was a putative tannase/esterase (Pa_5_7020; GenBank accession number CAP65121). Growth of both P. anserina wild-type and CDH^Δ strains on wheat straw favored the secretion of lignocellulose-acting enzymes (e.g., AA1 and AA5) and CAZymes targeting plant cell wall components (e.g., GH3, GH6, GH7, and GH74). Among them, five harbored a CBM1 module, i.e., a GH7-CBM1 putative cellobiohydrolase (Pa_3_730; CAP61105), a putative GH74-CBM1 xyloglucanase (Pa_4_7820; CAP66717), a GH6-CBM1 cellobiohydrolase (Pa_0_1250; CAP60942) (43), a GH131-CBM1 broad-specificity glucanase (Pa_3_10940; CAP61309) (44), and, in the wild type, PaCDH1 (Pa_7_2650; CAP68427) (12). As expected, PaCDH1 was absent from the secretome of the CDH^Δ mutant strain (Table 1). The main striking difference between the wild-type and the CDH^Δ mutant secretomes concerned the GH3 family, which consists mainly of putative β-glucosidases. The CDH^Δ mutant secretome contained five different GH3 enzymes (CAP61089, CAP64656, CAP68372, CAP73569, and CAP60165), while the wild-type secretomes only contained the most abundant one (CAP61089) (Table 1). This observation could be explained by the fact that expression of β-glucosidases is induced by cellobiose, as observed in the related fungus Neurospora crassa (45). Therefore, CDH could participate in the regulation of the expression of β-glucosidases in P. anserina. Moreover, it has been suggested that CDH could enhance cellulase activity by relieving product inhibition of cellobiohydrolases through the oxidation of cellobiose to cellobionic acid (46). P. anserina may compensate for the lack of CDH with the secretion of β-glucosidases to favor GH6 and GH7 cellobiohydrolase activity by relieving product inhibition. This mechanism might in part explain the weak phenotypes of the CDH deletion mutants.

TABLE 1.

Comparative analysis of the CAZymes identified in the wild-type and CDH^Δ mutant secretomes^a

GenBank accession no.	Functional annotation	CAZy annotation	Total no. of spectra
			Wild-type strain	Mutant CDHΔ strain
CAP65121.1	Tannase and feruloyl esterase (pfam07519)		152	230
CAP61636.1	Putative aromatic peroxygenase		38	14
CAP68535.1	Catalase		32	35
CAP61105.1	Cellobiohydrolase	GH7-CBM1	48	12
CAP61378.1	FAD oxidoreductase		44	11
CAP68327.1	Beta-1,3-glucanase	GH55	14	32
CAP64719.1	Putative laccase	AA1_3	27	9
CAP61089.1	Beta-glucosidase	GH3	23	14
CAP60942.1	Cellobiohydrolase	GH6-CBM1	30	4
CAP66717.1	Xyloglucanase	GH74-CBM1	22	10
CAP73016.1	β-1,3-Glucanosyltransglycosylase	GH72-CBM43	27	3
CAP61475.1	Carboxyl ester hydrolase		14	9
CAP70294.1	Glucoamylase	GH15-CBM20	7
CAP61309.1	Beta-glucanase	GH131-CBM1	13	2
CAP60434.1	Copper radical oxidase	AA5_1		14
CAP66761.1	Glycanase	GH16	13	2
CAP65628.1	Copper oxidoreductase		13
CAP64656.1	Beta-glucosidase	GH3		8
CAP68372.1	Beta-glucosidase	GH3		5
CAP59879.1	α,α-Trehalase	GH37	5	6
CAP61625.1	Unknown function		5	2
CAP66532.1	Beta-1,3-glucanase	GH55	10
CAP61276.1	Unknown function		8	2
CAP65464.1	FAD oxidoreductase			3
CAP68427.1	Cellobiose dehydrogenase (PaCDH1)	AA3_1-AA8	9
CAP62084.1	FAD oxidoreductase		8
CAP70307.1	Beta-glucanase	GH17	8
CAP68194.1	Carboxyl ester hydrolase		8
CAP65285.1	GMC oxidoreductase	AA3_2		6
CAP61511.1	Peptidase s28 family		5
CAP65657.1	Beta-glucanase	GH17	3	3
CAP61726.1	Putative glutaminase (pfam08760)		4	3
CAP60040.1	Unknown function		5	2
CAP73527.1	Peptidase a4		2	2
CAP73079.1	Unknown function		2	3
CAP65694.1	Unknown function		3
CAP73947.1	Unknown function		4	2
CAP66439.1	Putative dehydrogenase		6
CAP65311.1	Unknown function			5
CAP68130.1	Chitinase	GH18-CBM50		2
CAP71075.1	Unknown function		2	3
CAP68366.1	Unknown function		2
CAP67351.1	Unknown function			3
CAP61335.1	FAD oxidoreductase		2	3
CAP61538.1	Unknown function		2	2
CAP68226.1	Protease s28			3
CAP73569.1	Beta-glucosidase	GH3		4
CAP70690.1	Putative aldose 1-epimerase		4
CAP72727.1	Glycanase	GH16	2	2
CAP61110.1	Unknown function		2
CAP70412.1	Glucoamylase	GH15-CBM20	3
CAP70942.1	Aspartic protease		3
CAP60165.1	Beta-glucosidase	GH3		3
CAP64871.1	GMC oxidoreductase	AA3	3
CAP73573.1	Unknown function			2
CAP65364.1	Unknown function		2
CAP62325.1	GMC oxidoreductase	AA3_2	2
CAP67022.1	FAD oxidoreductase		2

^aProteins are listed according to their putative function and CAZy annotation. Total spectra of unique peptide sequences identified by LC-MS/MS in each secretome are indicated.

Cellulose degradation potential of Podospora anserina secretomes.

To further evaluate the importance of CDH in the degradation of cellulose, wild-type and mutant secretomes were tested for their capacity to degrade cellulose. Both secretomes were able to efficiently convert cellulose into soluble products yielding significant amounts of glucose (DP1) and cellobiose (DP2). Glucose yields were more than 2-fold higher in the case of the CDH^Δ mutant secretome (10.0 μM) compared to the WT secretome (4.24 μM) (Fig. 5A). Addition of the AA9 LPMO from P. anserina (PaLPMO9H) (13) boosted the conversion of cellulose in both cases (wild-type and CDH^Δ mutant secretomes). Reaction mixtures contained oxidized cello-oligosaccharide products (C-4- and C-1/C-4-oxidized products), in agreement with the regioselectivity of PaLPMO9H (Fig. 5B). As no external electron donor was added to the enzymatic reaction, these results provide evidence that redox partners other than CDH are present in the CDH^Δ mutant secretomes, triggering oxidative cleavage of cellulose by PaLPMO9H. One can suggest gallic or tannic acids as a potential candidate. Indeed, these acids have been shown to be alternative electron donors for LPMOs (40), and the putative tannase/esterase that could be involved in the production of these molecules is present in the mutant secretome. We have thus tested the ability of tannic acid to restore a higher fertility to the CDH^Δ mutant by adding various quantities of this product to M4 plates. As a positive control we used ascorbate, a universal electron donor, and as a negative control we used dehydroascorbate (d-ascorbate), the oxidized form of ascorbate. As expected, ascorbate promoted a restoration of CDH^Δ fertility in a dose-dependent manner, while d-ascorbate did not (Fig. S6). Tannic acid was not able to restore fertility of CDH^Δ (Fig. S6), suggesting that it may not be the electron donor compensating for a lack of CDH on complex biomass. It was even toxic at high doses for P. anserina (Fig. S6, compare the fertility of the wild-type cross at 0.02 g/liter of tannic acid with that at lower concentrations). The genome of P. anserina encodes many other potential candidates that could act as electron donors (30). For example, the flavo-oxidase from the AA3_2 subfamily (Pa_5_5180; CAP65285) identified in the CDH^Δ mutant secretome (Table 1) could provide electrons to LPMOs, therefore triggering oxidative cellulose degradation. Phylogenetic prediction of the function of this flavo-oxidase is hampered by the small number of characterized members of the AA3_2 subfamily (47). Our findings are in agreement with recent findings showing that AA3_2 GMC oxidoreductases can serve as extracellular electron sources for LPMOs (38, 39), thus extending the array of fungal redox partners in filamentous fungi. Overall, the present data suggest that compensatory mechanisms indeed enable P. anserina to cope with lack of CDH for the most part.

FIG 5.

Enzymatic degradation of cellulose using fungal secretomes. Chromatograms highlight soluble products identified after enzymatic treatment with wild-type (WT) and CDH triple mutant (CDH^Δ) secretomes alone (A) and secretomes supplemented with PaLPMO9H (B). (A) WT, CDH^Δ, and PASC represent control conditions. The inset highlights the identification of minor oxidized and nonoxidized products. (B) Oxidized oligosaccharides are identified only under conditions where PaLPMO9H is incubated with either WT or CDH^Δ mutant secretome. Oxidized cello-oligosaccharides were identified based on previous analysis (14).

Conclusions.

Analysis by targeted gene deletion of the CDH genes of P. anserina has shown that these enzymes are important for degradation of crystalline cellulose, but that their absence does not result in a dramatic decrease in the ability to use complex biomasses, such as wood shavings or miscanthus, in which crystalline cellulose may represent a minor fraction. The efficiency of the mutant CDH^Δ secretome to degrade cellulose appeared similar to that of the wild type, except that twice as much cellobiose was obtained as expected from the lack of conversion by CDH. Analysis of oxidized products indicated that LPMOs are still active, suggesting that additional partners able to provide electrons to LPMOs are present in CDH^Δ. Based on its increased quantity in the CDH^Δ secretome, the AA3_2 flavo-oxidase Pa_5_5180 (CAP65285) may be a good candidate for such a task. These data underscore the complex mechanisms of plant biomass degradation by fungi in which many activities, some potentially redundant, act in concert to achieve efficient degradation. This is particularly important for coprophilous fungi, such as P. anserina, because they likely encounter biomasses of diverse origins, which have been more or less extensively digested by herbivores, but also by the other competing fungi that grow and fructify in succession on dung.

MATERIALS AND METHODS

Strains and growth conditions.

The strains used in this study derived from the S wild-type strain (48) used for sequencing (30, 32). Standard culture conditions, media, and genetic methods for P. anserina have been described (29, 49). The M2 medium has the following composition: 0.25 g/liter KH₂PO₄, 0.3 g/liter K₂HPO₄, 0.25 g/liter MgSO₄·7H₂O, 0.5 g/liter urea, 0.05 mg/liter thiamine, 0.25 μg/liter biotine, 2.5 mg/liter citric acid, 2.5 mg/liter ZnSO₄, 0.5 mg/liter CuSO₄, 125 μg/liter MnSO₄, 25 μg/liter boric acid, 25 μg/liter sodium molybdate, 25 μg/liter iron alum, dextrine 5 g/liter, 12.5 g/liter agar. M4 medium has the same composition except that dextrin is replace by crystalline cellulose (microgranular CC31 cellulose powder; catalog no. 401050; Whatman). M0 has the same composition as M2 except that it lacks dextrin, which can be replace by the same amount/weight of autoclaved alkali-sulfonated lignin, wood shavings, or shredded miscanthus. Alkali-sulfonated lignin was purchased from Sigma-Aldrich (catalog no. 471003). Perithecium numbers were estimated using dotcount (http://reuter.mit.edu/software/dotcount/).

Gene deletions and phenotypic analysis.

The PaCDH1, PaCDH2, and PaCDH3 genes were deleted as described in reference 29 using resistance markers to Geneticin (50), hygromycin B (51), and phleomycin (51), respectively. After verification by Southern blotting (see Fig. S1 in the supplemental material), one verified mat⁺ and one verified mat-deleted strain were picked for phenotypic analyses. Double and triple mutants were constructed by genetic crosses, and genotypes were ascertained thanks to the resistance markers. At least three independently constructed PaCDH1^Δ PaCDH2^Δ and CDH^Δ mutants were analyzed for fertility on M4. All presented the fertility defect. The primers used are described in Table S1. Phenotypes were analyzed as described previously for fertility on various carbon sources (35–37), for paper degradation (52), for constitutive peroxide and superoxide production (41), and for hyphal interference (42).

Phylogenetic analysis.

CDH genes of P. anserina were identified by BLAST using various fungal CDHs as the query. Alignment was made with MAFFT (53) and manually refined. This alignment was used to construct a phylogenetic tree using the maximum likelihood method (PhyML software) (54) and transferred to the iTOL server for visualization (55). Bootstrap values are expressed as percentages of 100 replicates. Signal peptides were predicted using SignalP 4.1 at http://www.cbs.dtu.dk/services/SignalP/ and Predisi at http://www.predisi.de.

Secretome preparation and CDH activity assay.

On the basis of previous studies, the fungal cultures were grown in a liquid medium containing 15 g liter⁻¹ (based on the dry matter) wheat straw (Triticum aestivum Apache, France) as a carbon source, 2.5 g liter⁻¹ maltose as a starter, 1.842 g liter⁻¹ diammonium tartrate, 0.5 g liter⁻¹ yeast extract, 0.2 g liter⁻¹ KH₂PO₄, 0.0132 g liter⁻¹ CaCl₂, and 0.5 g liter⁻¹ MgSO₄. Fungal cultures were prepared in biological triplicate using 200-ml baffled flasks containing 100 ml of culture medium. Cultures were inoculated with mycelial fragments from five fungal disks (4-mm diameter) crushed in 1 ml minimum medium, using a FastPrep-24 system (MP Biomedicals, Illkirch, France) set to 40 m s⁻¹ for 60 s. Liquid cultures were incubated at 25°C with orbital shaking at 140 rpm (Infors HT, Switzerland). All of the cultures were stopped at 4 days after inoculation. The culture broths (secretomes) were harvested and pooled (total volume, 250 to 300 ml), filtered (using a 0.2-μm-pore-size polyethersulfone membrane; Vivaspin; Sartorius, Germany), diafiltered, and concentrated (Vivaspin polyethersulfone membrane with a 10-kDa cutoff; Sartorius) in 50 mM acetate solution buffer, pH 5, to a final volume of 5 ml and then stored in aliquots at −20°C until use. The total amount of protein was assessed using Bradford assays (protein assay dye reagent concentrate; Bio-Rad, Ivry, France) with a bovine serum albumin (BSA) standard that ranged from 0.2 to 1 mg ml⁻¹. CDH enzymatic activity was assayed as described in reference 13.

Saccharification assays.

The concentrated secretomes were tested for their ability to hydrolyze phosphoric acid-swollen cellulose (PASC) suspension (0.1%, wt/vol) prepared in 50 mM acetate buffer, pH 5. The enzyme reactions were performed in 2-ml tubes and incubated in a thermomixer (Eppendorf, Montesson, France) at 50°C and 850 rpm. After 24 h of incubation, all of the samples were boiled at 100°C for 10 min to stop the enzymatic reaction and then centrifuged at 16,000 rpm for 15 min at 4°C to separate the soluble fraction from the remaining insoluble fraction before carbohydrate determination. Assays were performed as triplicate independent experiments. The reaction mixture (300 μl liquid volume) contained 3 μg of total protein of P. anserina secretome. For supplementation of secretomes with recombinant P. anserina LPMO, 2 μM purified recombinant PaLPMO9H (13) was added. Assays were performed as triplicate independent experiments.

Analysis of oxidized and nonoxidized oligosaccharides.

Monosaccharides, oligosaccharides, and their corresponding aldonic acid forms generated after PASC and Avicel cleavage were analyzed by ionic chromatography (high-performance anion-exchange chromatography) as described by Bennati-Granier et al. (13), using nonoxidized oligosaccharides (Megazyme) as standards. Corresponding C-1-oxidized standards (from DP2 to DP6) were produced from nonoxidized cello-oligosaccharides by CDH treatment as described by Bennati-Granier et al. (13). All assays were carried out in triplicate.

Proteomic analysis of secretomes.

Short SDS-PAGE runs (precast 4 to 12% Bis-Tris mini gels; Invitrogen, France) were performed, allowing proteins diafiltered from pooled biological triplicates of secretomes (10 μg) to migrate to a length of 0.5 cm, and gels were stained with Coomassie blue (Bio-Rad, Marnes-la-Coquette, France). Each one-dimensional electrophoresis lane was cut into two slices of gel (2 mm in width), and protein identification was performed using PAPPSO (Plateforme d'Analyze Protéomique de Paris Sud-Ouest) platform facilities. In-gel digestion was carried out according to a standard trypsinolysis protocol. Gel pieces were washed twice with 50% (vol/vol) acetonitrile (ACN), 25 mM NH₄CO₃ and incubated in the presence of 10 mM dithiothreitol (DTT) for 1 h at 56°C. After cooling, the supernatant was removed and the samples were incubated with 55 mM iodoacetamide at room temperature in the dark. Gel plugs were washed with ACN and then dried in a vacuum speed concentrator. Digestion was performed for 8 h at 37°C with 200 ng of modified trypsin (Promega, Charbonnières-les-Bains, France) dissolved in 25 mM NH₄CO₃. Tryptic peptides were first extracted with 50% (vol/vol) CAN and 0.5% (vol/vol) trifluoroacetic acid (TFA) and then with pure ACN. Peptide extracts were dried in a vacuum speed concentrator (Thermo Fisher Scientific, Villebon sur Yvette, France) and suspended in 25 μl of 2% (vol/vol) ACN, 0.05% (vol/vol) TFA, and 0.08% (vol/vol) formic acid. High-performance liquid chromatography (HPLC) was performed on a NanoLC-Ultra system (Eksigent, Les Ulis, France). Trypsin digestion products were first concentrated and desalted on a precolumn cartridge (PepMap 100 C₁₈; 0.3 by 5 mm; Dionex, Thermo Fisher Scientific) with 0.1% HCOOH at 7.5 μl min⁻¹ for 3 min. The precolumn cartridge was connected to the separating column (C₁₈; 0.075 by 0.15 mm; Biosphere Nanoseparations, Nieuwkoop, The Netherlands), and the peptides were eluted with a linear gradient from 5 to 35% ACN in 0.1% HCOOH for 40 min at 300 nl min⁻¹. On-line analysis of peptides was performed with a Q-exactive mass spectrometer (Thermo Fisher Scientific, USA), using a nanoelectrospray ion source. Ionization (1.8-kV ionization potential) was performed with a stainless steel emitter (30-μm inner diameter; Thermo Electron, Villebon sur Yvette, France). Peptide ions were analyzed using Xcalibur 2.1 (Thermo Scientific, Villebon sur Yvette, France) with the following data-dependent acquisition steps: step 1, full MS scan (mass-to-charge ratio [m/z], 400 to 1,400; resolution, 70,000); step 2, MS/MS (normalized collision energy, 30%; resolution, 17,500). Step 2 was repeated for the eight major ions detected in step 1. Dynamic exclusion was set to 40 s. The raw mass data were first converted to mzXML format with the ReAdW software (SPC Proteomics Tools, Seattle, WA). Protein identification was performed by querying MS/MS data against databases, together with an in-house contaminant database, using the X!Tandem software (X!Tandem Cyclone, Jouy en Josas, France) with the following parameters: one missed trypsin cleavage allowed, alkylation of cysteine and conditional oxidation of methionine, and precursor and fragment ion set at 2 ppm and 0.005 Da, respectively. A refined search was added with similar parameters, except that semitryptic peptides, possible N-terminal acetylation, and histidine mono- and dimethylations were searched. All peptides matched with an E value lower than 0.05 were parsed with X!Tandem pipeline software. Proteins identified with at least two unique peptides and a log(E value) of lower than −2.6 were validated.