Retracted and Republished from: “Substrate-Specific Differential Gene Expression and RNA Editing in the Brown Rot Fungus Fomitopsis pinicola”

ABSTRACT

Wood-decaying fungi tend to have characteristic substrate ranges that partly define their ecological niche. Fomitopsis pinicola is a brown rot species of Polyporales that is reported on 82 species of softwoods and 42 species of hardwoods. We analyzed gene expression levels of F. pinicola from submerged cultures with ground wood powder (sampled at 5 days) or solid wood wafers (sampled at 10 and 30 days), using aspen, pine, and spruce substrates (aspen was used only in submerged cultures). Fomitopsis pinicola expressed similar sets of wood-degrading enzymes typical of brown rot fungi across all culture conditions and time points. Nevertheless, differential gene expression was observed across all pairwise comparisons of substrates and time points. Genes exhibiting differential expression encode diverse enzymes with known or potential function in brown rot decay, including laccase, benzoquinone reductase, aryl alcohol oxidase, cytochrome P450s, and various glycoside hydrolases. Comparing transcriptomes from submerged cultures and wood wafers, we found that culture conditions had a greater impact on global expression profiles than substrate wood species. These findings highlight the need for standardization of culture conditions in studies of gene expression in wood-decaying fungi.

IMPORTANCE All species of wood-decaying fungi occur on a characteristic range of substrates (host plants), which may be broad or narrow. Understanding the mechanisms that allow fungi to grow on particular substrates is important for both fungal ecology and applied uses of different feedstocks in industrial processes. We grew the wood-decaying polypore Fomitopsis pinicola on three different wood species—aspen, pine, and spruce—under various culture conditions. We found that F. pinicola is able to modify gene expression (transcription levels) across different substrate species and culture conditions. Many of the genes involved encode enzymes with known or predicted functions in wood decay. This study provides clues to how wood-decaying fungi may adjust their arsenal of decay enzymes to accommodate different host substrates.

INTRODUCTION

Wood (lignocellulose) is one of the most abundant carbon pools in terrestrial ecosystems. Decomposition of wood is a critical component of the carbon cycle and may be exploited in the production of biofuels and other technologies. The most abundant and best-studied wood-decaying organisms are Agaricomycetes (Basidiomycota) (1–5). Most wood-decaying Agaricomycetes can be placed into one of two major decay categories: (1) white rot, in which all components of plant cell walls are degraded, including the highly recalcitrant lignin fraction, and (2) brown rot, in which lignin is modified but not appreciably removed (6–8). Phylogenomic analyses have demonstrated that brown rot fungi have evolved repeatedly from white rot ancestors (9–11).

Evolution of brown rot is associated with extensive losses of genes encoding plant cell wall-degrading enzymes, including peroxidases, laccases, and other ligninolytic oxidases, as well as carbohydrate-active enzymes (CAZymes), such as glycoside hydrolases (GHs), carbohydrate esterases (CEs), and lytic polysaccharide monooxygenases (LPMOs), coupled with reductions in cellulose-binding modules (CBM1s) associated with diverse CAZymes. Nevertheless, brown rot fungi cause rapid loss of strength and mass in wood substrates (6, 7, 12–14). The model brown rot fungus Rhodonia placenta (here referred to as Postia placenta, to maintain continuity with prior literature) appears to attack wood cell walls in a two-step process, in which reactive oxygen species are first deployed to effect an oxidative pretreatment of lignocellulose, followed by action of GHs that break down cellulose and hemicellulose into assimilable sugars, leaving lignin as a modified polymeric residue (15–17).

Most species of wood-decaying Agaricomycetes have characteristic substrate ranges that define their ecological activity in forest ecosystems and which can be useful for identification purposes (18–20). Brown rot species have been suggested to preferentially decay softwoods (gymnosperms, particularly Pinaceae), although many also attack hardwoods (angiosperms) (21, 22). The hemicellulose and lignin composition of hardwoods and softwoods differ and may affect substrate preferences by fungi. The hemicellulose of hardwoods is enriched in glucuronoxylan, whereas the main hemicellulose of softwoods is galactoglucomannan, and the lignin fraction of softwood is mainly guaiacyl lignin, whereas hardwoods have a larger (but variable) fraction of syringyl lignin (23–25). Secondary metabolite profiles also differ, with softwoods containing large amounts of terpenes and other extractives and hardwoods being highly variable in extractive compounds (24).

The mechanism(s) of brown rot decay remain uncertain, but hydroxyl radicals have been implicated as the diffusible oxidants (26, 27). Fenton chemistry (H₂O₂ + Fe²⁺ + H⁺ → H₂O + Fe³⁺ + ·OH) has been proposed, and three models are often cited (4, 28). A central role for cellobiose dehydrogenase had been proposed, but it is now clear that some brown rot fungi that efficiently cause extensive decay, such as P. placenta and Wolfiporia cocos, lack the gene for this enzyme (29). Alternatively, certain low-molecular-weight glycopeptides (GLPs) are thought to reduce extracellular Fe³⁺ in white and brown rot fungi (30–33). Perhaps the most widely held model involves hydroquinone cycling (34–36), which generates both Fenton reactants through the activities of a quinone reductase and oxalate-Fe³⁺ chelates. Several studies have suggested that laccases may also be involved in hydroquinone oxidation (37–39). Regardless of the mechanism(s), brown rot is thought to involve sequential oxidation and hydrolysis of lignocellulose.

Numerous studies have demonstrated the impact of culture conditions on transcript and protein accumulation levels in wood decay fungi (29, 40–42), including comparisons of expression on defined media (e.g., glucose, maltose, or crystalline cellulose) versus wood as a carbon source or under aerobic or anaerobic conditions (43–48). However, there have been relatively few side-by-side comparisons of gene expression on different wood species (43, 49–57).

Several prior transcriptomic analyses concluded that wood decay fungi express similar sets of plant cell wall-degrading enzymes on both hardwood and softwood, although there may be disparities in the expression of individual genes on different substrates (43, 50, 51, 53, 56). For example, Vanden Wymelenberg et al. (56) found that 47 genes in the model white rot fungus Phanerochaete chrysosporium were upregulated >2-fold on aspen versus pine, including 13 genes encoding GHs, while in P. placenta 164 genes showed differential expression on the two substrates (roughly equal numbers upregulated on each substrate). The P. placenta genes upregulated on pine included 15 genes encoding cytochrome P450s, which are thought to play a role in the degradation of complex wood extractives (29, 58, 59). Similarly, Gaskell et al. (50) detected three P450s upregulated on pine versus aspen in another brown rot species, Wolfiporia cocos. A comparison of the genome of P. chrysosporium with the closely related Phanerochaete carnosa demonstrated an expansion of P450s in the latter species, which is almost exclusively collected on conifer wood (54). Moreover, P. carnosa was shown to transform a higher fraction of phenolics from heartwood than P. chrysosporium (54), consistent with the view that P450s play a role in detoxifying wood extractives. Couturier et al. (43) also found that eight genes encoding P450s in another white rot fungus, Pycnoporus coccineus, were differentially expressed on pine versus aspen. Genes encoding other enzymes active in lignocellulose digestion that have been shown to be differentially expressed on hardwoods versus softwoods include those encoding LPMOs, manganese peroxidases (MnP), lignin peroxidases (LiP), GH10 xylanases, and GH78 α-l-rhamnosidase (43, 50, 52).

It is difficult to make generalizations from the results of previous studies comparing gene expression on different substrates because their experimental conditions and organisms are highly variable. Most studies have used submerged milled wood substrates (43, 50, 52, 53, 56, 60), but others have used solid wood wafers (49) or sawdust (54, 55). Sampling times in studies using submerged substrates range from 3 days (43) to 16 days (53), while studies using sawdust or wood wafers have sampled at 10 days (49) to 42 days (54). About a dozen different tree species have been used as experimental substrates. The most commonly used hardwood is aspen (Populus tremuloides and P. grandidentata), including transgenic and hybrid lines with various lignin contents (49, 50, 60), but sugar maple (Acer saccharum) and yellow birch (Betula alleghaniensis) have also been employed (51, 52, 54). Softwood substrates include fir (Abies balsamea), three species of spruce (Picea abies, P. glauca, and P. rubens), and five pines (Pinus contorta, P. halepensis, P. resinosa, P. strobus, and P. sylvestris) (43, 50–54, 56). Yakovlev et al. (55) compared results on heartwood, reaction wood, and sapwood of P. abies and P. sylvestris.

The most commonly studied fungal species in analyses of substrate-specific gene expression are white rot species of the Polyporales, particularly P. chrysosporium and P. carnosa (49, 51, 52, 54, 56, 60). Other white rot species that have been studied include Pycnoporus cinnabarinus and Dichomitus squalens (43, 53), which are both in the Polyporales, and Heterobasidion annosum, which is in the Russulales (55). Two brown rot species of Polyporales, P. placenta and W. cocos, have been used in studies comparing gene expression on different tree species, specifically aspen and pine (49, 50, 56).

Here, we present an analysis of substrate-specific responses in Fomitopsis pinicola, which is a brown rot wood-decaying species of Polyporales that is reported to occur on 82 species of gymnosperms (softwoods), most commonly on pines and spruces, as well as 42 species of angiosperms (hardwoods), including aspen, birch, and cherry (61). This diverse range of substrates suggests that F. pinicola may have the ability to adjust its arsenal of decay enzymes according to the chemistry and anatomy of different hosts. F. pinicola, P. placenta, and W. cocos are members of the informally named “Antrodia clade,” which includes diverse brown rot species. The Antrodia clade is strongly supported as a monophyletic group (10), suggesting that the brown rot decay mechanisms of F. pinicola, P. placenta, and W. cocos are homologous.

We obtained results at one time point on three substrates (pine, spruce, and aspen) in submerged cultures and at two time points on two substrates (pine and spruce) on wood wafers. Thus, our results allow us to assess the impact of substrate wood species versus culture conditions on transcription.

RESULTS

Expression profiling of F. pinicola grown in submerged culture.

Cultures of F. pinicola were grown in submerged cultures supplemented with ground aspen, pine, or spruce as the sole carbon source for 5 days. Owing to the accessibility of the substrate, rapid growth of wood decay fungi is typically observed under these conditions (29, 47, 49, 60, 62–64). Transcript levels and patterns of expression observed here support the involvement of numerous genes in lignocellulose degradation.

The highest transcript levels corresponded to the same genes in all three substrates. Not surprisingly, putative “housekeeping” genes such as short-chain dehydrogenase/reductases (SDR), ADT/ATP carriers, and a FAD/NAD domain ranked among the 30 highest FPKM (fragments per kilobase per million) values (Table 1; see also Table S1). Genes encoding glycoside hydrolases and carbohydrate esterases (GH18, GH128, GH78, CE4, and CE6), broadly categorized as hemicellulases, were also highly expressed, especially those with predicted secretion signals. This sharply contrasts with white rot fungi that express high levels of cellulose-attacking hydrolases (GH6s and GH7s) and lytic polysaccharide monooxygenases (LPMOs). Oligopeptide transporter and extracellular protease may reflect the importance of nitrogen scavenging in wood, whereas extracellular lipases likely facilitate triglyceride utilization.

TABLE 1.

Thirty most abundant transcripts in submerged culture containing ground pine as the sole carbon source^a

Pro ID	Putative functionb	Conditions						Protein
		Pine FPKM	Pine order	Spruce FPKM	Spruce order	Aspen FPKM	Aspen order	No. of TMHs	SP
Fompi_1030775	AA3_3, alcohol oxidase	12,993	1	10,068	1	11,821	1
Fompi_1152048	Aquaporin-like	8,946	2	4,319	4	7,387	2	6
Fompi_1039841	NAD-dependent formate dehydrogenase	6,013	3	7,407	2	5,516	3
Fompi_49484	CipC protein	3,552	4	4,951	3	3,315	4
Fompi_160683	GH18 protein	2,936	5	2,705	6	3,003	5		+
Fompi_1084629	Mannoprotein	2,811	6	2,639	7	2,870	6		+
Fompi_1022702	Glutamine-dependent formaldehyde dehydrogenase	2,588	7	2,896	5	2,452	8
Fompi_1023896	ABC transporter	2,554	8	1,967	14	2,591	7
Fompi_49935	NAD(P) binding SDR	2,236	9	2,257	9	1,912	11
Fompi_1090124	Endosulfine-domain protein	2,105	10	1,338	20	2,167	9
Fompi_1023614	Acid protease	1,816	11	2,240	10	2,129	10	1	+
Fompi_126349	MFS general substrate transporter	1,633	12	1,250	23	1,758	12	12
Fompi_1021547	Lipase*	1,631	13	848	36	941	30		+
Fompi_1168917	NAD-dependent formate dehydrogenase	1,564	14	1,905	15	1,445	15
Fompi_1022193	GH128 protein	1,529	15	1,514	17	1,527	13		+
Fompi_1147363	Catalase	1,484	16	1,336	21	1,172	22
Fompi_1039147	General substrate transporter	1,417	17	2,204	11	1,253	19	12
Fompi_159299	CE4-CBM50	1,367	18	1,449	18	1,419	17		+
Fompi_1023676	Peptidoglycan binding	1,357	19	1,448	19	1,489	14		+
Fompi_1022010	OPT oligopeptide transporter	1,313	20	869	33	968	27	16
Fompi_129535	FAD/NAD(P)-binding domain	1,292	21	1,742	16	1,207	20
Fompi_1131909	Aldehyde dehydrogenase	1,190	22	1,251	22	1,089	26
Fompi_1021548	Lipase*	1,171	23	1,055	28	780	34	1	+
Fompi_145400	GH78	1,150	24	773	43	1,320	18		+
Fompi_1119394	CE6	1,070	25	1,053	29	1,180	21	1	+
Fompi_98853	GH16	1,064	26	916	31	1,140	24	1
Fompi_1022048	Eukaryotic ADP/ATP carrier	1,029	27	1,145	26	1,122	25	3
Fompi_1025554	SDR	1,012	28	890	32	950	29	2
Fompi_59695	Hexose transporter	1,012	29	2,592	8	1,429	16	11
Fompi_161955	SDR	962	30	1,187	25	880	31

^aExcludes genes designated “hypothetical” and those with no significant similarity to NCBI NR accession numbers. GH, glycoside hydrolase; CE, carbohydrate esterase; TMHs, predicted transmembrane helices; SP, predicted secretion signal; AA, auxiliary activities as described in the Carbohydrate Active Enzymes database (http://www.cazy.org/) (88, 89); SDR, short-chain dehydrogenase/reductase.

^b*, Lipase genes tandemly linked ∼2 kb apart.

Hierarchical clustering analysis based on expression levels revealed three groups corresponding to the substrates (see Fig. S1). Transcriptome profiles of F. pinicola with pine are more similar to those with aspen than spruce, which is surprising, because spruce and pine are softwoods (Pinaceae), whereas aspen is a hardwood. To verify transcriptome results, we performed qRT-PCR analyses of two differentially regulated genes. Consistent with RNA-seq, the expression patterns of F. pinicola with pine are more similar to those with aspen, and the lowest transcript levels of thaumatin-like protein (Fompi_110401) and a peptidase (Fompi_1113467) were observed in colonized pine and spruce, respectively (see Fig. S2).

Of the 13,885 genes predicted in the F. pinicola genome, 91 genes were differentially expressed (4-fold change and P < 0.05) among the three pairwise comparisons of substrates (aspen versus pine, aspen versus spruce, and pine versus spruce) in submerged culture (Fig. 1; see also Table S1). Consistent with the global expression pattern, fewer genes showed differential expression in pairwise comparisons between aspen and pine than in the other two pairwise comparisons (Fig. 1). GO enrichment analysis of upregulated genes on each substrate reveal very different terms for cellular component (CP), biological process (BP), and molecular function (MF) (Fig. 2), with none shared by all three substrates.

FIG 1.

Heatmap showing hierarchical clustering of 11 genes with predicted functions in wood decay with a 4-fold change (FDR < 0.05) in pairwise-substrate comparisons at day 5. The scale above the map shows log₂-based signals under central row. Protein IDs and their enzyme names are indicated on the right side of the heatmap. Transporters are labeled in blue, and wood-decay enzymes are labeled in red. The bottom column designations refer to replicate libraries.

FIG 2.

GO enrichment of upregulated genes from each substrate after 5-day-old submerged incubation. Categories of GO terms are labeled on the right side of the heatmap. Substrates are labeled at bottom of heatmap. Dark blue shading indicates presence, and light blue indicates absence. The genes number in each GO term was labeled in dark blue.

Eleven DEGs among the three pairwise comparisons encode enzymes with predicted functions in wood decay (Fig. 1, names in red). The expression levels of these genes were roughly clustered based on their putative function. These include laccase, benzoquinone reductase and aryl-alcohol oxidase (Fig. 1) possibly involved in the generation of hydroxyl radical via Fenton reactions (34–36, 39). In particular, benzoquinone reductase has been implicated in the redox cycling of Fe and H₂O₂, key Fenton reactants (35), and the process of hydroquinone oxidation may be augmented by laccases (39). While generally considered an oxidative process in brown rot, hydrolytic enzymes may also participate in cellulose degradation (65). Supporting this, a predicted endo-1,4-glucanase (Fig. 1) was expressed. Typical of brown rot fungi, this GH5_5 family protein lacks a cellulose binding module, and no exo-cellobiohydrolases occur within the F. pinicola genome. Thus, the importance of cellulose hydrolysis by conventional enzymes remains unclear. Differential expression analyses also revealed four genes encoding P450s, which are distributed among all three pairwise-substrate comparisons (Fig. 1). These may be involved in intracellular metabolism of low-molecular-weight breakdown products of lignin and resins (66, 67).

A previous study of W. cocos (50) indicated that membrane-bound proteins are potentially important for carbohydrate metabolism during wood decay. We found that 8% to 22% of DEGs from each pairwise comparison have transmembrane (TM) domains, including six transporters (Fig. 1), which may facilitate movement of small solutes across cell membranes. In addition, two proteins have functions related with methylation (Fompi_47533) and phosphorylation (Fompi_55544) (see Table S1), suggesting that modification of proteins might be involved during the decay process in submerged culture.

Expression profiling of F. pinicola grown on wood wafers.

Submerged cultures and colonized wafers share many of the more highly expressed genes (Tables 1 and 2), including some that are likely involved in lignocellulose degradation. Among these are several glycosidases, a lipase (Fompi_1021547) potentially involved in fatty acid catabolism, and an alcohol oxidase (Fompi_1030775) similar to the methanol oxidase of Gloeophyllum trabeum (68). The latter enzyme has been implicated in the generation of peroxide from methanol during brown rot demethylation of lignin. The presence of proteins, as well as modifications of lignin and wood extractives, after 10 and 30 days of postinoculation was confirmed by time-of-flight secondary ion mass spectrometry (ToF-SIMS; for ToF-SIMS methods and results, see the supplemental material). The fungal treated wood samples, on both day 10 and day 30, were distinguished from the corresponding untreated samples by the abundance of protein-specific peaks, as well as the depletion of many extractive peaks, including 43, 55, and 67 Da (see Fig. S3). A closer look at the loadings of principal-component analysis also indicated the removal of generic aromatic peaks (see Fig. S3).

TABLE 2.

Thirty most abundant transcripts in colonized spruce and pine wafers after 30 days^a

Pro ID	Putative function	Conditions				Protein
		Spruce-30D FPKM	Spruce order	Pine30 FPKM	Pine30 order	No. of TMHs	SP
Fompi_1152048	Aquaporin-like	5,523	1	6,517	1	6
Fompi_1030775	AA3_3, alcohol oxidase	4,056	2	5,301	2
Fompi_128521	l-Fuculose-phosphate aldolase	3,490	3	1,397	15
Fompi_1021547	Lipaseb	3,192	4	3,496	3		+
Fompi_1131909	Aldehyde dehydrogenase	1,808	5	1,586	8
Fompi_1090124	Endosulfine-domain-containing protein	1,537	6	1,561	9
Fompi_160683	GH18	1,493	7	1,451	13		+
Fompi_1025554	SDR	1,487	8	1,555	11	2
Fompi_1022048	Mitochondrial carrier	1,454	9	1,503	12	3
Fompi_1147363	Catalase	1,423	10	1,478	13
Fompi_148786	CE4	1,383	11	751	50		+
Fompi_86900	MFS general substrate transporter	1,367	12	961	34	12
Fompi_1044587	Plasma membrane proteolipid 3	1,365	13	989	30	2
Fompi_1022193	Glycoside hydrolase family 128 protein	1,361	14	1,637	5		+
Fompi_1022498	Ribosomal protein S25	1,341	15	1,560	10
Fompi_126349	MFS general substrate transporter	1,263	16	1,055	24	12
Fompi_1025379	Peptidyl-prolyl cis-trans isomerase	1,245	17	1,003	28
Fompi_113951	Ribosomal protein 60S	1,234	18	1,183	19
Fompi_1021745	Polyubiquitin	1,232	19	1,354	16
Fompi_1039841	NAD formate dehydrogenase	1,208	20	1,632	6
Fompi_1022753	RNA-binding domain	1,161	21	969	33
Fompi_127599	Actin 1	1,153	22	1,223	17
Fompi_1112040	Histone H3	1,146	23	1,184	18
Fompi_1023676	CBM50	1,088	24	995	29		+
Fompi_1025222	Manganese superoxide dismutase	1,078	25	627	63
Fompi_56132	Acid protease	1,023	26	1,138	21
Fompi_1028183	Acetohydroxy acid isomeroreductase	992	27	692	58
Fompi_1142989	Ubiquitin/40S ribosomal protein fusion	983	28	1,024	27
Fompi_84886	Thaumatin-like protein	957	29	572	74		+
Fompi_56163	GH5_9	955	30	822	42		+

^aExcludes genes designated “hypothetical” and those with no significant similarity to NCBI NR accession numbers. GH, glycoside hydrolase; CE, carbohydrate esterase; TMH, predicted transmembrane helices; SP, predicted secretion signal; AA, auxiliary activities as described in the Carbohydrate Active Enzymes database (http://www.cazy.org/) (88, 89); SDR, short-chain dehydrogenase/reductase.

^bLipase genes tandemly linked ∼2 kb apart.

Time-dependent gene expression.

Differentially expressed transcripts corresponding to 320 and 776 genes were identified by pairwise comparisons on pine and spruce, respectively (Fig. 3A; see also Table S2). DEGs showed different trajectories on different substrates. On pine, there were far fewer upregulated genes than downregulated genes at day 30 compared to day 10, whereas on spruce there were slightly more upregulated genes than downregulated genes on day 30 versus day 10 (Fig. 3A).

FIG 3.

Differential expression of F. pinicola on different wood wafers (pine and spruce) after 10-day and 30-day incubations. (A) Relative percentage of upregulated DEGs on different substrates and at different time points. (B) GO enrichment of upregulated genes from each condition in each pairwise comparison. Four pairwise comparisons are labeled on the right side of the heatmap. Dark blue shading indicates presence, and light blue indicates absence. The number of genes in each GO term was labeled in dark blue. (C) Distribution in relative proportion of wood decay CAZymes (GH, CE, and PL), redox enzymes potentially involved in Fenton chemistry (AA families, HTP, POD, CRO, OXO, GLP, and QRD), and cytochrome P450s at different time points (C) and on different substrates (D). For each category, only the upregulated genes at one time point/substrate relative to the other time point/substrate was considered.

We performed enrichment analyses of upregulated DEGs from different time points on each substrate and found the enriched GO terms to be highly variable along the time course (Fig. 3B). The influence of incubation time is also pronounced among relative proportion of TM proteins (Fig. 4). MFS-1, cytochrome P450, and oligopeptide transporter are enriched after a 10-day incubation on pine but disappear after a 30-day incubation (Fig. 4B). Conversely, on spruce MFS-1 and cytochrome P450 are enriched at the later stage (Fig. 4C).

FIG 4.

Annotations of DEGs on wood wafers having TM domains. (A) Distribution in relative proportion of TM proteins at different time points (left) and on different substrates (right). For each category, only the upregulated gene at one time point/on one substrate relative to the other time point/substrate was considered. (B to E) showing the details for stacked bars in the panel A. The font size in the word cloud is directly related to the word (protein) frequency in each panel. The colors of words are consistent with the conditions used in panel A.

To evaluate the roles of carbohydrate-active enzymes and nonenzymatic oxidation reactions during the time course, we counted the number of CAZymes (GH, CE, and PL), various redox enzymes potentially involved in Fenton systems (AA family members, HTPs, oxalate metabolism, and iron homeostasis), and cytochrome P450s among the DEGs. For each category, only the upregulated gene at one time point relative to the other one was considered. On each substrate, the number of CAZymes decreased with increased incubation time, while the number of lignin-degrading enzymes increased (Fig. 3C; see also Table S3). In contrast, the cytochrome P450 transcript accumulation patterns are not consistent between substrates (Fig. 3C; see also Table S3). On pine, the trend of cytochrome P450 is consistent with the CAZymes, whereas the cytochrome P450 on spruce is consistent with that of lignin enzymes.

Based on comparisons to P. chrysosporium iron reducing glycopeptides (31, 33), 10 F. pinicola homologs were identified. Transcripts corresponding to Fompi_1027255 decreased over time in both pine and spruce. This expression pattern would be expected where initial stages of decay relied heavily on oxidative depolymerization of cellulose whereas later stages might shift toward hydrolytic mechanisms.

Substrate-dependent gene expression.

We compared expression patterns on pine and spruce substrates at the same time points. A total of 294 DEGs were identified by substrate comparisons after a 10-day incubation, of which 199 were upregulated on pine, and 95 were upregulated on spruce (Fig. 3A; see also Table S2). After a 30-day incubation, 135 DEGs were identified on pine versus spruce, of which only 18 were substantially upregulated on pine (Fig. 3A; see also Table S2).

GO enrichment analysis of substrate-dependent upregulated genes after a 10-day incubation revealed six different biological processes and one different cellular component between pine and spruce. Most biological processes are enriched on spruce (Fig. 3B). The difference in enriched GO terms is diminished in the 30-day comparisons. After 10-day incubations, TM proteins, cytochrome P450, sugar transporter, and oligopeptide transporter are enriched on pine relative to spruce, while MFS-1 is enriched on spruce relative to pine after 30-day incubation (Fig. 4). CAZymes, AA enzymes, and cytochrome P450s were also explored between substrates. After a 10-day incubation, spruce-dependent upregulated genes include 5-fold more genes encoding CAZymes and one-third as many genes encoding cytochrome P450s relative to pine. However, there was no change in AA enzymes (Fig. 3D; see also Table S3). After a 30-day incubation, spruce-dependent upregulated genes included slightly more genes encoding AA and other CAZyme family members, compared to pine (Fig. 3D; see also Table S3).

Cross-culture method comparisons.

We performed hierarchical clustering of gene expression patterns using samples from different substrates (pine versus spruce), time points (day 5, day 10, and day 30) and culture methods (submerged versus wood wafers). Samples from the same time points and culture methods clustered with each other, regardless of substrate (see Fig. S4 and S5). Samples on submerged culture at day 5 were more similar to samples from wood wafers at day 30, than wood wafer samples at day 10 (Mann-Whitney U test, P < 0.00001) (see Fig. S4). A neighbor-joining analysis of samples further indicated that the samples at day 5 were more distant from samples at day 10 than at day 30 (Fig. 5).

FIG 5.

Cross-culture method comparisons. A neighbor-joining tree with branch lengths inferred using (1 – Spearman’s rho) for all pairs of gene expression profiles is shown.

DISCUSSION

Fomitopsis pinicola exhibits substrate-dependent and time-dependent differential gene expression.

Fomitopsis pinicola produces similar sets of wood-degrading enzymes on both hardwood and softwood substrates, as do other wood decay fungi (47, 50). Nevertheless, there are differences in the expression of specific genes, yielding substrate-specific expression profiles (Fig. 1; see also Table S1). Genome and transcriptome analyses of other brown rot species have suggested that they employ nonenzymatic Fenton chemistry to generate hydroxyl radicals, which are generated by various redox systems (49, 50, 69, 70). In this study, differentially expressed decay-related oxidoreductases included laccase, AA3 (aryl alcohol oxidase), benzoquinone reductase, Fe(III)-reducing glycopeptides (GLP), and oxalate oxidase/decarboxylases (OXO) (Fig. 1 and 3C and D; see also Table S3). Glycoside hydrolases are also important for brown-rot decay, and we found 16 kinds of upregulated GHs (GH1, GH3, GH5, GH16, GH17, GH18, GH28, GH30, GH31, GH43, GH45, GH78, GH79, GH92, and GH128) from different substrates and time points (Fig. 1 and 3C and D; see also Table S3). The biological roles of some GHs are uncertain. For example, GH18 and GH16 are related to lysis of fungal mycelium, but recent studies in white-rot fungi show that the two gene families are differentially expressed during wood decay (49, 71), suggesting potential roles in wood decay. In particular, GH16 enzymes may attack β-1,3-glucans or xyloglucans, which play significant roles in hemicellulose degradation (72, 73). Along the time course on wood wafers, upregulated CAZymes are greater at day 10 than at day 30, whereas upregulated AA enzymes are greater at day 30 than at day 10 (Fig. 3C and D). This is not consistent with recent observations in P. placenta (17), where greater numbers of upregulated redox enzymes occurred at early stages of decay. A recent study suggests that lignocellulose oxidation by F. pinicola fluctuates for 3 months in submerged culture supplemented with spruce wood (74).

Unknown “hypothetical” proteins also contribute to substrate-specific responses. For instance, 50% (146) of the DEGs (294) between pine and spruce after a 10-day incubation lack clear annotations according to InterPro prediction. Among these, 44 proteins (30%) were predicted to have secretion signals, and 47 (32%) have TM domains, suggesting that they may contribute to extracellular processes, including plant cell wall decomposition, or transport of solutes across fungal cell membranes (74).

Pine and spruce “softwood” substrates elicit distinct gene expression profiles.

Transcriptomes from 5-day-old submerged cultures showed that global gene expression profiles on pine were more similar to those from aspen than those from spruce (Fig. 1; see also Fig. S1). This was surprising, because pine and spruce are both softwoods in the Pinaceae, whereas aspen is a hardwood. These findings suggest that it may be misleading to generalize from gene expression results on individual wood species to all “softwoods” (gymnosperms) or “hardwoods” (angiosperms). Moreover, prior studies using different genotypes of aspen (49, 60) or different wood types from pine and spruce (i.e., heartwood, sapwood, and tension wood [55]) demonstrate that gene expression may vary according to the genotype of the host species, or the specific kind of wood that is used. Subtle differences in wood anatomy, physical properties and/or chemical composition may substantially impact transcript profiles. Thus, to maximize comparability among studies it will be necessary not only to select a uniform set of species (perhaps focusing on plants with the greatest potential as feedstocks for biofuel production) but also to consider genetic diversity within species and the anatomical and chemical properties of wood samples.

Large differences exist in wood components for pine and spruce versus aspen. Lignin content is higher (27 to 30%) compared to aspen (22%), while cellulose and hemicellulose are approximately the same. However, the hemicellulose composition is largely acetylated glucuronoxylans in hardwoods and acetylated galactoglucomannans in softwoods (24, 75). Other relatively minor wood components, such as extractives, may have a greater impact on specific gene expression as shown for the white rot fungus Phlebiopsis gigantea (76).

Brown rot Polyporales have elevated expression of certain P450s on pine versus aspen substrates.

Fomitopsis pinicola is the third brown rot species in the Antrodia clade of the Polyporales to be used for comparative transcriptomic analyses on different wood substrates. Vanden Wymelenberg et al. (56) and Gaskell et al. (50) analyzed W. cocos and P. placenta, respectively, and both studies compared aspen and pine substrates in submerged cultures sampled at 5 days, as in the present study. Under these conditions, in all three studies there were few genes upregulated (>4-fold, false discovery rate [FDR] < 0.05) on pine versus aspen, including 19 genes in P. placenta, 5 genes in W. cocos, and 12 genes in F. pinicola (Fig. 1; see also Table S1). In each species, the upregulated transcripts on pine include genes encoding cytochrome P450s (10 in P. placenta, 3 in W. cocos, and 1 in F. pinicola), suggesting that these may contribute to a common response to pine relative to aspen. The precise function of these P450s remains to be established, but depletion of small lignin derivatives or extractive compounds observed with ToF-SIMS is consistent with a role for these enzymes in metabolism of toxic breakdown products (see the supplemental material).

Culture methods greatly affect gene expression.

Transcriptomic studies of wood decay in fungi usually employ submerged culture methods (29, 47, 50, 64, 77), wood wafers (49) or sawdust (54, 55). Of these methods, solid wafers may best mimic “natural” conditions in which the porosity of sound wood excludes many enzymes during incipient decay (78, 79). However, steam sterilization likely impacts substrate accessibility, and this pretreatment is no doubt more pronounced when the substrate is ground and dispersed as in submerged cultures.

Global analysis of gene expression profiles on pine and spruce in submerged cultures at 5 days or wood wafers sampled at 10 and 30 days indicates that culture conditions have a greater impact on overall transcriptome profiles than substrate species, and the global gene expression pattern at day 5 in submerged substrates is more similar to that on wood wafers at day 30 than on wood wafers at day 10 (see Fig. S4). The number of DEGs reflects the sensitivity of F. pinicola in recognizing substrates. Among the three sampling points, F. pinicola at day 5 in submerged culture has the fewest DEGs. Thus, F. pinicola in submerged culture exhibits the least sensitivity to substrates among the conditions that we employed. It is not clear whether this effect is a consequence of the time of sampling or the culture conditions. Additional studies with overlapping time points in both submerged cultures and wood wafers could help discriminate the relative effects of time of sampling versus culture conditions.

Conclusions.

Fomitopsis pinicola, a wood-decaying Agaricomycete with a broad range of softwood and hardwood hosts, exhibits differential gene expression on different wood substrates (aspen, pine, and spruce) and under different culture conditions (submerged substrates versus wood wafers). Many of the DEGs identified here encode enzymes with known or predicted functions in plant cell wall decomposition, suggesting that F. pinicola can modulate its decay apparatus according to substrate. However, culture conditions (submerged ground wood versus wood wafers) and sampling times had a greater impact on overall gene expression profiles than did wood species, which highlights the need for standardization of experimental parameters in studies of wood decay mechanisms. Comparisons of gene expression on pine versus aspen in 5-day-old submerged cultures of F. pinicola, P. placenta and W. cocos, which are all brown rot species of the Antrodia clade (Polyporales), indicate that genes encoding some cytochrome P450s may be upregulated on pine. To assess the generality of these results, it will be necessary to study gene expression in additional species (ideally represented by multiple strains) of brown rot Polyporales on aspen, pine and spruce substrates.

MATERIALS AND METHODS

Culture conditions.

Fomitopsis pinicola FP-58527 was obtained from the Forest Mycology Center, USDA Forest Products Laboratory (Madison, WI). Originally isolated from Pinus ponderosa, a draft genome sequence of this strain was analyzed by Floudas et al. (9), and the version 3 assembly is publicly available through the JGI portal (https://genome.jgi.doe.gov/Fompi3/Fompi3.home.html). Two culture methods were used to explore the transcriptome of F. pinicola on different substrates. Two-liter flasks containing 250 ml of basal salt media were supplemented with wood ground in Wiley-mill with a #10 mesh screen. Each flask contained wood of quaking aspen (Populus tremuloides), loblolly pine (Pinus taeda), or white spruce (Picea glauca) as the sole carbon source. The basal medium was prepared as described previously (50). Following autoclaving at 121°C for 15 min, triplicate cultures for each substrate were inoculated with mycelium scraped from malt extract agar (2% [wt/wt] malt extract, 2% [wt/wt] glucose, 0.5% peptone, 1.5% agar) and placed on a rotary shaker (150 rpm) at 22 to 24°C. Five days after inoculation, mycelia and pellets of various sizes were collected by filtration through Miracloth (Calbiochem) and stored at −80°C. In addition to submerged cultures, the mycelium was also cultured on wood wafers of the same three species in solid-block jar microcosms for solid-state culturing. Wood wafers were created from the same individual wood samples that were used to obtain milled substrates used in submerged cultures. The soil mixture for the microcosm was prepared using two parts of loam soil, two parts of vermiculite, and one part of peat moss that was wetted to capacity and placed in glass jars (473 ml). Two wood feeder strips corresponding to each type of wood were placed together on the top of the soil mixture and autoclaved twice for 1 h at 121°C with a 24-h period in between sterilizations. A 5-mm-diameter plug of media containing mycelium was placed at the corner of the feeder strips. When the mycelium covered the feeder strips completely, thin (10 × 15 × 1 mm), longitudinally cut wood wafers of each type of wood were placed on the feeder strips and incubated at 22°C. The wood wafers were removed after 10 and 30 days snap-frozen in liquid nitrogen and stored at −80°C. There were three replicates for each treatment.

RNA extraction and RNA-seq library construction.

Prior to extraction, colonized ground wood and wafer samples were pulverized under liquid nitrogen in a SPEX 6775 mill (Metuchen, NJ). A modified version of the method of Miyauchi et al. (80) was used for RNA purification from the SPEX-milled wafer. The samples were transferred to Oak Ridge tubes containing 12 ml of TRIzol. To disperse the material, the tubes were vortexed, and the contents were drawn three to five times into a polypropylene pipette. Chloroform (2.4 ml) was added with a glass pipette, and the suspension was briefly vortexed. After standing for 2 to 3 min, the tubes were centrifuged in the prechilled JA-17 rotor at 40,000 × g for 15 min at 4°C. With minimal disturbance to the interface, the aqueous phase (6 ml) was transferred to a new Oak Ridge tube, and 0.6 volumes (3.6 ml) of isopropyl alcohol were added. The contents were gently mixed by inverting several times, and then the tubes were incubated at −20°C for 1 to 2 h. The nucleic acids were pelleted by centrifugation at 40,000 × g for 10 min at 4°C. After careful decanting, the pellet was washed twice with 75% ethanol and air dried. The RNA was resuspended in 887.5 μl of diethyl pyrocarbonate (DEPC)-treated water and transferred to a sterile Eppendorf tube. The sample was DNase treated with Qiagen’s RNase-free DNase set by adding 100 μl of RDD buffer and 12.5 μl of DNase (3 U/μl). After 20 min, another 3 U of DNase was added. RNA was precipitated by adding 330 μl of 8 M LiCl, followed by incubation at −20°C for at least 1 h. The Eppendorf tubes were then centrifuged in prechilled JA-18.1 at 12,000 rpm (∼18,000 × g) for 20 min, and the pellet was resuspended in 150 μl of isopropyl alcohol. The samples were then centrifuged in JA-18.1 at 12,000 × g for 5 min and washed twice with 1 ml of 70% ethanol. Droplets of remaining ethanol were aspirated to hasten air drying. The dry pellet was resuspended in 25 μl of DEPC-water and placed at 55°C for 5 min to dissolve. All RNA samples were quantified by a Qubit fluorometer (Thermo Fisher Scientific) and quality checked with a 2100 BioAnalyzer (Agilent), respectively.

Plate-based RNA sample prep was performed on the Perkin-Elmer Sciclone NGS robotic liquid handling system using the Illumina TruSeq Stranded mRNA HT sample prep kit utilizing poly(A) selection of mRNA according to the protocol outlined by Illumina in their user guide and using the following conditions: total RNA starting material was 1 μg per sample, and eight cycles of PCR were used for library amplification. The prepared libraries were quantified using KAPA Biosystem’s next-generation sequencing library qPCR kit and run on a Roche LightCycler 480 real-time PCR instrument. The quantified libraries were then multiplexed and prepared for sequencing on the Illumina HiSeq sequencing platform utilizing a TruSeq Rapid paired-end cluster kit, v4. Sequencing of the flow cell was performed on the Illumina HiSeq2000 sequencer using HiSeq TruSeq SBS sequencing kits, v4, following a 1 × 101 indexed run recipe.

Identification of differentially expressed genes.

Raw reads were filtered and trimmed using the JGI QC pipeline. Using BBDuk (https://sourceforge.net/projects/bbmap/), raw reads were evaluated for sequence artifacts by kmer matching (kmer = 25), allowing one mismatch, and any detected artifact was trimmed from the 3′ end of the reads. RNA spike-in reads, PhiX reads and reads containing any Ns were removed. Quality trimming was performed using the phred trimming method set at Q6. Finally, following trimming, reads under the length 25 nucleotides were removed. Filtered reads from each library were aligned to the haploid F. pinicola FP-58527 genome using HISAT (81) (see Table S5). featureCounts (82) was used to generate the raw gene counts using gff3 annotations (https://genome.jgi.doe.gov/Fompi3/Fompi3.info.html). Only primary hits assigned to the reverse strand were included in the raw gene counts (-s 2 -p –primary options). FPKM normalized gene counts were calculated by Cufflinks (83). edgeR (84) was subsequently used to determine which genes were differentially expressed between pairs of conditions. The parameters used to call a DEG between conditions were a P value of <0.05 and a log₂-fold change of ≥4. SignalP 4.1 (85) was used to search for secretory signal peptides in DEGs proteins. TMHMM 2.0 (86) was used to predict and characterize TM domains in proteins encoded by DEGs, with default parameters. Functional categories enriched with DEGs were identified using FungiFun2 with an FDR of <0.05 (87).

Data availability.

Transcriptome data set have been deposited in the NCBI Sequence Read Archive, accession numbers SRP140951 to SRP140968 (see Table S4 for details).

PUBLISHER’S NOTE: The American Society for Microbiology and Applied and Environmental Microbiology would like to inform the readers that this article is a corrected and republished version of https://doi.org/10.1128/AEM.00991-18, which was retracted (https://doi.org/10.1128/AEM.00330-21). As stated in the retraction notice, after publication, the authors performed additional analyses and “discovered that the RNA editing results (but not the gene expression results) are artifactual.” Thus, one of the main conclusions of the original paper can no longer be supported, but the conclusions around the “substrate-specific differential gene expression” continue to be validated and described in this paper. Therefore, the authors were allowed to submit an entire, corrected manuscript for consideration to the journal. The manuscript was peer reviewed and subsequently accepted.

PUBLISHER’S NOTE

The American Society for Microbiology and Applied and Environmental Microbiology would like to inform the readers that this article is a corrected and republished version of https://doi.org/10.1128/AEM.00991-18, which was retracted (https://doi.org/10.1128/AEM.00330-21). As stated in the retraction notice, after publication, the authors performed additional analyses and “discovered that the RNA editing results (but not the gene expression results) are artifactual.” Thus, one of the main conclusions of the original paper can no longer be supported, but the conclusions around the “substrate-specific differential gene expression” continue to be validated and described in this paper. Therefore, the authors were allowed to submit an entire, corrected manuscript for consideration to the journal. The manuscript was peer reviewed and subsequently accepted.