sulphureus ATCC 52600 genome does not resemble typical brown-rot fungi
Genomic sequencing of filamentous fungi followed by transcriptomic and proteomic approaches has been widely employed to understand the strategies of microorganisms to degrade plant biomass (22–28). Overall, the L. sulphureus ATCC 52600 genome revealed only subtle differences compared to the previously sequenced L. sulphureus var. sulphureus v1.0 (21), indicating that the strains might have undergone some changes in their ecological niches to shape their genomes to the environmental conditions. Our phylogenetic analysis, providing high resolution on the evolutionary history of organisms by considering whole-genome information (28), complements the previous phylogeny of the order Polyporales (10). The phylogenetic tree (Figure 1) strongly supports monophyletic clades for the families within the order Polyporales. L. sulphureus ATCC 52600 clusters with L. sulphureus var. sulphureus v1.0 and W. cocos giving further support to the existence of the family Laetiporaceae Jülich, as proposed previously by Justo and colleagues and currently present in Mycoguide, but retrieved as an invalid name in MycoBank and Index Fungorum.
The genomic CAZyme content in both L. sulphureus strains and the closely related brown-rot Polyporales shows a fairly typical number of GHs, CEs, PLs, and GTs compared to W. cocos, P. placenta and F. radiculosa, whereas a lower number of CAZymes, particularly GHs, were observed in comparison with Fomitopsis pinicola (Fomitopsidaceae). In turn, L. sulphureus ATCC 52600 shows a higher AA content than in the other genomes (9). Additionally, the genome reveals several similarities with other brown-rot genomes associated with the evolutionary reductions and losses in key enzymes involved in biomass breakdown, especially cellulases and lignin-modifying enzymes (3). Accordingly, it presents a reduced number of genes coding CAZymes from the families GH1, GH3, GH5, GH7, GH10, AA9, and CE1 along with the absence of GH6, GH11, AA3_1, CBM1, and CE15 (Additional file 1: Table S1).
Considering these reductions/absences, other enzymes may also be necessary to achieve an effective breakdown of cellulose/hemicellulose, such as the identified AA9 and AA14 LPMOs. AA9s perform oxidative cleavage on cellulose and other glucans with great importance in lignocellulose degradation (29), presenting an average number of 3 genes in Polyporales genomes (3,25). The recently established family AA14 also groups LPMOs that are widespread in fungi. Within the order Polyporales, white-rot strains present 4.5 genes on average, whereas brown-rot strains present 2.5 genes. This reduction pattern can also be added to the other gene reductions associated with the evolution of the brown-rot lifestyle (30). One AA14 member characterized from the white rot Pycnoporus coccineus presents oxidative activity on xylans of xylan-coated cellulose fibers (30), and its sequence aligns with the L. sulphureus ATCC 52600 AA14 LPMO. One of the sequences, however, lacks the amino terminus that can be attributed to a gap in the genome sequencing analysis (Additional file 2: Figure S3).
Regarding the enzymes involved in the oxidative mechanism, AA3_1 CDHs are absent in L. sulphureus, as verified in P. placenta, W. cocos, and F. pinicola (31). In turn, a large number of genes coding for AA3_2 (aryl alcohol oxidase and glucose 1-oxidase) and AA3_3 (alcohol oxidase) were identified, and the products H2O2 (reduction of oxygen by the oxidases) and hydroquinones (reduction of quinones) can support other enzymes or reactions that are important for the deconstruction of lignocellulose (32). Similarly, AA5_1 glyoxal oxidases and AA6 benzoquinone reductase, which are also involved in the generation of Fenton reagents (33–35), were identified (Additional file 1: Table S1). Notably, the absence of CDH may also suggest the presence of other redox partners for AA9 and AA14 LPMOs, such as AA3_2 flavoenzymes (36) and GMC oxidoreductases, among others (37), or the peroxide produced might be driving LPMOs reaction (38).
The L. sulphureus genome also revealed some distinctions in the enzymatic lignocellulolytic repertoire. For example, the well-known lack of cellulases in brown-rot fungi is generally attributed to a reduced number of GH6/GH7 CBHs (31,39,40), which are recognizably absent in the genomes of brown-rot Polyporales (41). Our sequencing, however, identified one putative GH7 CBH (g8442) in the L. sulphureus ATCC 52600 genome, in accordance with a GH7 CBH previously identified in the secretome of L. sulphureus growing on CMC (19). Sequence analysis shows these enzymes sharing more than 90% identity, and the phylogeny using predicted and characterized fungal CBHs reveals 65% similarity with other fungal CBHs. The sequence identity is higher compared to basidiomycete CBHs (44%) than to the characterized counterparts (33%) that are mainly from ascomycetes (Additional file 2: Figure S4). Additionally, analysis of 42 fungal genomes indicates that brown rots generally have a reduced number of GH45, in a 3:1 ratio in comparison with white rots (9). Our initial search parameters identified one putative GH45 (g10751), coinciding with a GH45 (ID 174393) previously identified in the L. sulphureus secretome (19). These sequences share 92.5% identity, having an expansin domain predicted by InterPro v.78.1 (42), despite the previous classification as GH45 class C (19). Expansins are indeed closely related to GH45 endoglucanases and have been widely found in brown-rot strains (9), performing an important function in reducing biomass recalcitrance, thus increasing the deconstruction of lignocellulose in synergism with cellulases (43).
Lignin degradation and the presence and importance of different lignin-active enzymes in brown rots is a matter of debate, but it is widely recognized that brown rots present a reduced number of laccases and absence of PODs class II in comparison with white-rot strains (8). L. sulphureus ATCC 52600 has AA1_1 and AA1_3 laccases, similarly to F. pinicola, P. placenta, and W. cocos (44). Additionally, 13 predicted PODs were identified in the L. sulphureus ATCC 52600 genome, and two of them presented AA2 domain predicted by dbCAN. InterPro annotation classified one of them as an intracellular POD class I, while the other (g11846) was classified as a fungal ligninase/POD class II. The enzyme presents a predicted SP, and a BLAST search retrieved 87% and 66% identity with PODs class II from L. sulphureus var. sulphureus v1.0 and W. cocos MD-104SS10 v1.0, respectively. POD class II has been reported as a single copy in P. placenta, W. cocos and F. pinicola genomes (31), and the P. placenta peroxidase (Ppl44056) was classified as a basal peroxidase, not closely related to LiP and MnP (45). Laccases in Polyporales are multigenic (46) and have been characterized as functional enzymes in P. placenta and F. pinicola (47–49), playing a role in wood decay performed by P. placenta (47). Significant lignolysis has been observed in Gloeophyllum trabeum (Gloeophyllales) and P. placenta without considering the involvement of PODs class II (6,50). Nevertheless, the biological importance or the precise role of these PODs II found specifically in L. sulphureus and other closely related brown rots are uncertain since these enzymes have not been characterized to date.
Insights of the L. sulphureus ATCC 52600 biomass deconstruction mechanism
Several omics studies analyzing brown-rot fungi with significant taxonomic and niche distances such as W. cocos, F. radiculosa, P. placenta, G. trabeum, and Serpula lacrymans (Boletales) cultivated in different conditions show the common presence of a two-step mechanism employed in biomass deconstruction (8,24,45,51–57). The initial oxidoreductive step persists for 48 h (8), which can be correlated with both the observed slow growth of P. placenta in cellulose and spruce (53) and the L. sulphureus growth and glucose consumption in liquid medium (Additional file 2: Figure S1C).
Based on these studies, our transcriptomics analysis performed in a short cultivation period reveals a series of upregulated genes related to the oxidative mechanism induced by recalcitrant in natura sugarcane bagasse (Figure 3 and Additional file 1: Table S2). The most highly upregulated transcripts include alcohol dehydrogenase, cytochrome P450, aldo/keto reductases, and redox genes involved in the generation of hydrogen peroxide, while hydroquinone dehalogenase is involved in hydroquinone production which, in turn, initiates Fenton reaction by the carrying of Fe3+ (58). Moreover, the presence of AA6 quinone reductases points to this enzyme as the main component in the quinone redox cycle supporting Fenton chemistry, as previously observed in P. placenta (45), while also playing a role in the detoxification process (5). Such observations are consistent with a biodegradative role of Fenton chemistry occurring during early cultivation, as verified in other brown-rot transcriptomes (9,59,60).
Regarding CAZymes (Figure 3A and Additional file 1: Table S2), previous brown rot transcriptomic studies also revealed a small set of cellulases and a very similar and general set of hemicellulases with predicted activity on glucans and mannans (8,45). The upregulation of some cellulases and hemicellulases supports the existence of inducing mechanisms, which rely mostly on substrate exposure and availability, occurring even at the beginning of the degradation process. Additionally, two AA1 laccases-encoding genes were upregulated suggesting the microorganism's ability to partially oxidize lignin substructures. On the other hand, two other AA1 laccases as well as two non-CAZy peroxidases were downregulated, so the importance of ligninases for this fungus remains unclear (Figure 3A and Additional file 1: Table S2). Transcripts of AA9 and AA14 LPMOs were upregulated in the transcriptome, while not being detected in the proteomes, corroborating the concept of LPMOs being naturally produced by fungi during early biomass degradation (59,61). The biological importance of LPMOs for brown-rot fungi remains unclear since their secretion has only been identified in G. trabeum growing on lignocellulose (45). Furthermore, the early upregulation of different pectinases observed in L. sulphureus is similarly verified in P. placenta and G. trabeum. Pectinases act by facilitating the access of other enzymes to the plant cell wall components after pectin removal (8,53).
In contrast to the transcriptome, the secretome data (7-day cultivation) reflects the late hydrolytic decay profile (53), which is supported by the absence of AAs in the secretome produced on Avicel. Comparative analysis of the secretomes allows the identification of a core set of constitutive CAZymes, comprising some GHs with predicted activity on cellulose and a wide diversity of GHs acting on glucans, xylan, mannans, trehalose, starch, and chitin (Additional file 1: Table S3). Apart from the xylan-active enzymes, the hemicellulases set is very similar to the profile observed in the transcriptome. This wide-range enzymatic core allows the fungus to obtain energy sources from substrates with diversified composition, providing an increase in survival capability under different environmental conditions.
Regulatory mechanisms take place after the sensing and transport of inducers, resulting in the secretion of a series of regulated CAZymes directed to substrate degradation. Here, comparing grass and wood substrates, which typically present different compositions (62,63), differences in the secretion of specific enzymes were mainly observed in the cultivation of L. sulphureus on SCB, which indeed resulted in the highest number and diversity of secreted and upregulated proteins. This result may be related to substrate recalcitrance and pretreatment (64,65).
Endoglucanases were poorly secreted by L. sulphureus, apparently playing a minor role in cellulose degradation, despite the importance of processive endoglucanases in brown rots (66). Two GH3 β-glucosidases were upregulated on SCB while GH7 CBH was upregulated on Avicel. These data, in addition to the basal secretion of some CAZymes found in all evaluated substrates, show that CBH is inducible and it is not under catabolite repression, as verified for the endoglucanase from G. trabeum (67) or cellulases from P. placenta (53). However, the gene encoding GH7 CBH was not differentially expressed (transcriptome - early stage), and the secretion of endoglucanases and β-glucosidases, as well as oxidative agents, can compensate for this absence, characterizing less effective biomass decay in early stages (4,6).
In addition to the constitutive hemicellulases, a diversity of mannanases and glucanases were upregulated at both early and late response to biomass degradation and can be explained by a natural preference of brown rots for softwoods (1,3,7,56). There is evidence that hemicellulose loss progresses faster than cellulose loss in coniferous wood decay caused by G. trabeum and F. pinicola (68). Additionally, our secretome data clearly show that L. sulphureus targets hemicellulose as part of the hydrolytic late response. Several enzymes active on xylan, the main hemicellulose found in grasses, were secreted, i.e., one GH10 xylanase is upregulated on Avicel and SCB, while the production of another GH10 xylanase is constitutive. Also, one beta-xylosidase is upregulated on SCB, while two α-L-arabinofuranosidases are widely secreted on the polymeric substrates. Corroborating this orientated inducible mechanism, transcripts of most of these enzymes do not show early upregulation; rather, one GH30 xylanase and one GH51 arabinofuranosidase are downregulated. This result shows that L. sulphureus shifts its metabolism to the degradation of grass, however other brown rots from the Antrodia clade have been reported to be inefficient in the degradation of corn stalk (69).
This hemicellulase-enriched profile was also reported for L. sulphureus ATCC 52600 secretome produced on CMC (19). Despite that, the saccharification analysis showed higher glucan conversion than xylan conversion (Figure 5). A similar result was previously reported using L. sulphureus ATCC 52600 secretome for sugarcane bagasse hydrolysis, in which the xylan conversion was around 50% of the glucan conversion (70). The lower biomass conversion efficiency, when compared to commercial cocktail, could be related to slow growth rate and glucose consumption results (Additional file 2: Figure S1C). These results allow us to hypothesize that slow growth and low sugar consumption is an adaptive mechanism reducing the advantages of competitors that demand sugar and grow at high rates.
In Figure 6, an overview of the L. sulphureus strategies for biomass deconstruction is proposed based on the multi-omics data. Our results are consistent with a temporal two-step oxidative-hydrolytic mechanism for the degradation of lignocellulose, while also demonstrating that this fungus does not resemble typical brown-rot fungi in many aspects, thus contributing to the weak dichotomy between white- and brown-rot strains, as previously proposed (25). Additional data applying biological approaches such as gene deletion and analysis of wood decay, as well as biochemical characterization of the enzymes would contribute to addressing this question.