Report

The Plant Cell Wall–Decomposing Machinery Underlies the Functional Diversity of Forest Fungi

See allHide authors and affiliations

Science  05 Aug 2011:
Vol. 333, Issue 6043, pp. 762-765
DOI: 10.1126/science.1205411

This article has a correction. Please see:

Abstract

Brown rot decay removes cellulose and hemicellulose from wood—residual lignin contributing up to 30% of forest soil carbon—and is derived from an ancestral white rot saprotrophy in which both lignin and cellulose are decomposed. Comparative and functional genomics of the “dry rot” fungus Serpula lacrymans, derived from forest ancestors, demonstrated that the evolution of both ectomycorrhizal biotrophy and brown rot saprotrophy were accompanied by reductions and losses in specific protein families, suggesting adaptation to an intercellular interaction with plant tissue. Transcriptome and proteome analysis also identified differences in wood decomposition in S. lacrymans relative to the brown rot Postia placenta. Furthermore, fungal nutritional mode diversification suggests that the boreal forest biome originated via genetic coevolution of above- and below-ground biota.

Many Agaricomycete fungi have been sequenced to date (1), permitting comparative and functional genomic analyses of nutritional niche adaptation in the underground fungal networks that sustain boreal, temperate, and some subtropical forests (2). Through the sequencing of the brown rot wood decay fungus Serpula lacrymans, we conducted genome comparisons with sequenced fungi, including species representing each of a range of functional niches: brown rot and white rot wood decay, parasitism, and mutualistic ectomycorrhizal symbiosis.

Only 6% of wood-decay species are brown rots (3), but being associated with conifer wood (4), they dominate decomposition in boreal forests. Their lignin residues contribute up to 30% of carbon in the organic soil horizons (5). Long-lived (6) and with capacity to bind nitrogen and cations (7), these phenolic polymers condition the nutrient-poor acidic soils of northern conifer forests.

Brown rot wood decay involves an initial nonenzymic attack on the wood cell wall (8), generating hydroxyl radicals (∙OH) extracellularly via the Fenton reaction:Fe2+ + H2O2 + H+ → Fe3+ + ∙OH + H2OHydrogen peroxide is metabolically generated by oxidase enzymes such as glyoxal oxidases and copper radical oxidases. The hydroxyl radical has a half-life of nanoseconds (8) and is the most powerful oxidizing agent of living cells. However, we do not know how it is spatially and temporally targeted to wood cell wall components. Divalent iron is scarce in aerobic environments, where the fungus is obligate and the trivalent ion is energetically favored. Phenolates synthesized by brown rot fungi, including S. lacrymans (9), can reduce Fe3+ to Fe2+. Such phenolates may be modified lignin derivatives or fungal metabolites (10). After initial bond breakages in the cellulose chain, side chain hemicelluloses (arabinan and galactan) are removed, followed by main chains [xylan and mannan (11)], with subsequent hydrolysis of cellulose by synergistic glycoside hydrolases (GHs). Residual lignin is demethylated. In contrast, white rot fungi decompose both cellulose and lignin, with free radical attack theorized to break a variety of bonds in the lignin phenylpropanoid heteropolymer.

S. lacrymans is in the Boletales, along with several ectomycorrhizal lineages (Fig. 1A) (12). S. lacrymans is thus phylogenetically distant from brown rot Postia placenta (Polyporales) (13), as well as other sequenced ectomycorrhizal fungi (14, 15), parasites, and white rot wood decomposers (16). We estimated divergence dates in fungal phylogeny using the data set of Binder et al. [supporting online material (SOM), molecular clock analyses] (17), with two well-characterized fungal fossils that were used to calibrate the minimum ages of the marasmioid (Fig. 1A, node 10) and suilloid clades (Fig. 1A, node 11). The estimated age of the split between Serpula and its ectomycorrhizal sister-group Austropaxillus (53.1 to 15 million years ago) (Fig. 1A and table S11) suggests that transition from brown rot saprotrophy to mutualistic symbiosis occurred after rosids (Eurosids I) became widespread (Fig. 1A) (18). Diversification in fungal nutritional modes occurred alongside diversification of angiosperms and gymnosperms, as these fungi are currently associated with members of both gymnosperms (Pinaceae) and angiosperms (18).

Fig. 1

Molecular phylogeny and lignocellulose-active gene evolution in the Agaricomycetes. (A) Chronogram of Agaricomycetes inferred from a combined six-gene data set by use of Bayesian relaxed molecular clock analyses. Time divergence estimates (in millions of years) are presented as 95% highest posterior density (HPD) node bars in light blue, which describe the upper and lower boundaries of time estimates, and as mean node ages (numbers in bars). The HPD of nodes that were calibrated with fossil ages are in red, and the Serpula-Austropaxillus split is highlighted by a black node bar. The numbering of nodes in bold type corresponds to the tMRCA statistics (time to most recent common ancestor) summarized in table S11. (B and C) Patterns of gene duplication and loss in (B) 12 lignocellulose-active CAZy and (C) 7 oxidoreductase gene families estimated by means of gene tree–species tree reconciliation analysis (fig S3). Red, blue, and black branches indicate lineages with net expansions, net contractions, or no change in copy number, respectively. Numbers at nodes and along branches indicate estimated copy numbers for ancestral species and ranges of gains and losses, respectively, estimated by using 90 and 75% bootstrap thresholds for gene trees in reconciliations. Bars indicate copy numbers in sampled genomes.

S. lacrymans comprises two subgroups that diverged in historic time (19), S. lacrymans var. shastensis, which is found in montane conifer forest, and S. lacrymans var. lacrymans, which is a cause of building dry rot. Two S. lacrymans var. lacrymans complementary monokaryons (haploids of strain S7), S7.9 (A2B2) and S7.3 (A1B1) (20), were sequenced via Sanger and 454 pyrosequencing, respectively. The genome of S. lacrymans S7.9 was 42.8 megabase pairs (Mbp), containing 12,917 gene predictions (21).

We analyzed 19 gene families of enzymes for lignocellulose breakdown: carbohydrate active enzymes (CAZys; www.cazy.org) (22) (GHs and carbohydrate esterases) and oxidoreductases (table S9). Losses and expansions in these families were compared across 10 fungi, including Agaricomycetes, with a range of nutritional modes (Fig. 1, B and C, and table S9). Convergent changes in enzyme complement were found in the two independently evolved brown rot species, with parallels in the ectomycorrhizal Laccaria bicolor (fig. S3 and table S9). The inferred most recent common ancestor of the Agaricales, Boletales, and Polyporales is predicted to be a white rot with 66 to 83 hydrolytic CAZy genes and 27 to 29 oxidoreductases (Fig. 1, B and C). Brown rot and ectomycorrhizal fungi have the fewest hydrolytic CAZy genes. Brown rot fungi have the fewest oxidoreductases, not because of gene losses but because of gene duplications in white rot species.

Both brown rot and ectomycorrhizal fungi lacked class II peroxidases, which are used by white rot fungi in depolymerizing the lignin matrix of wood and unmasking usable cellulose embedded within it. This family was expanded in the white rots Coprinopsis cinerea, Phanerochaete chrysosporium, and Schizophyllum commune, with 29, 43, and 24 genes, respectively, with only 19 each in S. lacrymans and P. placenta. Oxidoreductases conserved in brown rot fungi included iron and quinone reductases and multicopper oxidases (fig. S3 and table S8). Absence of ligninolysis in brown rots raises the question of how they achieve pervasive cellulolysis in wood with the lignin matrix intact.

GH gene families had parallel patterns of losses and expansion in both brown rots and ectomycorrhizas. CAZy families GH5 (endoglucanases, hydrolyzing cellulose) and GH28 (pectinases, hydrolyzing intercellular cohesive polysaccharides in plant tissues) were expanded in both brown rot species, in which they might facilitate intercellular enzyme diffusion, and retained in L. bicolor, in which they might facilitate intercellular penetration of living roots. Both brown rot species lacked GH7 (endoglucanase/cellobiohydrolase CBHI), and GH61 genes—with unknown function but recently implicated in oxidative attack on polysaccharides (23)—were reduced. GH6 (cellobiohydrolase CBHII) and cellulose-binding modules (CBM1), which were absent from P. placenta (13), were present in S. lacrymans. One CBM was associated with an iron reductase in a gene (S. lacrymans S7.9 database protein ID, 452187) originally derived from a cellobiose dehydrogenase (fig. S5).

The general utility of the conserved suite of GH genes in wood decay by S. lacrymans was supported through transcriptomic and proteomic analysis. Carbohydrate-active enzymes accounted for 50% of proteins identified (table S14), and 33.9% of transcripts regulated greater than 20-fold by S. lacrymans growing on pine wood as compared with glucose medium (fig. S4). Cellulose-, pectin-, and hemicellulose-degrading enzymes (GH families 5, 61, 3, and 28) were prominent, and GH5 endoglucanase (S. lacrymans S7.9 database protein ID, 433209) and GH74 endoglucanase/xyloglucanase (S. lacrymans S7.9 database protein ID, 453342) were up-regulated greater than 100-fold.

We conclude that brown rot fungi have cast off the energetically expensive apparatus of ligninolysis and acquired alternative mechanisms of initial attack. Wood decomposition by S. lacrymans may involve metabolically driven nonenzymatic disruption of lignocellulose with internal breakage of cellulose chains by highly localized ∙OH radical action. Mycelia in split plates mimicking realistic nutrient heterogeneity (fig. S1) produced variegatic acid (VA), an iron-reducing phenolate (Fig. 2, A to C), via the Boletales atromentin pathway, which was recruited in S. lacrymans for the Fenton’s reaction. The genome was rich in secondary metabolism genes (table S15), including a putative atromentin locus (24). Mycelium imports amino acids to sites of wood colonization (25), which is consistent with observed up-regulation of oligopeptide transporters on wood (table S12). Localizing variegatic acid production to well-resourced parts of the mycelium could enhance Fenton’s chemistry in contact with wood.

Fig. 2

(A) Proposed chemical reaction demonstrating iron redox cycling by S. lacrymans secondary products. (B) Iron reduction capacity of S. lacrymans ethyl acetate extracts (60% variegatic acid, 15% xerocomic acid) measured with the Ferrozine assay (21) and compared with 2,3- dihydroxybenzoic acid (DHBA), a redox chelator used to stimulate Fenton systems. (C) Comparison of HPLC chromatograms of S. lacrymans ethyl acetate extracts as a function of nitrogen supply. Red trace, nitrogen rich medium (+N); black trace, nitrogen-depleted minimal medium (–N). The identity of the compounds was confirmed with mass spectrometry and by their ultraviolet-visual spectrum (1, variegatic acid; 2, xerocomic acid; 3, atromentic acid).

Wood colonization is presumably followed by coordinated induction of the decay machinery revealed in the wood-induced transcriptome (Fig. 3 and fig. S4). GHs and oxidoreductases accounted for 20.7% of transcripts, accumulating greater than fourfold on wood relative to glucose medium (fig. S4 and table S12). Iron reduction mechanisms included an enzyme harboring a C terminal cellulose-binding module (S. lacrymans S7.9 database protein ID, 452187) (fig. S5) that is up-regulated 122-fold on wood substrate (fig. S4 and table S12). This enzyme, which is present in Ph. chrysosporium but absent from P. placenta (26), is a potential docking mechanism for localizing iron reductase activity, and hence ∙OH generation, on the surface of microcrystalline cellulose. Cellulose-targeted iron reduction, combined with substrate induction of variegatic acid biosynthesis, might explain the particular ability of brown rot fungi in Boletales to degrade unassociated microcrystalline cellulose without the presence of lignin (27).

Fig. 3

Schematic overview of the proposed mechanism of wood decay by S. lacrymans. Scavenging mycelium colonizes a new food source, inducing VA production and expression of oxidoreductase enzymes, which drive hydroxyl radical attack on the lignocellulose composite. CAZy gain access to the weakened composite structure and break down accessable carbohydrates. Cellulose-binding iron reductase targets ∙OH-generating Fenton’s reaction on cellulose chains, releasing chain ends for hydrolysis and assimilation. IR, iron reductase; HQ, hydroxyquinones; CBM, cellulose-binding module.

Thus, comparative genomics helps us understand the molecular processes of forest soil fungi that drive the element cycles of forest biomes (28). Sequenced forest Agaricomycetes revealed shared patterns of gene family contractions and expansions associated with emergences of both brown rot saprotrophy and ectomycorrhizal symbiosis. In Boletales, loss of aggressive ligninolysis might have permitted brown rot transitions to biotrophic ectomycorrhiza, which is promoted in soils impoverished in nitrogen by brown rot residues, and by the nutritional advantage conferred by the connection to a mycorrhizal network. S. lacrymans and other fungi cultured with conifer roots (29) ensheath Pinus sylvestris roots with a mantle-like layer (fig. S6), suggesting nutrient exchange.

The chronology of divergences in extant fungal nutritional mode (Fig. 1A) matches the predicted major diversification in conifers (18), suggesting that the boreal forest biome may have originated via genetic coevolution of above- and below-ground biota.

Supporting Online Material

www.sciencemag.org/cgi/content/full/science.1205411/DC1

Materials and Methods

SOM Text

Figs. S1 to S6

Tables S1 to S15

References (30–89)

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. Acknowledgments: J. Schilling, University of Minnesota, and D. Barbara, University of Warwick, critically reviewed the manuscript; T. Marks designed graphics; and B. Wackler and M. Zomorrodi gave technical assistance. Assembly and annotations of S. lacrymans genomes are available at www.jgi.doe.gov/Serpula and DNA Data Bank of Japan/European Molecular Biology Laboratory/GenBank, accessions nos. AECQB00000000 and AEQC00000000. The complete microarray expression data set is available at the Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/) accession no. GSE27839. The work was conducted by the U.S. Department of Energy Joint Genome Institute and supported by the Office of Science of the U.S. Department of Energy under contract DE-AC02-05CH11231. Further financial support is acknowledged in the supporting online material on Science Online.
View Abstract

Navigate This Article