Proceedings of the Royal Society B: Biological Sciences
You have accessResearch articles

Cross-kingdom interactions and functional patterns of active microbiota matter in governing deadwood decay

Witoon Purahong

Witoon Purahong

Department of Soil Ecology, UFZ-Helmholtz Centre for Environmental Research, Theodor-Lieser-Str. 4, D-06120 Halle (Saale), Germany

[email protected]

Contribution: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing

Google Scholar

Find this author on PubMed

,
Benjawan Tanunchai

Benjawan Tanunchai

Department of Soil Ecology, UFZ-Helmholtz Centre for Environmental Research, Theodor-Lieser-Str. 4, D-06120 Halle (Saale), Germany

Bayreuth Center of Ecology and Environmental Research (BayCEER), University of Bayreuth, 95440, Bayreuth, Germany

Contribution: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing

Google Scholar

Find this author on PubMed

,
Sarah Muszynski

Sarah Muszynski

Institute for Bioanalysis, Coburg University of Applied Sciences and Arts, 96450 Coburg, Germany

Contribution: Formal analysis, Investigation, Methodology, Writing – review & editing

Google Scholar

Find this author on PubMed

,
Florian Maurer

Florian Maurer

Institute for Bioanalysis, Coburg University of Applied Sciences and Arts, 96450 Coburg, Germany

Contribution: Formal analysis, Investigation, Writing – review & editing

Google Scholar

Find this author on PubMed

,
Sara Fareed Mohamed Wahdan

Sara Fareed Mohamed Wahdan

Department of Soil Ecology, UFZ-Helmholtz Centre for Environmental Research, Theodor-Lieser-Str. 4, D-06120 Halle (Saale), Germany

Department of Botany, Faculty of Science, Suez Canal University, Ismailia 41522, Egypt

Contribution: Data curation, Formal analysis, Validation, Writing – review & editing

Google Scholar

Find this author on PubMed

,
Jonas Malter

Jonas Malter

Institute for Bioanalysis, Coburg University of Applied Sciences and Arts, 96450 Coburg, Germany

Contribution: Formal analysis, Investigation, Writing – review & editing

Google Scholar

Find this author on PubMed

,
François Buscot

François Buscot

Department of Soil Ecology, UFZ-Helmholtz Centre for Environmental Research, Theodor-Lieser-Str. 4, D-06120 Halle (Saale), Germany

German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Deutscher Platz 5e, D-04103 Leipzig, Germany

Contribution: Resources, Supervision, Writing – review & editing

Google Scholar

Find this author on PubMed

and
Matthias Noll

Matthias Noll

Bayreuth Center of Ecology and Environmental Research (BayCEER), University of Bayreuth, 95440, Bayreuth, Germany

Institute for Bioanalysis, Coburg University of Applied Sciences and Arts, 96450 Coburg, Germany

[email protected]

Contribution: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing

Google Scholar

Find this author on PubMed

Abstract

Microbial community members are the primary microbial colonizers and active decomposers of deadwood. This study placed sterilized standardized beech and spruce sapwood specimens on the forest ground of 8 beech- and 8 spruce-dominated forest sites. After 370 days, specimens were assessed for mass loss, nitrogen (N) content and 15N isotopic signature, hydrolytic and lignin-modifying enzyme activities. Each specimen was incubated with bromodeoxyuridine (BrdU) to label metabolically active fungal and bacterial community members, which were assessed using amplicon sequencing. Fungal saprotrophs colonized the deadwood accompanied by a distinct bacterial community that was capable of cellulose degradation, aromatic depolymerization, and N2 fixation. The latter were governed by the genus Sphingomonas, which was co-present with the majority of saprotrophic fungi regardless of whether beech or spruce specimens were decayed. Moreover, the richness of the diazotrophic Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium group was significantly correlated with mass loss, N content and 15N isotopic signature. By contrast, presence of obligate predator Bdellovibrio spp. shifted bacterial community composition and were linked to decreased beech deadwood decay rates. Our study provides the first account of the composition and function of metabolically active wood-colonizing bacterial and fungal communities, highlighting cross-kingdom interactions during the early and intermediate stages of wood decay.

1. Introduction

Deadwood is a major reservoir of photosynthetically fixed carbon (C) and nutrients in forest ecosystems that are highly resistant to decay [1]. The contribution of wood to total current C stock in the world's forests is estimated to reach 73 Pg C or 8% of the global C pool [2]. Mineralization of deadwood into simple compounds is carried out by a broad range of saprotrophs, including insects, fungi and bacteria, which can share such compounds with other heterotrophic organisms [35]. For instance, saprotrophic fungi are supported during wood decay by bacterial community members that provide bio-available nitrogen (N) sources [3,4,6,7]. Atmospheric N2 fixing microorganisms (diazotrophs) from the bacterial orders Burkholderiales and Rhizobiales were strongly correlated with principal fungal deadwood decomposers, especially at the initial stage of decomposition [7]. N-rich fungal enzymes and fruiting bodies were also found to be positively correlated by the presence of diazotrophs [6].

Different abiotic factors shape the deadwood microbiota in forest ecosystems. At plot level, deadwood quantity and quality on forest floor, dominant tree species, forest structure and age, soil physico-chemical variables, soil pH, temperature, humidity, elevation, plot location, slope and micro-climate were correlated with richness and distribution patterns of wood associated fungi [8,9]. However, only micro-climatic conditions and forest age were significantly correlated to wood-associated bacterial community composition [7,10]. In addition, the dominant tree species in forests supported deadwood decomposition of the same dominant tree species, which was mediated by a defined fungal community without significant impact on bacterial community composition [11]. At the individual wood piece level, wood species and anatomy (deciduous versus conifer), wood compartments (sapwood versus heartwood), wood physico-chemical properties such as wood density, size, water content, pH, C and N content, C : N ratio, lignin content explained the fungal and bacterial community patterns [9,12,14]. While these wood physico-chemical variables are tree-species-specific, they also change along the succession of decomposition [15]. Biotic factors such as fungal-fungal, bacterial-bacterial and fungal-bacterial interactions affect deadwood microbiota and their functional patterns along the decomposition process [1618].

Although a significant amount of data and knowledge of wood-associated microbiota with Next-Generation Sequencing (NGS) were gathered, the majority of such studies to date are based on total genomic DNA [19,20]. The total genomic DNA pool consists of both intracellular DNA (represents the intact potentially living cells, including both active and dormant cells) and extracellular DNA (represents the DNA originates from both cell lysis and active secretion) [21]. Thus, the present microbial community datasets derived from total DNA represent both active and inactive (dead or dormant) microbes [12,21] of historic and present microbial community compositions [12], including those community members that used to live within the trees before tree death such as endophytes, ectomycorrhizal and arbuscular mycorrihzal fungi, animal parasite and lichens [8]. Diverse saprotrophic fungal communities were frequently detected in deadwood but mostly without addressing their activity status [22]. In addition, bacterial community compositions (including diazotrophs) in deadwood were correlated to biological N2 fixation, however without information as to which part of the microbial community was active during N2 fixation [6,7,23]. The wood degrading enzymes are known to have a complex structure and require a large amount of resources [1,8,24], thus microbes have to be very active and capable of translocating many resources to produce such enzymes. Community-level analyses of metabolically active microbial communities on a nucleic acid-based level are well established and can be performed by extractions of RNA [9] and also newly synthesized DNA (obtaining by Bromodeoxyuridine (BrdU) labelling and immunocapture assay) [25]. Shifts in microbial community composition to environmental parameters were much faster on RNA than on DNA level [19], but concentration of rRNA and microbial growth attributes are not simply correlated, intracellular rRNA turnover is specific to strain and environmental conditions, and dormant cells may also contain rRNA [20]. In turn, BrdU-based approaches are long-term established in mammalian cell biology and medicine [26] and can clearly elucidate proliferating microorganisms substrate-independently in various environments [25,27]. However, a few bacteria are not capable of BrdU uptake or DNA incorporation [26].

In this experiment, BrdU labelling and immunocapture assay coupled with a NGS approach [28] was employed to investigate the colonization of active microbiota on gamma sterilized Norway spruce and European beech specimens after 1 year of exposure. Gamma sterilization was applied to inactivate the wood-inhabiting communities and thereby to remove potential priming effects [16,29]. Therefore, the colonization from surrounding environments was envisaged.

We aimed to (i) describe and compare the richness and community composition of metabolically active wood microbiota (including both bacteria and fungi), (ii) disentangle the ecological and environmental factors that correlate with richness and community composition of the metabolically active deadwood microbiome, (iii) investigate the general relationships between the active bacteria and fungi as well as their co-occurrence network patterns, and (iv) link microbiome richness, taxonomically resolved identity and microbiome composition with microbial-mediated ecosystem functions such as C, N and phosphorus (P) contents, acquisition of hydrolytic enzyme activities and lignin-modifying enzyme activities and mass loss. Wood tree species and forest type played a significant role in shaping richness and community compositions of overall genomic DNA approach, and we hypothesize that holds also true for metabolically active wood microbiota. We expected that especially diverse diazotrophs and co-occurring fungal saprotrophs are highly active in both spruce and beech wood. We also hypothesized that high hydrolytic enzyme activities are linked to accelerated wood decomposition rates, which are related to the active fraction of both bacteria and fungi. By contrast, we assumed that lignin-modifying enzymes, especially general peroxidase and manganese peroxidase, are more related to active fungal community members.

2. Material and methods

(a) Exploration sites, experimental design and sample preparation

Investigated forest plots were part of the German Biodiversity Exploratories [30] and located in the UNESCO Biosphere Reserve Swabian Alb in the southwest of Germany. The field experiment was conducted on 16 forest plots (each one hectare) of the exploratory of which eight plots are located in a Fagus sylvatica and eight plots in a Picea abies dominated forest site. Furthermore, investigated plots differed in forest management intensity (ForMI) as introduced earlier [31]. We excluded the results from Norway spruce and European beech specimens in some plots due to low DNA quality and quantity. The final experimental design consisted of specimens of two tree species × 2 forest types × 6 independent replicate plots.

Sapwood deadwood specimens of one F. sylvatica and one P. abies tree trunk were cut in 50 × 25 × 15 mm. Each specimen was gamma sterilized at 70 kGy by Synergy Health Radeberg GmbH (Radeberg, Germany) to inactivate endophytes as explained earlier [32]. Afterward specimens were dried at 103 ± 2°C for 48 h, and dry weight was determined precisely under kiln-dried conditions.

In May 2017, both of two F. sylvatica and two P. abies deadwood specimens were placed in mesh bags on top of the forest ground of each forest plot with 10 cm distance between each specimen to obtain the same environmental conditions, and after 370 days of exposure specimens were collected. One specimen was directly treated with 0.1 ml of 10 mM BrdU on top for 48 h at room temperature as explained earlier [25], while the other specimen was shock frozen with liquid N at the field site to measure mass loss, wood pH, enzyme activities, C and N content and its isotopic signatures. After BrdU treatment, specimens were shock frozen with liquid nitrogen and stored at −80°C until further use.

Mass loss and wood pH measurement was carried out as explained recently [32]. Enzyme activities such as acid phosphatase, β-glucosidase, cellobiohydrolase, manganese peroxidase, N-acetylglucosaminidase, laccase, peroxidase and xylosidase were assessed to differentiate C, N and P acquisition as explained earlier [24,28]. Content of C and N and its isotopic signatures was measured from milled sub-samples by a vario EL III element analyser combined with an Isoprime 100 stable isotope ratio mass spectrometer (EA-IRMS) (Elementar Analysensysteme GmbH, Langenselbold, Germany).

(b) Analyses of metabolically active wood microbiota using BrdU labelling technique, Illumina sequencing and statistics

DNA was extracted from the BrdU-treated specimens using Quick-DNA Fecal/Soil Microbe Miniprep Kit (Zymo, California, USA). BrdU-labelled DNA from the metabolically active microbial cells was isolated from each total DNA extract by a BrdU immune-capture approach as described previously [25,28] and in the electronic supplementary material. Details of the construction of the bacterial and fungal amplicon libraries, Illumina sequencing and sequence data processing can be found in the supplementary material.

The effects of tree species of specimen on bacterial and fungal ASV richness and community composition were analysed using t-test and permutational multivariate analysis of variance (PERMANOVA), based on Jaccard distance measure. Correlations between various factors and the co-occurrence network patterns of the microbial communities were analysed using Spearman's rank correlation. The correlation matrix was visualized and the networks were extracted using Cytoscape v. 3.8.0. For full details of material and methods, and statistical analyses, see electronic supplementary material.

3. Results

(a) Active community composition and richness in deadwood

High diversity (bacterial: 16 phyla (44 classes); fungi: 15 classes) with moderate number of functions (nine functional groups for both bacteria and fungi) were detected with uneven distribution in beech and spruce deadwood (figures 1 and 2). Gammaproteobacteria (highly contributed by Stenotrophomonas, Massilia, Rhodanobacter, Luteibacter) and Alphaproteobacteria (highly contributed by Sphingomonas, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium (ANPR)) highly dominated the bacterial communities in both tree species whereas Acidobacteriae, Actinobacteria, Bacteroidota and Mollicutes were frequently detected bacterial classes (electronic supplementary material, figure S1; genus level in figure 1a). These phyla except Mollicutes were observed very frequently (electronic supplementary material, figure S1). Chemoheterotrophic bacteria (saprotrophs) were the most dominant functional group, followed by bacteria involved with N cycling (ureolysis, diazotrophs, nitrate respiration and reduction) (figure 1b). Diazotrophic bacteria were detected in all specimens and they even dominate/co-dominate the majority of bacterial functions in both tree species (beech = average 34.3 ± 6.0% (mean ± s.e.), spruce = average 18.5 ± 1.4% (mean ± s.e.)) (electronic supplementary material, figure S2a). This dominance of potential diazotrophic bacteria was reflected by a diverse average species richness of 28.5 ± 2.4 and 35.8 ± 2.5 (mean ± s.e.) ASVs in beech and spruce deadwood, respectively (electronic supplementary material, figure S2b). Members of the genus Sphingomonas were the most frequent detected diazotrophic bacteria in beech (7.8 ± 2.0%) and spruce (10.3 ± 1.1%) (mean ± s.e.) deadwood blocks. In addition, ANPR were detected in average 1.9% both in spruce and beech wood blocks (electronic supplementary material, figure S2c).

Figure 1.

Figure 1. Active bacterial community composition at genus level retrieved from decayed beech and spruce deadwood in relative sequence read abundances (a), and their assigned metabolic functions (b). Av = average, B = beech wood, F = spruce wood, BCP = Burkholderia-Caballeronia-Paraburkholderia. The relative sequence read abundances of ASVs assigned to N fixing bacteria, and their richness as well as the relative sequence read abundances of ASVs assigned to the Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium complex (ANPR) are presented in electronic supplementary material, figure S2. Details on class level and presence/absence data can be found in electronic supplementary material, figure S1. (Online version in colour.)

Figure 2.

Figure 2. Active fungal community composition at genus level retrieved from decayed beech and spruce deadwood in relative sequence read abundances (a), and their assigned ecological functions (b). Av = average, B = beech wood, F = spruce wood. Details on class level and presence/absence data can be found in electronic supplementary material, figure S3. (Online version in colour.)

Agaricomycetes (mainly Mycena spp.) highly dominated the fungal communities in specimens of both tree species (average relative abundances = 60 ± 11% (beech) and 69 ± 10% (spruce) (mean ± s.e.)) followed by Sordariomycetes (average relative abundances = 19 ± 7% beech and 19 ± 7% spruce (mean ± s.e.) (electronic supplementary material, figure S3, genus level in figure 2a). Fungal saprotrophs were the most dominant functional groups while other functional groups were rarely detected in the metabolically active fungal communities (figure 2b). There were 26 plant pathogenic fungal ASVs detected with low relative abundances (average 0.6 ± 0.3% (mean ± s.e.), ranging from 0–6.5%) (figure 2b; electronic supplementary material, table S1).

(b) Comparison of the richness and community composition of metabolically active microbiota

Tree species did not significantly impact on richness of the active bacterial and fungal communities (figure 3a,b). Bacterial and fungal richness varied greatly, especially fungal richness (ranging from 3 to maximum 110 ASVs). Bacterial and fungal community compositions also strongly differed among all samples (electronic supplementary material, figures S1d and S3d). The similarity among all the samples was below 30% and 20% for bacteria and fungi. The majority of the deadwood samples were composed each of one to three ASVs that contributed to more than 80% of the relative sequence read abundances (figure 2a). Nevertheless, we found that bacterial community composition in beech specimens was significantly different from those in spruce specimens (electronic supplementary material, figure S1d). The effect of tree species on bacterial community composition was consistent in both beech and spruce forests (F = 1.08–1.12, p = 0.023–0.026). The bacterial genera Sphingomonas, Mucilaginibacter, Rhodanobacter and Halomonas were frequently detected in both tree species. Stenotrophomonas and Luteibacter were enriched in beech specimens, whereas Massilia and Puia were significantly enriched in spruce specimens. Fungal community compositions differed only for spruce specimens incubated in beech and spruce forest (electronic supplementary material, figure S3d). Members of the genus Mycena were the most dominant in the active fungal community compositions (figure 2a).

Figure 3.

Figure 3. Active bacterial (a) and fungal richness (b) and networks retrieved from decayed beech (c) and spruce deadwood (d). Fungi (pink nodes; Angustimassarina (A), Capronia (Ca), Chaetosphaeria (Ch), Cladophialophora (Cp), Cladosporium (Cs), Cyphellophora (Cy), Exophiala (Ex), Trichoderma (T), Mollisia (Ml), Mortierella (Mr), Mycena (My), Phialocephala (P), Umbelopsis (U)), general bacteria (yellow nodes) and diazotrophic bacteria (orange node; Acidiphilium (Ac), Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium (ANPR), Burkholderia (Bu), Derxia (De), Enterobacterales (En), Hyphomicrobium (Hy), Novosphingobium (No), Sphingomonas (Sp), Pseudomonas (PS)) are indicated. Additional information to respective ASV numbers of selected ASVs can be found in electronic supplementary material, table S1.

(c) Factors corresponding with community composition of metabolically active deadwood microbiota

C and N contents were significantly correlated to both bacterial and fungal community compositions associated with beech specimens (electronic supplementary material, table S2). Furthermore, the plot factors such as soil temperature and ForMI also significantly contributed to effect both community compositions. Bacterial and fungal community compositions in spruce specimens were only significantly correlated to soil pH for bacteria and ForMI for fungi (electronic supplementary material, table S2).

(d) Relationships and co-occurrence network patterns of the active microbiota

Richness of active bacterial and fungal ASVs was correlated when both beech and spruce specimens were considered together (R = 0.64, p < 0.001). However, richness of active bacteria and fungi present in spruce specimens was significantly correlated (R = 0.84, p < 0.001) whereas such correlation in beech specimens was not significant. Relationships between fungi and bacteria strongly differed between both tree species (figure 3c,d). The same fungal genera mostly co-occurred with different bacterial taxa. Nevertheless, we observed that diazotrophic bacteria were associated with most fungal saprotrophs, including the main fungal decomposers in this study (Mycena spp.) (compare figure 3c,d with figure 2b).

(e) Link between microbiome richness, composition and microbial-mediated ecosystem functions and processes

Even with similar decomposition time and environmental frame, deadwood samples showed a high variance in mass loss reflecting different decay stages. Beech specimens had average mass loss of 25 ± 4% (mean ± s.e.) whereas spruce specimens had average mass loss of 9 ± 2% (mean ± s.e.) (electronic supplementary material, table S3). Significant correlations between mass loss and enzyme activities were observed in specimens of both tree species. High beech mass loss was significantly correlated to high cellobiohydrolase, β-glucosidase and peroxidase activity, while in spruce mass loss was only positively correlated with cellobiohydrolase (electronic supplementary material, table S4). In addition, patterns of wood degrading enzyme activities in both tree species were highly correlated with wood-physiochemical properties, species richness and relative abundances of specific microbial taxa (electronic supplementary material, table S5). N content in beech specimens was significantly correlated with enzyme activity patterns whereas spruce specimens were significantly correlated to C content (electronic supplementary material, table S5). Mucilaginibacter, ANPR, Hypocreaceae, Trichoderma, Halomonas, Gammaproteobacteria and Bacteroidota were significantly correlated with enzyme activity patterns in beech while Mucilaginibacter, Sphingomonas, Umbelopsis and Umbelopsidaceae in spruce specimens. Among these bacterial taxa, Mucilaginibacter is the only genus significantly correlated with enzyme activity patterns in both deadwood species and the diazotrophic bacteria from ANPR were strongly correlated with enzyme activity patterns only in beech specimens (electronic supplementary material, table S5).

Individual enzyme activities were mostly negatively correlated with richness and/or relative abundances of the majority of bacterial taxa but positively with fungal taxa in beech specimens (figure 4a). Especially Bdellovibrio known as bacterial predators were highly negatively correlated to enzyme activities (5 out of 8 enzymes) in beech specimens (figure 4a). Furthermore, richness of Bdellovibrio spp. significantly corresponding with the bacterial community composition in beech specimens (r2 = 0.43, p < 0.05). Xylosidase and N-acetyl-glucosaminidase activities were positively correlated in spruce specimens with richness/relative abundances of Acidobacteria (including Granulicella) and fungi. All correlated taxa were different between both tree species, and only the relative abundances of Strophariaceae (Tubaria and Hypholoma) and peroxidase activity was similar (figure 4a). The majority of plot factors and wood physico-chemical properties were negatively correlated with individual enzyme activities (figure 4b). However, higher N contents had the most prominent positive correlations to acid phosphatase and mass loss both in beech wood and Xylosidase and N-acetyl-glucosaminidase in spruce wood. Overall, hydrolytic enzyme activities were correlated to active bacteria and fungi whereas lignin-modifying enzyme were positively correlated to active fungi (Strophariaceae).

Figure 4.

Figure 4. Heat map of Spearman's rank correlations between richness versus relative abundance of microbial taxa (a), wood physico-chemical properties and plot factors (b) on wood mass loss versus enzyme activities in Fagus sylvatica or Picea abies. Light and dark green colours of the heat map indicate significant positive correlations (p < 0.05) and pink and red colours indicate significant negative correlations (p < 0.05). Full enzymatic names are presented in the Material and methods section.

Mass loss of beech specimens was positively correlated to richness/relative abundances of ANPR (ρ = 0.81, p < 0.001), fungi (Strophariaceae and Herpotrichiellaceae) as well as C and N content but negatively to Bdellovibrio (ρ = −0.65, p < 0.05) (figure 4a; electronic supplementary material, figure S4). N content in beech specimens was positively correlated with ANPR and δ15N signature (electronic supplementary material, figure S4). In spruce specimens, mass loss was positively linked to richness/relative abundances of Gammaproteobacteria and negatively to Actinobacteria and wood pH (figure 4a).

4. Discussion

Recent studies based on DNA sequencing demonstrated that total bacterial communities in beech and spruce specimens are co-dominated by diverse bacterial phyla including, members of the classes Alphaproteobacteria, Betaproteobacteria, Gammaproteobacteria, Acidobacteriae, Actinobacteria and Bacteroidota [10,13,14]. However, our results based on active bacterial fraction revealed that only Gammaproteobacteria, Alphaproteobacteria and Bacteriodetes governed the deadwood communities. Moreover, the dominant members of the Gammaproteobacteria and Alphaproteobacteria were also consistent during decay of spruce and beech specimens of our study, regardless of respective decay stage. Although other dominant bacterial phyla, which were previously detected with DNA sequencing approaches, were not dominant in the active bacterial communities, but they were still frequently detected [10,13,14]. Members of the genera Stenotrophomonas, Massilia, Rhodanobacter, Luteibacter, Sphingomonas and ANPR were highly represented in the active bacterial communities. While the majority of these dominant bacteria genera are known for their potential function in C cycle such as cellulose degraders (Stenotrophomonas, Massilia, Luteibacter) [3335] and catabolizing aromatic lignin depolymerization (Sphingomonas) [36], some of them are also known to potentially play important roles in N cycling. Majority of members of Sphingomonas and ANPR are able to fix atmospheric N2 [37], Rhodanobacter spp. show denitrification capabilities [38] and Luteibacter spp. are efficient chitin degraders [39]. Moreover, ANPR was dominant bacterial fraction in the fruiting bodies of saprotrophic Cantharellales indicating a close relationship between both [40]. Thus, our results demonstrate that apart from bacteria involved in C cycling, the bacteria involved in N cycling are highly active in deadwood. In fact, based on bacterial functional analyses, we showed that these N cycling bacteria were potentially drivers for C and N cycling during deadwood decay, including N fixation, ureolysis, nitrate respiration and reduction, and N respiration. It is noteworthy that potential diazotrophic bacteria represent in average 18.5% and 34% of total detected bacterial sequences in spruce and beech specimens, respectively. Moreover, Acidobacteriae (Granulicella) and Bacteroidota (Mucilaginibacter) may also highly contribute in the decomposition of specimens as members of both genera are known as efficient cellulose degraders [39]. The relative abundance of Mucilaginibacter was highly correlated with enzyme activity patterns in both beech and spruce specimens (electronic supplementary material, table S5), indicating that released cellulose from the wood fibres was metabolized by these bacteria. However, such correlations should be carefully interpreted as bacterial occurrence is not always linked to environmental enzyme activities. Interestingly, in our work we also detected relatively high richness of obligate predatory bacteria Bdellovibrio spp. and found that their high richness and increased relative abundances negatively impacted both enzyme activities and decomposition rates in beech specimens. Prey range of Bdellovibrio spp. is relatively narrowly specific to gram-negative bacteria [41]. Frequently detected gram-negative bacteria genera (Pseudomonas, Rhizobium) detected in our study that were classified as potential diazotrophic bacteria are known to be among the potential preys of Bdellovibrio spp. An indirect study showed that predator–prey interactions were known in some ecosystems to impact on microbial ecosystem processes and services [42], but this is the first time we detected such phenomenon during deadwood decay. Direct consequences of the obligate predator Bdellovibrio spp. could shift the bacterial community composition (figure 1), which leads to decreased deadwood decay rates. Deadwood has a high bacterial biomass (2 × 1012 gene copy numbers g−1) [43,44], which is more than enough to enable predatory lifestyle of Bdellovibrio spp. with need of a minimal concentration of 7 × 105 prey cells ml−1 [45,46]. Future studies should address the prey turnover and its consequences for the microbial part in ecosystem services of deadwood decay.

The early decomposition process of deadwood is mediated by fungal endophytes which used to live within the trees before tree death [29]. On the one hand, these fungal endophytes play a direct role as early decomposers to decompose the wood. This aspect has been demonstrated in both beech and spruce wood, which endophytes (beech: Xylaria and spruce: Phialocephala) change their functionality from endophytes to saprotrophs (soft-rot fungi) [8,12]. On the other hand, the endophytes indirectly regulate the wood decomposition rate through the priority effects as initial community determines the subsequent community successions [16,29]. We demonstrated in this study that eliminating the endophytes from specimens can cancel the priority effects of endophytes, thereby enabling opportunities for diverse fungi to colonize deadwood. Despite similar deadwood physico-chemical properties, surrounding environment conditions (including ForMI) and decomposition time, each specimen from each tree species used in this study colonized by different active fungal and bacterial community compositions and their decomposition rates differed greatly. Some wood blocks have reached the late stage of decomposition whereas some wood blocks are still at the very early phase. In addition, ForMI also significantly affected active fungal community composition (electronic supplementary material, table S2), which was already shown in a previous study from the same sites [47]. However, we could show that this shift in the active fungal community composition was correlated with a lower mass loss, lower peroxidase and manganese dependent peroxidase activities compared to forest sites with low ForMI (figure 4). This phenomenon is strongly different from deadwood that contains initial endophytes, which often resulting in homogenization of the subsequent communities depending on the dominant endophytic taxa [8,16,29].

Our study demonstrates that saprotrophs are the most active fungal functional group in specimens whereas other functional groups are also active but their contributions were relatively minor. This can be explained by two observations: (i) fungal saprotrophs gain competitive advantage in specimens as compared with other functional groups due to the sufficient nutrient source availability and their abilities to produce wood-degrading enzymes; and (ii) other fungal functional groups, especially mycorrhizal fungi, plant pathogens and animal pathogens, are slow growing or partly in dormant phase to conserve their energy and waiting for their suitable hosts [48,49]. Fungal endophyte is the only functional group that can substantially compete for deadwood substrate resources with fungal saprotrophs. Although they are eliminated from the initial community, our work showed that endophytes from surrounding environments can re-colonize and proliferate into deadwood. This again demonstrates the ability of fungal endophytes to switch their lifestyle as saprotrophs (commonly as soft rot) [8,29].

The presence of diazotrophs in deadwood logs of spruce [6,50] and beech [6,22] has been previously shown, and the overall majority of these diazotrophs in deadwood were characterized as asymbiotic and free-living diazotrophs. The cleavage and reduction of atmospheric N2 molecule by the nitrogenase system is a very energy-consuming process of the diazotrophs [3]. In our beech specimens, we observed not only a decrease in the C/N ratio based on the removal of the absolute C content but also an increase in absolute N content, indicating an active input of N into the specimen. N in wood is among the most important macronutrients required for fungi to produce enzyme to degrade wood [3,7]. Indeed, we found in our study that N content was highly correlated with mass loss in beech wood. Moreover, there were strong linear relationships of δ15N signature increase and total N content of beech specimens as well as δ15N signature increase and sequent abundances of diazotrophs (ANPR) (electronic supplementary material, figure S4), indicating a linear relationship. As the δ15N signature was getting less negative and more close to δ15N (0‰) of atmospheric N2 [51], a higher proportion of microbial fixed atmospheric N2 was present in deadwood and therefore other sources to enrich absolute N content were less likely. A very recent study showed by metagenomics and metatranscriptomics linked to the N-cycling genes that old and young beech deadwood were dominated by transcripts of N fixation [22], which supports our findings of a very active N fixing bacterial community during beech deadwood decay. Interestingly, the richness of the diazotrophic part of the bacterial community composition, at 20–30% (electronic supplementary material, figure S2a), was very high compared with recent transcriptomic analyses [22], however both analysis types has its drawbacks in over- and underestimating present status. Network analyses showed a clear link between the diazotrophic taxa and fungi (figure 3c and d). The majority of the 13 active mainly saprotrophic fungal taxa were linked to members of the diazotrophic genus Sphingomonas. A cross-kingdom association between ectomycorrhizal and arbuscular mycorrhizal fungi and Sphingomonas was shown several times [52,53], and the presence of Sphingomonas enriched the system with additional N [52]. Moreover, Sphingomonas were mainly found in the early stage of deadwood decay [54], indicating that they are of high importance to initiate accelerated deadwood decay by atmospheric N addition like in our study. The wood-rotting fungal taxa Mollisia was linked to five ASVs, which were all assigned to Sphingomonas (figure 3c). Furthermore, Angustimassarina was linked to Sphingomonas in both spruce and beech deadwood decay, indicating a tree-species-independent partnership between both cross-kingdom taxa (compare figure 3c,d). However, some diazotrophic genera share co-presence with more than one fungal taxa. For instance, Derxia was associated in the co-occurrence network with both endophytic Phialocephala and saprotrophic Chaetosphaeria or Burkholderia with three saprotrophic fungal taxa Trichoderma, Mycena and Umbelopsis (figure 3), indicating a free diazotrophic life style with changing partners. The mass loss of spruce specimens involved a greater richness and diversity of diazotrophic bacteria, indicating that the initial deadwood decay just started and superior taxa will outcompete others in a later decay stage. Moreover, there was no such correlation between increase of δ15N signature and mass loss in spruce deadwood decay, indicating that N acquisition was not high enough to accelerate deadwood decay.

5. Conclusion

The spotlight on the metabolic active fraction during beech and spruce deadwood decay discovered a less diverse fungal and bacterial taskforce with clearly defined functional assignments. Similar biological patterns of colonizers were found between beech and spruce, but different patterns of microbial interactions and functions. Mass loss of deadwood was conducted by saprotrophic fungi and their respective hydrolytic and lignin-modifying enzyme activities. So few active fungi were accompanied by a selected bacterial community that is capable for cellulose degradation, aromatic depolymerization and N cycling including diazotrophs. The latter diazotrophic activity is underlined by N enrichment in the deadwood and an increase of the δ15N signature, which was closely linked to the abundance of ANPR. Moreover, members of the genus Sphingomonas were co-present with most saprotrophic fungi regardless of whether beech or spruce specimens were decayed, highlighting their importance in ecological functioning in the early stage of deadwood decomposition. Further research should employ proteome analyses, ideally in deadwood decay stages, to get more insight into the interaction patterns of the deadwood decaying community.

Ethics

Fieldwork permits were issued by the responsible environmental office ‘Regierungspräsidium Tübingen (Schwäbische-Alb)’ (according to §7BbgNatSchG).

Data accessibility

The raw 16S rRNA gene and ITS region sequences are publicly available and deposited at https://www.bexis.uni-jena.de/ddm/data/Showdata/30944 and https://dx.doi.org/10.5061/dryad.g79cnp5rs [55]. Plot, enzyme and wood physico-chemical property data: https://dx.doi.org/10.5061/dryad.g79cnp5rs [55].

The data are provided in electronic supplementary material [56].

Authors' contributions

W.P.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, resources, supervision, visualization, writing—original draft, writing—review and editing; B.T.: formal analysis, investigation, methodology, writing—original draft, writing—review and editing; S.M.: formal analysis, investigation, methodology, writing—review and editing; F.M.: formal analysis, investigation, writing—review and editing; SFM.W.: data curation, formal analysis, validation, writing—review and editing; J.M.: formal analysis, investigation, writing—review and editing; F.B.: resources, supervision, writing—review and editing; M.N.: conceptualization, formal analysis, funding acquisition, investigation, methodology, resources, supervision, visualization, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

The work has been funded by the DFG Priority Program 1374 ‘Infrastructure-Biodiversity-Exploratories' (grant no. NO834/5-4). The next generation sequencing cost was funded by personal research budgets of W. Purahong from the UFZ-Helmholtz Centre for Environmental Research.

Acknowledgements

We thank the staffs of the three Exploratories. Full acknowledgements can be found in electronic supplementary material.

Footnotes

These authors contributed equally to this study.

Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5958617.

Published by the Royal Society. All rights reserved.

References