Function of sesquiterpenes from Schizophyllum commune in interspecific interactions

Wood is a habitat for a variety of organisms, including saprophytic fungi and bacteria, playing an important role in wood decomposition. Wood inhabiting fungi release a diversity of volatiles used as signaling compounds to attract or repel other organisms. Here, we show that volatiles of Schizophyllum commune are active against wood-decay fungi and bacteria found in its mycosphere. We identified sesquiterpenes as the biologically active compounds, that inhibit fungal growth and modify bacterial motility. The low number of cultivable wood inhabiting bacteria prompted us to analyze the microbial community in the mycosphere of S. commune using a culture-independent approach. Most bacteria belong to Actinobacteria and Proteobacteria, including Pseudomonadaceae, Sphingomonadaceae, Erwiniaceae, Yersiniaceae and Mariprofundacea as the dominating families. In the fungal community, the phyla of ascomycetes and basidiomycetes were well represented. We propose that fungal volatiles might have an important function in the wood mycosphere and could meditate interactions between microorganisms across domains and within the fungal kingdom.


Introduction
Wood contains a high amount of lignin and is therefore difficult to decompose. Therefore, colonization with saprotrophic fungi is considered as the key driver for wood decomposition. The respective basidiomycetes are able to degrade lignin by secretion of lignocellulolytic enzymes, e.g. laccases and other peroxidases [1,2]. Furthermore, bacteria have been reported to colonize wood [3] and to contribute either indirectly to wood decay by competition with fungi or directly by decomposing cellulose and hemicelluloses [4], and it has been indicated that bacteria are also involved in lignin degradation [5]. Saprophytic fungi and bacteria are in competition for the same resources and it is likely that antagonistic interactions influence the community structure, which in turn has an effect on the rate of wood decomposition and nutrient cycling [6].
Wood decomposition involves chemical [7], as well as structural changes [8] affecting the permeability through formation of pores, and it provides a heterogenous environment for the Fungi were initially cultivated on complex yeast extract medium (CYM) with 1.8% agar at 28˚C for several days [24]. Sesquiterpene extraction was performed from S. commune liquid cultures grown for 3 days at 150 rpm and 28˚C. Bacterial strains were cultivated and maintained on Standard I (StdI; Carl Roth, Karlsruhe, Germany) and incubated for 1-2 days at 28˚C. Bacterial swarming activity was determined on soft StdI containing 0.6% agar [25].

Quantification and extraction of fungal sesquiterpenes
Fungal volatile analysis of S. commune surface cultures was carried out using solid phase microextraction (SPME). Headspace volatiles were collected over a period of 1 hour from S. commune cultivated on CYM in polystyrene Petri dishes. The SPME fibers and the holder were obtained from Supelco (Bellefonte, PA, USA). A divinylbenzene-carboxen-polydimethylsiloxane (50/30 μm DVB/CAR/PDMS) fiber was used. The fibers were first conditioned according to the manufacturer´s instructions. Prior to sampling, the Petri dish was opened and accumulated volatiles were removed by a gently stream of nitrogen. The SPME fiber was inserted into the volatile collection unit through a hole in the Petri dish before enclosing it with polyethylene terephthalate foil (PET; Toppits Bratschlauch, Minden, Germany).
For quantification of sesquiterpenes, a closed loop stripping system was used as previously described [26]. In brief, the closed loop stripping system consists of two stainless steel tubes placed over a fungal culture. The collection unit was enclosed with polyethylene terephthalate foil and the tubes were connected to a rotary vane pump (model G 12/02 EB, Denver, Fürstenfeldbruck, Germany) equipped with a stainless steel housing for the trap [27]. The volatiles emitted by the fungi were continuously collected over a period of 3 hours on small charcoal traps (1.5 mg of charcoal, CLSA-Filter; Gränicher, Daumazan sur Arize, France) and eluted in two steps with 50 μL (2 x 25 μl; sample 1) and 150 μl (sample 2) dichloromethane containing an internal standard (dodecane, 50 μg ml -1 ; Merck, Darmstadt, Germany). Measurement was performed using five biological replicates for each strain.
For gas chromatography-mass spectrometry (GC-MS) analysis, a TRACE MS device (Thermo-Finnigan, Bremen, Germany) equipped with a ZB5 column (15 m, 0.25 mm I.D, 0.25 μm film thickness and 10 m guard column; Phenomenex, Aschaffenburg, Germany) was used. Mass spectra were recorded in electron impact (EI) mode at 70 eV, covering a mass range from m/z 33-450 Da. Volatiles were eluted under programmed conditions: 45˚C (2 min isotherm), followed by heating at 10˚C min -1 to 200˚C and at 30˚C min -1 to 280˚C, using helium (1.5 ml min -1 ) as the carrier gas. The GC injector (split ratio 1:10), transfer line and ion source were set at 220, 280 and 200˚C, respectively. Bisabolene isomers were identified by comparing their Kováts retention indices [28] with those from reference databases and a synthetic bisabolene mix standard by injecting an n-alkane standard solution C 8 -C 20 in hexane (Fluka, Steinheim, Germany; see S1 Fig in S1 File).
The statistical analyses were performed with unpaired Student's t-test.

Exposure to fungal volatiles
To study the effect of S. commune volatiles, fungi and bacteria were cultivated in two-compartment Petri dishes (Greiner, Frickenhausen, Germany). The physical barrier in these Petri dishes excludes physical interaction, but allows for signaling exclusively via the gas phase. Fungal agar plugs of 2 mm diameter were used for inoculation on CYM plates closed with Parafilm (Bemis, Neenah, USA) prior to incubation for 8 days at 28˚C. Control plates were prepared similarly using sterile growth medium in the second compartment. Analysis of fungal growth was documented using the SPOT advanced software (version 4.6, Diagnostic instruments, Houston, USA). To study the effect of S. commune volatiles on bacteria, S. commune was precultivated on CYM for 5 days at 28˚C before bacterial strains were inoculated after growth for 1-2 days in StdI at 120 rpm and 28˚C, washing in 0.8% NaCl, and inoculated at an optical density OD 600 0.6 diluted 10 −5 in 0.8% NaCl. Incubation took place for 1-2 days at 28˚C with five biological replicates. The activity of S. commune volatiles was determined by comparing the cfu ml -1 of bacteria exposed to fungal volatiles to that of bacteria exposed to sterile CYM. Bacterial swarming activity was determined on StdI soft agar [25]. The effects of sesquiterpenes on fungal and bacterial growth were evaluated using two-compartment Petri dishes with indicator fungi or bacteria in one compartment on CYM or StdI, respectively. The second compartment was equipped with sterile paper disks (6 mm diameter; Carl Roth, Karlsruhe, Germany) and loaded with the selected sesquiterpene bisabolene (30 μg; Alfa Aesar, Karlsruhe, Germany) or (-)-(1R,2S)-β-bisabolol (2 mg) corresponding to a final concentration of 0.6 mg l -1 bisabolene and 40 mg l -1 (-)-(1R,2S)-β-bisabolol. Pure sesquiterpenes were dissolved in n-hexane (GC-ultra grade, Carl Roth, Karlsruhe, Germany) to obtain stock solutions. Calculation assumed the complete compound evaporation from the filter paper and corresponds to a 6-day-cocultivation period. As a control, n-hexane (GC-ultra grade, Carl Roth,Karlsruhe, Germany) was used. The sealed plates were incubated for several days at 28˚C. Bacterial swarming motility assayed on StdI soft agar. The activity of the pure compounds was determined by comparing the diameter of bacterial swarming to that of bacteria exposed to n-hexane. Data are shown as average values with standard deviation. The statistical analyses were performed with unpaired Student's t-test.

Community analyses
Wood was sampled from Salix sp. in Jena, Germany (N 50˚55' 16.4", E 11˚34' 54.1", altitude 149 m) in 2016 adjacent to fruiting bodies of S. commune. The upper 3 cm of wood were removed; the surface sterilized with ethanol, and a sterile drill was used. The core was stored at 4˚C for further analysis. DNA extraction was performed from 1 g wood (fresh weight), which has been flash-frozen in liquid nitrogen and ground to a fine powder using the MoBio Soil DNA Extraction Kit (MoBio, Carlsbad, USA). Extraction was performed three times and DNA was pooled to ensure a representative wood community. DNA was purified using the OneStep PCR Inhibitor Removal Kit (Zymo Research, Irvine, USA) and sequenced (MR DNA, Shallowater, USA). The 16S rDNA hypervariable regions V1 to V3 were addressed using the primer pair whoi341 (5'-CCT ACG GGN GGC WGC AG) and whoi785 (5'-GAC TAC HVG GGT ATC TAA TCC) for bacteria and ITS1 using the primers ITS1F (5'-CTT GGT CAT TTA GAG GAA GTA A) and ITS2 (5'-GCT GCG TTC TTC ATC GAT GC) for fungi [29]. Data analysis was performed using the MR DNA analysis pipeline, including joining of sequences after removing short sequences (<150 bp). Operational taxonomic units (OTUs) were defined by 97% similarity. The Illumina MiSeq data and relevant sequences can be found under Bio-Project PRJNA662029 (NCBI at http://www.ncbi.nlm.nih.gov). Rarefaction curves, Shannon diversity index and Simpson diversity index were generated using Past 3.15 [30].

Bacterial and fungal community in wood degradation
Wood samples were collected from Salix sp. showing early signs of wood degradation adjacent to a fruiting body of S. commune. Only a small proportion was cultivated on StdI (1.71 x 10 11 cfu per g of wood) and TSB medium (3.54 x 10 6 cfu per g of wood). Successful isolation was achieved for Gammaproteobacteria, with the phyla Pseudomonadaceae and Rhodanobacteraceae, or Actinobacteria (mainly Microbacteriaceae and Intrasporangiaceae; S2 Table in S1 File). Aside from S. commune, no other fungal species could be cultivated confirming both its presence and its suppression of other wood rotting fungi (S3 Table in S1 File). The altogether low numbers of cultivatable species suggested the occurrence of interspecies competition in wood, which restricts bacterial or (other) fungal growth and might lead to the isolation of presumably adapted organisms.
To gain a better insight into the community structure, in addition to the necessary cultivation approach yielding strains for subsequent testing of volatile effects on co-occurring organisms, the fungal and bacterial community composition was analyzed in a culture-independent approach. A total of 160,145 high quality filtered bacterial and 35,956 high quality filtered fungal read pairs were obtained from wood samples. However, the rarefaction curves did not approach saturation, which indicates that the diversity is high, and yet undetected microorganisms are present in the community (S2 Fig and S4 Fig in S1 File).

Effect of S. commune volatiles on fungal and bacterial growth
The low number of cultivatable, presumably adapted organisms, may bias the investigation of interspecies interaction. Moreover, secondary metabolite profiles of S. commune change during life stage, with specifically dikaryons producing a variety of VOCs of different chemical classes [21]. To investigate the effect of S. commune volatiles on the growth of other basidiomycetes (Pleurotus ostreatus, Ganoderma lucidum, Flammulina velutipes and Kuehneromyces mutabilis) and bacteria (Pseudomonas fluorescens, Erwinia amylovora and Serratia marcescens), co-cultivation in two-compartment Petri dishes excluding physical contact was performed. In these experiments, S. commune wildtype monokaryon and dikaryon did not influence the growth of G. lucidum, F. velutipes and K. mutabilis, and only slightly inhibited P. ostreatus (S4 Fig in S1 File).
Since S. commune thn mutants show a similar VOCs pattern like a S. commune dikaryon with increased production of sesquiterpenes including (-)-(1R,2S)-β-bisabolol, β-bisabolene and (E)-γ-bisabolene, a more intense response might be expected. Indeed, all tested fungi responded to volatiles of three different thn mutants with a strong mycelial growth reduction (Fig 2). This effect was least pronounced with G. lucidum which significantly responded only to two thn strains, S. commune 4-39 thn and S. commune W22 thn. Since we observed a strong effect of S. commune volatiles on the tested indicator fungi, we next investigated whether also bacteria might response to fungal volatiles. Despite the absence of direct contact, S. commune thn strains exerted a strong growth reduction in P. fluorescens and E. amylovora; again no effect was observed using wildtype S. commune VOCs (Fig 3). Only slight inhibition was scored for S. marcescens (S5 Fig in S1 File). However, specifically with the latter bacterium, a reduced swarming activity was scored (Fig 4), which might impact colonization of timber. Again, these effects were more prominent when S. marcescens has been exposed to volatiles of the thn mutant strains of S. commune as compared to the wildtype monokaryotic strain.

Effects of sesquiterpenes on fungal and bacterial growth
The presence of sesquiterpenes in the volatile profile of S. commune thn mutants was considered potentially responsible for the growth inhibition of the tested fungi and bacteria, because their absence in S. commune wildtype monokaryon correlated with reduced effects. The lack of an authentic standard prompted us to extract (-)-(1R,2S)-β-bisabolol directly from S. commune thn cultures, while for β-bisabolene and (E)-γ-bisabolene, a synthetic bisabolene mixture was available. We tested the sesquiterpenes for their antimicrobial activity, with the bisabolene mixture

PLOS ONE
added at a concentration corresponding to a 6 day co-cultivation period. This led to growth inhibition in G. lucidum, F. velutipes and K. mutabilis (see Fig 6). The influence of (-)-(1R,2S)-β-bisabolol was even stronger, and a growth reduction for all tested fungi was observed.
With respect to the tested bacteria, neither (-)-(1R,2S)-β-bisabolol, nor the mixture of bisabolenes had an effect on bacterial growth that would mimic the effect of S. commune thn volatiles ( S6 Fig in S1 File). Thus, additional VOCs must be necessary for the effect exerted on bacterial growth. However, the bisabolene mix as well as (-)-(1R,2S)-β-bisabolol significantly reduced swarming motility in S. marcescens (see Fig 4).

Discussion
S. commune is adapted to the early colonization of wood infecting dead and living trees by utilizing enzymes that have been described for an intermediate wood decay mechanism between brown and white rot [32]. Here, we show that beneath fruiting bodies of S. commune, a specific microbial community was present that is associated with the wood degradation process. The presence of S. commune was confirmed, dominating the zone of wood decay and suppressing other fungi. Thus, S. commune likely could influence other members of the community. The co-occurring bacterial community represented presumably adapted strains. As a means of influencing the community structure, volatiles were identified.
As a means of study, thn mutants producing higher amounts of (-)-(1R,2S) β-bisabolol [22] were used to test for specific effects in interspecific interactions proven by pure sesquiterpene compounds. The composition of the volatile profile depends on growth condition and is highly variable during culture age, most likely due to shifts in the metabolism [33]. The production of sesquiterpenes and C8-volatiles seems to be linked to mating involving the signature compounds 1-octen-3-ol and 3-octanone, as well as (-)-(1R,2S) β-bisabolol for S. commune dikaryons [21]. Thus, sesquiterpenes are relevant for the local microbial community.
Sesquiterpenes are known fungal VOCs, and wood decay fungi, including Lentinus lepideus, Polyporus pinicola, Resinicium bicolor, H. fasciculare have been shown to produce sesquiterpenes [9,16,34]. Moreover, several plant species release a complex mixture of terpenes, including monoterpenes, sesquiterpenes, and diterpenes in response to wounding representing an important defense strategy or communication signal [35]. Matricaria recutita contains high amounts of α-bisabolol and extracts are used traditionally as herbal medicine against common human ailments [36]. In fungi the antimicrobial properties of sesquiterpenes were linked to defense response [10,16], and other VOCs, like C8 volatiles, benzaldehyde and benzoic acid were shown to be involved in intra-kingdom and inter-domain interactions [37]. Subtle changes in growth rate, morphological alterations and pigment production can effect community composition [16]. We could show a strong decrease in mycelia growth, accompanied by a reduced formation of aerial mycelium, with P. ostreatus. In Fusarium oxysporum not only inhibiting the growth of antagonistic fungi, but also affecting aerial mycelial growth, modulating mycelial surface hydrophobicity, and influencing the expression of virulence genes could be associated with the sesquiterpenes α-humulene and β-caryophyllene [18]. The growth inhibition was also strong with K. mutabilis, while G. lucidum was less effected, in line with the strain specific (-)-(1R,2S) β-bisabolol and (E)-γ-bisabolene production. In the natural environment, fungi are not completely enclosed within the substrate and sesquiterpenes may not reach the concentrations measured in this study. However, individual sesquiterpenes were tested in appropriate doses to allow for comparison with co-cultivation experiments. The sesquiterpenes (-)-(1R,2S) β-bisabolol and (E)-γ-bisabolene inhibited the growth of all tested fungi, except for P. ostreatus being sensitive only to (-)-(1R,2S) β-bisabolol, which is in line with species specific susceptibility to VOCs [17].
Essential oils from plants containing sesquiterpenes show an antimicrobial activity against a wide range of bacterial species, which can be related to their ability to disrupt the structure of microbial cell membrane [36]. Also volatile-mediated interactions between soil bacteria and fungi have been described [10,38]. In our study, bacteria seemed to be less impaired by S. commune volatiles, and the responsible compound impairing growth of P. fluorescens and E. amylovora could not be identified. Here additional VOCs, like esters and ketones could be relevant [22]. S. marcescens was not affected in growth, but showed lower motility, also when exposed to individual sesquiterpenes. This is in accordance with terpenes manipulating bacterial biofilm formation and motility [25]. The sesquiterpene alcohol α-bisabolol has been reported to inhibit prodigiosin pigment synthesis, biofilm formation and swarming motility in S. marcescens [39]. Swarming motility is controlled by quorum sensing and involves developmental modifications as well as metabolic changes [40], suggesting that sesquiterpenes interfere in the highly regulated cell-to-cell-communication process in bacteria.
Although pigmentation has been reported for VOC-mediated interaction [9,16], this seems not the case for S. commune where direct physical contact was found to be prerequisite for pigment production [12]. Sesquiterpenes are lipophilic, targeting the membrane, and they can facilitate the passage of other toxins through membranes leading to a synergistic effect [35]. This would indicate that also membrane stability control could account for species specificity.
The tested sesquiterpenes strongly inhibited the growth of antagonistic fungi, reinforcing our hypothesis that sesquiterpenes contribute to the observed dominance of S. commune on infected wood. A high bacterial diversity was seen in early wood decay, with Proteobacteria dominating. The often restricted nitrogen availability in decayed wood can be overcome by the help of nitrogen-fixing bacteria, like Rhizobium giardinii, which has been detected in the community [41]. A high abundance of Alpha-and Gammaproteobacteria, Actinobacteria, Chitinophagia and Cytophagia was found inhabiting this habitat. A similar community composition has been reported before in decayed pine wood [41] and these phyla were also affiliated with the white-rot fungus H. fasciculare [3]. We observed a dominance of Pseudomonadaceae and Sphingomonadaceae, which are saprophytic plant colonizing bacteria that are known to improve plant growth by outcompeting pathogens, and they might be also be involved in lignocellulose degradation [4,42]. Acidobacteria are well adapted to acidic conditions and were shown as being abundant in decaying wood [3,41]. However, in our study members of this phylum were underrepresented in these samples.
Interestingly, we found comparatively high numbers of OTUs of Actinobacteria in the mycosphere of S. commune. These Gram-positive bacteria are predominately found in soil and are known for their ability to colonize wood and-at least partially-degrade lignin and cellulose [4,41]. We isolated mostly wood colonizing Actinobacteria, including Janibacter sp., Frigoribacterium sp. and Curtobacterium sp., which has been described as a fungal growth promoting bacterium [43].
The cultivation provided conditions selecting for fast growing bacteria and fungi [44], which explains the differences of community analyses between the two methods. Nevertheless, both methods show different bias and it is essential to compare the results to obtain comprehensive insight into community structures. In addition, in order to test strains co-occurring with S. commune, we needed isolates from the same piece of wood. In conclusion, S. commune suppression of growth of competing species likely led to a selection for adapted microorganisms [3] that might enhance species depending on metabolic cross-feeding or other dependencies on co-occurring species involved in wood decay [45]. The reduce bacterial swarming motility of S. marcescens might also be a selective pressure, since swarming is a means to colonize new wood surfaces.
Thus we could highlight sesquiterpenes for their impact on community structure in the wood mycosphere.