Wood-inhabiting fungal communities in woody debris of Norway spruce (Picea abies (L.) Karst.), as reflected by sporocarps, mycelial isolations and T-RFLP identification

Authors


  • Editor: Julian Marchesi

Johan Allmér, Department of Forest Mycology and Pathology, Swedish University of Agricultural Sciences, PO Box 7026, S-750 07 Uppsala, Sweden. Tel.: +46 18 672794; fax: +46 18 673599; e-mail: Johan.Allmer@mykopat.slu.se

Abstract

Wood-inhabiting fungi play a key role in forest ecosystems and constitute an essential part of forest biodiversity. We therefore examined the composition and abundance of wood-inhabiting fungi by three methods: sporocarp counts, mycelial culturing and direct amplification of internal transcribed spacer terminal restriction fragment length polymorphism from wood combined with sequencing of reference rDNA. Seven-year-old slash piles left after a thinning were analyzed in a 50-year-old Norway spruce plantation. Fifty-eight fungal species were detected from the piled branches and treetops. More species were revealed by sporocarp counts and cultured mycelia than by direct amplification from wood. In principle, sporocarp monitoring may reveal all fruiting taxa, but it poorly reflects their relative abundance in the wood. In contrast, terminal restriction fragment length polymorphism will record the most frequent fungal taxa in the wood, but it may overlook uncommon taxa. Culturing mycelia from wood gives a bias towards species favoured by the cultural medium. The results demonstrate the advantage and the limitations of these methods to be considered in analyses of fungal communities in wood.

Introduction

Dead wood forms the substrate for about 30% of the 25 000–30 000 Fennoscandian multicellular forest species and is thus a key component for biodiversity in these forests (Siitonen, 2001; Dahlberg & Stokland, 2004). The term dead wood includes a wide and heterogeneous variety of types, such as dead standing trees, logs, stumps and roots from different woody bushes and trees of varying diameter and degree of decay. Species that at some stage of their life cycle are dependent on dead wood are referred to as saproxylic (Speight, 1989). The most species rich groups of saproxylic organisms in Fennoscandia are insects and fungi, represented by at least 3000 and 2500 taxa, respectively (Dahlberg & Stokland, 2004).

The ecological knowledge of the saproxylic fungi almost exclusively relies upon observations of sporocarps e.g. (Renvall, 1995; Lindblad, 1997; Dahlberg & Stokland, 2004; Norden et al., 2004). However, such investigations do not reveal the entire species richness present as mycelia and do not even reveal the relative activity of the fruiting species, because fruiting is affected by environmental factors and restricted in time. The relative proportion of resources allocated to sporocarps and for mycelial growth evidently varies among species, and not all fungi may produce visible sporocarps. For example, many fungal species were present as mycelia inside Norway spruce (Picea abies (L.) Karst) logs without producing sporocarps (Käärik & Rennerfelt, 1957; Johannesson & Stenlid, 1999; Gustafsson, 2002). Wood-inhabiting fungi may also be detected by culturing mycelia from wood on selective media, and by biochemical, chemical and immunological analyses (Käärik & Rennerfelt, 1957; Stalpers, 1978). However, these methods are time-consuming and some saproxylic fungi might not even be culturable e.g. (Käärik & Rennerfelt, 1957; Rayner & Boddy, 1988). Recently, PCR-RFLP methods have been employed not only to identify fungi from mycelial cultures, but also to discover and identify fungal presence directly in wood (Johannesson & Stenlid, 1999; Jasalavich et al., 2000; Adair et al., 2002; Högberg & Land, 2004). One obstacle has been to develop efficient methods for extracting amplifiable fungal DNA from wood, e.g. Johannesson & Stenlid (1999). A promising and efficient method to analyze samples with multiple species is terminal RFLP (T-RFLP) (Liu et al., 1997; Lord et al., 2002). T-RFLP targeting the internal transcribed spacer (ITS) region of rRNA genes has been used to characterize arbuscular mycorrhizal communities effectively (Vandenkoornhuyse et al., 2003) as well as the presence of ectomycorrhizal hyphae in soil (Dickie et al., 2002; Landeweert et al., 2003; Edel-Hermann et al., 2004), but such studies on saproxylic fungal communities are not yet reported. Forthcoming work employing T-RFLP and other molecular methods will undoubtedly increase our knowledge of relationships among fungal species richness, relative biomass and activity between fruiting and mycelial growth in wood.

Studies of dead wood and of saproxylic organisms have so far mainly been confined to coarse woody debris (CWD), dead wood defined to have a diameter of >10 cm. Such studies suggest that (1) species richness of saproxylic fungi is strongly correlated with the amount of CWD available, and that (2) the presence of threatened and rare wood-inhabiting fungi critically relies upon CWD (Bader et al., 1995; Kruys & Jonsson, 1999; Kruys et al., 1999; Sippola et al., 2001; Pentilläet al., 2004). However, fine woody debris (FWD), with a diameter of <10 cm, may also be important for the existence of certain species and the total species richness of wood-inhabiting fungi (Kruys & Jonsson, 1999; Norden et al., 2004). As yet, little is known of fungal species exploiting FWD and to what degree the same species utilize both CWD and FWD (Renvall, 1995; Kruys & Jonsson, 1999; Norden et al., 2004). However, sporocarp studies from Scots pine and Norway spruce in Finland suggest that only a part of the total wood-inhabiting fungal community is associated with FWD (Renvall, 1995; Sippola & Renvall, 1999).

Forest management, including clear cutting, prevention of forest fires, cutting of firewood and measures against insect attacks, has resulted in drastically decreased amounts of dead wood in Fennoscandia since the mid-19th century (Siitonen, 2001). The amount of dead wood in managed spruce forests is 1–10% of that in natural forests (Fridman & Walheim, 2000; Siitonen, 2001). In managed forests, almost all CWD is removed during logging. Thus, slash (FWD) and stumps left from forest operations may be increasingly important for saproxylic organisms in managed forest landscapes.

The primary objectives of this study were to compare different approaches for characterizing fungal saproxylic communities by (1) sporocarp monitoring, (2) identification of fungal mycelia cultured from wood by sequence analysis and (3) direct amplification of fungal DNA from wood, for which T-RFLP was employed. In addition, we characterized and compared fungal communities in branches, tops and stumps of Norway spruce left after thinning.

Material and methods

Study site

The study was conducted in one of the experimental areas (S3-Hamerheden) described by Thor & Stenlid (2005), located in east central Sweden (60°15′N, 16°45′E, 55 m asl), about 150 km northwest of Stockholm. The studied site comprised a 50-year-old plantation of Norway spruce (Picea abies), thinned 7 years prior to the study. Slash (branches and tops) from the thinning was left in piles, each relating to individual trees.

Sampling

In October 2001, slash piles from twenty trees were randomly selected within an area of 0.5 ha located in the centre of the stand. At each pile, five branches were randomly selected in addition to the top. The samples were kept at 4°C and processed within a week. For each branch the length, basal and top diameter were recorded. The volume for each branch was calculated according to Smalian's formula (Husch et al., 1982). For the comparison of fungal communities on FWD and CWD, we included the data of fungal occurrences from 65 stumps collected in the fall of 2000 at the same stand to represent the CWD (Table 1). The stump data was originally collected for a study of eventual effects on the mycodiversity in stumps as a result of the biocontrol agent Rootstop (Vasiliauskas et al., 2005). The mycelial cultures from the stump data included in Table 1 were isolated and identified with the same methods used by Vasiliauskas & Stenlid (1998).

Table 1.   The number of records as reflected by sporocarps, sequenced cultures on branches and tops and ITS T-RFLP patterns from branches in slash heaps (n=20) of Norway spruce 7 years after thinning
SpeciesGenbank Accession no.BranchesTopsStumps
SporocarpsMycT–RFLPSporocarpsMycSporocarpsMyc
  1. The average number of branches colonized within each heap is reported within parentheses (n=5). Myc, sequenced cultures. Stump data are taken from Vasiliauskas et al. (2005) and included for comparison. The stump data include occurrences detected by mycelial cultures and sporocarps (n=65).

  2. ITS, internal transcribed spacer; T-RFLP, terminal restriction fragment length polymorphism.

Zygomycetes
 Mortierella sp. 1DQ0082221 (1)
 Mortierella ramanniana4
 Unidentified sp. 370AY7812142
Ascomycetes and Deuteromycetes
 Ascocoryne cylichniumDQ0082264 (1.3)1 (1)15
 Ascocoryne sarcoidesAY7812167
 Ascocoryne sp. 1DQ0082253 (1)
 Ascomycetes sp. 1DQ0082411 (1)
 Ascomycetes sp. 2DQ0082341 (1)1
 Ascomycetes sp. 4DQ0082332
 Ascomycetes sp. 61 (1)
 Chaetosphaeria sp. 419DQ0082361 (1)11
 Chalara sp.1DQ0082242 (1)
 Chalara sp. 400AY7812202
 Chalara sp. olrim510DQ0082433 (1)1
 Cladophialophora sp. 407AY7812171
 Cylindrocarpon sp. 1DQ0082391 (1)
 Dactylaria sp. 414AY7812211
 Fusarium lateritiumAY7812221
 Geomyces pannorumAY7812231
 Geomyces sp. 378AY7812361
 Hypoxylon serpensAY7812261
 Lecythophora hoffmanniiDQ0082283 (1.3)11
 Lecytophora sp. 15DQ0082291 (1)6
 Lecytophora sp. 22AY7812291
 Leptodontidium elatiusAY78123012
 Mollisia minutellaDQ0082421 (1)
 Nectria fuckelianaAY7812311
 Nectria viridescensDQ0082303 (1)1 (1)11
 Phialocephala dimorphosporaAY6063096
 Phialocephala sp. 18DQ0082402 (1)1
 Phialocephala sp. 35DQ00822317 (1.8)731
 Phialophora fastigiataDQ0082272 (1)13
 Phialophora malorumAY7812334
 Phoma sp. 552DQ0082352 (1)
 Phomopsis quercellaDQ0082321 (1)
 Sarea resinaeAY7812371
 Trichoderma polysporumDQ00823118 (3)1125
 Unidentified sp. 6AY7812401
 Unidentified sp. 19AY7812411
 Unidentified sp. 365AY7812424
 Unidentified sp. 3661
 Unidentified sp. 3733
 Unidentified sp. 3881
 Unidentified sp. 401AY7812441
Basidiomycetes
 Amylostereum areolatumAY7812451
 Amylostereum laevigatumAY7812462
 Antrodia serialis1
 Athelia bombacina2 (1)
 Athelia decipiens3 (1)1
 Athelia epiphylla3 (1.3)3
 Botryobasidium botryosum2 (1)2
 Botryobasidium subcoronatum3 (1.3)1
 Botryobasidum angustisporum1 (1)
 Ceraceomyces eludens1 (1)1
 Ceratobasidium cornigerum2 (1)
 Ceriporiopsis sp. 4AY7812502
 Collybia butyraceaAY7812511
 Collybia sp. 9AY7812521
 Coniophora sp. 359AY7812533
 Fibulomyces mutabilis1 (1)
 Ganoderma applanatum1
 Gloeocystidiellum ochraceumAY7812541
 Gymnopus sp. 406 1
 Heterobasidion parviporumDQ0082383 (1.3)220
 Hyphoderma argillaceum3 (1)3
 Hyphoderma obtusiforme1 (2)1
 Hyphoderma praetermissum4 (1.3)2
 Hyphodontia aspera1 (1)2
 Hyphodontia breviseta1 (1)3
 Hyphodontia subalutacea1 (1)1 
 Hypholoma capnoidesDQ0082164 (1.5)12012
 Hypochnicium lundelliiDQ00821819 (4)5 (1.2)3 (1.3)1411
 Kuehneromyces mutabilisAY781258 34
 Marasmius androsaceusAY781260 11
 Mycena epipterygiaAY781261 2
 Mycena sp. 4
 Mycena sp. 403DQ00821715 (1.5)18 (1.4)21
 Panellus mitis3 (1.3)5
 Peniophora incarnataAY7812631
 Phanerochaete laevis1
 Phanerochaete rimosaAY7812651
 Phanerochaete sordidaAY7812661
 Phellinus viticola1
 Phlebia livida1
 Phlebia sp. JA1DQ0082449 (1.2)
 Phlebiella vagaDQ0082192 (2.5)2 (1)43
 Phlebiopsis gigantea3
 Pholiota sp.1
 Pholiota spumosaAY7812682
 Piloderma croceum3
 Postia caesia29
 Postia fragilis1
 Postia stiptica3
 Postia tephroleuca121
 Recinicium bicolor2 (1)422
 Sistotrema brinkmaniiDQ0082203 (1)5 (1)230
 Sistotrema sp. 1DQ0082211 (2)
 Skeletocutis bigutulata1
 Skeletucutis amorpha1
 Steccherinum litschaueri11
 Stereum sanguinolentumAY7812721010
 Trichaptum abietinum33
 Trichosporon porosumAY7812741

A total of 100 branches and 20 tops from 20 piles were analyzed completely for presence of sporocarps and subsampled for presence of mycelia in the wood. Mycelial cultures were made from three wood discs for each branch and top – 360 wood discs in total. Samples for direct amplification of fungal DNA were obtained from the branches only.

Wood discs 1.5 cm thick were sliced from the base, middle and top of each branch. Small samples of wood (5 × 10 × 15 mm3) were cut out from the centre of the branch discs using a sterilized knife. The wood samples were divided into two replicate parts: one for culturing of mycelia and one for direct extraction of fungal DNA. Wood samples for direct DNA extraction were kept in 1.5 mL microcentrifuge tubes and frozen at −40°C. Wood discs 10 cm thick were sliced from the base, middle and top of each top. Samples of wood (5 × 5 × 10 mm3) were cut out from each disc at the location where different decay zones were visible, and used for culturing of mycelia. Procedures for fungal isolation from the wood samples and subculturing of fungal strains were similar to those performed in earlier studies by Vasiliauskas & Stenlid (1998) and Vasiliauskas et al. (2005). The fungal strains were grouped according to mycelial morphology, and two to seven randomly selected mycelia from each group were further identified using sequencing of the ITS of the nuclear rDNA. In one of the morphological groups, which included dark septate fungi, all specimens were sequenced.

All sporocarps on the studied branches and tops were collected and identified to the species level. Voucher specimens are preserved in the private collection of Johan Allmér and reference specimens of all species are deposit in the herbarium of the Swedish Museum of Natural History in Stockholm. The nomenclature follows Hansen & Knudsen (1997), except for Heterobasidion parviorum (Niemelä & Korhonen, 1998).

Identification of fungi from mycelial cultures

Mycelia were removed from the Hagem agar (Stenlid, 1985) with a sterilized knife. Approximately 1 g of mycelia was transferred to separate microcentrifuge tubes. DNA was extracted using a CTAB method as described in Vasiliauskas et al. (2005).

The ITS 1 and 2, including the 5.8S region of nuclear rDNA, was amplified using the primers ITS1 and ITS4 (White et al., 1990). PCR amplification was performed using a CR PC-960G Gradient Thermal Cycler (Corbett Research Pty Ltd, Sydney, Australia). Initial denaturation at 95°C for 5 min was followed by 35 cycles with denaturation at 95°C for 30 s, primer annealing at 50°C for 30 s and primer extension at 72°C for 30 s, with a final primer extension at 72°C for 7 min. In the PCR reactions, final concentrations of 2 mM MgCl2, 200 μM dNTP, 0.2 μM of each primer, 0.3 U μL−1 Redtaq (Sigma, St Louis, MO, USA) and reaction buffer as recommended by the manufactures was used. Amplified products were separated with 2% agarose gel electrophoreses stained with ethidium bromide and visualized under UV light. Before sequencing, the PCR products were purified with QIAquick 8 PCR purification kit (Qiagen, Hilden, Germany). Sequences were run with Beckman Coulter CEQ 8000 Genetic Analysis System with DTCS Quick Start Mix (Beckman Coulter, Fullerton, CA, USA).

Sequences were edited in the module SeqMan of Lasergene (DNA-Star, Madison, WI, USA). BLAST searches were preformed using BioEdit's internal blast function (BioEdit, version 5.0.0, North Carolina State University, NC, USA) together with a reference sequence database at the Department of Forest Mycology and Pathology. The sequences from our sequence database are also available from GenBank The database is a compilation of sequences submitted to GenBank from different projects at our department. Differences in base pairs were checked manually for each sequence that did not have a 100% match and were evaluated if it was a true base pair difference or not.

Direct identification of fungi from branch wood samples

The wood samples were freeze-dried and ground with a ball mill. A 100 mg quantity of pulverized wood sample was transferred to a new 1.5 mL microcentrifuge tube and 800 μL 2% CTAB was added. The samples were vortexed for 20 s and put in a heating block at 65°C for 2 h. The samples were centrifuged at 16 000 g for 20 min and the supernatant transferred to a new 1.5 mL microcentrifuge tube and precipitated with isopropanol. The DNA was dissolved in 50 μL autoclaved H2O, then cleaned with GENCLEAN III kit (QBiogene, Carlsbad, CA, USA) to get rid of potential PCR inhibitors. ITS1 and 2, including the 5.8S region of nuclear rDNA, were amplified using the primers ITS1F and ITS4 (Gardes & Bruns, 1993) labelled with WellRED dyes D4-PA and D3-PA, respectively (Proligo, Boulder, CO, USA). PCR amplification was performed using an Applied Biosystems GeneAmp PCR System 2700 Gradient Thermal Cycler (Applied Biosystems, Foster City, CA, USA) as described above, except for the annealing temperature, which was 55°C. The PCR products were digested with TaqI and CfoI separately. The T-RFLP patterns were analyzed with a Beckman Coulter CEQ 8000 Genetic Analysis System using CEQ DNA Size Standard Kit-600.

To allow for identification of unknown T-RFLP patterns, sequenced mycelial cultures were also cut with TaqI and CfoI. Further, samples with unknown T-RFLP patterns that only had one visible band on the agarose gel were sequenced and matched against our reference sequence database at the Department of Forest Mycology and Pathology. The resulting fragment sizes from mycelial cultures and identified direct amplified T-RFLP patterns were used as a reference database. The fragment sizes from the unknown T-RFLP patterns were compared to the reference database using tramp (Dickie et al., 2002). To account for the accuracy of the base calling during analysis in the Beckman Coulter CEQ 8000 Genetic Analysis System and within taxa variation of the ITS region, a threshold level for fragment size in tramp was set to ±2 bases.

From a subset of five branches, the three subsamples were run both as 15 separate samples and as five pooled samples from each branch to see if the ITS T-RFLP method was sensitive enough to detect the same number of species when pooled. Because no differences in the sensitivity to detect fungal species could be observed, subsamples from all remaining branches were pooled.

Results

Fungal community as defined by sporocarps

In total, 113 sporocarps representing 21 species were recorded from 94 branches in 19 piles. (Table 1). Nineteen of the species were basidiomycetes, comprising 95% of the sporocarps. The 20 tops all carried sporocarps, altogether 93 specimens representing 26 basidiomycete species. The most frequently encountered species was the corticoid fungus Hypochnicium lundellii, where the hymenium frequently covered the major part of the underside of the branches. It was recorded on branches in 19 of the piles, with an average occurrence of four out of five examined branches per colonized pile and on 14 of the tops.

Fungal community as defined by cultured mycelia

In total, 263 different mycelia were successfully cultured from 86 branches in 18 piles. These mycelia were distinguished into 65 morphological groups, 22 with more than one mycelial culture and 43 single cultures. DNA sequencing of these mycelial morphological groups revealed 26 species, of which 18 species were ascomycetes, seven species were basidiomycetes and one species was a zygomycete (Table 1). Fifty-three mycelial cultures were obtained from 18 tops. These cultures were grouped into 21 morphological groups, two with several cultures and 19 with single cultures. DNA sequencing revealed 14 species, of which 11 of the species were ascomycetes, and three species were basidiomycetes. The two most frequent species were Trichoderma polysporum and Phialocephala sp. 35 (Table 1). Only one mycelial morphological group was revealed by sequencing to consist of more than one taxon. It consisted of two species, Phialocephala sp. 35 and 18, which did not differ by mycelial morphology. This finding, with further collections of Phialocephala and with ecological notes and molecular characterization, is further reported by Menkis et al. (2004).

Fungal community as defined by T-RFLP-analyses

Of the 100 pooled branch-samples analyzed, fungal DNA was detected in 91 samples (i.e. branches). Of these, 83 samples were successfully analyzed with ITS T-RFLP and in total five species detected (Table 1). The most frequent species was a yet-unidentified Mycena species occurring in 18 piles and on average on 1.4 out of five examined branches in these piles. The second most frequent species was a yet-unidentified Phlebia species occurring in nine piles and on average on 1.2 branches out of five examined branches in these piles. This species was only detected with the ITS T-RFLP method.

Comparison between different methods to detect fungal communities

In total, 58 wood-inhabiting fungal species were recorded from the branches and tops, 31 species using sporocarps, 27 species using sequence analysis of mycelial cultures from wood and five species when fungi were directly identified from wood by T-RFLP (Table 1, Fig. 1a). One species, H. lundellii, was detected by all methods (Table 2). Four out of the five species detected by ITS T-RFLP were also recorded in mycelial cultures. One species was only recorded by ITS T-RFLP.

Figure 1.

 Comparison of the number of fungal species recorded as (a) sporocarps, cultured mycelia and from direct T-RFLP and (b) from branches, tops or stumps. In (a), the numbers show the number of species documented with either method, individual or in combination with two or three methods. In (b), the numbers show the number of species recorded exclusively from one, two or all three wood fractions.

Table 2.   The number of branches where the same species were detected by more than one of the three methods studied: sporocarps, cultured mycelia from wood and directly amplified fungal DNA from wood
SpeciesDetected with all methodsDetected as sporocarps and from cultured myceliaDetected as sporocarps and from ITS-T-RFLPDetected from cultured mycelia and from ITS-T-RFLP
  1. The number within parentheses indicates the total number of records by each of the two methods in the column. (In all three methods, n=100).

  2. ITS, internal transcribed spacer; T-RFLP, terminal restriction fragment length polymorphism.

Ascocoryne cylichnium0001 (5 : 1)
Hypochnicium lundellii36 (76 : 6)1 (76 : 4)0
Mycena sp. 40300019 (23 : 26)
Nectria viridescens0001 (3 : 1)
Phlebiella vaga01 (5 : 2)00
Sistotrema brinkmannii02 (3 : 5)00

Only three out of the 31 species recorded by sporocarps were recovered as mycelia in the wood of branches. H. lundellii was fruiting in 19 of the piles and on 14 tops and recovered as mycelia cultures from five piles and as T-RFLP pattern from three piles (Table 2). Sistotrema brinkmannii was recorded as sporocarps in three piles and as mycelia in five piles, and Phlebiella vaga was recorded as sporocarps in two piles and as mycelia in two piles. H. lundellii was detected as mycelia from 6% of the branch samples and with direct ITS T-RFLP in 4% of the samples, suggesting that it occupies about 5% of the wood volume (Table 3).

Table 3.   A comparison between the mean surface area and volume for individual branches, treetops and individual stumps investigated by the sporocarp survey, mycelial culturing and ITS T-RFLP detection monitoring
MethodMean surface area (cm2)Mean volume (cm3)Number of sample unitsNumber of species detected
  1. The numbers within parentheses in the column showing the number of sample units refer to the number of branches where fungi were detected with the respective method. Standard deviations for mean volume and mean surface area were ±70% and ±40% for branches and treetops, respectively. Stump volume refers to volume above the ground.

  2. ITS, internal transcribed spacer; T-RFLP, terminal restriction fragment length polymorphism.

Branches
 Sporocarps462214100 (94)31
 Mycelial cultures0,25100 (86)28
 ITS T-RFLP0,15100 (91)5
Tops
 Sporocarps4652739620 (20)26
 Mycelial cultures0,2520 (18)14
Stumps
 Sporocarps137159796514
 Mycelial cultures0,256553

None of the most frequently cultured species was recorded as sporocarps on the branches or tops. However, four of the fungal species identified from the cultured mycelia were also detected directly from the wood. The proportions between number of detected basidiomycete and ascomycete taxa differed between the methods: 30 : 1 in the sporocarp survey, 1 : 3 with the mycelial cultures and 2 : 3 with direct ITS T-RFLP detection.

The species richness detected from mycelial cultures was similar to what was recorded from sporocarp monitoring (Tables 1 and 3). The overlap of species detected by the different methods was low (Fig. 1a).

Comparison of fungi colonizing branches, tops and stumps

In total, 58 fungal taxa were identified in this study. Of these, 17 species were only recorded on branches and nine species only on tops (Fig. 1b). When comparing the occurrences of species on branches and tops from this study with the stump data included in Table 1, it is evident that the overlap of species between FWD and CWD were low at this site. Eleven species were detected from all three wood fractions.

Discussion

This study demonstrates a clear detection bias using different measures to survey wood-inhabiting fungi. At a first glance, the correspondence of the saproxylic community among the methods was low. Only three out of 31 fruiting species of the branches and tops were recovered as mycelia in the wood and 25 species were found only as mycelia inside the wood (Table 1, Fig. 1a). Similarly, only five fungal taxa of 64 detected in the stumps were recorded both as sporocarps and mycelia (Table 1). However, there are several reasons for this nonconformity, as discussed below.

Part of the poor correspondence between detection measures is due to the different sampling size. Clearly, many more wood samples would have been needed to obtain a more reliable picture of the fungal communities using direct amplification of ITS T-RFLP. In order to get a broader picture of the actual fungal communities in the wood, more wood discs within each wood unit would have been required. Furthermore, and maybe more importantly, wood samples from the exterior of the wood discs should also have been included. Sporocarp inventories easily survey the whole surface of investigated samples, whereas detection from cultured mycelia or by direct DNA amplification can for practical reasons only be made from a small fraction of the wood volume – in this study, less than 0.1% (Table 3). The number of species recorded from a particular sample will depend on the size of the sample, how common the species are and their distribution in the substrate from which the sample is taken (Taylor, 2002). Certainly, saproxylic fungi differ in fitness for different wood qualities – for example, the interior or exterior part of the wood and the degree of decomposition – and thus vary in frequency among those. Accordingly, records of sporocarps will embrace all potential wood qualities of a studied substrate, whereas detection from the interior of wood critically depends upon the number of samples and the sampling strategy. Mycelial domains physically extend more easily longitudinally along the fibres than transversally in wood (Rayner & Boddy, 1988). Some species are also extensively distributed in wood units, whereas other species typically have localized and small mycelial domains (Rayner & Boddy, 1988). In future studies using wood samples, we thus recommend subsamples from larger wood volumes to be analyzed. When working with mycelial cultures, it is most convenient to work with small wood samples that are cut out from sections of wood or wood discs and put on agar plates for culturing (Vasiliauskas & Stenlid, 1998). The number of wood samples should be enough to cover for eventual differences between the interior and exterior parts of the wood section and sampled in a standardized manner to allow for statistical treatment. Similarly, the wood sections should be taken at intervals that will reflect differences along the substrate (Gustafsson, 2002). When using T-RFLP, we believe that it would be better to take drill samples by drilling through the wood surface into the centre of the wood from different directions along a circle corresponding to a wood section. This would mean fewer samples to be analyzed per wood section because it is possible to analyze samples with multiple species with T-RFLP (Lord et al., 2002). However, this design is only suitable if the objective is to describe what species are occurring in the wood because it is not possible to tell from where in the sample a certain species is originating.

Mycelial cultures revealed six times more species than were found by ITS T-RFLP from wood samples (Table 1, Fig. 1b). The most frequently recorded species from the mycelial cultures, Trichoderma polysporum was not detected by ITS T-RFLP. Probably it was present as conidia or minute mycelial units in the wood and successfully growing when put on the agar plates. Possibly the DNA from its conidia or mycelia in the wood was too insignificant to be detected by the ITS T-RFLP, as earlier reported for mycorrhizal fungi (Dickie et al., 2002). Some of the ascomycete species found in this study (e.g. T. polysporum) might not be wood degraders at all; rather, they utilize low-molecular-weight compounds such as monosaccharides and amino acids or organic compounds exudated by wood degraders, as discussed by Lindahl & Olsson (2004). It is notable that Phlebia sp. 1 was recorded in 11% of the branches by ITS T-RFLP, but not detected at all from mycelial cultures. This observation might be an example of a species incapable of growing on the nutrient medium used. Similarly, a comparison of DNA and culture-based detection of fungi from ericoid mycorrhizal roots revealed Sebacina DNAs to predominate, despite their never being detected from cultures from the same roots (Allen et al., 2003). Clearly, DNA and culture-based detection reveal different aspects of fungi in the wood than sporocarp monitoring.

Other factors explaining the poor congruence between methods is that sporocarp surveys may miss mycelia in wood if not conducted at the appropriate fruiting time (Rayner & Boddy, 1988; Gustafsson, 2002). Certain saproxylic ascomycetes also predominantly fruit during spring and are thus missed in fall surveys (Norden et al., 2004). In this study, we recorded a number of basidiomycetes only as mycelia in the wood that commonly produce sporocarps (e.g. Mycena sp. 1) (Table 1). Fungal fruiting is environmentally controlled, and thus sporocarp inventories need to be repeated during a season and over years to gain a more comprehensive picture of fruiting fungal species. An additional factor to consider is the fungal succession of wood, where later-arriving species successively replace established species (Rayner & Boddy, 1988). Different species may also allocate varying amounts of resources for reproduction and dispersal in relation to mycelial growth, as has been discussed in ectomycorrhizal fungi (Gardes & Bruns, 1996). Moreover, some ascomycete taxa are overlooked because they rely entirely or predominantly upon asexual reproduction by microscopic conidia.

Saproxylic fungi can be classified as unit-restricted (dependent on propagules as a mode of arrival and establishment on a new substrate) or as nonunit-restricted (being able to colonize substrates through hyphal cords or rhizomorphs) (Rayner & Boddy, 1988). The mode of establishment is critical for the amount of resource captured, where colonization via hyphal cords is more effective in occupying resources than propagules (Holmer & Stenlid, 1996). In our study, only three species found as sporocarps were also detected as mycelia in the wood, H. lundellii, P. vaga and S. brinkmannii (Table 2). At least two of these species, P. vaga and S. brinkmannii, can be classified as nonunit-restricted group because they form conspicuous rhizomorphs and are commonly found fruiting on debris in the organic layer in coniferous forests (J. Allmér, personal observations). H. lundellii rarely forms rhizomorphs in connection to the sporocarps, making its nonunit-restricted growth habit less obvious. However, it commonly fruits on branches, twigs, cones and mosses in the organic layer, which indicates that it belongs to this group (Allmér, personal observations). In this study, the sporocarps of H. lundellii were covering larger parts of most of the branches in the piles, suggesting it to be widely distributed as mycelia in the wood. However, it was only detected in about 5% of the wood samples. This could be a result of the fact that H. lundellii mainly grows in the exterior of the wood, whereas the analyzed wood samples were derived from the centre of the branches. Most of the fruiting species in this study only occurred on one to three of the 20 piles of branches studied (Table 1), indicating that the probability of finding fruiting species in the wood samples was very low. Presumably, many of these species are unit-restricted, and thus have a more scattered distribution among the substrates.

Fungal species richness, as detected from sporocarps, are reported to be higher on CWD than on FWD when equal numbers of wood units were compared in two Swedish wood-inhabiting fungal community studies (Kruys & Jonsson, 1999; Norden et al., 2004). However, these studies also reported an opposite pattern with higher species richness on FWD when equal wood surface was compared. It is possible that mycelia from certain fungal species utilize both FWD and CWD, but only fruit on CWD as the amount of energy from FWD is not sufficient. One such example is Heterobasidium parviporum, recorded as mycelia in the branches, but not fruiting (Table 1). It commonly fruits on affected trees and is reported to fruit almost exclusively on spruce stumps or wind-thrown trees in managed forests (Korhonen & Stenlid, 1998; Pentilläet al., 2004). The polypore species Antrodia serialis, Phellinus viticola, Postia spp., Skeletocutis biguttulata and Trichaptum abietinum, with conspicuous sporocarps recorded from tops and stumps, were not recorded from the branches (Table 1). Probably, saproxylic fungi with larger sporocarps like these polypores may typically require larger volumes of wood to be able to produce sporocarps. However, all the recorded fruiting species in this study have been reported as sporocarps from both FWD and CWD, though with different frequencies (Dahlberg & Stokland, 2004) (J. Allmér, personal observation). Furthermore, certain wood-inhabiting fungi (i.e. tomentelloid and corticoid fungi, e.g. Piloderma croceum) are not saprotrophic but are ectomycorrhizal, and only use the wood as support for their fruiting structures (Erland & Taylor, 1999; Tedersoo et al., 2003) (Table 1).

Many saproxylic fungi preferentially occur on CWD and are red-listed in Fennoscandia, owing to declining and rare populations as a consequence of decreasing amount and quality of CWD, (Bendiksen et al., 1998; Gärdenfors, 2000; Rassi et al., 2001). No saproxylic fungus confined to FWD is considered threatened and red-listed, but this may be a result of the fact that fungi-utilizing FWD is as yet inadequately studied. The increasing use of logging residues and slash for bioenergy purposes in Fennoscandia (Anonymous, 2002) will reduce the amount of dead wood in these forests. Yet the importance of FWD for saproxylic organisms and to what degree an increased removal of forest residues will affect their populations is poorly known. During normal forestry practice, stumps are left after cutting, representing a substantial amount of dead wood. Thus, stumps might presumably function as refuges for species expelled from logging residues. However, our study indicates that if there would be a decline in species when removing logging residues it could have negative effects on the species richness and also on community structure at a stand level because there was little overlap between branches, tops and stumps (Fig. 1b). However, the eventual effects will largely depend on how significant this fraction of dead wood is for saproxylic fungi and to what degree it host a particular set of fungal species. One apparent step to advance the knowledge of the importance of FWD and CWD would be to compile and develop the information of wood quality requirements of individual saproxylic fungi to be analyzed with calculations of the amount of different wood qualities available through a forest generation and at a landscape perspective.

Acknowledgement

This work was financially supported by the Swedish Energy Agency.

Ancillary