How do attached crown parts and branches contribute to the diversity of saproxylic fungi and beetles in downed and decaying spruce trees?

A significant proportion of forest-dwelling species in boreal forests are saproxylic, i.e., dependent on deadwood. To safeguard deadwood-associated diversity in managed forest landscapes, it is important to understand how substrate preferences and specialization structure saproxylic species communities across different deadwood resource types. In this study, we investigated the diversity and associations of saproxylic fungi and beetles at the scale of entire trees to understand how different tree parts contribute to species diversity. To do this, we sampled species assemblages in trunks (d > 15 cm), tops (d 5–10 cm) and branches (d < 5 cm) of 31 fallen Norway spruce trees. Fungal assemblages were investigated with DNA metabarcoding from wood samples, and beetles were surveyed by bark peeling and sieving. Our results showed that, fungal and beetle assemblages were clearly differentiated between trunks and branches. In the tops, fungal community composition was intermediate between trunks and branches, whereas beetle species composition was more closely aligned with trunks. Trunks and branches both harbored specialized fungal and beetle species, but no species were identified as specialists of tops. Fungal and beetle richness were lowest in branches, and fungal richness peaked in tops. Substrate specialization of saproxylic species at the scale of individual trees imply that deadwood restoration in managed forests should prioritize whole-tree retention instead of partial retention such as artificial high stumps or pruned logs.

Methods
Study sites were located in Norway spruce (Picea abies) dominated forests in the cities of Espoo, Helsinki, Vantaa, Lahti and Tampere in southern Finland. At each site, we sampled wood-inhabiting fungi and saproxylic beetles in one or two fallen spruce trees. We selected tree trunks in early stages of decomposition. In all trees, fruiting-bodies of wood-decay fungi were visibly developing on the surface, but most of the stem surface was still covered by bark. Fungal and beetle sampling was conducted between August and November 2022.

We collected wood samples for the analysis of wood-inhabiting fungi by drilling from the basal section of the downed trunk, from the top of the tree, and from the branches. The basal part was sampled at five equidistant points from 50 cm to 850 cm distance from the base on both vertical sides of the trunk. The top was sampled at five equidistant points on one side of the stem along a section where stem tapered from 10 cm to 5 cm in thickness. Branch samples were collected from five branches randomly picked at roughly equidistant points along the trunk from the lowest branching point to 10 cm stem thickness. From each branch, we sampled a segment 20-35 cm outward from the trunk.

At each drilling point, we exposed a fresh wood surface with a knife and extracted wood with a 6 mm drill head down to a maximum depth of 5 cm. Drill dust was collected directly into a sterilized polyethylene bags yielding one sample from each tree compartment per one tree. Samples were stored in -20 °C, and later ground into fine dust in a bead beater in 50 ml capsules with 8 mm steel bead for 2 min in 30 Hz. We DNA sterilized tools with bleach solution between samples when extracting and processing the wood samples.

DNA was extracted from ca. 90 mg of wood powder and eluted into 50 µl final elution volume with DNeasy Plant Pro extraction kit according to the manufacturer’s instructions. For the analysis of wood-inhabiting fugal communities by metabarcoding, the internal transcribed spacer 2 of the nrRNA coding region was amplified with primers ITS3-2024F (GCATCGATGAAGAACGCAGC) and ITS4-2409R (TCCTCCGCTTATTGATATGC). Indexed amplicons were sequenced with Illumina NovaSeq 6000 (paired-end 250 bp).

Sequence reads were demultiplexed, and index and primer sequences were removed from paired-end reads. Subsequent sequence processing was done with vsearch v2.18.0. R1 and R2 reads were quality filtered and assembled. Assembled reads were chimera filtered de novo. Remaining reads were clustered into operational taxonomic units (OTUs) with 98% similarity threshold. OTUs were taxonomically assigned with 80% confidence cutoff using Naïve Bayesian Classifier trained with UNITE (v9) database in mothur v1.36.1. For further analyses, we converted OTU read abundances into presence-absence data by applying 1% relative abundance threshold.

We sampled saproxylic beetles at the basal end of the trunk, at the middle part of the trunk, at the top of the tree, and from the branches. At each sampling point along the trunk and top, we sampled a fixed 1 m2 area of bark. For branch sampling, we picked five branches per tree randomly along the length of the tree and sampled from the total bark area available on those branches.

Beetles were collected by peeling and sifting bark. We identified adult beetles to species morphologically and larvae and pupae by sequencing the cytochrome c oxidase subunit I barcode gene region with the primers LCO1490 (GGTCAACAAATCATAAAGATATTGG) and HCO2198 (TAAACTTCAGGGTGACCAAAAAATCA). We used BOLD as the reference database for barcode sequence identification. In addition, species were recorded based on visual observations of characteristic exit holes and larval galleries in the wood.