Litter chemistry of common European tree species drives the feeding preference and consumption rate of soil invertebrates, and shapes the diversity and structure of gut and faecal microbiomes

Litter chemistry of common European tree species drives the feeding preference and consumption rate of soil invertebrates, and shapes the diversity and structure of gut and faecal microbiomes

unspecified saprotrophs.Our study suggest that litter quality is a strong driver of feeding preference and consumption rates as well as composition of bacterial and fungal communities in gut and faeces of two species representing the main groups of litter feeding soil fauna in European forests.

Introduction
Soil fauna play a significant role in a litter decomposition and transformation by shredding the litter and thus making the litter substrate more available for the primary decomposers (Frouz, 2018;Coq et al., 2022).For example, between 30 and 50% of net primary production is returned to the soil via feeding activity of various groups of soil fauna (Wardle et al., 2004;Wall et al., 2008;García-Palacios et al., 2013;Frouz, 2018).Soil microfauna (nematodes, tardigrades and rotifers) contribute indirectly to litter decomposition via a grazing effect on bacterial and fungal communities (Petersen and Luxton, 1982;Sohlenius and Boström, 1999;Cesarz et al., 2013;Xiong et al., 2018), while mesofauna and macrofauna contribute directly to the litter decomposition via their consumption activity and indirectly by modifying conditions for bacterial and fungal decomposers (Frouz, 2018;Joly et al., 2020;Prescott and Vesterdal, 2021;Coq et al., 2022).In addition, a study by Frouz et al. (2007) reported that litter fragmentation by soil macrofauna during consumption of leaves provided suitable substrates for a variety of fungal and bacterial decomposers, which resulted in increased decomposition rate of the litter substrate.
Aboveground vegetation shapes the foraging activity of litter feeding macrofauna, particularly via the chemical composition of leaf litter, which in turn affects its palatability (Wardle et al., 2004;Ferreira Quadros et al., 2014;Frouz, 2018).Tree species producing leaf litter with low lignin content, low C/N ratio and high concentrations of base cations were reported to have fast decomposition rates, which are expected to be associated with high palatability for soil meso-and macrofauna (Frouz, 2018;Prescott and Vesterdal, 2021).Terrestrial isopods and millipedes represent important groups of litter feeding macrofauna for litter decomposition in temperate forests (David and Gillon, 2002;Gerlach et al., 2014;Joly et al., 2020;Coq et al., 2022).Despite soil animals are considered to be generalists (Tajovsky, 1992;David et al., 2001;Loginova and Busargina, 2005;Gerlach et al., 2014;Ardestani et al., 2019), many soil animals show a feeding preference allowing them to better adapt to available food resources (Heděnec et al., 2013;Steinwandter and Seeber, 2020).However, the specific factors shaping the feeding preference of soil animals still require further studies, e.g., the extent to which substrate quality controls the feeding ecology of soil animals (Frouz, 2018;Ardestani et al., 2019).
The processing and digestion of recalcitrant litter require special extracellular hydrolytic and oxidative enzymes such as beta-glucosidase and peroxidases for decomposition (Kadamannaya and Sridhar, 2009;Frouz, 2018).Soil macrofauna species are unable to produce extracellular enzymes but instead host a variety of microbial symbionts in their digestive system to decompose recalcitrant compounds such as cellulose and lignin using their own extracellular enzymes (Berg et al., 2004;Kostanjsek et al., 2004;Knapp et al., 2010;Coq et al., 2022).Faeces produced by litter feeding macrofauna play an important role in stabilizing of soil organic matter in soil (Coq et al., 2022;Joly et al., 2018).For example faeces under controlled laboratory conditions showed higher C loss (40.0%) than that of litter (26.6%) (Joly et al., 2018).Similarly, the N dynamics switched from net immobilisation (7.7%) in litter to a net release (14.6%) in faeces (Joly et al., 2018).A recent study by Joly et al. (2020) showed a C loss among faeces types ranging from 21.2% of initial C for faeces from Armadillidium vulgare faeces derived fed by horse chestnut to 42.7% for faeces from Porcellio scaber fed by beech litter.These complex processes rely on highly diverse gut and faecal microbiome (Knapp et al., 2010;Delhoumi et al., 2020;Arora et al., 2022), but the link between diet quality and diversity and structure of gut and faecal microbiome of soil animals remains poorly understood.
Recent progress in amplicon sequencing methods has uncovered a vast diversity of bacterial and fungal species in the digestive tract of litter feeding fauna (Kostanjsek et al., 2004;Delhoumi et al., 2020;Arora et al., 2022;Zheng et al., 2022a), but the effect of consumption of different quality leaf litters on the diversity, structure and relative abundance of various taxonomic and functional groups of bacteria and fungi in gut and faeces of soil macrofauna remains unknown.To bridge this knowledge gap, we designed a mesocosm experiment with the most common species of litter feeding macrofauna in western Europe: the terrestrial isopod Oniscus asellus and the millipede Glomeris marginata.We used leaf litter from six common European tree species producing different quality leaf litter as food resources: the broadleaves beech (Fagus sylvatica L.), pedunculate oak (Quercus robur L.), lime (Tilia cordata L.), sycamore maple (Acer pseudoplatanus L.), ash (Fraxinus excelsior L.) and the conifer Norway spruce (Picea abies (L.) Karst.).We hypothesize that O. asellus and G. marginata would exhibit higher consumption rate and higher percentage of consumed litter of ash, maple and lime than litter of beech, oak and particularly Norway spruce (H1).We expect tree species differences to be linked to litter chemical composition, i.e., that low concentrations of lignin and cellulose and high concentrations of nutrients will increase the consumption rate and percentage of consumed leaf litter for both O. asellus and G. marginata (H2).We hypothesize that bacterial and fungal communities in gut will differ from microbial communities in faeces as well as between two faunal species (H3).Finally, we hypothesize that the relative abundance of bacterial and fungal genera in gut and faeces of O. asellus and G. marginata will vary among tree species with different quality foliar litter (H4).

Sampling of leaf litter and soil animals
Approximately 300 g of freshly fallen leaf litter from ash, maple, lime, beech oak and fallen needles from Norway spruce were collected from the top layer of forest floor in a common garden experiment in Viemose Skov (Denmark) in October 2018 (Zheng et al., 2022b).Leaf litter from each tree species was oven dried at 55 • C for 48 h in hot air-flow oven.Oven dried leaf litter was stored in paper bags for chemical analyses and mesocosms experiments.Approximately 100 individuals of O. asellus and G. marginata were collected from forest floor of a beech stand adjacent to the common garden experiment in Viemose Skov in July and August 2019.The both animal species were kept separately in translucent plastic boxes (20 cm long, 10 width and 5 cm high) with 2 cm of water saturated sand layer in bottom covered by 2 cm layer of oven-dried beech leaves to create proper microclimatic conditions.Animals in plastic boxes were sprayed by 10 ml of tap water once per week to keep stable moisture conditions.The plastic boxes with animals were kept in a dark room at 20 • C. Animals were fed once or twice per week according to their feeding activity by an oven dried mixture of leaf litter collected at the Viemose site.Plastic boxes with animals were maintained once per month to remove faeces and to add a new litter layer to ensure suitable microhabitat.

Litter chemistry analyses
Litter pH in 0.01 M CaCl 2 was measured at the ground litter:solution ratio of 1:5 and analysed with a Radiometer combination-electrode GK2401 (Radiometer, Copenhagen, Denmark).The total organic carbon (TOC) and total nitrogen (TN) concentrations were measured on P. Heděnec et al. ground samples with an elemental analyser (Thermo Fisher Scientific, Waltham, MA, USA) after air-drying to constant mass.Total concentrations of elements in litter were determined after microwave-assisted digestion in concentrated HNO 3 , and the digests were subsequently analysed for total element contents by ICP-OES (PerkinElmer Optima 3000XL).Concentrations of lignin, hemicellulose and cellulose were measured according to National Renewable Energy Laboratory (NREL) procedures (Sluiter et al., 2010).Water and ethanol extraction were performed on a Soxhlet apparatus.Concentrations of cellulose and hemicellulose were analysed using an Ultimate HPLC (Thermo Fisher Scientific Inc., Waltham, MA USA).Lignin was determined as the dry weight of the samples (after acid hydrolysis) taking the ash content into account (Sluiter et al., 2010).Differences in litter chemistry among various tree species taken from the Viemose Skov common garden site is described in Table 1.

Consumption and feeding preference tests
Ten individuals of O. asellus or G. marginata were added to translucent plastic mesocosms (20 cm long, 20 width and 5 cm high) with 2 cm layer of water saturated sand on the bottom.In total, 0.3 g of oven dried ash, maple, lime, beech and oak leaves needles were randomly distributed on the inner peripheral part of each mesocosm.Because of small size of needles, 1 g of oven dried Norway spruce needles was randomly distributed on the inner peripheral part of each mesocosm together with broadleaves litter.In total, 6 replicates with 10 individuals of O. asellus and G. marginata with leaves from ash, maple, lime beech, oak and needles from Norway spruce were incubated separately in a dark room at 20 • C for 10 days.Plastic boxes were sprayed by 10 ml of tap water on the first, fourth and eighth day of the incubation to keep stable moisture conditions.Unconsumed leaf litter was collected after ten days, oven dried and weighed.The percentage of consumed litter was calculated as the difference between litter mass before and after incubation divided by the initial mass.Consumption rate was calculated as the difference between mass before and after incubation divided by the number of incubation days and by the total number of animals.

Diet effect test
Six individuals of O. asellus or G. marginata were added to translucent plastic mesocosms (5 cm in diameter and 10 cm in high) with 2 cm of water saturated sand layer on bottom and approximately 5 g of oven dried ash, maple, lime, beech, oak and 10 g needles of Norway spruce.Each mesocosm with individual leaf litter consisted of 3 replicates so in total 18 mesocosms were used for each faunal species.Animals were incubated in a dark room at 20 • C for two weeks.Plastic mesocosms were sprayed by 10 ml of tap water on the first, fourth, eight and twelfth day to keep stable moisture conditions.The survival of the two faunal species was good during incubations.Mesocosms with oak, maple and Norway spruce had an average mortality of 1-2 individuals (both species) per mesocosms while both species propagated in ash, maple and lime mesocosm and increased by 1-2 individuals per mesocosm.Animals that died during incubation were excluded from further measurements.At the end of the experiment, animals were killed by adding 70% ethanol and immediately processed for gut dissection.The entire digestive tract including gut content from survived individuals per mesocosm was dissected using dissection needles and tweezers under aseptic conditions and pooled to make one composite sample per mesocosm.Approximately one gram of faeces was collected at the end of the incubation from each mesocosm to form one composite sample per mesocosms.Gut and faeces samples were stored at − 20 • C prior to DNA extraction.Unconsumed leaf litter was collected after two weeks, oven dried and weighed.The percentage of consumed litter was calculated as the difference between litter mass before and after incubation divided by the initial mass.

Processing of sequence data and bioinformatics analyses
The SEED pipeline version 2.1.2was used for filtering and trimming of sequence reads obtained from Illumina MiSeq sequencer (Větrovský et al., 2018).The reads were merged into paired end sequences with at least 30 bp overlap (Větrovský et al., 2018).All sequences with

Table 1
Chemical composition of leaf litter sampled in the Viemose Skov common garden site.
Litter TOC (mg g − 1 ) TN (mg Ca (mg g − 1 ) Fe (μg g ambiguous bases and average base quality scores lower than 30 were removed from the dataset.Sequences without primers and identifiers as well as sequences with mismatched identifiers were also removed. Remaining sequences were sorted into samples according to the MID sequences.Chimeric sequences were detected using algorithm UCHIME included in USEARCH 7.0.1090(Edgar et al., 2011) and deleted.Chimera free sequences were clustered using UPARSE implemented within USEARCH 7.0.1090(Edgar, 2013) at a 97% similarity level.From each cluster, the most abundant sequence was selected as a representative sequence for subsequent analysis.All singletons and chimeric sequences were removed.Bacterial sequences were clustered using BLAST against the local SILVA database (Yilmaz et al., 2014).The extraction of the ITS2 region was processed by ITSx (Bengtsson-Palme et al., 2013).The non-fungal ITS2 sequences were removed from the dataset and fungal ITS2 sequences were clustered using BLAST search against the local UNITE database (Nilsson et al., 2019).In total, approximately 800000 bacterial (16S rRNA gene) and 2900000 fungal (ITS2) singleton-free sequence reads were clustered to 4900 bacterial and 3100 fungal OTUs, respectively.The bacterial and fungal sequence reads were resampled to 2000 bacterial and 7000 fungal sequences per sample, respectively.After removal of rare OTUs with relative abundance lower than 0.1%, 260 bacterial and 170 fungal OTUs were maintained.Categorization of bacteria into oligotrophs and copiotrophs was based on available literature (Fierer et al., 2007).Categorization of fungi into functional groups was based on the FungalTraits database (Põlme et al., 2020).Two samples affiliated to faeces from G. marginata were omitted due low sequence numbers.The obtained raw OTU tables including assigned taxonomy are available in the supplementary file (Supplementary file 2).

Statistical analyses
Two-way analyses of variance (ANOVA) followed by Tukey HSD test was used to test effects of different qualities of leaf litter on the consumption rate, percentage of consumed litter and diversity of bacterial and fungal microbiome in gut and faeces of O. asellus and G. marginata.Alpha diversity indices (OTU richness, Shannon index and Pielou index) of bacterial and fungal microbiomes were calculated using the vegan package (Oksanen et al., 2012).Pearson's correlation coefficient was used to test correlations between litter chemistry and consumption rate, percentage of consumed litter and diversities of bacterial and fungal microbiomes.The relative importance of litter chemistry for consumption rate and percentage of consumed litter was tested using random forest regression with the "rfPermute" package (Liu et al., 2020).Analysis of dissimilarity was used to test effects of different leaf litters on bacterial and fungal communities in gut and faeces of both animal species based on Hellinger transformed Bray-Curtis distance of OTUs using the "adonis" function (PerMANOVA) in the "vegan" package (Oksanen et al., 2012).To visually interpret community dissimilarity, non-metric multidimensional scaling (NMDS) ordination was conducted.The "envfit" function was used to test correlations (p < 0.01) of litter chemistry with the Hellinger-transformed Bray-Curtis distance of bacterial and fungal communities (Oksanen et al., 2012).The relative abundances of bacterial and fungal genera were analysed, correlated and visualized by principal component analyses (PCA).All statistics and graphics were performed in R-Studio (www.r-project.org)using "vegan" and "ggplot2" packages (Oksanen et al., 2012;Wickham, 2016).

Effect of litter chemistry on feeding preference and consumption rate
The consumption rate of O. asellus in mesocosms with six litter types increased in the order beech, Norway spruce, oak < maple, lime < ash (Fig. 1A), and consumption rates of G. marginata increased in the order Norway spruce < beech, oak < maple, lime, ash (Fig. 1B).The percentage of litter consumed by O. asellus increased in the order: Norway spruce < beech < oak < lime, maple < ash (Fig. 1C), and a very similar order was observed for G. marginata: Norway spruce < beech < oak, maple < lime, ash (Fig. 1D).The percentage of consumed litter in mesocosms with six litter types together indicated that O. asellus preferred ash over oak, beech and Norway spruce (Fig. 1C), while G. marginata preferred ash and lime over beech and Norway spruce (Fig. 1D).When the two faunal species were fed by a single litter type, the preferences for tree species were slightly different from those observed in experiments with all six litters present.The percentage of litter consumed by O. asellus increased in the order Norway spruce < maple < beech, oak, ash < lime (Fig. S1A), and the percentage of a single litter type consumed by Glomeris marginata increased in the order Norway spruce < maple < oak < lime, ash, beech (Fig. S1B).
The consumption rate and percentage of litter consumed by O. asellus increased with concentrations of Mg, S and K, but decreased with concentrations of Fe, P, TOC, lignin, cellulose and hemicellulose (Fig. 2).Correlations with litter quality variables were similar for G. marginata except for no correlations with P, lignin and hemicellulose concentrations but a positive effect of litter pH (Fig. 2).The concentrations of Mg, S and K showed negative correlation with TOC, lignin and cellulose, while P showed a positive correlation with TOC, lignin and L:H:C ratios (Table S1).Finally, random forest analyses revealed concentrations of Mg and S in litter as the most important factors for high consumption rate and percentage of litter consumed by both O. asellus and G. marginata (Fig. 3A-D).In summary, consumption rate and percentage of consumed litter did differ significantly between the two animal species.

Litter chemistry effect on diversity of bacterial and fungal microbiomes in gut and faeces
The millipede G. marginata had higher bacterial OTU richness and Shannon index than the isopod O. asellus, while fungal OTU richness, Shannon index and Pielou index were similar for O. asellus and G. marginata (Table 2).Both faunal species had higher bacterial OTU richness, Shannon index and Pielou index in faeces than in the gut (Table 2).The type of foliar litter did not affect OTU richness, Shannon index and Pielou index of bacteria in gut and faeces of O. asellus and G marginata (Table 2).For both faunal species, feeding on different litter types significantly influenced OTU richness, Shannon index and Pielou index of the fungal microbiome in gut and faeces (Table 2).The fungal microbiome had lowest OTU richness, Shannon index and Pielou index in gut and faeces of O. asellus fed by ash litter (Table 2).In contrast, the fungal microbiome in gut and faeces of O. asellus fed by Norway spruce, maple and lime litters had higher OTU richness, Shannon index and Pielou indices.The gut and faeces of G. marginata fed with ash and spruce litter had lower fungal OTU richness than for G. marginata fed with lime, maple and beech.The bacterial community across both faunal species was characterized by higher overall OTU richness (59.4 ± 18.9) than the fungal community (25.7 ± 5.1) (Table S2).Finally, the bacterial and fungal OTU richness in gut and faeces of both faunal species showed positive as well as negative correlations with pH, and concentrations of TOC, TN, Mg, P, S, K, Ca, Fe, lignin, cellulose and hemicellulose (Table 3).

The effect on community structure of bacterial and fungal microbiomes in gut and faeces
Bacterial and fungal community structures in gut and faeces of G. marginata were more clustered than for O. asellus (Fig. 4A-B).Moreover, bacterial and fungal communities of both animals were more clustered in faeces than in gut (Fig. 4A-B).The ordering of OTUs based on Hellinger-transformed Bray-Curtis distance showed divergent community structures for bacteria and fungi between O. asellus and G marginata (Table S3).The bacterial community structure differed significantly between gut and faeces of O. asellus but not among the six litter types (Table 4).In addition, there was no significant effect of litter chemistry on the bacterial community structure in gut and faeces of O. asellus (Fig. 5A).In contrast, the fungal community structure in gut and faeces of both faunal species was significantly affected by litter type, i.e. being more separated between broadleaves and Norway spruce (Table 4; Fig. 5C-D).The bacterial communities in the gut of G. marginata differed among animals fed by needle litter from Norway spruce and animals fed by litter from the broadleaved trees.The bacterial communities in spruce-fed animals was associated with TOC, TN and P (Fig. 5B), while the bacterial community in the gut of G. marginata fed by broadleaves was associated with Mg.Glomeris marginata fed by various litter types differed in the fungal communities in both gut and faeces (Fig. 5D).Fungal communities in gut and faeces of G. marginata fed by ash, maple and lime differed from fungal communities in gut and faeces of animals fed by beech, oak and Norway spruce.In addition, fungal communities in gut and faeces of O. asellus fed by beech and oak were separated from fungal communities in animals fed by ash, maple and lime (Fig. 5C).Fungal communities in both animals fed by Norway spruce, beech and oak were associated with high Fe, P, TOC, TN, lignin, cellulose and hemicellulose concentrations, while communities in gut and faeces fed by ash, maple and lime were associated with high concentrations of Mg, S, Ca and K in litter.

Diet effect on relative abundance of bacterial and fungal genera in gut and faeces
The relative abundances of bacterial and fungal genera in gut and faeces of O. asellus and G. marginata differed according to the food source (Fig. 6A-D).Ash, maple and lime differed from beech, oak and spruce.Bacterial and fungal genera in gut and faeces of O. asellus and G. marginata fed by ash, maple and lime were related to high concentrations of K, Ca, and Mg, while fungal and bacterial genera in gut and faeces of O. asellus and G. marginata fed by beech, oak and spruce were associated with TOC, TN, cellulose and lignin concentration.For bacteria, the relative abundance of Streptomycetes was associated with cellulose and hemicellulose, while Pseudomonas was more related to Mg and S as in litter of ash and maple (Fig. 6A-B).For fungi, the relative abundance of Trichoderma was related to cellulose concentration as in litter of Norway spruce, while the relative abundance of Talaromyces was associated with high P and K concentration as in oak litter (Fig. 6C-D).The gut and faeces of O. asellus had high relative abundances of the classes Alphaproteobacteria, Gammaproteobacteria and the phylum Bacteroidota (Fig. 7A).The gut of G. marginata showed high relative abundance of the phylum Bacteroidota while faeces showed high relative abundance of Alphaproteobacteria and Gammaproteobacteria (Fig. 7B).Moreover, O. asellus and G. marginata hosted different bacterial genera in their gut and faeces.For example, the gut of O. asellus had high relative abundance of the genus Ricketsiella while faeces showed high relative abundance of Methylotenera and Pseudomonas (Fig. 7C).The gut of G. marginata had high relative abundance of unclassified Proteobacteria, Bacteroidota and Verrucomicrobiota, while faeces showed high relative abundance of Merthylothenera, Pseudomonas and Rhizobium (Fig. 7C).We found that gut and faeces of both O. asellus and G. marginata were dominated by copiotrophic bacteria (Figs.S2A-B).The relative abundances of copiotrophs and oligotrophs did not differ in gut and faeces of O. asellus and G. marginata fed by the six litter types.
The gut and faeces of O. asellus were dominated by Ascomycota, but feeding by lime and spruce litter resulted in increased relative abundance of Basidiomycota and Mucoromycota in gut and faeces respectively (Fig. 8A).Similarly, feeding of G. marginata by spruce needles resulted in increased relative abundance of Mucoromycota in gut and faeces (Fig. 8B).In addition, gut and faeces of O. asellus and G marginata harboured different fungal genera.For example, feeding of O. asellus by ash, maple and oak litter increased the relative abundance of Talaromyces in gut and faeces, while feeding by spruce needles resulted in increased relative abundance of Trichothecium in gut and Trichoderma in faeces (Fig. 8C).In addition, the gut of O. asellus fed by beech showed a high relative abundance of Lactarius.Gut and faeces of G. marginata fed by litter from ash, maple and lime had increased relative abundance of Aspergillus, while gut and faeces of G. marginata fed by litter from beech oak and Norway spruce was high in relative abundance of Talaromyces (Fig. 8C).Feeding with spruce litter resulted in high relative abundance of Trichoderma in gut and faeces of G. marginata.The most dominant functional groups of fungi in gut and faeces of both animal species were various groups of saprotrophs, of which the most dominant groups were unspecified saprotrophs, litter saprotrophs and soil saprotrophs (Figs.S3A-B).Furthermore, we also found mycoparasites, animal parasites and plant pathogens.We also found sequences affiliated to ectomycorrhizal fungi in the gut of O. asellus fed by beech litter.Mycoparasites and soils saprotrophs were most abundant in gut and faeces of O. asellus and G. marginata fed by Norway spruce.

Effect of litter chemistry of food preference
Our laboratory incubation experiment with litter from various tree species revealed strong feeding preferences for both faunal species.When offered six litter types Oniscus asellus preferred ash over oak, beech and Norway spruce while G. marginata preferred ash and lime over beech and Norway spruce.This finding aligns with field-based results from the same common garden reported by Peng et al. (2022), showing a higher abundance of millipedes and isopods in soils beneath ash, maple and lime.When the two faunal species were fed by a single litter type, the preferences for tree species were slightly different from those observed in experiments with all six litters present.Less maple was consumed in spite of favourable chemical composition, and we suggest that this could be due to large leaf area and higher concentration of  Values represents mean ± standard error of the mean.Different letters indicate statistically different mean values (p < 0.05).Two-way ANOVA followed by Tukey HSD Test; *p < 0.05, **p < 0.01, ***p < 0.00.NA -Not available.

Table 3
Correlation of litter chemistry and diversity of bacterial and fungal community in gut and faeces of Oniscus asellus and Glomeris marginata.Bonferroni corrections was used to adjust p value of multiple correlations.
P. Heděnec et al. phenolic compounds in comparison to other leaf litter (Barbehenn et al., 2005).We also hypothesize that ontogenetic differences between individuals can highly affect consumption rate.For example, the juveniles from the same species can consume more litter per unit of body mass than adult individuals (Frouz, 2018;Ardestani et al., 2019).We provide strong evidence that differences in chemical composition of leaf litter play a significant role for litter consumption differences among common European tree species.We suggest that palatability of foliar litter based on high concentration of S and Mg and low lignin and cellulose concentrations a key driver of feeding preference.A parallel study from alpine grassland by Steinwandter and Seeber (2020) uncovered that the millipedes Cylindroiulus meinerti and Cylindroiulus fulviceps preferred more palatable high quality (low C/N ratio, low lignin content) litter from Geranium sylvaticum over more recalcitrant (high C/N ratio, high lignin content) grass litter from Dactylis glomerata and shrub litter from Vaccinium vitis-idaea.We did not find any significant differences in consumption rate or percentage of consumed litter between O. asellus and G. marginata.Both soil fauna species in our study are approximately similar in size, and we assume that they share same or similar niche (Ardestani et al., 2019).
The results of the food preference experiment revealed a significant effect of litter chemistry on consumption rate as well as percentage of consumed litter.We suggest that the consumption rate of foliar litter is stimulated by low concentration of lignocellulosic compounds and high concentrations of nutrients.High concentrations of cellulose and lignin can negatively affect food preference due to lower palatability (Rief et al., 2012;Ferreira Quadros et al., 2014).High palatability may also be attributed enzymatic reactions during digestion that would be enhanced by high nutrient concentrations in litter (Berg et al., 2004;Šustr et al., 2020).We also suggest that microbiota associated with leaf litter can affect palatability and food preference and feeding activity of soil animals (Crowther and A'Bear, 2012;Heděnec et al., 2013).For example, recent studies by by Heděnec al. (2020) and Zheng et al. (2022)b revealed highest microbial biomass C, bacterial biomass and fungal biomass in forest floor of ash in comparison to maple, lime, oak and Norway spruce.Finally, we take into an account that keeping soil animals under controlled laboratory conditions may result in lower feeding performances or unusual behaviour compared to field conditions (Steinwandter and Seeber, 2020).

Diet effect on diversity and composition of gut and faecal microbiome
Our results revealed higher diversity of bacteria in gut and faeces of G. marginata than O. asellus, suggesting that the bacterial community is linked to faunal species (Zheng et al., 2022a).Higher diversity of bacterial microbiome in gut of G. marginata suggested that more complex gut compartments provide specific microhabitats for bacteria ( Šustr et al., 2020).In contrast fungal diversity in gut and faeces did not differ between animal species, suggesting fungal diversity is more closely linked with other factors such as litter chemistry (Korkama-Rajala et al., 2008;Algora Gallardo et al., 2021).Furthermore, our results showed that faeces harboured higher bacterial and fungal diversity than gut for both animals fed by various leaf litter.We suggest that soil fauna is also able to selectively digest microorganisms in the gut and thus reduce the diversity of gut microbiome (Frouz et al., 2003).It has been suggested that litter conversion into faeces enhanced organic matter lability during passage through the digestive tract (Joly et al., 2020;Coq et al., 2022) and thereby faeces provide more nutrients for large numbers of bacterial and fungal species associated with faecal pellets.In addition, the inner part of faeces provides a habitat for anaerobic microbes, while the surface of faeces harbours aerobic microbes (Kostanjsek et al., 2004).Faeces hosts bacteria and fungi from the gut but also receives bacteria and fungi from the surrounding environment such as air or soil.
The community structure of bacterial and fungal communities strongly deviated between O. asellus and G. marginata, suggesting high host-specificity among various litter feeding fauna (Zhang et al., 2021;   Zheng et al., 2022a).Litter from various tree species did not affect the structure of bacterial communities, but indeed affected the diversity of fungal communities in gut and faeces of the two animals.These results suggest that the diversity of bacterial communities in gut and faeces is ubiquitous (Barberán et al., 2014), while fungal communities are associated with litter quality to a greater extent (Tedersoo et al., 2014;Mrnka et al., 2020;Algora Gallardo et al., 2021).The random forest analyses suggest that litter chemistry affect fungal diversity via lignin cellulose content and concentration of nutrients, which in turn affect decomposability of leaf litter.We suggest that low quality leaf litter with  high lignin-cellulose content and low concentration of nutrients will have slow decomposition during passage of the digestive tract and will produce a large number of biopolymers as by-products of litter decomposition processes (Cepáková and Frouz, 2015), which provide substrate for a diverse fungal community (Korkama-Rajala et al., 2008).The structure of bacterial communities in gut and faeces of G. marginata diverged between litter from broadleaves and Norway spruce.The fungal community indicated even stronger divergence between various litter types.We suggest that variation in nutrient, lignin, and cellulose concentrations among different litter types support specialization among litter-inhabiting microbes and also promote variation in substrate preference (Algora Gallardo et al., 2021).
A large number of fungal and bacterial genera were observed in gut and faeces of O. asellus and G. marginata fed by various of foliar litter.We suggest that bacterial or fungal taxa can be passed on from the phyllosphere (microbes on plant surface), plant endosphere (microbes in plant  tissue) or from the soil surface (Jumpponen and Jones, 2010;Martínez-Romero et al., 2021).Studies by de Souza et al. (2016) andTrivedi et al. (2020) showed that Proteobacteria and Bacteroidota are dominant groups of bacteria in the phyllosphere and endosphere, and Koivusaari et al. (2019) reported high relative abundances of Ascomycota in the leaf endosphere.This corresponds well with our results showing high relative abundances of Ascomycota as well as Proteobacteria and Bacteroidota in gut and faeces of both animal species.The gut microbiome can probably also be inoculated from faeces via coprophagy, which is common in isopods as well as millipedes (David and Handa, 2010).Gut and faeces of O. asellus and G. marginata had high relative abundances of Proteobacteria, suggesting these bacteria are likely dominant in the gut system of many soil faunal groups (Delhoumi et al., 2020;Zhang et al., 2021;Zheng et al., 2022a).In addition, we suggest that relative abundances of bacterial and fungal genera are shaped by litter chemistry.We hypothesize that Proteobacteria are involved in the lignocellulose degradation (Liu et al., 2019) and then in mediating the functional role of terrestrial isopods and millipedes as litter decomposers and regulators of nutrient cycling in soil ecosystems (Delhoumi et al., 2020;Šustr et al., 2020).
We found that the gut of G. marginata had high relative abundance of Bacteroidota, suggesting Bacteroidota are linked with the digestive tract and are responsible for degradation of complex polysaccharides such as cellulose (Thomas et al., 2011).Fungal taxa found in our study are similar to those found in other studies related to gut microbiota related invertebrate (Chakraborty et al., 2020;Větrovský et al., 2020) and consists of a wide spectrum of common saprotrophic fungi, suggesting that a number of saprotrophic fungal strains may have a symbiotic relationship with soil fauna (Chakraborty et al., 2020).Finally, we suggest that gut and faeces of both O. asellus and G. marginata provide microhabitats for highly diverse bacterial and fungal groups, which are associated with decomposition of gut content but also able to interact with hosts as pathogens as well as among each other as competitors.
Feeding of O. asellus by ash, maple and oak litter showed increased relative abundance of Talaromyces in gut and faeces while feeding by Norway spruce litter showed increased relative abundance of Trichothecium in gut and Trichoderma in faeces.Talaromyces, Trichothecium and Trichoderma are fungi widely associated with plants (Ownley et al., 2010;Kakvan et al., 2013), therefore we suggest these fungi can be also survive in the gut of invertebrates (Nicoletti and Becchimanzi, 2022).In addition, the gut of O. asellus fed by beech litter showed high relative abundance of Lactarius suggesting spores or fungal mycelia can be swallowed and persist in gut (Vašutová et al., 2019;Chakraborty et al., 2020).We suggest that some mycoparasitic fungi such as Trichoderma or antagonistic for other fungi such as Talaromyces can also form symbiosis with invertebrates to enhance resistance for insect pathogens (Nicoletti and Becchimanzi, 2022).Finally, we hypothesize that gut and faeces of both O. asellus and G. marginata provide microhabitat for highly diverse bacterial and fungal taxonomical and functional groups, which are associated with decomposition of gut content but can also interact with hosts as a pathogens as well as between each other as competitors.

Conclusions
This study combined a mesocosm experiment with amplicon sequencing to provide an integrated view of the effect of litter quality on consumption of soil animals as well as on gut and faecal microbiomes in two key macrofauna species.Foliar litter from ash, maple and lime was consumed to greater extent than foliar litter from oak, beech and Norway spruce.This was explained by distinct differences in chemical composition of the foliar litter.Faeces of both animals hosted a higher diversity of bacterial and fungal communities than in their guts, suggesting passage by the digestive tract enhanced organic matter lability and thus increased diversity of bacterial and fungal species in faeces.High diversity of bacteria in gut and faeces of G. marginata suggested that more complex gut compartments provide specific microhabitats for bacteria.The fungal diversity in gut and faeces did not differ between animal species, suggesting fungal diversity is more closely linked with quality of the foliar litter as a substrate.Bacterial and fungal genera in gut and faeces may in part be passed on from the phyllosphere or from the soil surface.Feeding by foliar litter from different tree species revealed changes in relative abundances of bacterial and fungal phyla and genera via variation in nutrients, lignin, and cellulose concentrations in the six litter types.Finally, we suggest further research should investigate the possible transfer of microbiome from phyllosphere and endosphere to gut and faeces of soil invertebrates.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Litter consumption (A-B) and percentage of litter consumed (C-D) by the isopod O. asselus (A-C) and the millipede G. marginata (B-D).Bars represent mean values ± standard error of the mean.Different letters represent statistical homogenous groups (p < 0.05).

Fig. 2 .
Fig. 2. Heat map of Pearson correlations among chemical properties of leaf litter from six tree species in the Viemose common garden experiment and litter consumption by the isopod O. assellus and the millipede G. marginata (*p < 0.05, **p < 0.01, ***p < 0.001).COA -Consumption rate by O. asellus; POA -Percentage of litter consumed by O. asellus; CGM -Consumption rate by G. marginata; PGM -Percentage of litter consumed by G. marginata.

Fig. 3 .
Fig. 3.The relative importance of litter chemistry on consumption rate and percentage of consumed litter by Oniscus asellus (A, B) and Glomeris marginata (C, D).Estimate permutation p-values for random forest importance metrics: 0.05 (*), 0.01 (**) and 0.001 (***).The + andsigns refer positive and negative effect of litter chemistry on consumption rate and percentage of consumed liter.

Fig. 5 .
Fig. 5.The non-metric multidimensional scaling of bacterial (A-B) and fungal (C-D) communities (Hellinger-transformed Bray-Curtis distance) in gut and faeces of Oniscus asellus (A-C) and Glomeris marginata (B-D) fed by litter from different tree species.

Fig. 6 .
Fig. 6.The PCA ordination plots of the relative abundances of bacterial (A-B) and fungal (C-D) genera in gut and faeces of Oniscus asellus (A-C) and Glomeris marginata (B-D) fed by different quality leaf litter.

Fig. 7 .
Fig. 7. Relative abundance of bacterial phyla (A-B) and genera (C-D) in gut and faeces of Oniscus asselus (A,C) and Glomeris marginata (B,D) fed by various leaf litter.

Fig. 8 .
Fig. 8. Relative abundance of fungal phyla (A-B) and genera (C-D) in gut and faeces of Oniscus asselus (A,C) and Glomeris marginata (B,D) fed by various leaf litter.

Table 2
Alpha diversity of bacterial and fungal community in gut and faces of the isopod Oniscus asellus and the millipede Glomeris marginata fed by foliar litter from different tree species.

Table 4
Dissimilarity analyses of bacterial and fungal communities (Hellinger transformed Bray-Curtis distance) in gut and faeces of Oniscus asellus and Glomeris marginata fed by different leaf litter.