Meiofauna promotes litter decomposition in stream ecosystems depending on leaf species

Abstract Litter decomposition, a fundamental process of nutrient cycling and energy flow in freshwater ecosystems, is driven by a diverse array of decomposers. As an important component of the heterotrophic food web, meiofauna can provide a trophic link between leaf‐associated microbes (i.e., bacteria and fungi)/plant detritus and macroinvertebrates, though their contribution to litter decomposition is not well understood. To investigate the role of different decomposer communities in litter decomposition, especially meiofauna, we compared the litter decomposition of three leaf species with different lignin to nitrogen ratios in litter bags with different mesh sizes (0.05, 0.25, and 2 mm) in a forested stream, in China for 78 days. The meiofauna significantly enhanced the decomposition of leaves of high‐and medium‐ quality, while decreasing (negative effect) or increasing (positive effect) the fungal biomass and diversity. Macrofauna and meiofauna together contributed to the decomposition of low‐quality leaf species. The presence of meiofauna and macrofauna triggered different aspects of the microbial community, with their effects on litter decomposition varying as a function of leaf quality. This study reveals that the meiofauna increased the trophic complexity and modulated their interactions with microbes, highlighting the important yet underestimated role of meiofauna in detritus‐based ecosystems.

Meiofauna may significantly contribute to litter decomposition in the following ways. First, they can act as microdetritivores by directly consuming leaves. For example, harpacticoid copepods effectively remove organic matter, fungi, and bacteria that accumulate in the debris (Perlmutter & Meyer, 1991). Meanwhile, the meiofauna can also indirectly affect litter decomposition by changing the microbial decomposition of leaves. Some meiofauna, such as rotifers and nematodes, feed on the associated biofilms. The grazing may alter the microbial abundance and/or the community structure, increasing or decreasing the microbial decomposition activity of microbes (Chambord et al., 2017).
The common method used to assess the litter decomposition process is based on litter bags of two mesh sizes (i.e., coarse and fine mesh to separate the decomposition of macrofauna and microbes; Graça, Barlocher, & Gessner, 2005), which may blur the contributions of microbes and meiofauna. To date, their interaction with microbial communities and the contribution of meiofauna to the litter decomposition have not yet been elucidated. Moreover, the chemical characteristics of leaf litter (e.g., initial N and lignin contents and the lignin:N ratio) are known to affect microbial and invertebrate colonization and activity, consequently affecting litter decomposition (Alonso, González-Muñoz, & Castro-Díez, 2010;Sales, Gonçalves, Dahora, & Medeiros, 2014). Invertebrates can significantly consume the high-quality leaf litter, whereas the decomposition of low-quality leaf litter may be under the synthetic action of microbes and invertebrates (Raposeiro, Ferreira, Gea, & Gonçalves, 2018;Santonja et al., 2018). Thus, leaf chemical qualities may modify the contributions of meiofauna and other decomposers to the decomposition process, though the exact mechanisms remain unclear.
This study aimed to assess the effects of different decomposer communities, especially meiofauna, on the decomposition of leaves with varied chemical qualities. Three different mesh sizes were used in the decomposition bags to establish a gradient of increasing trophic and community complexity: (a) microbes; (b) microbes and meiofauna; (c) microbes, meiofauna, and macrofauna. We hypothesized that the presence of meiofauna and macrofauna should trigger different effects on the community structure of microbes, and these effects on litter decomposition should vary as a function of leaf quality. Specifically, for the high-quality leaf species, meiofauna should promote the litter decomposition mainly through direct consumption as macrofauna. With the decreasing quality of leaves, the direct consumption of meiofauna and their interaction with microbes together should promote the litter decomposition.

| Study site and selected species
This study was conducted in a forested headwater stream, lo- Harms., Fagus longipetiolata Seem., and Machilus leptophylla Hand.-Mazz (Qian et al., 2016). Leaves from three riparian forest species were selected for the field experiment that considered leaf chemical characteristics and their dominance at the study site (i.e., Alangium chinense, Liquidambar acalycina, and Machilus leptophylla). The leaf characteristics were determined from litter material of the original batches, including carbon and nitrogen (TOC/TN auto-analyzer, Shimadzu TOC-L CPH, TNM-1, Japan), phosphorus (molybdenum blue photometric method after basic digestion with sulfuric acid and 30% hydrogen peroxide, Jones, 2001), lignin (acid-detergent fiber method, Graça et al., 2005), toughness (tensile force required to tear the blade in unit width using a tensiometer (ZP-50, Aigu, HongKong, China;Pérez-Harguindeguy et al., 2000).

| Water characteristics
The physico-chemical characteristics of the stream water were measured at each sampling date with litter bags retrieved during the experimental period. The temperature, pH, conductivity, and dissolved oxygen of the stream water were measured in situ using a

| Experimental design
The litter decomposition experiment was carried out between October 2017 and January 2018. Leaves were collected immediately after abscission in September 2017 and dried at 60°C for 48 hr.
Three grams (3 ± 0.1 g) of each leaf species was enclosed in the bags (15 × 10 cm). The litter bags were sprayed with deionized water after being weighed to prevent leaf breakage during handling and transport to the field. Three different mesh sizes (0.05, 0.25, and 2 mm) were chosen to restrict organisms of different body size: fine mesh (0.05 mm) mainly allowed access for microbes. Although small nematodes and rotifers can pass through 0.05 mm mesh, the fine mesh can restrain access to most meiobenthos (like chironomids, copepods, and oligochaetes). The medium mesh (0.25 mm) allowed access for microbes and meiofauna. The entire decomposer community, including microbes, meiofauna, and macrofauna, can enter the coarse mesh (2 mm). The increasing mesh size was expected to induce an increase in the faunal community complexity and trophic complexity (Bradford, Tordoff, Eggers, Jones, & Newington, 2002). With this approach, the microbial contribution to decomposition was assumed to correspond to the mass loss in fine-mesh bags, and the difference in mass loss between different sized meshes was interpreted to be equivalent to invertebrate-mediated decomposition.
A total of 108 litter bags (3 species × 3 mesh sizes × 4 replicates × 3 times) were tied to nylon lines that were anchored randomly to the stream bed using house bricks in shallow riffles. Water depth at base flow was 5-30 cm at the place where the litter bags were exposed. Four replicates for the treatments in species and mesh size were retrieved on days 21, 47, and 78. The three collection dates were selected, according to the early, middle, and late stage for the species that decomposed fastest. Each litter bag was sealed individually in one plastic bag when they were still immersed, before being transferred in a cool box to the laboratory. The leaf from each bag was gently rinsed with distilled water to remove other material and then put into vacuum freeze-dryer. From each bag, a subsample (weighing 0.6 g) was taken for microbial measurement and stored at F I G U R E 1 The study site (red spot) located in Jinfo Mountain, south-western China −80°C for further analyze. The weight of the subsample was taken into consideration when the litter decomposition rate was calculated. The remaining litter material was dried (60°C for 48 hr) and weighed to determine the final dry mass.

| Fungal biomass
The fungal biomass on the leaves was estimated by ergosterol concentration of freeze-dried, pulverized leaf subsamples (Graça et al., 2005). Lipid extraction was carried out in KOH methanol at 80°C for 30 min. Extracted lipids were purified using solid-phase extraction (SPE) cartridges (Supelclean™ LC-18 SPE Tubes 500 mg), and ergosterol was eluted with isopropanol. The ergosterol concentration was quantified with high-pressure liquid chromatography (HPLC, Agilent 1100), with 100% methanol as the mobile phase (1.4 ml/min, column temperature 33°C, measuring absorbance at 282 nm). Finally, the results were expressed as mg of fungal biomass per gram of leaf litter dry mass (DM). The data were converted to fungal biomass by the average conversion factor of 5.5 mg of ergosterol per gram of fungi (Gessnert & Chauvet, 1993).

| Microbial community structure
Fungal and bacterial communities were estimated using high-throughput sequencing methods. Amplicons were pooled in equimolar amounts and paired-end sequenced on the Illumina MiSeq platform (Illumina) at Majorbio Bio-Pharm Technology Co. Ltd. Raw fastq files were demultiplexed, quality-filtered by Trimmomatic and merged by FLASH according to the criteria, detailed in Qin et al. (2018). Operational taxonomic units (OTUs) were clustered based on a 97% identity cut-off using UPARSE (version 7.1 http://drive5.com/upars e/), and chimeric sequences were identified and removed using UCHIME. The taxonomy of each 16S/ITS gene sequence was analyzed by the RDP Classifier algorithm (http://rdp.cme. msu.edu/) against the Unite database for fungal sequences and the Silva database for bacterial sequences. Unclassified OTUs were annotated using basic local alignment search tool (BLAST) searches of National Center for Biotechnology Information (NCBI), USA GenBank's nonredundant nucleotide database. The naming of OTUs was based on the best BLAST hits.

| Benthic meiofauna and macrofauna
Benthic meiofauna and macrofauna were sampled at the time when litter bags were put in and then retrieved. At each time, three samples covering most microhabitats present in riffles in the study reach were collected for macrofauna and meiofauna using two successive nets with mesh sizes of 250 and 50 µm. Samples were preserved in 75% ethanol and transferred to the laboratory. Macrofauna and meiofauna were sorted and identified to the lowest possible taxonomic level when possible (mainly to genus or species level), except for Oligochaeta (Class level), Diptera (Family or subfamily level), Copepods (Subclass level), Nematodes (Phylum level), and Ostracoda (Class level). Functional feeding groups were assigned according to Majdi Traunspurger Richardson and Lecerf (2015), Morse, Yang, and Tian (1994), and Tachet, Richoux, Bournaud, and Usseglio-Polatera (2002).

| Statistical analyses
Litter decomposition rates (k) were determined using the negative exponential model: where M t is the remaining mass at time t (in days), and M o is the initial mass (Graça et al., 2005). The leaf mass remaining (%), denoted as R, was calculated according to the formula: The effect or relative contribution of microbes, meiofauna, and macrofauna, denoted as E, to leaf mass loss denoted as L, was quantified based on the following formulas (Chen & Wang, 2018;Seastedt, Todd, & James, 1987): where L (microbes) is the mass loss from fine-mesh bags; L (meiofauna) is the difference in mass loss between the fine-and medium-mesh bags; L (macrofauna) is the difference in mass loss between the medium-and coarse-mesh bags; and L (total) is the mass loss in the coarsemesh bags. Microbial Shannon diversity indexes (H′) were calculated based on the number of OTUs using the online Majorbio cloud platform (www.major bio.com).
A three-way ANOVA was conducted to determine the effect of leaf species, mesh size, time, and their interactions on leaf mass remaining and fungal biomass. A two-way ANOVA was performed to determine the effect of leaf species and mesh size on litter decomposition rates (k). A one-way ANOVA was applied to compare differences for litter decay (litter decomposition rates and mass remaining) and microbial parameters (fungal biomass, and fungal and bacterial abundance) among mesh size at each time for each species.
LSD post hoc tests were performed when the ANOVAs detected a significant difference. For all parametric analyses, normality and homogeneity were respected. R software was used for all statistical analyses.

| Initial litter characteristics
The three leaf species differed markedly in terms of their initial characteristics (Table 2)

| Litter decomposition rates
Litter decomposition rates (k) varied 10-fold (from 0.0022 to 0.033/ day) (Figure 2), with significant differences between litter types (M. lep- sizes (F 2,27 = 9.0, p < .001). The interactions between mesh size and species for the decomposition rates were significant (species × mesh, F 4,27 = 3.6, p = .02; Table 3). The mean decomposition rates for all three litter species showed generally increasing trends with mesh size (Figure 2a), implying that the increasing community complexity enhanced the litter decomposition. The decomposition rates for all three leaves in the coarse-mesh bags were significantly higher than those in the fine-mesh bags (LSD test, p < .001). In particular, for L. acalycina, the decomposition rates in the fine-mesh bags were significantly lower than those in the medium-mesh bags (meiofauna and microorganisms present), suggesting a positive meiofaunal effect (LSD test, p = .02).
Moreover, its decomposition rates in medium-and coarse-mesh bags were not significantly different (LSD test, p > .05) (Figure 2a).

| Contribution of biotic communities to litter decomposition
The

| Mass remaining
Mesh size effects for leaf remaining mass of leaves significantly differed over time (mesh × time, F 4,81 = 10.5, p < .001). Specifically, the mesh effects on the mass loss of A. chinense were only significant on day 78, with the highest values in coarse-mesh bags (F 2,9 = 9.5, p = .006). For L. acalycina, on day 21, its mass loss significantly increased (coarse-mesh > fine-mesh > medium-mesh, LSD test, p < .05), while on days 47 and 78, the mass losses in medium-mesh were significantly greater than those in fine-mesh bags (LSD test, p = .05, p = .03) and no significant difference was detected between the mass loss in coarse-and medium-mesh bags (LSD test, p > .05). For M. leptophylla, the mass loss in coarse-mesh bags was significantly greater than that in fine-mesh bags on day 78 (LSD test, p = .001), while the mass losses in fine-and medium-mesh bags were not significantly different (LSD test, p > .05; Figure 3).  biomass was the significantly higher on day 47 than on other dates (F 2,68 = 6.5, p < .01). Specifically, for a given species and date, the associated fungal biomass for A. chinense in medium-and coarsemesh bags was significantly lower than for fine-mesh bags on day 47 (LSD test, p = .03, p = .002). For L. acalycina, the associate fungal biomass significantly increased with mesh size on day 47, by onefold for medium-mesh bags and twofold for coarse-mesh bags compared to fine-mesh (fine-mesh < medium-mesh < coarse-mesh, LSD test, p < .01). This increasing trend was still seen day 78 (fine-mesh < medium-mesh < coarse-mesh, LSD test, p < .05). For M. leptophylla, the associated fungal biomass peaked in the coarse-mesh on days 21 and 47 (F 2,29 = 4.5, p = .04; F 2,29 = 6.0, p = .02), though it did not significantly differ among mesh types on day 78 (F 2,29 = 0.7, p > .05;

Figure 4).
A total of 55,846 OTUs were found based on ITS gene sequencing. The fungal community on leaf material was dominated F I G U R E 2 (a) Litter decomposition rates (k) (day −1 ) (mean ± SE, n = 4) for different mesh-size litter bags for the three leaf species. For a given species and date, significant differences among the mesh sizes are indicated by different letters (one-way ANOVA and LSD test, p < .05); (b) Contribution of different biotic communities to leaf mass loss in litter bags after the 78 days of immersion

TA B L E 3
Summary of ANOVA testing of the effects of leaf species and mesh size on litter decomposition rates (k); effects of leaf species, mesh size, and time on mass remaining and fungal biomass F I G U R E 3 Mass remaining (%) (mean ± SE, n = 4) of different mesh-size litter bags for the three leaf species over the experimental period. For a given species and date, significant differences among the mesh sizes are indicated by the stars (*p < .05, **p < .01, ***p < .001) by taxa belonging mainly to the Ascomycota phyla, which made up 97.5% of all recorded OTUs, followed by Rozellomycota  (Figures 5 and 7

| Bacteria
A total of 41,794 OTUs were found based on sequence analyses of the leaf-associated bacterial 16S rRNA gene pool. The Proteobacteria were found to be the major bacteria phylum present in the litter bags (64.05) with γ-Proteobacteria being the most abundant class for all three leaves and mesh sizes ( Figure 6). In addition, the bacterial

| Meiofauna and macrofauna
A total of 16 meiofaunal taxa were found with a mean density of 9,529 ± 2,658 ind./m 2 during the whole study (Table S1)

| Effect of initial litter quality
The different decomposition rates across leaf species may be mainly due to their initial chemical characteristics (Ágoston-Szabó et al., 2016;Bruder, Schindler, Moretti, & Gessner, 2014). In general, the contents of C, N, P elements and lignin in leaves are important factors affecting decomposition rates (Santschi, Gounand, Harvey, & Altermatt, 2018).
The relative order of decomposition rate among the three leaf types and Ocotea sp.) (Li, Ng, & Dudgeon, 2009;Ligeiro, Moretti, Gonc, & Callisto, 2010), probably because of the relatively higher temperature in tropical regions (Li et al., 2009). In particular, the decomposition of

| Relative contributions of biotic communities
In this work, microbes were major contributors to litter decomposition.
The presence of meiofauna and macrofauna significantly increased the leaf decomposition rates, when compared to a single microbial treatment. This suggests the crucial role played by trophic complexity with regard to litter decomposition in decomposer communities (Santschi et al., 2018;Stocker et al., 2017). The effect of invertebrates on the decomposition process became increasingly important with decreasing leaf quality. High-quality leaf species are important carbon sources for microbial consumers, while low-quality species are important for invertebrates as substrata for attachment and eventually as a source of particulate organic matter (Ardón & Pringle, 2008).
The meiofauna can affect litter decomposition in several key ways: (a) as microshredders for the direct consumption of leaves; (b) grazing and selective feeding on fungal communities; and (c) bioturbation effects on the leaf-associated microenvironment (Mathieu, Leflaive, Ten-Hage, De Wit, & Buffan-Dubau, 2007;Perlmutter & Meyer, 1991;Traunspurger, Bergtold, & Goedkoop, 1997). In this study, the meiofauna significantly improved the decomposition rates of A. chinense and L. acalycina, rather than on M. leptophylla.
Regarding the low-quality species (M. leptophylla), the meiofauna significantly affected neither the leaf decomposition nor the associated fungal biomass. Compared to the macrofauna, the meiofauna had a reduced direct consumption of low-quality leaves (Nolen & Pearson, 1993).
The presence of meiofauna significantly enhanced the decomposition of high-and medium-quality leaves, but in different ways.
Specifically, the meiofauna's grazing effect decreased the fungal biomass for high-quality leaves but increased the fungal biomass and diversity of medium-quality leaves. A positive relationship was seen between fungal biomass and litter decomposition, so that a reduced fungal biomass may have led to the slow litter decomposition (Duarte, Pascoal, Alves, Correia, & Cássio, 2010;Pascoal & Cássio, 2004).
The top-down effect of the small chironomid larvae on fungal biomass and respiratory activity were also found for Salix alba, another high-quality leaf species (Ágoston-Szabó et al., 2016). Nevertheless, the meiofauna stimulated fungal biomass of medium-quality leaves, probably because the moderate grazing on aging microorganisms can keep the fungal community active and increase the fungal demand for nutrients from leaves (Chen & Wang, 2018;Lillebø, Flindt, Pardal, & Marques, 1999;Piot, Nozais, & Archambault, 2013). Taking together, for high-quality leaves, the direct consumption by meiofauna outweighed the potential negative effect on decomposition These findings, with regard to the meiofauna, were the leaf quality dependent. Different qualities of leaf type support different microbial communities (Gulis, 2001;Jabiol & Chauvet, 2012) and may also support different colonized meiofauna. Meanwhile, the meiofauna can change the fungal composition by preferentially feeding on some taxa. For instance, fungivorous nematodes (e.g., from the Aphelenchida) have very specific fungal diets (Dighton, Zapata, & Ruess, 2000). The selective feeding of meiofauna could reduce the growth of some fungi taxa, while promoting others, thereby reducing the competition for resources between different fungi species (Chambord et al., 2017). Indeed, the meiofauna reduced the relative abundance of Alatospora acuminata and

T. Marchalianum but increased T. elegans and Penicillium italicum
that showed specific ligninolytic activities (Hofmann et al., 2016;Osono, 2007). Because of the complexity and importance of the feeding preference for meiofauna, like those of the shredding macroinvertebrates, further studies are warranted (Canhoto & Graça, 2008;Mora-Gómez et al., 2016). Moreover, the competition be-  For medium-quality leaf species, the macrofauna could maintain the positive interactions between meiofauna and leaf-associated fungi, though this may have been offset by the predation of macrofauna on the meiofauna (Ptatscheck et al., 2020).

| Effect of macrofauna
For low-quality leaf species, the role of the macrofauna is even more pronounced. The macrofauna's direct consumption produced fragments of plant material that could allow microorganisms to enter the nutritious internal tissues. The fragments can be more readily utilized by meiofauna and microbes, favoring decomposition (Santonja et al., 2018). In addition, the increased abundance and diversity of microbes in the presence of macrofauna may also contribute to decomposition by improving the palatability of leaves for the macrofauna.

| CON CLUS ION
With regard to increased trophic complexity, this study provides evi- Overall, for detritus-based ecosystems, these findings imply that the meiofauna and trophic diversity in the decomposer community are crucial for litter decomposition and subsequent nutrients dynamics.

ACK N OWLED G M ENTS
The authors acknowledge the financial support provided by Returnees (CX2017120).

CO N FLI C T O F I NTE R E S T
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.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available in the Dryad Digital Repository at http://doi.org/10.5061/dryad. cfxpn vx3g.