Fungal succession on the decomposition of three plant species from a Brazilian mangrove

Leaf decomposition is the primary process in release of nutrients in the dynamic mangrove habitat, supporting the ecosystem food webs. On most environments, fungi are an essential part of this process. However, due to the peculiarities of mangrove forests, this group is currently neglected. Thus, this study tests the hypothesis that fungal communities display a specific succession pattern in different mangrove species and this due to differences in their ecological role. A molecular approach was employed to investigate the dynamics of the fungal community during the decomposition of three common plant species (Rhizophora mangle, Laguncularia racemosa, and Avicennia schaueriana) from a mangrove habitat located at the southeast of Brazil. Plant material was the primary driver of fungi communities, but time also was marginally significant for the process, and evident changes in the fungal community during the decomposition process were observed. The five most abundant classes common to all the three plant species were Saccharomycetes, Sordariomycetes, Tremellomycetes, Eurotiomycetes, and Dothideomycetes, all belonging to the Phylum Ascomycota. Microbotryomycetes class were shared only by A. schaueriana and L. racemosa, while Agaricomycetes class were shared by L. racemosa and R. mangle. The class Glomeromycetes were shared by A. schaueriana and R. mangle. The analysis of the core microbiome showed that Saccharomycetes was the most abundant class. In the variable community, Sordariomycetes was the most abundant one, mainly in the Laguncularia racemosa plant. The results presented in this work shows a specialization of the fungal community regarding plant material during litter decomposition which might be related to the different chemical composition and rate of degradation.


Scientific Reports
| (2022) 12:14547 | https://doi.org/10.1038/s41598-022-18667-x www.nature.com/scientificreports/ dynamics in this habitat depends on the physical and chemical properties of the substrate 1 . Thus, it is of great importance to have additional information about fungi diversity and distributions to improve the predictions about the diversity of fungal communities associated with different plant species around the world 9 . Succession is the process of change in community composition over time in an ordinated manner 12,14 . Studies of fungi decomposition showed this process many marine ecosystems [15][16][17] . For the fungal communities, the decomposition process of leaf litter generally occurs in three distinct phases: early, intermediate, and late decomposers 18 . However, the degradation of leaves in mangroves is dynamic and generally occurs in a matter of a few weeks 1 which can alter this organized process. Therefore, the aim of this study was to evaluate and compare the diversity of fungi community during the decomposition process of three mangrove plant species from Cananéia mangrove, Brazil. We set the assays using microcosms conditions to follow up the decomposition process of the three plant species during four different periods. Therefore, this study hypothesized that during leaf decomposition the different plant species select for a specific fungal community structure along the process and this selection reflect on differences in the role of the observed populations.

Material and methods
Site description. Samples were collected from one mangrove forest in the city of Cananéia in São Paulo state, Brazil. The collection of plant, sediment and water was approved by the Biodiversity Authorization and Information System (#65580) and the study complies with local and national guidelines. The mangrove located in Cananéia (Can) (25° 05′ 03″ S-47° 57′ 75″ W) a pristine area with little human influence. Samples were taken during the low tide. Most mangroves of the south coast of Brazil have homogeneous vegetation composed of Rhizophora mangle, Laguncularia racemosa, and Avicennia schaueriana. Plant species identification was made by Marta A. Moitinho and Rodrigo G. Taketani based of marked morphological differences between the species and the material was not deposited in a herbarium.
Sample collection. Fresh mature leaves that did not present any sign of lichen or lesion of any sort were carefully collected directly from these three plant species (avoiding contact with sediment, water or human DNA). Samples were collected from 5 at least individual trees fond in a 100 m 2 area and leaves were mixed to form a compost sample from each specie. Five kilograms of the top 5 cm of sediments were collected in each mangrove sample was a compost of three locations within the area were leaves were collected. 5 l of brackish estuarine water was also sampled in the proximities of the sampling sites. The material was immediately placed in sterile bags and transported to the laboratory for immediate processing and microcosm implementation.
Microcosm's construction. The experiment was conducted in a completely randomized design consisting of 3 plant species × 4 time periods × 3 repetitions totalizing 36 samples. Microcosms were built using 50 ml penicillin flasks (50 mm diameter) to evaluate the fungal dynamic throughout the organic matter decomposition process. Microcosms were constructed in three replicates, and each one contained 20 g of sediment, eight disks of leaves randomly selected (1.7 cm diameter) from one of the three plant species without any previous treatment in order to decrease external influences in the treatments and simulate mangrove ecosystems. 2 ml of brackish water was placed on top of the sediments. They were incubated at 24 °C and monitored for 60 days with subsamples taken at 7, 15, 30, and 60 days. Samples were randomly spread in the incubator and microcosms were selected at random for sampling. At each sampled date, four leaf disks were collected and used to extract the DNA 14 . With the intent to have the acclimatization and stabilization in the microcosms, by the mixture of the fresh leaves along the sediment and water simulating the environmental conditions found in the mangrove sediment, no subsample was taken at time zero, and the first sample on day 7 was considered as a parameter.
Fungi communities' analysis using ITS region sequencing. The samples of decaying leaves were submitted to DNA extraction using the PowerSoil DNA Isolation Kit (MoBio Carlsbad, USA) according to the manufacturer's instructions. These extractions were made by transferring 0.25 g of plant material directly from the microcosms into the PowerBead tubes provided in the Power Soil. Quality and quantity of the extracted DNA were evaluated by Nanodrop spectrophotometer (Thermo Scientific 2000 spectrophotometer) and in 2.4% agarose gel.
The samples were PCR amplified using the primers set ITS1f 19 and ITS2 20 , along with tags of five nucleotides with the forward primer to distinguish the samples, and to generate the amplicons of the ITS rRNA gene. The amplification mix contained 18 µl of water; 2.5 µl of PCR buffer; 3.3 µl of MgCl 2 ; 0.5 µl od dNTPs; 0.5 µl of Taq DNA Polymerase (Invitrogen, São Paulo, Brazil); 0.25 µl of each primer and finally 0.5 µl of DNA. Cycling conditions were an initial denaturation of 94 °C for 3 min; 30 cycles of 94 °C 30 s; 57 °C for 45 s; and 72 °C for 1 min; and a final 2 min extension at 72 °C.
The PCR product was purified with the Promega Wizard® SV Gel and PCR Clean-Up System Kit, quantified by Qubit 2.0 Fluorometer (Life Technologies) and Sequencing was performed on an Ion Torrent PGM system of Life Technologies using the Ion 316™ Chip. The enrichment phase was performed by using the One-Touch 2 device with the Ion Sequencing 400 Kit according to the manufacturer's instruction (Life Technologies) 14 . Raw sequencing data obtained from the PGM were processed using QIIME software (Quantitative Insights Into Microbial Ecology) following the tutorial described previously.
Processing of the sequences. Raw sequencing data obtained from the PGM were processed using QIIME software (Quantitative Insights Into Microbial Ecology) 21 following a modified version of the 454 Overview Tutorials using the UNITE database 22 . The identification and removing of chimaeras were made with the following commands "identify_chimeric_seqs.py" and "filter_fasta.py" respectively. The Otus were picked www.nature.com/scientificreports/ with the command "pick_otus.py", and the selection of the representative Otus was made with the "pick_rep_set. py" command. The taxonomies were assigned with "assign_taxonomy.py", and the OTU tables were assembled with "make_otu_table.py". The normalization values were obtained with the command "biom summarize-table". After the initial filtering steps, we obtained 314,786 million reads with an average of 8744.056 reads per library (min. 4196.0 and max. 18,777.0). After removing unclassified microorganisms not belonging to the Fungi reign, the reads count dropped to 11,646 taxa distributed in 36 samples and seven variables (3 plant species × 4 time periods × 3 repetitions). After the rarefaction step, the taxonomic table remained with 6378 taxa among five taxonomic ranks that were used in downstream analyses.

Statistical analyses.
We evaluated the composition of the whole community and the core and variable portions of fungi communities during the decomposition process of three plant species from Cananéia Mangrove. To improve the understanding from the 26,739 representative OTUs picked, only 11,648 were classified in the kingdom Fungi and were retained for analysis. The α diversity was calculated considering Shannon and Simpson Indexes and richness was analyzed by Chao1, both calculated in Phyloseq Package on R Environment 23 . The Permutational Analysis of Variance was performed using function Adonis 24 , with Bray Curtis distance matrix and 999 permutations in order to identify the contribution of the factors (Plant and Time) in the fungi community structure. Principal coordinates analysis (PCoA) was performed to summarize the variation of the community structure in each plant (colours) and time (shapes). Bar plots with the dominant taxonomic groups, (i.e., ten most abundant phyla) were generated to identify the contribution of different Phyla in each treatment 25 . The ternary plots were created with Package ggtern 26 , considering the relative average of each Phylum.
Core communities are described as the organisms that are widely distributed in a determined environmental type despite variations within it 27 and at the same time, variable communities are organisms which appearance varies along with a given environmental parameter. The identification of the core and variable communities was made by a probabilistic method as described by Gumiere, 2018 28 using the R software version 3.2.2 (http:// www.r-proje ct. org/). The size of the core and variable communities and also the taxonomic compositions were verified using bar plot graphics. Similarly, the identification of specialists and generalists were performed using the EcolUtils R package 29 according to Levin's definition of niche breadth 30 and 100 permutations. FunGuild 31 was used to assign fungal phylotypes to a functional guild where possible.
Nucleotide sequence accession numbers. Fungal ITS rRNA gene sequences obtained in this study are publicly available in the NCBI server under the SRA accession number: SRP151571.

Results
Alpha and beta-diversity profiles. The alpha diversity metrics analysis indicated that CHAO1 richness estimator did not differ significantly along time for all the three studied plant species, but in T7 it was lower in Avicennia shaueriana compared to the other species (Fig. 1a). The Shannon diversity index also did not show significant differences along time for A. shaueriana and L. racemosa; however, R. mangle showed a drop on T15 followed by a new increase in diversity in T30 and T60. L. racemosa showed highest Shannon diversity compared to the other species mainly on T7 and T30 (Fig. 1b). Simpson index showed the same trend (Fig. 1c).
The role of the plant species as the main factor contributing to the fungal community structure during the decomposition process is also evident in the Principal Coordinated analysis (PCoA). Clear separation of the samples based on plant material were observed despite no separation of samples based on time (Fig. 2a). The compositional plot of PCoA shows a higher number of OTUs from the phyla Ascomycota and Basidiomycota in all plant species (Fig. 2b). The Permanova analysis showed that plant material was the most significant factor affecting the plant decomposition mediated by the fungi community in Cananéia mangrove (R 2 = 0.41, p = 0.001) with time also being not significant for the process (R 2 = 0.07 considering p > 0.05).
Community composition. The six predominant fungal phyla in the three plant species at Cananéia mangrove were Ascomycota, Basidiomycota, Chytridiomycota, Glomeromycota, Incertae sedis and Zygomycota. The plant material from different species selected different taxa during the process of degradation. The most abundant classes common to all the three plant species were Saccharomycetes, Sordariomycetes, Tremellomycetes, Eurotiomycetes, Dothideomycetes, as well as unidentified classes belonging to the Phylum Ascomycota and fungi from unidentified Phylum (Fig. 3).
Saccharomycetes was the predominant class in all four stages in A. shauerianna. After 15 days, this group presented a drastic reduction while there was a considerable increase in Sordaryomicetes and the unclassified groups of the Phylum Ascomycota. In the samples that followed, Eurotiomycetes and other minor classes presented an increase in abundance. The decomposition process in L. racemosa was more dynamic in the final phases (time 30 and 60). Sordariomycetes was the most abundant class until the 30th day. The initial stages (7 and 15 days) presented similar abundances of the classes Sordariomycetes, Saccharomycetes, Tremellomycetes and unclassified classes from the Phylum Ascomycota. In the samples from the 30th day, Sordariomycetes and Eurotiomycetes increased their abundance while Saccharomycetes and Tremellomycetes decreased. In the last sampling time (60th day), the most abundant classes (Sordariomycetes and Saccharomycetes) had very similar proportions while Unclassified Ascomycota, Eurotiomycetes, Dothideomycetes and Incertae sedis had an increase in their proportions compared to the former samples. Similarly, Sordariomycetes was the most abundant in all samples except in the final phase (60th day) in R. mangle (Fig. 3). This plant also presented a substantial increase in the abundances of Eurotiomycetes during at time 60. Sordariomycetes considerable increased its abundance in time 15 compared to the others. In the Last period, Saccharomycetes and Sordaryomicetes had equal abundances, while Dothideomycetes and unclassified Ascomycota were slightly increased.  Role of fungi on the community. Unclassified fungi accounted for more than 50% of the organisms in the core community in all the three plant species. The most abundant class in the core community was Saccharomycetes (with more than 10% in all samples) and Sordariomycetes both belonging to the Phylum Ascomycota. Sordariomycetes were expressive more abundant in R. mangle (34%) than in L. racemosa (8%) and A. schaueriana (2%). The phylum Basidiomycota occurred in lower abundance being represented by the class Tremellomycetes (less than 5% for all samples) (Fig. 4).
Higher proportions of unclassified microorganisms (more than 50%) also were observed in the variable fungi community followed by the class Sordariomycetes that had higher proportions in L. racemosa (40%) than in A. schaueriana (16%) and R. mangle (17%). As opposed to the core community, the class Saccharomycetes occurred in lower abundance in the variable communities (less than 3% for all samples). The other variable classes that composed the variable communities were Eurotiomycetes, Dothideomycetes (phylum Ascomycota) and Tremellomycetes (phylum Basidiomycota) (Fig. 4).
Despite the variation in the frequencies of core (or variable) OTUs over time, some classes were more abundant in one group rather than the other. Saccharomycetes, Tremellomycetes, and Sordariomycetes were dominants in the core group, while the Sordariomycetes, Eurotiomycetes, Dothideomycetes and unclassified classes were abundant in the variable group.
The separation of OTUs into generalists and specialists (sensu Levins), indicates that there were more OTUs that could be significantly (p < 0.05) assigned to the latter than the former (Fig. 5). Saccharomycetes were the most abundant in both groups, followed by Sordariomycetes. However, within the specialists, the Sordariomycetes was less abundant in A. shaueriana than the other plant material. Also, Tremellomycetes was the third most frequent class among the generalists and was almost undetected among specialists indicating that most of the individuals of this class found in this samples must have a large niche breadth.
In terms of relative abundance, A. shaueriana decaying leaves had the highest abundance of generalists than the other material while the specialist's frequency was more stable. Also, a temporal variation could be observed for the former while the latter presented low variation. The three plant materials showed a high abundance of generalists the seventh day (and fiftieth day for L. racemosa), with a sharp decline after that and a steady increase in the last two samples.
The identification of the functional guilds to which the detected OTUs were assigned 31 also indicated that a difference among the different decaying material and time of sampling (Fig. 6). A. shaueriana had an overall higher proportion of saprotrophs from different groups than the remaining leaves. Interestingly, the other plant material had a high proportion of OTUs assigned as a plant pathogen. Most of these OTUs were identified as belonging to the genus Pilidiella and Pestalotiopsis. The Pilidiella was only identified in L. racemosa while Pestalotiopsis was detected in both.

Discussion
In this study, we aimed to provide an analysis of the fungal community involved in leaves decomposition from three mangrove plant species, in Cananéia mangrove, Brazil. The molecular identification revealed that the decomposition process in the Cananéia mangrove is composed of a diverse community of fungi, and this result is in agreement with previous study of the fungal diversity in this environment 3 . Previous studies demonstrated the succession process of fungal communities during decomposition of leaf litter material from different plants and different locations like parks vegetation 32,33 ; woodland stream 12 ; temperate forests 34,35 and tropical seasonal forest 36 . During the decomposition of plant material in a temperate forest, fungal biomass increased rapidly from low values to a maximum point at the 60th day of the experiment 34 .
The main factor contributing to the fungal community structure during the decomposition process was the plant species. A previous study with the endophytic fungi community showed that plant species was a key driver of the fungal community 3 . However, differently from what was observed to the endophyte's community 3 , in the decomposition process, L. racemosa was the plant that presented the lowest numbers of populations of endophytic fungi, followed by Rhizophora mangle. A. schaueriana was the plant species with the highest community of endophytic fungi.
Previously studies have already shown that in mangroves, the fungal colonization of plants, the decomposition process, and the succession stages vary according to the kind of mangrove substrate like plant species 16,37 . Possibly this differences in community composition and assembly are affected by leaf chemical characteristics 37 . For instance, species of Rhizophoraceae have a high concentration of tannin, a phenolic substance that can inhibit fungal growth 37 , A. schaeuriana present high concentration of salt on their leaves 38,39 , while L. racemosa presents less salt secretion on their leaves than A. schaueriana 38,40 and less recalcitrant compounds than R. mangle 14,41 . The chemical content of these plants differs, and this influences how palatable the material is to each microorganism 41 . It also influences the bacterial community composition and succession process 14 .
Core communities are described as the organisms that are largely distributed in a determined environmental type despite variations within it 27 ; these organisms can be crucial in the maintenance of this habitat. Among the fungal community, the core group of species exerts a significant influence on the turnover of litter in mangrove ecosystems 16 . Classes belonging to the Ascomycota were predominant in the core and variable community in the three studied plants, indicating that they can exploit multiple litter types. Saccharomycetes were the predominant class in the core group, with higher density in the R. mangle plant compared with the two other plants, while in the variable community Sordariomycetes were more abundant, mainly in the L. racemosa plant. Indicating that these groups have some degree of specificity to the plant substrate.
There is a conceptual difference between core taxa and specialist (as well as a variable and a generalist). While the core is an OTU that is expected to be found in every sample of a given treatment (as opposed to the variable that is not 28 ), a specialist is an OTU that has a smaller niche breadth (NB) than a determinist 30 . Thus, the characteristics of the environment, such as the resources present, would be key to determining ones  www.nature.com/scientificreports/ NB. Therefore, the detection of more generalist taxa in A. shauerianna indicates that either there is a variety of resources or that they are more labile. This observation is supported by the faster degradation rates of A. shauerianna 42 . This might also be the reason for the higher abundance of Sordariomycetes in the specialists' group L. racemosa and R. mangle. These OTUs might be related to the adapted or degraded recalcitrant materials such as lignin and tannins that are present in higher concentrations in these plants 42 . The Phylum Ascomycota is an essential group in the initial stages of the decomposition process of plant material in different environments 34,43,44 . They are one of the dominant phyla of fungi in Brazilian mangroves 3 and were predominant in the rhizosphere and bulk soil collected from the Red Sea grey mangroves (Avicennia marina) 45 . Members of this Phylum generally decompose cellulose selectively over lignin 46 and dominate the fungal community and are highly active (production of extracellular enzymes) in the early stages of litter decomposition 47 . Aquatic ascomycetes play important roles in wood and leaf litter degradation since basidiomycetes are found in lower quantities 11 . Ascomycetes appear to be better able to withstand the conditions prevailing in aquatic habitats, far more than basidiomycetes 48 .
The phylum Basidiomycota was described in mangrove ecosystems in lower quantities than the Ascomycota phylum, which corroborates with the results found in this study 3 . The phylum Basidiomycota also is constant during the decomposition process of leaf material, and are considered to have more ligninolytic abilities than the ascomycetes 46 ; and are correlated with the late stages of decomposition of leaf litter 34,46 .
The main classes of fungi found in this study were previously reported in different decomposition studies from different environments. Leotiomycetes, Sordariomycetes and Eurotiomycetes were the abundant classes during decomposition of litter from a forest of Austria 49 . Sordariomycetes, Saccharomycetes, Eurotiomycetes and Dothideomycetes classes were described during the decomposition process of wheat straw 50 . Ascomycota, Dothideomycetes, and Leotiomycetes dominated in the communities, and there was a high relative abundance of Sordariomycetes. They were also the most abundant groups on submerged litter decomposition in streams 51 . Sordariomycetes also was the most recovered class from aquatic environments from Mexico and the other classes like Saccharomycetes, Eurotiomycetes. Leotiomycetes, Dothideomycetes, Sordariomycetes, Tremellomycetes and Chytridiomycetes also were observed in this study 11 . Microbotryomycetes are Basidiomycete yeasts common to aquatic environments like mangroves. Agaricomycetes also were isolated from mangrove wood substrate and Incertae sedis are common fungi in decaying leaf litter in streams 48 . Chytridiomycota also is an abundant phylum described in aquatic environments 11 .
The high abundance of saprotrophic OTUs in this material indicates that the taxa found are likely involved in the degradation of the leaf material. It also indicates a lower abundance of cross feeders. However, the high abundance of OTUs related to the plant pathogenic Pilidiella and Pestalotiopsis fungi indicates that some abundant OTUs might not be active in the degradation of these materials. Nevertheless, despite the presence of many plant pathogens in these genera, several isolates were shown to degrade lignin 52,53 .
Altogether, our results indicated that the different plant materials selected for specific fungal communities. The Ascomycota dominated in the first stages of degradation and were later substituted by Basidiomycota. This change was reflected in the differences in the core and variable communities between plant materials and overtime, where we observed a similar contrast in the abundance of each group in the entire community.
In conclusion, the results presented in this work show that there is aspecialization of the fungal community while degrading different plant species of mangrove. This is associated with the different rates of degradation expected due to the chemical composition of the leaves. For instance, OTUs identified as Saccharomycetes are more abundant in samples of A. shauerianna that has lower concentrations of tannins and lignin. It also provides essential information about the frequency and composition of the main fungal classes of the communities associated with the degradation of three plants in the Brazilian mangrove forests.

Data availability
Fungal ITS rRNA gene sequences obtained in this study are publicly available in the NCBI server under the SRA accession number: SRP151571.