TY - JOUR
AU1 - Liang, Shiyou
AU2 - Wang, Chengpan
AU3 - Ahmad, Farhan
AU4 - Yin, Xuejie
AU5 - Hu, Yin
AU6 - Mo, Jianchu
AB - Introduction Herbivorous insects obtain energy and nutrients from their host plants. However, lignocellulose in plants biomass is highly recalcitrant to enzymes attack. To overcome this barrier, these insects have evolved specific feeding tactics [1] or digestive capacities such as masticating organs, gut structures, digestive enzymes, and symbiotic systems [2]. Another challenge is a wide variety of constitutive, or induced defense compounds. These chemical defenses can act as repellents, or antinutritives, or toxics, or attract predators [1, 3–5]. The herbivores have also adapted some strategies against these compounds. For example, the desensitization of insect proteins targeted by plant defenses or the modification of ingested plant chemicals [3]. The capacity of herbivores to break these two obstacles shape them to be specialists or generalists. Fungus-cultivating termites (Blattodea, Termitidae, Macrotermitinae) are successful herbivores, feed on various plant materials such as dry grass, leaf litter, root and ruminant dung [6, 7]. They contribute to more than 90% of plant biomass in arid tropics [8] through the efficient degradation of highly recalcitrant lignocellulosic materials. To this purpose, they cultivate special basidiomycete fungi (Agaricomycetes, Lyophyllaceae, Termitomyces) in their nests as an external “digestive” system called fungus-combs. The food processing in Macrotermitinae presents a complex age polyethism involving two gut passages [9–11] (Fig 1). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 1. The food processing pathway in fungus-cultivating termite (O. formosanus). Food forage: old workers collect the food from outside environment and pile in the colony; First gut passage: young workers ingest the foraged material along with nodules (conidia) from middle-aged fungus-comb and excrete primary feces on the top surface of the comb to establish fresh comb; Comb fermentation: fungus grows on the fresh comb, degrade it efficiently to convert it into middle-aged comb, where new nodules are produced [12, 13]. The comb turns to mature in 45 days [12]; Second gut passage: old workers ingest mature comb and excrete final feces. YW, young worker; OW, old worker. https://doi.org/10.1371/journal.pone.0232329.g001 The termite comb not only harbor fungus but also bacteria [14–16]. The lignocellulolytic potential of Termitomyces has been demonstrated by enzyme activity assays of culture in vitro [17, 18], RNAseq in situ [19], and genome sequencing [10]. However, the functions of the bacteria in comb are unclear. In wood-feeding termite, Fibrobacteres, Firmicutes, Bacteroidetes, the TG3 phylum, Elusimicrobia, Spirochaetes are the fiber associated lignocellulolytic bacteria [20–23]. Some studies indicated that bacteria in fungus comb have antimicrobial, polycyclic-degrading and cellulose-degrading activity [24–27]. Otani et al. [15] and Long et al. [16] described the presence of bacterial community in field-collected fungus-combs, but their studies are unable to clarify what kind of community composition are active at different decomposition stages of comb and how do they respond to different plant substrates? Therefore, we fed fungus-cultivating termite (Odontotermes formosanus) with three different kinds of lignocellulose diets such as leaves, branches, and mix food respectively in laboratory and sampled the fungus-combs at different stages of decomposition or ages. Further, we also explored the impact of three physicochemical indexes (moisture, ash content and organic matter) on bacterial diversity in fungus-combs. Materials and methods Termite colonies feeding and sample collection Three colonies of O. formosanus (height ≈ 6–8 cm; diameter ≈ 12–15 cm) harboring king(s) and queen(s) were collected from a forested area in Sanming, Fujian Province, China (117.90° E, 25.53° N). The private land owner gave permission for this collection. The colonies along with fungus-combs were wrapped in plastic film separately and transported to the lab within 12 h of excavation. All colonies were placed separately in plastic chambers (45 × 45 ×30 cm, length × width × depth) containing clay soil obtained from the area where the colonies were collected. The rearing systems were maintained in complete darkness at 26 ± 1°C and > 90% relative humidity. After the colonies became stable in a laboratory, they were allowed to feed on one of the three different kinds of food, including i) leaves or ii) branches of Quercus palustris, or iii) mix food collected from Laohe Mountain, Hangzhou, Zhejiang Province, China. All colonies were successful in building new fungus-combs. We divided the whole comb into three layers i) top or fresh comb, ii) middle or middle-aged comb, and iii) bottom or mature comb (Fig 1). The fungus-combs constructed by individual food were collected as described previously by Li et al. [9]. Briefly, dark brown color for fresh, yellowish-brown for middle-aged and gray for mature. For every sample, 3 g was used to measure the physicochemical indexes immediately, and 0.5 g was stored at—80°C for subsequent microbial diversity analysis. Measurement of physicochemical indexes The moisture, ash content and organic matter were measured from each sample. For moisture content measurement, a weighing bottle was washed, dried in the oven and weighed (m1). Approximately 0.5 g fungus-comb was put into the bottle and re-weighed (m2), dried in an oven at 105°C for 6 hours, and re-weighed (m3) [28]. In case of ash content, a ceramic crucible was washed and dried in a muffle furnace at 550°C for 4 hours and weighed (m4). Approximately 0.5 g of totally dried fungus-comb was put into the crucible and re-weighed (m5), re-dried at 550°C for 4 hours, and re-weighed (m6) [29]. The moisture, ash and organic matter content were determined as follows: DNA extraction Approximately, 0.15 g of each sample was ground in a 1.5 ml Eppendorf tube with 800 μl Tris-EDTA (100 mM Tris base, 50 mM EDTA, PH 8.0). The mixture was transferred to a 2 ml screw-cap tube and homogenized by bead-beating [30]. After that, the homogenate was mixed with 100 μl 5 mg/ml lysozyme and incubated at 37°C for 20 min. Next, 60 μl 32 mg/ml protease k and 40 μl 25% SDS, were added and incubated at 37°C for 40 min. Then, the lysate was mixed with 167 μl 5 M NaCl and 133 μl 10% CTAB, and incubated at 65°C for 10 min [31]. After the extraction by phenol-chloroform-isoamyl alcohol (25: 24: 1, v: v: v) from the lysate, the crude DNA was purified using a Qiagen DNeasy column following the protocol as described by the company. The purified DNA in the column was eluted by 100 μl AE buffer (Qiagen). Illumina sequencing The bacterial communities in fungus-combs were analyzed using 454 Illumina sequencing. The primer pair 338F (ACT CCT ACG GGA GGC AGC AG) and 806R (GGA CTA CHV GGG TWT CTA AT) [32] were used to amplify the V3—V4 region of bacterial 16S rRNA genes in the respective DNA samples. The amplified products were purified by the AxyPrep DNA kit (AXYGEN, China). The amplicons were sequenced on Illumina MiSeq by Shanghai Personal Biotechnology Inc., China. Data analysis The physicochemical indexes were analyzed through Duncan's new multiple range method using SPSS, v.22.0 (SPSS Inc., Chicago, IL) at α = 0.05. Raw sequences were quality filtered using the QIIME release 1.8.0 software [33] under the stringent conditions (reads ≥ 150 bp, no ambiguous bases, bases mismatched by 5’ primer ≤ 1 and a maximum number of homopolymers ≤ 8). Subsequent removal of chimeras was carried out through UCHIME [34] by uchime3_denovo. After reducing the redundancy of sequences and demarcating OTU (Operational Taxonomic Unit) with a similarity threshold of 97% using UCLUST [35], sequences with the highest abundance of each OTU were selected as the representative sequences of corresponding OTU. Representative sequences of each OTU were aligned and taxonomically classified in the reference database, SILVA [36] using MOTHUR [37]. A subsequent assignment to the termite-specific reference database (DictDb) [38] was performed with a confidence threshold of 60%. OTUs were simplified by removing the OTUs whose abundance was lower than 0.001% of total sequences [39]. QIIME was used to calculate and map the rarefaction curves based on 97% sequence similarity cutoff and calculate diversity indexes of each sample by uclust. The relative abundance of bacteria at different levels were calculated and mapped through R software (http://www.r-project.org). LEfSe (linear discriminant analysis effect size) analysis was performed to identify discriminatory taxa through the online platform, Galaxy (http://huttenhower.sph.harvard.edu/galaxy/). Heatmap of selected bacterial lineages, the constrained corresponding analysis (CCA) of the bacterial community and variation partition analysis for physicochemical indexes were implemented in R software (packages pheatmap and vegan). The high-quality sequences of the samples have been submitted to the NCBI Sequence Read Archive (BioProject PRJNA604886); for accession numbers, see S5 Table. Termite colonies feeding and sample collection Three colonies of O. formosanus (height ≈ 6–8 cm; diameter ≈ 12–15 cm) harboring king(s) and queen(s) were collected from a forested area in Sanming, Fujian Province, China (117.90° E, 25.53° N). The private land owner gave permission for this collection. The colonies along with fungus-combs were wrapped in plastic film separately and transported to the lab within 12 h of excavation. All colonies were placed separately in plastic chambers (45 × 45 ×30 cm, length × width × depth) containing clay soil obtained from the area where the colonies were collected. The rearing systems were maintained in complete darkness at 26 ± 1°C and > 90% relative humidity. After the colonies became stable in a laboratory, they were allowed to feed on one of the three different kinds of food, including i) leaves or ii) branches of Quercus palustris, or iii) mix food collected from Laohe Mountain, Hangzhou, Zhejiang Province, China. All colonies were successful in building new fungus-combs. We divided the whole comb into three layers i) top or fresh comb, ii) middle or middle-aged comb, and iii) bottom or mature comb (Fig 1). The fungus-combs constructed by individual food were collected as described previously by Li et al. [9]. Briefly, dark brown color for fresh, yellowish-brown for middle-aged and gray for mature. For every sample, 3 g was used to measure the physicochemical indexes immediately, and 0.5 g was stored at—80°C for subsequent microbial diversity analysis. Measurement of physicochemical indexes The moisture, ash content and organic matter were measured from each sample. For moisture content measurement, a weighing bottle was washed, dried in the oven and weighed (m1). Approximately 0.5 g fungus-comb was put into the bottle and re-weighed (m2), dried in an oven at 105°C for 6 hours, and re-weighed (m3) [28]. In case of ash content, a ceramic crucible was washed and dried in a muffle furnace at 550°C for 4 hours and weighed (m4). Approximately 0.5 g of totally dried fungus-comb was put into the crucible and re-weighed (m5), re-dried at 550°C for 4 hours, and re-weighed (m6) [29]. The moisture, ash and organic matter content were determined as follows: DNA extraction Approximately, 0.15 g of each sample was ground in a 1.5 ml Eppendorf tube with 800 μl Tris-EDTA (100 mM Tris base, 50 mM EDTA, PH 8.0). The mixture was transferred to a 2 ml screw-cap tube and homogenized by bead-beating [30]. After that, the homogenate was mixed with 100 μl 5 mg/ml lysozyme and incubated at 37°C for 20 min. Next, 60 μl 32 mg/ml protease k and 40 μl 25% SDS, were added and incubated at 37°C for 40 min. Then, the lysate was mixed with 167 μl 5 M NaCl and 133 μl 10% CTAB, and incubated at 65°C for 10 min [31]. After the extraction by phenol-chloroform-isoamyl alcohol (25: 24: 1, v: v: v) from the lysate, the crude DNA was purified using a Qiagen DNeasy column following the protocol as described by the company. The purified DNA in the column was eluted by 100 μl AE buffer (Qiagen). Illumina sequencing The bacterial communities in fungus-combs were analyzed using 454 Illumina sequencing. The primer pair 338F (ACT CCT ACG GGA GGC AGC AG) and 806R (GGA CTA CHV GGG TWT CTA AT) [32] were used to amplify the V3—V4 region of bacterial 16S rRNA genes in the respective DNA samples. The amplified products were purified by the AxyPrep DNA kit (AXYGEN, China). The amplicons were sequenced on Illumina MiSeq by Shanghai Personal Biotechnology Inc., China. Data analysis The physicochemical indexes were analyzed through Duncan's new multiple range method using SPSS, v.22.0 (SPSS Inc., Chicago, IL) at α = 0.05. Raw sequences were quality filtered using the QIIME release 1.8.0 software [33] under the stringent conditions (reads ≥ 150 bp, no ambiguous bases, bases mismatched by 5’ primer ≤ 1 and a maximum number of homopolymers ≤ 8). Subsequent removal of chimeras was carried out through UCHIME [34] by uchime3_denovo. After reducing the redundancy of sequences and demarcating OTU (Operational Taxonomic Unit) with a similarity threshold of 97% using UCLUST [35], sequences with the highest abundance of each OTU were selected as the representative sequences of corresponding OTU. Representative sequences of each OTU were aligned and taxonomically classified in the reference database, SILVA [36] using MOTHUR [37]. A subsequent assignment to the termite-specific reference database (DictDb) [38] was performed with a confidence threshold of 60%. OTUs were simplified by removing the OTUs whose abundance was lower than 0.001% of total sequences [39]. QIIME was used to calculate and map the rarefaction curves based on 97% sequence similarity cutoff and calculate diversity indexes of each sample by uclust. The relative abundance of bacteria at different levels were calculated and mapped through R software (http://www.r-project.org). LEfSe (linear discriminant analysis effect size) analysis was performed to identify discriminatory taxa through the online platform, Galaxy (http://huttenhower.sph.harvard.edu/galaxy/). Heatmap of selected bacterial lineages, the constrained corresponding analysis (CCA) of the bacterial community and variation partition analysis for physicochemical indexes were implemented in R software (packages pheatmap and vegan). The high-quality sequences of the samples have been submitted to the NCBI Sequence Read Archive (BioProject PRJNA604886); for accession numbers, see S5 Table. Results Effect of diets and fungus-comb ages on the physicochemical indexes The physicochemical indexes including moisture, ash content and organic matter from the fungus-combs built by leaves, branches, and mix food respectively, are summarized in Table 1 and Fig 2. The moisture content at different decomposition stages of the combs i.e., top, middle and bottom was significantly different (d.f = 8, F = 99.927, p<0.001), (d.f = 8, F = 62.247, p<0.001) and (d.f = 8, F = 9.595, p<0.014) among leaves, branches and mixed food, respectively (Table 1). Moreover, comparatively higher moisture was recorded from the comb constructed by leaves followed by mix food and branches. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. Comparison of the physicochemical indexes among fungus-combs built by different diets. The bar indicates mean±SD (n = 9), bar labeled by different letters of the same index are significantly different as determined by Duncan's new multiple range test (P < 0.05). https://doi.org/10.1371/journal.pone.0232329.g002 Download: PPT PowerPoint slide PNG larger image TIFF original image Table 1. Comparison of the physicochemical indexes at different decomposition stages of fungus-combs. https://doi.org/10.1371/journal.pone.0232329.t001 The bacterial community structure among different fungus-combs The whole comb of mix food, branches and leaves occupied 496, 760 and 266 OTUs, respectively, however, 721 OUTs is common among all (Fig 3A). In case of comb parts, the OTU cluster analysis clearly divided the bacterial community into two groups (Fig 3B). The groups are formed top and middle (group 1) and bottom of all combs (group 2), with bacterial community similarity. The bacterial communities showed similarities at top and middle comb but they were entirely different from bottom combs, irrespective of food (Fig 3B and 3C). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 3. Bacterial community profiles at different decomposition stages of fungus-combs. (A) Common bacteria OTU Venn diagram of fungus-combs built from different foods (OTUs whose abundance lower than 0.001% of total sequences were removed). (B) Community similarities based on weighted Unifrac metric; (C) The constrained corresponding analysis (CCA) of the bacterial community and variation partition analysis for three physicochemical indexes. Whereas TL, ML, BL, TB, MB, BB, TM, MM, and BM described in Table 1. The Denoising stats of OTUs are listed in S6 Table. https://doi.org/10.1371/journal.pone.0232329.g003 The microbial diversity indexes (Simpson [40] and Shannon [41]), estimated community richness (ACE [42] and Chao1 [43]) (Table 2) described that the average highest diversity was recorded from branches followed by mix food and leaves (Table 2). For comb parts, it increased in the middle and dropped in the bottom comb (Table 2). Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Bacterial diversity indexes at different decomposition stages of fungus-combs. https://doi.org/10.1371/journal.pone.0232329.t002 At the phyla level, the most prominent first five out of seven bacterial groups comprised the members of Bacteroidetes, Proteobacteria, Firmicutes, Spirochaetae, Actinobacteria, Planctomycetes, and Acidobacteria (Fig 4A, for details, see S1 Table) were recorded in the top and middle of all combs. However, the relative abundances of Bacteroidetes, Planctomycetes, and Spirochaetes were reduced or even eliminated at the bottom of the combs, while Proteobacteria, Acidobacteria, and Actinobacteria were increased. The Firmicutes remained stable in all stages of the combs. At the family level, the abundances between the top and middle were similar but they were different from the bottom combs. There were also strong differences in the relative abundances of lineages among all fungus-combs across ages and food types (Figs 4B and 5). For example, Rhodospirillaceae and Burkholderiaceae preferred the bottom comb of oak leaves (BL), while Comaonadaceae preferred the bottom comb of oak branches and mix food (BB and BM). LEfSe analysis of fungus-combs across different decomposition stages indicated that specific bacterial groups may contribute to specific decomposition stage of the comb (S1 and S2 Figs). For instance, Alistipes was the most discriminatory genus in fresh combs, while Burkholderia-Paraburkholderia was the most discriminatory genus in mature combs. Meanwhile, a detailed taxonomic classification is attached to supplementary material (S2 Table). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 4. The bacterial community structure at different decomposition stages of fungus-combs. (A) The bacterial community structure at the phyla level. (B) The community composition at the family level. https://doi.org/10.1371/journal.pone.0232329.g004 Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 5. Heatmap of selected bacterial lineages at different decomposition stages of fungus-combs. There are four groups of bacterial lineages: A, enriched (RA > 5%) in fresh comb of three food treatments; B, enriched (RA > 5%) in mature comb of three food treatments; C, showed preference in fresh comb of specific food treatment(s); D, showed preference in mature comb of specific food treatment(s). T, M, and B described in Table 1. https://doi.org/10.1371/journal.pone.0232329.g005 Effect of physicochemical indexes on bacterial community structure To link the bacterial community with physicochemical indexes in combs, the constrained corresponding analysis (CCA) was performed (Fig 3C). The length of the lines with arrows represent the contribution of the corresponding factors to the variation in the bacterial community structure. The moisture content contributed the most followed by the organic matter and ash content. In case of diets, there was a positive interaction between bacterial diversity and organic matter (Tables 1 and 2, Fig 3C). Contrary to this, moisture and ash contents have a negative relation with the bacterial community. For example, moisture and ash contents were highest in leaves comb followed by mix food and branches. Opposite to this, organic matter and bacterial community were higher in branches followed by mix food and leaves. For comb stages, the organic matter had a strong positive interaction with bacterial diversity followed by moisture and ash contents. For example, among all combs, we generally recorded the highest bacterial diversity in the middle combs where organic matter was also high, but moisture and ash contents were not too high as compared to the bottom. On the basis of these results, we suggested that too high moisture and ash content, and low organic matter negatively affect the bacteria diversity. Effect of diets and fungus-comb ages on the physicochemical indexes The physicochemical indexes including moisture, ash content and organic matter from the fungus-combs built by leaves, branches, and mix food respectively, are summarized in Table 1 and Fig 2. The moisture content at different decomposition stages of the combs i.e., top, middle and bottom was significantly different (d.f = 8, F = 99.927, p<0.001), (d.f = 8, F = 62.247, p<0.001) and (d.f = 8, F = 9.595, p<0.014) among leaves, branches and mixed food, respectively (Table 1). Moreover, comparatively higher moisture was recorded from the comb constructed by leaves followed by mix food and branches. Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. Comparison of the physicochemical indexes among fungus-combs built by different diets. The bar indicates mean±SD (n = 9), bar labeled by different letters of the same index are significantly different as determined by Duncan's new multiple range test (P < 0.05). https://doi.org/10.1371/journal.pone.0232329.g002 Download: PPT PowerPoint slide PNG larger image TIFF original image Table 1. Comparison of the physicochemical indexes at different decomposition stages of fungus-combs. https://doi.org/10.1371/journal.pone.0232329.t001 The bacterial community structure among different fungus-combs The whole comb of mix food, branches and leaves occupied 496, 760 and 266 OTUs, respectively, however, 721 OUTs is common among all (Fig 3A). In case of comb parts, the OTU cluster analysis clearly divided the bacterial community into two groups (Fig 3B). The groups are formed top and middle (group 1) and bottom of all combs (group 2), with bacterial community similarity. The bacterial communities showed similarities at top and middle comb but they were entirely different from bottom combs, irrespective of food (Fig 3B and 3C). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 3. Bacterial community profiles at different decomposition stages of fungus-combs. (A) Common bacteria OTU Venn diagram of fungus-combs built from different foods (OTUs whose abundance lower than 0.001% of total sequences were removed). (B) Community similarities based on weighted Unifrac metric; (C) The constrained corresponding analysis (CCA) of the bacterial community and variation partition analysis for three physicochemical indexes. Whereas TL, ML, BL, TB, MB, BB, TM, MM, and BM described in Table 1. The Denoising stats of OTUs are listed in S6 Table. https://doi.org/10.1371/journal.pone.0232329.g003 The microbial diversity indexes (Simpson [40] and Shannon [41]), estimated community richness (ACE [42] and Chao1 [43]) (Table 2) described that the average highest diversity was recorded from branches followed by mix food and leaves (Table 2). For comb parts, it increased in the middle and dropped in the bottom comb (Table 2). Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Bacterial diversity indexes at different decomposition stages of fungus-combs. https://doi.org/10.1371/journal.pone.0232329.t002 At the phyla level, the most prominent first five out of seven bacterial groups comprised the members of Bacteroidetes, Proteobacteria, Firmicutes, Spirochaetae, Actinobacteria, Planctomycetes, and Acidobacteria (Fig 4A, for details, see S1 Table) were recorded in the top and middle of all combs. However, the relative abundances of Bacteroidetes, Planctomycetes, and Spirochaetes were reduced or even eliminated at the bottom of the combs, while Proteobacteria, Acidobacteria, and Actinobacteria were increased. The Firmicutes remained stable in all stages of the combs. At the family level, the abundances between the top and middle were similar but they were different from the bottom combs. There were also strong differences in the relative abundances of lineages among all fungus-combs across ages and food types (Figs 4B and 5). For example, Rhodospirillaceae and Burkholderiaceae preferred the bottom comb of oak leaves (BL), while Comaonadaceae preferred the bottom comb of oak branches and mix food (BB and BM). LEfSe analysis of fungus-combs across different decomposition stages indicated that specific bacterial groups may contribute to specific decomposition stage of the comb (S1 and S2 Figs). For instance, Alistipes was the most discriminatory genus in fresh combs, while Burkholderia-Paraburkholderia was the most discriminatory genus in mature combs. Meanwhile, a detailed taxonomic classification is attached to supplementary material (S2 Table). Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 4. The bacterial community structure at different decomposition stages of fungus-combs. (A) The bacterial community structure at the phyla level. (B) The community composition at the family level. https://doi.org/10.1371/journal.pone.0232329.g004 Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 5. Heatmap of selected bacterial lineages at different decomposition stages of fungus-combs. There are four groups of bacterial lineages: A, enriched (RA > 5%) in fresh comb of three food treatments; B, enriched (RA > 5%) in mature comb of three food treatments; C, showed preference in fresh comb of specific food treatment(s); D, showed preference in mature comb of specific food treatment(s). T, M, and B described in Table 1. https://doi.org/10.1371/journal.pone.0232329.g005 Effect of physicochemical indexes on bacterial community structure To link the bacterial community with physicochemical indexes in combs, the constrained corresponding analysis (CCA) was performed (Fig 3C). The length of the lines with arrows represent the contribution of the corresponding factors to the variation in the bacterial community structure. The moisture content contributed the most followed by the organic matter and ash content. In case of diets, there was a positive interaction between bacterial diversity and organic matter (Tables 1 and 2, Fig 3C). Contrary to this, moisture and ash contents have a negative relation with the bacterial community. For example, moisture and ash contents were highest in leaves comb followed by mix food and branches. Opposite to this, organic matter and bacterial community were higher in branches followed by mix food and leaves. For comb stages, the organic matter had a strong positive interaction with bacterial diversity followed by moisture and ash contents. For example, among all combs, we generally recorded the highest bacterial diversity in the middle combs where organic matter was also high, but moisture and ash contents were not too high as compared to the bottom. On the basis of these results, we suggested that too high moisture and ash content, and low organic matter negatively affect the bacteria diversity. Discussion Bacteria and the ectosymbiotic fungus (Termitomyces) share the habitat, fungus-comb [14–16]. Termitomyces produce various enzymes to degrade the plant material [10, 17, 19], while bacteria may contribute to decompose the biomass of fungus-comb in different pattern. Similar to termites, the leaf-cutting ants maintained fungus garden in their nests [44]. The biologists suggested that the functions of bacteria in the fungus garden of leaf-cutting ants are lignocellulose degradation, nitrogen fixation, nutrient supplementation, and antibiotic production [45–49], but whether these bacteria play similar roles in fungus-comb of termite, remains unclear. We present the detailed comparison of bacterial communities and physicochemical indexes among different decomposition stages in fungus-combs of termite (O. formosanus). We found that communities were similar among all combs but there were marked shifts in bacterial composition across comb ages over time. Moreover, the compositions of physicochemical indexes were different among all combs, as well as among all comb stages. Besides, bacterial diversity had a strong relationship with physicochemical indexes. Below, we discuss the consequences of these findings. The shift of dominant bacterial groups across the age-groups of combs The dominant phyla and families were very similar at the same decomposition stages of all combs (S1 Table and Figs 4 and 5), may have a basic function of maintaining fungus-comb. To assign the role of bacteria in comb, we assumed that bacteria lineages enriched in the top comb may be shifted from termite gut and have complementary or similar roles in the comb as well as in the gut [15], while the community composition in the bottom comb may have a significant role only in a comb. The dominant phyla in fungus-comb (see for details, S1 Table) are inconsistent with the previous studies on fungus cultivating termite combs and guts [15, 16, 50]. These phyla are important for carbohydrate metabolism, reductive acetogenesis, fungal cell wall degradation [10, 51, 52] and cello-oligomer degradation [53], which may help termite to get protein and fungal enzymes to cleave lignin [10]. As shown in S4 Table and Fig 5, the genera which were dominant in the “comb-age specific” bacterial families (Group A and B) and in “food-type specific” bacterial families (Group C and D) showed similar preferences with the corresponsive families for the same combs. For example, Rikenellaceae and its dominant genus, Alistipes both preferred the top combs. According to LEfSe analysis (S1 and S2 Figs), these genera were included all discriminatory genera across different aged combs by, except Nitrospirillum of Rhodospirillaceae which was not a discriminatory genus and it was only dominant in the bottom comb of leaf diet. Porphyromonadaceae and Rikenellaceae are the most dominant families of Bacteroidetes in top comb, metabolize carbohydrates and produce organic acids like acetate, succinate, and butyrate [54, 55], to create an acidic environment for Termitomyces growth and better utilization of cellulose. Alistipes and BCf9-17, the members of Rikenellaceae, were the discriminatory genera in the top combs. These are gut‑specific bacterial lineages in fungus-cultivating termite [38, 50]. Furthermore, Alistipes has a positive correlation with protein diets [38, 50, 56], whereas, the functions of BCf9-17 are poorly understood. Spirochaetae and Planctomycetes were abundant both in top and middle comb, but they were scarce in bottom comb (S1 Table and Fig 4A). The traditional Planctomycetes are chemoorganoheterotrophs [57]. The discriminatory genus in top combs is an uncultured bacterium belong to an uncultivated group, vadinHA49 (S1 and S2 Figs), enriched in fungus-cultivating termite gut [58]. The specific functions of these microbes are out of the box. Treponema belonging to Spirochaetae was the discriminatory genus in the top combs (S2 Table). These bacteria are important for acetogenesis from H2 plus CO2 and nitrogen fixation in fungus-comb, however, it has been suggested that the abundance of nitrogen fixation bacteria is high in termite gut [59–64]. The potential of nitrogen fixation in fungus-comb requires further detailed investigation. Proteobacteria and Actinobacteria were enriched in bottom comb, which is comparable with the previous works on fungus cultivating termite combs and guts [15, 65]. It has been suggested that Proteobacteria might assist in detoxification or food digestion [66–70], while Actinobacteria play a key role of defense through inhibiting the invasive fungus, Pseudoxylaria in fungus-cultivating termites [71]. The enrichment of Actinobacteria in mature comb might be from the surrounding soil by termites [65]. But, it is yet to be investigated that how are these bacteria regulated by termites? The genera Nitrobacter, Ochrobactrum, and Burkholderia-Paraburkholderia are the dominant members of Bradyrhizobiaceae, Brucellaceae, and Burkholderiaceae, respectively, belonging to Proteobacteria phyla, were discriminatory in bottom comb. It has been suggested previously that Nitrobacter are nitrite-oxidizing bacteria [72, 73]. Ochrobactrum microbes have lignolytic potential as well as these bacteria, also play a significant role in nitrite-oxidization [74–77]. Burkholderia-Paraburkholderia bacteria are capable to degrade aromatics [78]. Preference of bacterial lineages to different diets Bacterial lineages across different diets were compared at the family level (Fig 5). Rikenellaceae (Bacteroidetes) were high in the comb of leaves followed by branches and mix food combs. It has been suggested that Rikenellaceae metabolize carbohydrate actively [79]. BCf9-17 genera of Rikenellaceae was abundantly higher in top comb of leaves compared to top combs of branches and mix food i.e., 10.40%, 0.70%, and 1.60% respectively (S4 Table). This genus is termite-specific, but its functions are poorly studied [65]. The specific family of Firmicutes showed its preference to specific diet. For example, Ruminococcaceae and Lachnospiraceae are important for cellulose degradation [80, 81], showed their preferences to different diets. Ruminococcaceae were relatively lower in leaves comb as compared to branches and mix food. While Lachnospiraceae were high in leaves and mix food comb. This is might be due to that Ruminococcaceae mainly degrade lignin rich in branch and mix food diets, while Lachnospiraceae utilize cellulose more accessible in leaves. Moreover, Peptostreptococcaceae of Firmicutes has potential to degrade hydrocarbons [82, 83], and these bacteria showed less preference to the comb of leaves (Fig 5). This is consistent with Schauer et al. [84] who described that a high-fiber diet caused shifts in the diversity of dominant gut bacteria in cockroach. Furthermore, the discriminatory genus in Peptostreptococcaceae, Peptoclostridium, has the ability to breakdown and utilize plant fibers [85, 86]. In bottom combs, the families of Proteobacteria also presented different preferences to different diets. Burkholderiaceae and Rhodospirillaceae were in favor of leaves, while Bradyrhizobiaceae, Brucellaceae, and Comamonadaceae preferred branches and mix food (Fig 5). Considering the capacity of Burkholderiaceae to degrade aromatic compounds [25, 87–90], their enrichment in leaves comb might deal with a higher content of aromatic compounds [61, 91, 92]. Burkholderia-Paraburkholderia, was the most enriched genus of Burkholderiaceae, has aromatics-degrading potential as we mentioned before [78]. Species of Rhodospirillaceae diverse in saline environment [93–95], may have the potential of halotolerance. The Nitrospirillum was contributed almost all abundance of Rhodospirillaceae and has a vital role in nitrification [96]. Recent research suggested that Comamonadaceae in forest soils can degrade lignin substrate [97]. Delftia was the most abundant genus of Comamonadaceae, and it has the capacity to degrade lignin-derived aromatic compounds [98]. Brucellaceae and Bradyrhizobiaceae are most important for nitrogen-fixation [99, 100]. Ochrobactrum and Nitrobacter were the most enriched genera of these families respectively, and their nitrite-oxidizing potential has been discussed before. The relationship of bacterial community and physicochemical indexes in the comb The moisture content contributed the most to the variation in the bacterial community structure, followed by the organic matter and the ash content (Fig 3C). We recommend that specific amount of physicochemical indexes are required for better growth of bacterial diversity but excess moisture and ash contents, and low organic matter may harm their diversity (Fig 2 and Table 2). These findings are in the agreement with previous studies in soil [101–103]. Excess ash contents indicated the high salinity which can limit water availability [104] and it can also be toxic to metabolic activities [105]. Moisture can affect microbial activity by dissolving substrates [106]. Brockett et al. [107] reported that soil moisture and the potential lignocellulose-degrading activities of enzymes have a negative correlation. Thus, excess moisture could reduce the activities of the enzymes and lead to low nutrition availability in the comb. Soil organic matter content is often considered as a dominating factor of soil bacterial diversity [108, 109] because most soil micro-organisms rely on organic matter decomposition to obtain energy [110, 111]. Similarly, the effect of organic matter in comb can be explained by the availability of energy that most micro-organisms may obtain energy from organic matter in comb, and the low organic matter directly restricts the bacterial diversity and abundance. The shift of dominant bacterial groups across the age-groups of combs The dominant phyla and families were very similar at the same decomposition stages of all combs (S1 Table and Figs 4 and 5), may have a basic function of maintaining fungus-comb. To assign the role of bacteria in comb, we assumed that bacteria lineages enriched in the top comb may be shifted from termite gut and have complementary or similar roles in the comb as well as in the gut [15], while the community composition in the bottom comb may have a significant role only in a comb. The dominant phyla in fungus-comb (see for details, S1 Table) are inconsistent with the previous studies on fungus cultivating termite combs and guts [15, 16, 50]. These phyla are important for carbohydrate metabolism, reductive acetogenesis, fungal cell wall degradation [10, 51, 52] and cello-oligomer degradation [53], which may help termite to get protein and fungal enzymes to cleave lignin [10]. As shown in S4 Table and Fig 5, the genera which were dominant in the “comb-age specific” bacterial families (Group A and B) and in “food-type specific” bacterial families (Group C and D) showed similar preferences with the corresponsive families for the same combs. For example, Rikenellaceae and its dominant genus, Alistipes both preferred the top combs. According to LEfSe analysis (S1 and S2 Figs), these genera were included all discriminatory genera across different aged combs by, except Nitrospirillum of Rhodospirillaceae which was not a discriminatory genus and it was only dominant in the bottom comb of leaf diet. Porphyromonadaceae and Rikenellaceae are the most dominant families of Bacteroidetes in top comb, metabolize carbohydrates and produce organic acids like acetate, succinate, and butyrate [54, 55], to create an acidic environment for Termitomyces growth and better utilization of cellulose. Alistipes and BCf9-17, the members of Rikenellaceae, were the discriminatory genera in the top combs. These are gut‑specific bacterial lineages in fungus-cultivating termite [38, 50]. Furthermore, Alistipes has a positive correlation with protein diets [38, 50, 56], whereas, the functions of BCf9-17 are poorly understood. Spirochaetae and Planctomycetes were abundant both in top and middle comb, but they were scarce in bottom comb (S1 Table and Fig 4A). The traditional Planctomycetes are chemoorganoheterotrophs [57]. The discriminatory genus in top combs is an uncultured bacterium belong to an uncultivated group, vadinHA49 (S1 and S2 Figs), enriched in fungus-cultivating termite gut [58]. The specific functions of these microbes are out of the box. Treponema belonging to Spirochaetae was the discriminatory genus in the top combs (S2 Table). These bacteria are important for acetogenesis from H2 plus CO2 and nitrogen fixation in fungus-comb, however, it has been suggested that the abundance of nitrogen fixation bacteria is high in termite gut [59–64]. The potential of nitrogen fixation in fungus-comb requires further detailed investigation. Proteobacteria and Actinobacteria were enriched in bottom comb, which is comparable with the previous works on fungus cultivating termite combs and guts [15, 65]. It has been suggested that Proteobacteria might assist in detoxification or food digestion [66–70], while Actinobacteria play a key role of defense through inhibiting the invasive fungus, Pseudoxylaria in fungus-cultivating termites [71]. The enrichment of Actinobacteria in mature comb might be from the surrounding soil by termites [65]. But, it is yet to be investigated that how are these bacteria regulated by termites? The genera Nitrobacter, Ochrobactrum, and Burkholderia-Paraburkholderia are the dominant members of Bradyrhizobiaceae, Brucellaceae, and Burkholderiaceae, respectively, belonging to Proteobacteria phyla, were discriminatory in bottom comb. It has been suggested previously that Nitrobacter are nitrite-oxidizing bacteria [72, 73]. Ochrobactrum microbes have lignolytic potential as well as these bacteria, also play a significant role in nitrite-oxidization [74–77]. Burkholderia-Paraburkholderia bacteria are capable to degrade aromatics [78]. Preference of bacterial lineages to different diets Bacterial lineages across different diets were compared at the family level (Fig 5). Rikenellaceae (Bacteroidetes) were high in the comb of leaves followed by branches and mix food combs. It has been suggested that Rikenellaceae metabolize carbohydrate actively [79]. BCf9-17 genera of Rikenellaceae was abundantly higher in top comb of leaves compared to top combs of branches and mix food i.e., 10.40%, 0.70%, and 1.60% respectively (S4 Table). This genus is termite-specific, but its functions are poorly studied [65]. The specific family of Firmicutes showed its preference to specific diet. For example, Ruminococcaceae and Lachnospiraceae are important for cellulose degradation [80, 81], showed their preferences to different diets. Ruminococcaceae were relatively lower in leaves comb as compared to branches and mix food. While Lachnospiraceae were high in leaves and mix food comb. This is might be due to that Ruminococcaceae mainly degrade lignin rich in branch and mix food diets, while Lachnospiraceae utilize cellulose more accessible in leaves. Moreover, Peptostreptococcaceae of Firmicutes has potential to degrade hydrocarbons [82, 83], and these bacteria showed less preference to the comb of leaves (Fig 5). This is consistent with Schauer et al. [84] who described that a high-fiber diet caused shifts in the diversity of dominant gut bacteria in cockroach. Furthermore, the discriminatory genus in Peptostreptococcaceae, Peptoclostridium, has the ability to breakdown and utilize plant fibers [85, 86]. In bottom combs, the families of Proteobacteria also presented different preferences to different diets. Burkholderiaceae and Rhodospirillaceae were in favor of leaves, while Bradyrhizobiaceae, Brucellaceae, and Comamonadaceae preferred branches and mix food (Fig 5). Considering the capacity of Burkholderiaceae to degrade aromatic compounds [25, 87–90], their enrichment in leaves comb might deal with a higher content of aromatic compounds [61, 91, 92]. Burkholderia-Paraburkholderia, was the most enriched genus of Burkholderiaceae, has aromatics-degrading potential as we mentioned before [78]. Species of Rhodospirillaceae diverse in saline environment [93–95], may have the potential of halotolerance. The Nitrospirillum was contributed almost all abundance of Rhodospirillaceae and has a vital role in nitrification [96]. Recent research suggested that Comamonadaceae in forest soils can degrade lignin substrate [97]. Delftia was the most abundant genus of Comamonadaceae, and it has the capacity to degrade lignin-derived aromatic compounds [98]. Brucellaceae and Bradyrhizobiaceae are most important for nitrogen-fixation [99, 100]. Ochrobactrum and Nitrobacter were the most enriched genera of these families respectively, and their nitrite-oxidizing potential has been discussed before. The relationship of bacterial community and physicochemical indexes in the comb The moisture content contributed the most to the variation in the bacterial community structure, followed by the organic matter and the ash content (Fig 3C). We recommend that specific amount of physicochemical indexes are required for better growth of bacterial diversity but excess moisture and ash contents, and low organic matter may harm their diversity (Fig 2 and Table 2). These findings are in the agreement with previous studies in soil [101–103]. Excess ash contents indicated the high salinity which can limit water availability [104] and it can also be toxic to metabolic activities [105]. Moisture can affect microbial activity by dissolving substrates [106]. Brockett et al. [107] reported that soil moisture and the potential lignocellulose-degrading activities of enzymes have a negative correlation. Thus, excess moisture could reduce the activities of the enzymes and lead to low nutrition availability in the comb. Soil organic matter content is often considered as a dominating factor of soil bacterial diversity [108, 109] because most soil micro-organisms rely on organic matter decomposition to obtain energy [110, 111]. Similarly, the effect of organic matter in comb can be explained by the availability of energy that most micro-organisms may obtain energy from organic matter in comb, and the low organic matter directly restricts the bacterial diversity and abundance. Conclusion The current study gives a brief account of how bacterial community structure at different decomposition stages of fungus-comb respond to various lignocellulosic plant materials and physicochemical indexes. The shift of bacterial community across the fungus-comb over time indicated the adjustment of bacterial structure in the comb to deal with different kinds of diets (S3 Table, seeing the indicated functions). In addition, physicochemical indexes had been changed at different decomposition stages of comb and these also had a strong relationship with bacterial diversity. The extremely high moisture, low organic matter, and high ash content may reduce bacterial diversity. Furthermore, ruminants can only consume less lignified herbaceous plants such as soft grass [2]. Functional bacteria from the fungus-cultivating termites comb can be applied in the pretreatment of forestry and agricultural residues to improve their palatability and digestibility for ruminant animals. Supporting information S1 Table. Bacterial community structure at phyla-level across age-groups of different fungus-combs. https://doi.org/10.1371/journal.pone.0232329.s001 (XLSX) S2 Table. A detailed taxonomic classification of bacteria groups. https://doi.org/10.1371/journal.pone.0232329.s002 (XLSX) S3 Table. Indicated functions of predominant bacteria in fungus-combs. https://doi.org/10.1371/journal.pone.0232329.s003 (XLSX) S4 Table. Selected bacterial lineages at different decomposition stages of different fungus-combs. https://doi.org/10.1371/journal.pone.0232329.s004 (XLSX) S5 Table. NCBI accessions of samples. https://doi.org/10.1371/journal.pone.0232329.s005 (XLSX) S6 Table. Denoising stats of cluster. https://doi.org/10.1371/journal.pone.0232329.s006 (XLSX) S1 Fig. LDA scores observed for individual taxa that passed the LefSe significance threshold. Threshold on the logarithmic LDA score for discriminative features was 4.0. https://doi.org/10.1371/journal.pone.0232329.s007 (TIF) S2 Fig. Taxonomic cladogram of bacteria groups based on LDA scores. Only discriminant groups are presented. All datasets were submitted to the Sequence Read Archive of NCBI (http://www.ncbi.nlm.nih.gov; BioProject PRJNA604886). https://doi.org/10.1371/journal.pone.0232329.s008 (TIF) Acknowledgments We thank Zhengsen Chen for his assistance in collecting the termite colonies of O. formosanus and Mengyi Yang (Xiaoshan Institute of Termite Control) for the advice about rearing the colonies under lab conditions.
TI - Exploring the effect of plant substrates on bacterial community structure in termite fungus-combs
JF - PLoS ONE
DO - 10.1371/journal.pone.0232329
DA - 2020-05-01
UR - https://www.deepdyve.com/lp/public-library-of-science-plos-journal/exploring-the-effect-of-plant-substrates-on-bacterial-community-9Qarb2CCAK
SP - e0232329
VL - 15
IS - 5
DP - DeepDyve
ER -