Sociobiology

. Abstract Weaver ants ( Oecophylla smaragdina and O. longinoda ) are famous for their impressive nest constructions, where they weave together living leaves on their host plants. Also, they are known to protect tropical tree crops against insect pests and are thus being utilized for biological control. Nest numbers have been used to track weaver ant abundances in plantations to optimize their use, assuming the number of nests reflects ant numbers. In this study, we compared nest size, the density of ants, and the proportion of workers in nests between three host plants (cashew, citrus, and mango) and between two seasons (rainy and dry). Nest size, ant density, and worker proportion differed between host species, whereas only nest size and worker proportion showed seasonal differences and only on mango and citrus, respectively. On hosts with smaller nests (in increasing size order: cashew, citrus mango), ant densities in the nests were higher, and the proportion of workers in the nests seemed to increase during seasons where food production on host trees decreased or on hosts where more trophobionts were cultured inside nests. These results have implications for developing methods to estimate ant numbers and raise interesting questions about the function of weaver ant nests.


Introduction
Termites are typical social insects with various castes in their community, most of which are workers.The history of termites can be traced back to 250 million years ago, and over 3000 species are distributed worldwide.According to statistics, subterranean termite damage and control costs are estimated to approach $2 billion annually in the United States alone.During termite control, people have found that in order to achieve control of termite populations.The role of intestinal microorganisms in insect immunity has gradually become clear.
Many microbes live in the gut of insects and play many vital roles in their behavior, metabolism, diet, nutrition, and immunity .The vital activities of insects can only be discussed by considering intestinal symbiosis.Because of the dietary preferences, the composition of the intestinal symbiotes also differs.Due to the different composition of intestinal microorganisms, termites are divided into non-Termitidae termites and Termitidae, formerly called lower termites and higher termites.However, most research regarding termites and intestinal symbionts has focused on how the latter helps the former degrade lignocellulose quickly and effectively .Some intestinal microorganisms have been reported to participate in immunity among other insects.The pathogenic fungus Beauveria bassiana Vuillemin interacts with the gut microbiota Serratia marcescens Bizio to accelerate mosquito death.In addition, Serratia can secrete an enzyme to help mosquitoes fight malaria parasites.Intestinal flora directly mediated the Toll pathway to the anti-dengue virus.
Recently, the meta-transcriptome and metagenome have been used separately or both to investigate the role of uncultivated microorganisms in the gut or other organs of many species, including invertebrates and vertebrates .Research on the termite meta-transcriptome revealed the role of intestinal symbiotes and the expression of the host genes in cellulosic digestion or innate immunity in non-Termitidae termites.However, there are few reports about Termitidae immunity and intestinal flora.
In the current research, we infected the funguscultivated Termitidae Odontotermes formosanus Shiraki using entomopathogen Metarhizium robertsii St Leger & Roberts to resolve the potential function of intestinal bacteria in innate immune by metagenome and meta-transcriptome analysis.We reveal that intestinal flora potentially promotes GSH synthesis by upregulating gshA expression, thereby assisting termites in responding to entomopathogen infection.This finding holds significant theoretical value for deciphering termite immune response mechanisms and provides potential guidance for termite management and control strategies.

Termite collection and rearing
Workers of O. formosanus were chosen as the experimental subjects and gathered in Lion Mountain, the Huazhong Agricultural University, Hongshan District, Wuhan City, Hubei Province.The termites were raised in a Petri dish using 9 cm filter paper.The culture media was kept in a constant-temperature incubator at 25 °C and 85% humidity in the dark to preserve community stability.The filter paper was changed every two days.

Culture of fungal pathogen
The Shanghai Institute of Botany provided the M. robertsii (strain #2575) used in this experiment.M. robertsii spores were grown for 2-4 weeks on potato dextrose agar (PDA) and mixed with 0.1% Tween 80 to create a conidial solution stored at four °C.The spores were diluted to the desired concentration before the survival experiment to determine the time and concentration of infection.
Each infection and control group contained three biological replicates, each containing thirty termites in good health.The two treatment groups were in the dark for 48 hours in Petri dishes with a 9-cm diameter spread with filter paper and sprayed with either 1 mL of a 1 × 10 8 /mL spore suspension or 1 mL of sterile water.After treatment, the filter paper was removed, and the termite samples were dissected.

Termite gut dissection
Before dissection, the termites and forceps were cleaned with 75% ethanol to remove bacteria.1X PBS was used when dissecting the samples.The hindguts of termites were chilled in an ice bath.The termite guts were put into 1.5 mL Eppendorf tubes and stored at -80 °C to prevent RNA degradation.

Metagenome library preparation, sequencing, assembly, and annotation
Three biological samples were combined and sequenced.DNA was extracted using The E.Z.N.A.® Stool DNA Kit (Omega Bio-tek Inc., Norcross, USA) according to the manufacturer's recommendations for bacterial DNA extraction, which excluded lysis stages and was stored at -20 °C for future analysis.Using the Covaris M200 and the NEXTFLEX® Rapid DNA-Seq Kit, crude DNA was sheared at a length of 400 bp.The library was prepared with NovaSeq Reagent Kits (Illumina, San Diego, USA) and sequenced with paired 150bp reads.
Reads of the two libraries, including the control and infected group, were obtained from the sequencer.Low-quality reads were removed using Fastp v0.12.4.Contamination of the host genome (Coptotermes formosanus, Taxonomy ID: 36987) was also removed using KneadData (https://github.com/biobakery/kneaddata) with the default parameters.Cleaned reads were assembled by Megahit v1.2.9 with the default Kmer value, and assembly effectiveness was assessed with Quast 5.0.2.Prodigal v2.6.3 was used for ORF prediction with the option "-m -p meta" for predicting bacteria genes with no gaps.The Prodigal (-meta) results were processed using cd-hit v4.8.1 to eliminate the predicted gene redundancy.To obtain a solid taxonomic classification of species, reads were classified at multiple levels -phylum, class, order, family, genus, and species -by Kraken2 and Bracken [Citing: Kraken 2 (31779668), Bracken (36171387)].Additionally, Contigs from different phyla were retrieved using TBtools for further species annotation.Diamond aligned predicted genes to the eggnog 5.0 bacteria database under the e-value of 1e-5 while annotating the alignment result using the COG, KO, and CAZy databases.

Meta-transcriptome library preparation, sequencing, assembly, and annotation
Trizol was used to extract the total RNA.The Ribo-Zero rRNA Removal Kit was used to remove the rRNA from the total RNA (bacteria).Utilizing the TruSeqTM RNA Sample Prep Kit, cDNA was generated and used to create the meta-transcriptome libraries (Illumina, San Diego, USA).The NovaSeq Reagent Kit was used to sequence RNA fragments (Illumina, San Diego, USA).The library was prepared with NovaSeq Reagent Kit (Illumina, San Diego, USA) and sequenced with 150 bp paired reads.
The SAMSA2 pipeline was used to perform the conducted meta-transcriptome analysis.Quality control was undertaken using the default parameters of Trimmomatic v0.39 and FastQC v0.11.9.Cleaned reads were merged and used by PEAR to produce raw counts.Merged reads were then exported to SortMeRNA v2.1.4to remove ribosomal reads.MEGAHIT was used to assemble the cleaned reads, and the contig quality was evaluated using Quast 5.0.2.Kraken2 was used to annotate the contigs, and bracken was used for hierarchical categorization.For functional annotation, contigs from several phyla were retrieved using TBtools.For ORF prediction, Prodigal v2.6.3 was used with the option "-m -p meta" to predict genes without gaps.Additionally, the Prodigal (-meta) result was used to remove redundant genes and to convert the RNA to protein sequences by cd-hit (Citing: cdhit: 23060610).The protein sequence was annotated using KOBAS 3.0 software.

Glutathione content measurement
In order to demonstrate the positive role of intestinal bacteria in glutathione synthesis in the hindgut of termites, we first treated termites with 5% kanamycin (dissolve 1 g of kanamycin powder in 20 mL of distilled water) for 48 hours and then with M. robertsii for 48 hours.Each group consisted of 10 termites with three biological replicates.The posterior gut of the termite is dissected for GSH content detection.Refer to the instructions of the glutathione detection kit (Nanjing Jiancheng Bioengineering Institute, China) for the detection method.

Metarhizium robertsii infection effectively increases termite mortality level
Figure 1b shows a typical O. formosanus termite.The termites were infected with various concentrations of M. robertsii, and the termite survival rates were calculated at various time points, as per our prior study, to calculate the mortality rate.The termite survival at different spore concentrations is shown in Figure 1d.Before 48 hours, the mortality level in the four groups was higher than 85% but not significantly so.The survival rate of termites treated with suspended spores at 1×10 8 /mL decreased sharply at 48 hours and halved after 72 hours.
Additionally, it differs significantly from the CT group.Based on the aforementioned findings, the 48-hour time point was chosen as the best option for subsequent experiments.The growth of fungal spores on the termite surface and their transformation from white to green after two weeks of high-concentration M. robertsii infection demonstrated that infection was the cause of termite mortality (Fig 1c).

Summary of metagenome and meta-transcriptome data
The intestinal flora at the gene and transcription levels were analyzed, respectively, to clarify the variety of intestinal flora after infection (Table 1).Overall, there were 25,192,940 and 24,551,605 raw reads in the sequencing of metagenomes and 27,045,469 and 26,520,565 raw reads in the sequencing of meta-transcriptomes.The former number is controlled, and the latter is processed.In the two sequencing strategies, the total base of the raw data was 3804118840, 3682740750, 4083865819, and 4004605315.More than 98.20% of the raw data passed the filter, and the Q30 values of 93.66% demonstrated the high quality of the sequencing data.The meta-transcriptome was assembled independently in 313,422 and 290,833 contigs with 50.01%and 51.40% GC content in the control and infected treatment, respectively.The metagenome data were assembled separately in 291,977 and 282,923 contigs with 49.36% and 50.82%GC content in the control and infected treatment, respectively.N50 was 1475, 1396, 1795, and 1710 in the four libraries.Following contig assembly, most sequences were between 300 and 600 bp in length (Fig S1).As mentioned earlier, the findings showed that the constructed contigs were trustworthy and high-quality.

The change range of high-abundance flora under the infection of M. robertsii is smaller than that of low-abundance flora
After proving the reliability of the data, the species of the control group and the experimental group were annotated in the metagenome and meta-transcriptome data and then compared at three taxonomic levels (Fig 2).
In the metagenome, either before or after the infection of M. robertsii, the most dominant bacteria are Proteobacteria, Firmicutes, Bacteroides, and Actinobacteria (Fig 2a and 2b).After the infection, the relative abundance of the phyla except the Firmicutes increased slightly, while the Firmicutes decreased to a certain extent.The dominant flora in the metatranscriptome has analogous changes.Proteobacteria, after the infection of pathogenic fungi, the relative abundance in the gut has increased by nearly 0.5-fold.However, the relative abundance of other phylum has decreased to different degrees except for Proteobacteria, among which the relative abundance of Spirochaetes has decreased by nearly 70%, and Actinobacteria has also decreased by 60% (Fig 2b).
When the level of annotation is Genus or species, the bacteria existing in the meta-genome become low abundance in the meta-transcriptome, including Escherichia coli, Citrobacter, and Lactococcus (Fig 2c).There is no significant abundance in the meta-genome, but there is a relatively significant increase in the meta-transcriptome of bacteria, including Kosakonia, Enterobacter, Streptococcus, and Bacteroides.However, the activity of Streptomyces, Clostridium, and Bacillus decreased after the invasion of M. robertsii.The trend at the species level is similar to that at the genus level (Fig 2d).

The function of Spirochetes changed significantly after infection by M. robertsii
The abundance of intestinal microflora has changed significantly after the infestation in the previous section, but its function in the immune response is still unknown.Next, we split the sequence in the meta-transcriptome at the phylum level and analyzed the functional changes of the top five dominant bacteria in the termite gut before and after the infection of M. robertsii to termites in detail.
Figure 3 shows that the functional changes of Proteobacteria and Firmicutes before and after the infection were minimal or even constant.In contrast, the functional changes of Bacteroides, Actinobacteria, and Spirochaetes were more significant in termite resistance to M. robertsii.First, after termites were infected by M. robertsii, the transcriptional activity of Bacteroides in amino acid metabolism increased by about 8%.However, the transcriptional activity of the metabolism-related protein family decreased by about 5%, which indicates that the infection of M. robertsii promoted the amino acid metabolism of this phylum but led to a decline in the transcription of the metabolism-related protein.Secondly, Actinomycetes' carbohydrate and energy metabolism increased to a certain extent after the infection of M. robertsii, while the cofactor and vitamin metabolism decreased.There was no significant change in the functional items above the four phyla.
The transcription activity of spirochetes has changed significantly compared with the first four phyla.The infestation inhibited the genetic information process, energy, amino acid metabolism, folding, sorting, and degradation while significantly increasing the transcriptional activity of the genes, including signal and cellular processes, carbohydrate metabolism, and membrane transport processes.These functions were related to protein synthesis and secretion.

Bacteria have the potential to participate in termite resistance to infestation by upregulating the glutathione synthesis ratelimiting enzyme gene gshA
Figure 4 shows the significant differentially expressed genes, allowing us to evaluate and understand the potential function of bacteria in immunity at a deeper level.
The results of differentially expressed genes showed that many genes were significantly upregulated or downregulated after the infection of M. robertsii (Table 2, Fig 4a indicating that the gene with changed expression may participate in innate immunity.In order to evaluate the above speculation, we measured the GSH contents after M. robertsii and with or without 5% Kanamycin (Fig 4b).The result found that GSH contents in infestation were significantly higher than in infestation after antibiotic treatment, which means intestinal bacteria played positive roles in GSH synthesis.

Discussion
In this study, metagenome and meta-transcriptome approaches were combined and used to examine the role and mechanism of intestinal bacteria in termite host immunity.It was found that M. robertsii infestation significantly changed the gut flora and affected low-abundance flora.This effect is not only reflected in abundance but also affects functional gene transcription, the most obvious of which is Spirochete.The genes of microbial origin after infection reveal a possible mechanism that bacteria significantly upregulated the expression of glutamate-cysteine ligase and released it into the intestinal cavity to promote the synthesis of glutathione in the host or bacteria, thus combining with M. robertsii toxin to help the host resist M. robertsii infection.
It was discovered that intestinal invasion by foreign pathogens could be aided by the host immunity provided by the gut symbiont.The taxonomic analysis and functional annotation in these studies revealed that the abundance of symbiotic bacteria, genus, and expressed transcripts changed to accommodate the presence of metarhizium.In another study on mosquitoes, Serratia appeared to interact with invasive pathogens and hasten the death of the mosquitoes.In the present study, termite hindgut pathogens affected intestinal flora's diversity and composition.The relative abundance of Proteobacteria, Firmicutes, and Bacteroidetes was not transformed in metagenome level following infection, and low abundance phyla like Spirochaetes and Actinobacteria raised significantly (Fig 2a).It demonstrated that the imbalance of the microenvironment caused by M. robertsii infection is more conducive to the proliferation of non-dominant bacteria in the intestinal flora.At the same time, a different trend appeared in the meta-transcriptome level (Fig 2b).Combined with the above results, we found that the Actinobacteria and Bacteroides in the dominant bacteria group increased in the metagenome but decreased in the meta-transcriptome, indicating that although the proliferation led to the increase of its abundance, its activity was decreased.After the fungi infection, Proteobacteria increased in both omics, indicating that the invasion of M. robertsii promoted Proteobacteria's proliferation and metabolic activity.
Finally, the trend of change in Firmicutes is opposite to the former.Therefore, we infer that the bacteria with significantly enhanced metabolic activity may assist termites in resisting M. robertsii.Firmicutes and Bacteroidetes, on the other hand, decreased following pathogenic infection in this study.Enzymes that lyse the fungal cell wall appear mostly produced by Firmicutes and Bacteroidetes.Therefore, metarhizium infection decreased the content of these two bacteria, thereby reducing the threat of intestinal microorganisms to pathogens.Bacillus sp., which belongs to Firmicutes and is found in termites (Anacanthotermes), was discovered to produce a novel beta-1,4-glucanase that degrades fungal cell walls.Except for the above flora, the transcriptional abundance of Spirochetes decreased nearly three times after infection by M. robertsii.However, the function of change was the most abundant among all phyla (Fig 2 & 3).In our previous study, Spirochetes expressed the same trend during infestation from 12 to 72 hours.Spirochetes can undergo various metabolic processes, including acetate production, nitrogen fixation, and degradation of lignin phenols.Faced with intestinal imbalance, the ability to process cellulose of Spirochetes raised over one-fold and their advantages have been replaced by other antimicrobial tolerance gates with similar metabolic abilities, which are generally believed to enhance the resilience of microbial communities in the face of stress.
Additionally, Peterson et al. discovered that the bacteria in the termite hindgut potentially upregulate an amidohydrolase two gene, working in conjunction with the other symbiont and host endogenous antifungal enzymes to combat invaders during B. bassiana infection.Also, the potentially antifungal gene bacteria-derived gshA was upregulated 3.8-fold in termite hindgut after M. robertsii infestation 48 hours.Glutamate-cysteine ligase is the rate-limiting enzyme in glutathione biosynthesis and occurs in prokaryotes and eukaryotes.It positively affects the synthesis of glutathione and is inhibited by glutathione feedback.In many lower organisms, the GCL enzyme is a single polypeptide, but most eukaryotic GCL enzymes are heterodimeric complexes consisting of two distinct gene products.GSH and glutathione were well known for their protective abilities against the detrimental effects of oxidative stress within the human body and protection against infection by exogenous microbial organisms.When insects encounter a pathogen, organisms release many reactive oxygens (ROS) to kill them directly, including hydrogen peroxide, hydroxyl radical, superoxide anion, and other oxygen-free radicals.
On the one hand, glutathione can combine with the toxin released by the pathogen or insecticide and then be eliminated from the body.On the other hand, it can combine with reactive oxygen species to prevent the body from producing too many reactive oxygen species to damage its cell membrane.In addition, studies have shown that GSH can promote the participation of macrophages in cellular immunity.The GSH contents of the gut infestation group were found to be significantly higher than the group infestation with 5% kanamycin (Fig 4b).
Although bacteria in the hindgut potentially release GCL protein into epithelial cells to promote GSH total or directly release GSH into the gut lumen to participate immune, the exact bacteria species were not found in this study.Future research should concentrate on finding specific bacteria involved in GCL expression when termites are infected by M. robertsii, as well as investigating whether this expression may affect the production of termite antifungal peptides.

Conclusion
This study investigated changes in the quantity, species, and gene expression at the transcriptional level of termite gut microorganisms that encountered foreign pathogens in the host and entered the intestine.Infection with M. robertsii significantly changes the intestinal flora of the hindgut of termites, with a more significant impact mainly on the low-abundance flora.This effect is reflected in changes in abundance and the transcription of functional genes, with the most apparent impact being on genes belonging to the Spirochetes bacteria.Therefore, we have established a possible mechanism that bacteria significantly upregulate the expression of glutamate-cysteine ligase and release it into the intestinal cavity to promote the synthesis of glutathione in the host or bacteria, thereby binding to the toxin of M. robertsii to help the host resist infection.

Fig 1 .
Fig 1.The background of this experiment.(a): Normal termite Odontotermes formosanus in the filter paper, captured with iPhone 11. (b): Two weeks later, the termite died of fungal infection.(c): Survival of termites (n = 30) fed different spore concentrations per mL (AT).(CT termites were fed 1 mL sterile water.The experiments were performed in three biological replicates.1x10^8/mL spore suspension significantly increased termite death compared with other groups at 72 h [log-rank (Mantel-Cox) test, p < 0.0001].The experiments were performed in three biological replicates.The logrank test was used to assess the significance of differences between two survival curves using GraphPad Prism8 software.).

Fig 2 .
Fig 2. Comparison of relative abundance of microflora in the metagenome and meta-transcriptome sequencing.(a): an abundance of phylum in the metagenome.(b): an abundance of phylum in meta-transcriptome.(c): comparative heatmap of the abundance change of genus in MG/ MT after infestation.(d): comparative heatmap of the abundance change of species in MG/MT after infestation.MG_control: control sample in metagenome sequencing; MG_infected: infected sample in metagenome sequencing; MT_control: control sample in meta-transcriptome sequencing; MT_infected: infected sample in meta-transcriptome sequencing.

Fig 3 .
Figure4shows the significant differentially expressed genes, allowing us to evaluate and understand the potential function of bacteria in immunity at a deeper level.The results of differentially expressed genes showed that many genes were significantly upregulated or downregulated after the infection of M. robertsii(Table 2, Fig 4a).To understand the specific functions of these genes, we carried out a KEGG enrichment analysis.Most are related proteins in coding genetic information, including signal and cellular processes (FigS3).In particular, the gene gshA, glutamatecysteine ligase, participated in synthesizing glutathione (Fig S4), Table 2.The differential genes expressed in meta-transcriptome.Taxon Number up-regulated

Fig 4 .
Fig 4. Changes of bacterial abundance and differential genes in meta-transcriptome.(a): Differential expression genes in bacteria after metarhizium infection.(b): Results of experiments to estimate GSH content in termite's hindgut.The differential expression matrix is calculated by DESeq2 and |Log2Foldchange| >=1 means significant change.M. robertsii means eight termites infected by metarhizium for 48 hours, and M. robertsii + Kan + is eight termites infected by metarhizium after treatment with 5% kanamycin for 48 hours.

Fig S3 .
Fig S3.The pathway map of glutathione metabolism.

Table 1 .
Summary statistics for sequencing data from the Termitidae hindgut gut metagenome and meta-transcriptome.