The impact of allelochemicals on the differential expression of symbiotic bacteria in cotton aphids

LIU Ying, LIANG Ping-zhuo, LI Fen, MA Kang-sheng, CHEN Xue-wei, CHEN An-qi, LIANG Pei, GAO Xi-wu

Department of Entomology, China Agricultural University, Beijing 100193, P.P.China

Abstract Insects have developed a good adaptive mechanism in response to environmental stresses in the long-term evolution.They have developed a helpful metabolism system to resist plant allelochemicals. Insects also harbor different kinds of symbiotic bacteria, which provide them a competitive advantage. Here, using cotton aphid as an example, we investigated the effects of four plant allelochemicals on the differential expression of symbiotic bacteria based on transcriptome data.We also studied the composition of symbiotic bacteria and function on pathway level in three kinds of aphids. We found that the bacteria have a significant role in resisting the plant allelochemicals stress and host plant selection by aphids. These results should be useful to investigate the environmental adaption mechanism of aphids in the view of symbiotic bacteria.These results would offer a new insight for improving strategy of aphids and developing new pest control systems.

Keywords: Aphis gossypii, plant allelochemicals, differential expression, symbiotic bacteria

1. Introduction

Like all other living organisms, insects have to deal with various parasites and pathogens in the environment(Douglas 1998). In the evolutionary process, insects have developed a close relationship with the symbiotic bacteria.The symbiotic association is a general and an extensive phenomenon in nature. Aphid is a typical model organism for studying the symbiotic relationship between insect and bacteria (Hayneset al.2003). Bacteria satisfies the nutrient needs of the aphid providing essential amino acid,esters and vitamin that are not naturally produced by aphids(Birkleet al.2002; Akman and Douglas 2009). Bacteria also stimulate the immune system and signaling molecules,which influences the gene expression and physiology in aphids (Suet al.2015). Aphids, in turn, regulate the growth and reproduction of bacteria (Vigneronet al.2014).The most typical symbiotic bacterium in cotton aphid isBuchnera(Moran and Degnan 2006). It cannot survive when removed from aphid. Similarly, aphids suffer from sterility or death when deprived of the symbiont (Houk and Griffiths 1980). Besides this bacterium, aphids also harbor many other bacterial species. These symbiotic bacteria can affect various aspects of the host, including host-plant specialization, thermal tolerance, and resistance to drugs(Hayneset al.2003; MacDonaldet al.2011; Renozet al.2015).

Insects are the most varied animals on earth. There is hardly a plant that could avoid an insect invasion, while there is hardly an insect that could invade the whole plant. Most insects rely on the primary metabolites of plants to obtain basic nutrient material and other essential elements. On the other hand, the host plants are able to produce a diverse array of allelochemicals to defend itself against the damage due to insects. During this course, various detoxification mechanisms are induced in insect to resist and neutralize the plant defense (Pare and Tumlinson 1999a; Walling 2000;Nishida 2014). As an important role in the co-evolutionary relation with the insect, microbes have been involved in detoxification of defensive chemicals to resist the stress(Masonet al.2014). However, few researches have paid more attention on co-expression at the transcriptional level.

In our previous study, by analyzing transcriptome data and digital gene expression (DGE) profiles which the number was SRR4384533 in SRA database of NCBI, we found that when aphids are exposed to plant secondary chemicals such as 2-tridecanone, tannic acid, quercetin and gossypol, they show a differential gene expression.Some studies on transcriptome profiles of aphids have also been done in the past. In addition to the genome data of pea aphid, data for other aphids (e.g., cotton aphid, peach aphid, etc.) generated through high-throughput sequence technique have already been published (Burke and Moran 2011; Liet al.2013; Liuet al.2014; Yanget al.2014). The first transcriptome of cotton aphid was published in 2013(Liet al.2013) and, nowadays, cotton aphid data can be easily obtained from NCBI. Previous studies have primarily focused on the molecular mechanisms of ecological adaption (Liet al.2013), reproductive mode (Liuet al.2014),as well as wing polyphenism (Yanget al.2014). However,in the annotation step, we often face a problem. Compared to other insects, more unigenes were annotated as bacteria in aphids. Therefore, we doubt if any bacterial genes were sequenced in the sample and if so, how much of them?

In this study, we re-annotated the transcriptome sequences through a different method to obtain sequences related to bacteria in order to investigate the system of cotton aphids. By treating cotton aphids with the four important plant allelochemicals, we found that these chemicals not only impacted the insect but also acted on the endosymbiotic bacteria at the pathway level. Based on this finding, we may,therefore, infer that the structure of microbial community becomes an influence factor in host plant selection.

2. Materials and methods

2.1. Data sources

RNA-seq data were obtained from a HiSeqTM2000 Platform (Illumina, USA) as reported by Liet al.(2017).The sequence data have been deposited in SRA database of NCBI as accession no. SRR4366847. In the previous study, the aphid strain was grown on cotton leaves in the laboratory and combined equal numbers of samples from various developmental stages to extract total RNA. The mRNAs were enriched by oligo (dT) magnetic beads and fragmented into short fragments in the presence of divalent cations in fragmentation buffer. Using these cleaved, short fragments as templates, random hexamer primers were used to synthesize first-strand cDNA. Second-strand cDNA was generated using buffer, dNTPs, RNase H, and DNA polymerase I. Following end repair and adaptor ligation,short sequences were amplified by PCR and purified with a QIAquick PCR Extraction Kit (Qiagen, The Netherlands)and sequenced on a HiSeqTM 2000 platform (Illumina,USA). In the same project, Liet al.(2017) reported the DGE profile of the aphids treated with 2-tridecanone, tannic acid, quercetin, gossypol and the control group. In the treatment group, 5% sucrose solution contained 20 mg L–1allelochemicals was used, while in the control group, 5%sucrose solution was used.

In the mixed aphid transcriptome library, raw reads were assembled to unigenes using Trinity Program. The open reading frame (ORF) was predicted by using the Getorf tool(Tringeet al.2005). The clean tags of DGE were mapped to mixed transcriptome using RSEM (Li and Dewey 2011)with 1 nt mismatch. Then the RPKM value was measured in reads per kilobase of transcript sequence per million mapped reads. The false discovery rate (FDR) method was used to determine the threshold ofP-value by DESeq (Lenget al.2013). In this study, we set the threshold as FDR＜0.05 to judge the significance of differentiated gene expression.

2.2. Bacteria in cotton aphid

In order to obtain the annotation of species in cotton aphid,the ORF sequences of transcriptome data were submitted to GhostKOALA Program (ver. 2.0) in KEGG website (http://www.kegg.jp/ghostkoala/). The uploaded sequences were searched for homology using GHOSTX, a homology search tool which can detect remote homologues by using suffix arrays search against a non-redundant set of KEGG genes.Then, computation was performed to assignKnumbers and links to KEGG Mapper for inferring high-level functions.According to GhostKOALA, the sequences were annotated to analyze the taxonomic compositions and functions in the KEGG pathway.

By combining the expression analysis of cotton aphid transcriptome, the different expression of gene after allelochemicals treatment was acquired.

2.3. KEGG pathway restructuring

Pathway enrichment analysis was carried out using KEGG database. The unigenes, which were annotated as bacteria,were distributed to six classifications of KEGG pathway with differential expression information.

3. Results

3.1. DGE data source and taxonomic composition analysis in cotton aphid

In the mixed aphid transcriptome library, 28 555 unigenes were assembled by 53 763 866 raw reads using Trinity Program. The result of the expression abundance of transcripts suggested that more than 93% sequenced data were meaningful. Setting the selection criteria as the fold change over two in the differential expression, 171, 20, 84 and 44 genes were up-regulated, and 191, 218, 203 and 394 genes were down-regulated in the 2-tridecanone, tannic acid, quercetin, and gossypol treatment groups, respectively(Table 1).

The sequences in cotton aphid transcriptome were annotated by online software GhostKOALA in KEGG,which is a suitable tool to annotate microbial environment genomes. A total of 27 718 (98.30%) sequences were annotated by the taxonomic composition of KEGG with the GHOSTX score from 18 to 15 105. Among them,10 338 sequences received theKnumber assignment to perform the KEGG mapper analysis. From the taxonomic composition, the sequences were assigned to seven sections: animals, bacteria, protists, fungi, plants, archaea and viruses. The majority of sequences (11 654; 41.33%)were spread over animals, especially insects (Fig. 1). The bacteria also accounted for a big proportion due to huge group of symbiotic strains. Buchnera was the dominant bacterial community whileClostridiumandBacillus, both belonging to the Firmicutes, took the second and third places, respectively.

3.2. Differential expression of the bacterial section in cotton aphid transcript in response to different plant allelochemicals

We summarized their taxonomic composition and compared the four treatment groups by analyzing the differential expression data. Compared to the control group, 362, 238,287 and 438 sequences possessed significant difference in the groups treated with 2-tridecanone, tannic acid,quercetin and gossypol, respectively. Based on KEGG taxonomic annotation, the differential expression genes were classified into seven sections. In addition to animals,bacteria occupied the biggest section. In 2-tridecanonetreated group, 68 genes were annotated as bacteria and appeared significantly different in gene expression.Similarly, in tannic acid, quercetin and gossypol treated group, 63, 52 and 151 genes were, respectively, annotated as bacteria (Fig. 2).

The result of taxonomic composition showed 18.78,26.47, 18.12 and 34.47% differently expressed sequences annotated as bacteria for 2-tridecanone, tannic acid,quercetin and gossypol set, respectively (Table 2). The group treated with 2-tridecanone showed more genes with higher expression. The number of up-regulated genes was almost equal to the number of down-regulated genes in the group treated with quercetin. The majority of genes in the groups treated with both tannic acid and gossypol exhibited down-regulated genes. Particularly, in the gossypol treatment group, the expression of 143 genes was significantly down-regulated while only eight genes were upregulated. Both tannic acid and gossypol groups showed similar performance in expression tendency.

3.3. KEGG analysis of differently expression bacterial genes

We obtained theKnumber of cotton aphid transcriptome based on above analyses. Combining the result of taxonomic composition, we got bacterial genes from KEGG annotation. In total, 1 895 genes produced 1 205Knumber annotation, and 1 075 genes were assigned to 177 KEGG pathways. The pathways were divided into six categories:metabolism, cellular processes, environmental information processing, human diseases, genetic information processing and organismal systems.

Considering the significant differential expression of bacterial genes, we restructured the pathways to obtain the distribution in all kinds of classes (Fig. 3). The gossypol treatment group possessed the maximum differentialexpression genes and they were annotated in most pathways followed by tannic acid, 2-tridecanone, and quercetin treatment group. “Metabolic pathway” category was the most strongly represented in all the four treatment groups. The percentage of metabolic pathways represented in the four treatment groups were 93.75, 76.64, 76.92 and 78.09%, respectively. It is noteworthy that the pathway of‘genetic information processing’ also had an important role in tannic acid and gossypol treatment groups.

Table 1 The statistic of sequences expression after exposure to allelochemicals1)

Fig. 1 Statistic of taxonomic compositions analysis in cotton aphid transcriptome. The innercircle shows the proportion of transcriptome species annotation on the kingdom level, and the outer ring shows the classification on class level.

Fig. 2 Taxonomic composition of the different expression genes. The number in each column indicates the detail count of bacteria genes.

3.4. Differential expression of bacterial genes in pathways

Based on the analysis of the KEGG pathway, we found the distribution of higher and lower expression genes in various pathways. In the group treated with 2-tridecanone,almost all the higher expression genes were distributed in the metabolism pathway and they were found to participate in amino acid metabolism, xenobiotics biodegradation and metabolism, and biosynthesis of other allelochemicals.On the other hand, the decreased expression genes were involved in regulating few pathways including MAPK signaling pathway. The up-regulated gene was not linked to any pathways in the tannic acid treatment group. However,numerous down-regulated genes were annotated in various pathways. Besides their role in various metabolisms,they were also associated with the genetic information processing, including folding, sorting and degradation,replication and repair, transcription and translation. In the quercetin treatment group, only one over-expressed gene was annotated in the pathway, which was involved in glycan biosynthesis and metabolism. Metabolism and genetic information processing were still heavily affected. Gossypol treatment led to the annotation of more down-regulated genes than up-regulated genes. It covered all the pathway of first level classification in KEGG. Consistent with the results of tannic acid, gossypol treatment also caused the change in metabolism pathway as well as heavily impacted the genetic information processing classification as the down-regulation elements (Fig. 4).

Styrene degradation (ko00643) is a pathway of xenobiotic biodegradation and metabolism. The treatment with 2-tridecanone induced the formation of nitrilase (EC: 3.5.5.1)causing an increase in styrene degradation (Appendix A).The two genes in tannic acid treatment group were down-regulated and they affected the lysine biosynthesis(Appendix B).

Table 2 The statistic of differential expression sequences which were annotated as bacteria

4. Discussion

Production of allelochemicals is a key way by which plant defends itself against insect pest damage (Pare and Tumlinson 1999). In addition, secondary metabolites also possessed induction to the detoxification mechanism of insects (Yuet al.1979; Taoet al.2012; Giraudoet al.2015). Cotton aphid is a kind of insect that exhibits typical mutualism with the bacteria. Plant allelochemicals not only influenced the detoxification metabolism of the cotton aphid, but also affected its symbiotic strains. In this paper,we restudied the DGE data profiles to discover the series of effects. By analyzing the differential expression in control group and treatment groups by plant allelochemicals:2-tridecanone, tannic acid, quercetin, and gossypol, we found 68, 63, 52 and 151 genes were annotated as bacteria,respectively. These genes were involved in the regulation of many important metabolism pathways, cellular processes as well as genetic information processing. These four allelochemicals are universal plant secondary metabolites.Under a certain dosage, the allelochemicals may lead to untoward effect in insect. It may also induce the resistance mechanism in insects such as the detoxification system,which enables them to adapt to the host (Guenduez and Douglas 2009). As an important symbiotic companion, the bacteria were naturally influenced by allelochemicals. Of the four allelochemicals, gossypol had the greatest influence on cotton aphid bacteria.

Based on above evidence, we speculated that aphid enhances its adaptability to multitudinous host plants, either by reducing its sensitivity or by enhancing metabolism.Here, we pay close attention to the symbiotic bacteria that have cooperative coevolution with the aphids. The bacterial colony possess a strong ability to metabolize the exogenous chemicals by changing group structure and its self-proliferation in a short period. It is obvious that the bacterial population changed quickly after treating with secondary metabolites and caused a drastic change in the compound, enabling the aphid to adapt. Particularly, it had a quick response in the metabolism. On the other hand, the foreign genes coming from the symbiont bacteria may take part in the metabolism pathways in aphid, thus improving the ability to neutralize the influence of allogenic material.Bacterial genes were supplemented to the host and were integrated into the metabolism pathway.

Fig. 3 Metabolic pathways classification of bacterial genes.

To realize the relationship between bacterial content and host range, the pea aphid (Acyrthosiphon pisum) data were introduced. The genome sequences of pea aphid were obtained from the NCBI genome database and annotated using above-mentioned method employed for cotton aphid.All protein sequences (36 195) gained the taxonomic component note and 2 030 of them were classified into bacterial section. The distribution of taxonomic component is shown in Fig. 5. We chiefly aimed at the bacteria, the insect part of genome sequences was removed. In this part of data, the proportion of bacterial sequences was 36% in cotton aphid and 31% in pea aphid. The analyses of bacterial species of the two aphids revealed even more kinds of bacteria (Fig. 6). In total, there were 688 species of bacterial genes in cotton aphid transcript and 181 of them only existed in it. The sequences of pea aphid were annotated to 522 bacterial species. Among them, only 15 bacterial species existed. We inferred that the variety of host plant may relate to bacterial content in aphids because a majority of bacterial species provide greater facilitation for defensing allelochemicals in adaptation of the host-plants.Cotton aphid is a kind of omnivorous insect pest, which has an obvious selectivity for hosts and a good adaptability to cotton. Besides cotton, it also harms other crops and ornamental plants. The host plant of cotton aphid ranges up to 74 families and 285 species. In contrast, the host of pea aphid is less, only foraging in legumes and grasses. The feeding habit is exactly in agreement with the analysis and the distribution regularities of the bacteria. It suggests that the symbiosis may play an important role in the extension of plant hosts in aphids.

Fig. 4 Differential expression genes in response to different plant allelochemicals. Numbers in each column represent the number of bacterial genes.

Fig. 5 Taxonomic compositions analysis in cotton aphid and pea aphid after excluding insect group. A, Aphis gossypii. B,Acyrthosiphon pisum.

Fig. 6 Differences and similarities in bacterial species between cotton aphid and pea aphid. “Ago” indicates Aphis gossypii,and “Api” indicates Acyrthosiphon pisum.

5. Conclusion

We analyzed the relationship between aphid and its symbiotic bacteria under the stress of plant allelochemicals.The constituents and the main functions of bacteria were investigated based on the pathway level in three kinds of aphids. The results indicated that the symbiotic bacteria not only provide basic nutrients to host insects but also resist the secondary compounds from plants, by enhancing their metabolism and perfecting metabolism pathways to some extent. This symbiotic relationship, which improves the competitive ability of insects and protects them from deleterious compounds, is the result of long-term coevolution. Sufficient studies on the defense mechanisms of insect pests and symbiotic bacteria may be beneficial to understand the effects of plant allelochemicals. These results may offer a new insight for improving strategy of aphids, developing new pest control systems.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (31330064).

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm