Transcriptional responses of Daphnis nerii larval midgut to oral infection by Daphnis nerii cypovirus-23

Background Daphnis nerii cypovirus-23 (DnCPV-23) is a new type of cypovirus and has a lethal effect on the oleander hawk moth, Daphnis nerii which feeds on leave of Oleander and Catharanthus et al. After DnCPV-23 infection, the change of Daphnis nerii responses has not been reported. Methods To better understand the pathogenic mechanism of DnCPV-23 infection, 3rd-instar Daphnis nerii larvae were orally infected with DnCPV-23 occlusion bodies and the transcriptional responses of the Daphnis nerii midgut were analyzed 72 h post-infection using RNA-seq. Results The results showed that 1979 differentially expressed Daphnis nerii transcripts in the infected midgut had been identified. KEGG analysis showed that protein digestion and absorption, Toll and Imd signaling pathway were down-regulated. Based on the result, we speculated that food digestion and absorption in insect midgut might be impaired after virus infection. In addition, the down-regulation of the immune response may make D. nerii more susceptible to bacterial infections. Glycerophospholipid metabolism and xenobiotics metabolism were up-regulated. These two types of pathways may affect the viral replication and xenobiotic detoxification of insect, respectively. Conclusion These results may facilitate a better understanding of the changes in Daphnis nerii metabolism during cypovirus infection and serve as a basis for future research on the molecular mechanism of DnCPV-23 invasion. Supplementary Information The online version contains supplementary material available at 10.1186/s12985-021-01721-x.

as Cephonodes hylas Linnaeus, Ampelophaga rubiginosa Bremer & Grey, and Agathia lycaenaria Kollar. The genome of DnCPV-23 consists of ten segments of linear double-stranded RNA, referred to as genomic segments 1 (S1) to 10 (S10), in accordance with the fragments from longest to shortest [7]. Our previous research and unpublished data demonstrated that the virus could successfully replicate on the Sf9 [8] and Manduca sexta cell lines QB-MS 2-2 [9]. However, the molecular mechanism of the interactions between the new type cypovirus and its hosts remains unclear. It is necessary to identify the interactions between the virus and its hosts to achieve an in-depth understanding and reveal the exploitation potential of the virus for future insecticide development.
Recently, many studies in the field have generated large amounts of data using the aforementioned highthroughput approaches, from the silkworms or BmN cells infected with BmCPV, including (1) The possible host's RNAi response against BmCPV challenge in persistent and pathogenic Bombyx mori model was compared. During the pathogenic infection, it was found that higher level RNAi responses against BmCPV were observed, which further demonstrated the importance of RNAi as an antiviral mechanism [10]. (2) Gene expression profiles [11][12][13][14][15][16][17][18][19], DNA methylation [20], and lipidomic profile [21] of silkworm midgut or BmN cells after BmCPV infection were analyzed. These results suggested that many genes (for example, genes expressing Calreticulin, FK506-binding protein, and protein kinase c inhibitor gene, microRNAs, and activated protein kinase C) may play important roles in BmCPV replication. In addition, epigenetic regulation may influence silkworm-virus interaction, and BmCPV may modulate the lipid metabolism of cells for their self-interest.
Until now, the molecular mechanism underlying the midgut infection of DnCPV-23 is not clearly understood. Furthermore, since transcriptome analyses regarding D. nerii or DnCPV- 23 have not yet been performed, this study aims to fill this gap about the new type cypovirus. The data and analysis presented here provide insights into the possible mechanism of DnCPV-23 infection and host defense and a basis for future DnCPV-23 relevant studies.

Daphnis nerii larval midgut and virus stock
Newly wild-caught second instar larvae with a similar mass were used in this research investigation for the virus infection. Before infection, the D. nerii were supplied with 12-h day/night cycles under 50 ± 5% relative humidity conditions and were nurtured on oleander leaves at 27 ± 1 ℃ for three days. The midgut tissues were collected from four pathogenically infected larvae at 72 h [13,15] after feeding with DnCPV-23. The same tissues were also collected from three uninfected control larvae at the same time point. DnCPV was originally isolated from the larvae of D. nerii and propagated in D. nerii larvae [1]. The polyhedra suspension of DnCPV-23 utilized for infecting the D. nerii was stored at 4 °C in the dark.

Virus inoculation
In this study, the DnCPV-23 viral stock was suspended in distilled water at a concentration of 2 × 10 7 polyhedra/mL. Then, 100 μL of the viral suspension was spread evenly on one piece of oleander leaf measuring approximately 4 cm × 1.5 cm each in size. The leaf was then fed to four D. nerii larvae. The dose of infection was calculated as 2 × 10 6 polyhedra per larva. In addition, three control larvae were fed the same quantity of leaves treated with only distilled water. After approximately 12 h, fresh oleander leaves were used to feed the inoculated larvae after the DnCPV-23-inoculated leaves had been completely consumed.

Sample preparation
The midguts of both DnCPV-23-infected and control larvae were collected at 72 h post-inoculation by dissecting the larvae on ice. The isolated midgut was then quickly washed in 0.8% diethylpyrocarbonate (DEPC)-treated physiologic saline solution to remove the attached leaf pieces, and then frozen in liquid nitrogen [13,22].

RNA sequencing
All of the RNA-seq procedures were conducted by the Oebiotech Company (Shanghai, China). The total RNA was extracted from the D. nerii midgut tissue using TRIzol reagent (Invitrogen, USA) according to the manufacture's protocols. The RNA integrity and concentrations were checked using an Agilent 2100 Bioanalyzer (Agilent Technologies, USA). In addition, seven RNA samples (including three uninfected samples and four infected samples) with RNA integrity were used to construct the libraries. The cDNA libraries were prepared using a TruSeq RNA Sample Preparation Kit (Illumina, USA) according to the manufacturer's protocols. Thereafter, the obtained cDNA libraries were sequenced on the Illumina HiSeq2500 platform, which generated paired-end raw reads of 125 bp.

De novo assembly and functional annotation
The raw data was pretreated by discarding reads with adaptors and low quality (quality scores < 30). Then, the raw data was assembled using Trinity software with default parameters for de novo transcriptome assembly. Transcripts that were not shorter than 300 bp were used for subsequent analysis. To obtain the functional annotations of predicted protein-coding sequences, we searched against various databases, including the NCBI non-redundant (NR) protein, SwissProt, and euKaryotic Orthologous Groups (KOG) using Blastx with an E-value < 10 −5 . The top hit was utilized to assign gene names. Whereafter, the Gene Ontology (GO) annotations of the transcripts were then analyzed based on SwissProt annotations, and functional classifications were assigned by WeGO software. In addition, for the purpose of determining the biological pathways involved, the KEGG pathway was annotated based on the KEGG Orthology (KO) identifiers.

Differential gene expression analysis
RNA sequencing results from the two groups were mapped to the assembled transcriptome using bow-tie2 [23] and express [24]. The FPKM (fragments per kb per million reads) method [25] was utilized to calculate the expression levels of the unigenes, which eliminated the influencing effects of the different gene lengths and sequencing levels. The differences in the unigene expressions between the two groups were calculated with DESeq [26] and any significant differences were determined with P < 0.05 and an absolute value of log2 fold change > 1.

Real-time quantitative reverse transcription PCR (Real-Time qRT-PCR)
This study utilized qRT-PCR to analyze the expression level of DnCPV-23 S1, S10 genes of transcriptome samples, and verify the DEGs recognized by the RNA-seq. The total RNA was isolated from the samples of the transcriptomic analysis using TRIzol reagent (Life Technologies) and was then treated with DNase I (Fermentas, Glen Burnie, MD, USA). We reversely transcribed 1 μg of the total RNA per sample into complementary DNA (cDNA) using a PrimeScript RT Reagent Kit (Takara). Then, qRT-PCR was performed using Talent qPCR Pre-Mix SYBR Green (Tiangen, China) on a QuantStudio ™ 7 Flex Real-Time PCR System (Applied Biosystems ™ ). One cycle was added for melting curve analysis for all the reactions to verify the product specificity. The expression level of each gene relative to that of the RPL13 gene was calculated using the 2 −△△CT method [27]. All of the primers for the aforementioned target genes are listed in Table 1. Results are representative of two to three independent experiments.

Virus infection of the samples
Prior to the transcriptome analysis, qRT-PCR was used to detect the mRNA levels of the DnCPV-23 S1 and S10 genes in the infected and uninfected samples. The results showed that the infected group had been successfully infected based on the high relative expression of the viral gene mRNA compared with uninfected group (Fig. 1).

Transcriptome sequencing and assembly
The RNA-Seq data from the DnCPV-23-infected and control groups contained 346.39 million reads, and 334.60 million clean reads after trimming, among which 96.17 to 97.39% per sample were determined to be useful. The acquired clean reads were assembled into 31,696 unigenes (> 300 bp). The average length of these unigenes was 1347.61 bp, and the N50 length was 2348 bp; other information about these unigenes were shown in Table 2. This study then assembled 31.696 unigenes ranging from 301 bp to 32,420 bp. The total unigene length was 42,713,980.

Transcriptome annotation
A total of 31,696 assembled unigenes were searched against the public databases, including the NR, Swissprot, KOG, GO, and KEGG databases, among which 16,820 (53.1%) (Fig. 2) unigenes were annotated. The distribution patterns of the unigenes in the different databases were as follows: 16,615 unigenes in the NR database, 11,152 unigenes in the Swissprot database, 10,374 unigenes in the KOG, 10,468 unigenes in the GO, and 5501 unigenes in the KEGG databases (Table 3). Figure 2 shows the degree of overlap between the unigenes annotated in the different databases. It was found that 4353 (13.7%) unigenes overlapped in all five databases, while 12,390 (73.7%) unigenes overlapped in two or more databases.

Significant impacts of the viral infection on the hosts' transcriptome expressions
As shown in Fig. 3, the main component PCA1 had reached 41.56%, and the main component PCA2 had reached 27.23%. Therefore, the percentage total of the two was 68.79%, which accounted for a high proportion and represented the overall population to a large extent. This study's principal component analysis manifested a clear separation of the samples with the two treatments ( Fig. 3A), which indicated that the samples had good repeatability. The heat map of the gene expressions is presented in Fig. 3B. The results suggested that these DEGs could distinguish the samples. The results revealed that the viral infection could exert apparent influences on the midgut gene expressions. In addition, the transcriptome results showed that 1166 genes were down-regulated (accounting for 3.68% of the total assembled unigenes) and 812 genes (accounting for 2.56% of the total assembled unigenes) were up-regulated as a response to the DnCPV-23 infection (Fig. 3C).

Analysis of the differently expressed genes
In this study, KEGG function enrichment analysis was performed on the differential genes expressed in the DnCPV-23-infected and uninfected control groups to clarify the relevant biological pathways involved in the differential genes. Among all of the DEGs, 298 DEGs had KEGG annotations, of which 118 were up-regulated genes and 180 were down-regulated genes. According to the pValue of KEGG analysis of up-regulated and down-regulated signal pathways, we identified 20 most significant signal pathways each. These pathways play an important role in insect reproduction, immunity, digestion and absorption and xenobiotic metabolism and so on (Fig. 4).

qRT-PCR validation of DEGs
To verify the reliability of the transcriptome data and the DEG results obtained by RNA-seq, seventeen DEGs were selected for qPCR analysis. As shown in Fig. 5, the fold-change values of DnCPV_1 sample vs Mock_1 sample obtained in the qPCR analysis results were consistent with the values obtained by the RNA-seq for all of the selected genes.

Discussion
This study analyzed the transcriptome of the uninfected D. nerii midgut and the DnCPV-23-infected D. nerii midgut presented unique gene expression profiles induced by DnCPV-23 infection for the first time.
In addition, KEGG function enrichment analysis was performed on the differential genes expressed after DnCPV-23 infection. Compared with uninfected D. nerii midgut, the transcriptome profiles of the infected samples displayed universally changed transcript abundances for many pathways.    Based on the pValue of KEGG analysis regarding up-regulated and down-regulated signal pathways, we identified 20 most significant signal pathways each. Among these signal pathways, the retinol metabolism pathway, vitamin digestion, and absorption signal pathway were down-regulated, consistent with the transcriptome study about BmCPV infected midgut vs non-infected midgut [13]. In addition, protein digestion and absorption pathway way was down-regulated in accord with previous research [10]. DnCPV infection may destroy the functions of digestion and the absorption of midguts, which causes the disturbance of protein and amino acid metabolism in D. nerii [13,28]. Peptidoglycan recognition proteins (PGRPs) are pattern recognition molecules that are conserved from insects to mammals. PGRPs are the first receptors known to recognize, bind, or catalytically cleave the pathogenic microorganisms [29], PGRPs recognize bacteria and their unique cell wall component, eptidoglycan [30,31]. This study observed nine transcripts of D. nerii isoforms of PGRP genes. Six transcripts were found to be down-regulated in the infected D. nerii midgut. The most highly expressed and most dramatically downregulated was TRINITY_DN13195_c0_g1_i3_3, which was down-regulated by as much as 51-fold. The downregulation of PGRP expression can lead to a decrease in the ability of the D. nerii's innate immune system to recognize bacterial peptidoglycans (PGN), which may lead to D. nerii more susceptible to bacterial infections. In addition, BmPGRP-S2 was up-regulated upon BmCPV infection, overexpression of which can activate the Imd pathway and induce increased AMPs to enhance the antiviral capacity of transgenic silkworm against BmCPV [32]. Moreover, previous study demonstrates [33] that PGRPS2-1 and PGRPS2-2 can prevent BmCPV replication. Based on this work, was speculated that the down-regulation of PGRP was conducive to the replication of DnCPV-23.The gene CASP8 (KEGG gene name: caspase-8, Gene id: TRINITY_DN10280_c0_g1_ i1_3) (Dredd in Drosophila) was down-regulated more than two folds, and other caspase genes changed nonsignificantly. It is predicted to be involved in the cleavage of Relish, the Drosophila homolog of mammalian NF-κB, resulting in activating the immune-deficient pathway (IMD)-induced expression of antimicrobial peptides in response to Gram-negative bacteria [34][35][36], fungi and viruses [37]. Research performed by Li et al. proved BmDredd interacts with BmSTING to enhance antiviral signaling [38]. The down-regulation of this gene may be very important for DnCPV-23 to escape from the host innate immune system and replicate in the midgut. Our result conflicted with the work by Guo et al. [11]. We speculated the contradiction might be related to the different stages of virus-host interaction or the heterogeneity of different species against virues. The pathways and the genes mentioned above are listed in Table 4 (The expression of genes in each sample is shown in Additional file 1).
In this study, the up-regulation of glycerophospholipid metabolism was consistent with Zhang's research [21]. The up-regulation of this pathway may be related to the viral replication [39,40]. In addition, Glycine, serine and threonine metabolism were up-regulated in this transcriptome analysis. In the study by Wu et al., two genes related to this signaling pathway were upregulated and the other down-regulated. In our study, the expression levels of the phosphoserine phosphatase genes were significantly higher in DnCPV-23-infected midgut than in the non-infected group, suggesting that serine metabolism disorders were induced after DnCPV-23 infection. Expression of many UGT genes was up-regulated; UDP-glucuronosyltransferase (UGT) isozymes take endogenic and exogenic toxic substances as substrates, catalyze detoxification of many chemical toxins in our daily diet and environment by conjugation to glucuronic acid or glucose [41,42]. After DnCPV-23 infection, it was speculated that the D. nerii tended to strengthen the elimination of lipophilic endobiotics such as hormones and xenobiotics including phytoalexins and drugs conjugated by invertebrates and plants mainly with glucose [42] through promoting the transcription of UGTs by regulating the activities of nuclear-receptor family (CAR, PXR, FXR, LXR, and PPAR), the arylhydrocarbon receptor [43] or ubiquitous transcription factors (FOXA1, Sp1, and Cdx2)