Transcriptome analysis of the molecular mechanism underlying immunity- and reproduction trade-off in Locusta migratoria infected by Micrococcus luteus

Immune response and reproductive success are two vital energy-consuming processes in living organisms. However, it is still unclear which process is prioritized when both are required. Therefore, the present study was designed to examine this question arising for one of the world’s most destructive agricultural pests, the migratory locust, Locusta migratoria. Transcripts from the ovaries and fat bodies of newly emerged locusts were analyzed, using RNA-seq based transcriptome and qualitative real-time PCR, at 4 h and 6 d after being infected with the gram-positive bacteria Micrococcus luteus. Changes in the main biological pathways involved in reproduction and immunization were analyzed using bioinformatics. After 4 h of infection, 348 and 133 transcripts were up- and down-regulated, respectively, whereas 5699 and 44 transcripts were up- and down-regulated, respectively, at 6 d after infection. Moreover, KEGG analysis indicated that vital pathways related with immunity and reproduction, such as Insulin resistance, FoxO signaling, Lysosome, mTOR signaling, and Toll-like receptor signaling pathways were up-regulated. Among the differentially expressed genes, 22 and 17 were related to immunity and reproduction, respectively. The expression levels of PPO1 and antimicrobial peptide defensin 3 were increased (log2FC = 5.93 and 6.75, respectively), whereas those of VgA and VgB were reduced (log2FC = -17.82 and -18.13, respectively). These results indicated that locust allocate energy and resources to maintain their own survival by increasing immune response when dealing with both immune and reproductive processes. The present study provides the first report of expression levels for genes related with reproduction and immunity in locusts, thereby providing a reference for future studies, as well as theoretical guidance for investigations of locust control.

Introduction and qRT-PCR verification were used to predict and explore genes involved in immunity and reproduction, as well as to identify the molecular mechanisms underlying energy and resources trade-off in L. migratoria.

Insect rearing and experimental treatments
Eggs were collected from L. migratoria that were maintained in the insect laboratory at Hang Zhou Normal University (Zhejiang Province, China). The locust models were established by rearing the insects in each well-ventilated cage (50 × 50 × 50 cm) at densities of 200-300 insects per cage. The insects were reared on a 16 h photoperiod at 30 ± 2˚C and fed on fresh greenhouse-grown wheat seedlings [12][13]. Female locusts were collected and subject to parasite treatments within 12 hours after eclosion. The treated group was infected with the bacterial pathogen M. luteus (American Type Culture Collection, Manassas, VA, USA) which had been cultured to the logarithmic phase, according to the manufacturer's instructions by injecting the pronotum of each locust with 10 μl of bacterial culture (0.9 optical density at 600 nm) using a nanofil syringe (Shenggong, Shanghai, China). Meanwhile, the pronotum of each locust in the control group was stabbed using an alcohol-sterilized needle. For both the treatment and control groups, the wounding site of each insect was sealed using petroleum jelly to prevent exogenous natural infection. A diet of wheat seedlings and wheat bran was supplied in sufficient, but not excessive, amounts. Fat bodies and ovaries were collected from females in each group at 4 h and 6 d after injection, respectively, with two biological replicates per group. Furthermore, 30 ovaries from females in each group were dissected and weighed.

cDNA library construction and high-throughput transcriptome sequencing
Trizol (Invitrogen, Los Angeles, CA, USA) was used to extract total mRNA from mixed tissue samples (fat body and ovary) taken from individual insects of the control and treated groups at each time point, following the manufacturer's procedure. The quantity and purity of the total RNA were measured using a Bioanalyzer 2100 and RNA 6000 Nano LabChip Kit (RIN>7.0; Agilent, CA, USA). Poly(A) mRNA was isolated from~10 μg of each total RNA sample using poly-T oligos attached to magnetic beads (Thermo-fisher, MA, USA). The isolated mRNA was fragmented using divalent cations under elevated temperature and then used to construct a cDNA library, following the protocol for the Illumina RNA ligation-based method (Illumina, San Diego, CA, USA). Briefly, the fragmented RNA was dephosphorylated at the 3' end using phosphatase and phosphorylated at the 5' end using polynucleotide kinase. These samples were purified using the RNeasy MinElute Kit (Qiagen, Dusseldorf, Germany), following the manufacturer's instructions, ligated to a pre-adenylated 3' adapter, which enabled the subsequent ligation of a 5' adapter, and then subjected to both reverse transcription and PCR. The average insert size for the paired-end libraries was 300 bp (±50 bp). Finally, the library was subjected to single-end sequencing using an Illumina Hiseq 2000 by LC Sciences (Houston, TX, USA), following the vendor's recommended protocol.

Bioinformatic analysis
Raw data was filtered using Trimommatic software with default parameters. The clean reads that were mapped to the L. migratoria genome database (http://locustmine.org/index.html) were annotated using Locust Base-locust genome data, GO, KEGG, Nr, Nt, and Swissport. The number of perfect clean reads corresponding to each gene was calculated and normalized to the number of Reads Per Kilobase of exon model per Million mapped reads (RPKM) [14].

Quantitative real-time PCR analysis
To verify the RNA-seq data fifteen genes were selected. Primer 5.0 was used to identify appropriate primers (S1 Table), and the expression levels of the selected genes were normalized using the expression of actin. Quantitative real-time PCR (qRT-PCR) was performed using a Bio-Rad CFX96 Real-Time PCR Detection system (Bio-Rad, CA, USA) and Premix Ex Taq (SYBR Green) reagents (Takara, Dalian, China). The 20-μl qRT-PCR reactions were subject to a thermal profile of 95˚C for 3 min and then 40 cycles of 95˚C for 10 s and 60˚C for 30 s. The thermal melting profile was assessed using a final PCR cycle of 95˚C for 30 s, with temperature increasing constantly from 60 to 95˚C. Relative gene expression levels were calculated using the 2 -ΔΔCT method, with three replicates per sample.

Changes of ovaries in infected group and RNA-seq library analysis
We compared the weight of ovaries of infected group with control group. The weight of ovaries in infected group was significantly lighter than in control group (Fig 1). In RNA-seq experiment, single end reads has been obtained with the quality high enough for the gene expression analysis subsequently (Table 1). In order to investigate the patterns of gene expression related to locust immunity and reproduction, we analyzed control and infected groups at two time points, with two biological replicates each (eight samples in total). The L. migratoria, Nr, Nt, SwissProt, GO, and KEGG databases were used to annotate the transcripts ( Table 2). Most of the mRNAs had low RPKM values, typically between 30 and 80 (Table 3).

DEGs analysis
Firstly, the infected and control libraries for each time point were compared (Ml4 vs CK4 and Ml6 vs CK6) to identify genes that are differentially expressed at different times of infection. At 4 h after treatment, 348 and 133 genes were up-and down-regulated, respectively, in the infected group (S2 Table) and at 6 d after treatment, 5699 and 44 were up-and down-regulated (S3 Table). Secondly, the times points for the control and infected libraries were compared (CK6 vs CK4 and Ml6 vs Ml4) to identify genes that are differentially expressed during infection. For the control group, 24 and 1406 genes were up-and down-regulated, respectively, at 6 d after treatment, when compared to 4 h after treatment (S4 Table), and for the treatment group, 146 and 48 were up-and down-regulated (S5 Table). The number of the DEGs was summarized as columnar (Fig 2). All the DEGs were clustered to produce heat maps, the two replications of controls and treatments were clearly divided into two main clusters (Fig 3).

GO and KEGG analysis
For the up-regulated GO terms related to immunity (Fig 4A, S6 Table), the most frequently mapped transcripts were 'catalytic activity' and 'response to oxidative stresses.' In addition, a large number of transcripts were also mapped to functional groups related to fighting against external infections, such as 'Lysosome' and 'Endosome'. 'Immunity signal JNK cascade,' 'innate immune response,' 'defense response,' 'immune response,' and 'inflammatory response' were represented by seven, seven, six, five, and three transcripts, respectively. For up-regulated GO terms related to reproduction, six GO terms were significantly changed, including 'ovarian follicle cell development,' 'ovarian follicle cell stalk formation,' 'MCM complex,' 'ovarian nurse cell to oocyte transport,' 'female germline ring canal formation,' and 'negative regulation of TOR signaling.' For GO analysis of the transcripts up-regulated at 6 d in the control group, most of the DEGs were related to genes involved in individual growth ( Fig 5A). However, when comparing transcripts from the control and infected groups at 4 h after treatment and when comparing transcripts from the infected groups at 4 h and 6 d after treatment, none of the significantly changed GO terms were directly related to immunity or reproduction.
For the up-regulated transcripts mapped pathway ( Fig 4B, S7 Table), ten of them were involved in immunity and reproduction. The screening criteria used to determine the number  Table) was up-regulated in the infected group. In the control group, five down-regulated pathways were identified as involved in immunity and reproduction, including 'Endocytosis,' 'Lysosome,' 'Phagosome,' 'FoxO signaling pathway,' 'mTOR signaling pathway.' (Fig 5B, S9  Table). In the infected group, no pathways related to immune or reproductions were differentially expressed at 6 d after treatments, when compared to 4 h after treatment (S10 Table).

Immunity-and reproduction-related transcripts
The GO and KEGG analyses failed to identify any significant differences between the control and infected groups at 4 h after treatment. Therefore, in searching for genes involved in reproduction and immunity, mainly we have selected significantly expressed genes in Ml6 vs CK6 group.
Of the transcripts up-regulated at 6 d after treatment, 22 were annotated as immunityrelated genes (Table 4), and defensin 3 and PPO1 were increased the most (log 2 FC = 5.93 and Inf, respectively; where, 'Inf' indicates that the expression of the transcript in the control group was extremely low, so that the resulting ratio approached infinity). Two genes participated in Toll signaling, namely, Toll-9 receptor and Toll-like receptor 6, were up-regulated in the infected group, as were AP-1 complex subunit gamma-1, AP-2 complex subunit alpha, and AP-3 complex subunit mu-1, serine protease inhibitor 27A, and serine protease inhibitor 4-like protein. Seventeen reproduction-related transcripts were up-regulated, whereas both VgA and VgB were sharply down-regulated (log2FC = -17.82 and -18.13, respectively). Several genes related to JH metabolism, such as juvenile hormone acid methyltransferase, juvenile hormone epoxide hydrolase-like protein 5, and juvenile hormone-regulated gene, were also up-regulated in the infected group, as were insulin metabolism related gene and insulin-like receptor.

qRT-PCR validation of DEGs
Fifteen DEGs involved in either immunity or reproduction were randomly selected to verify the accuracy of the RNA-seq data using qRT-PCR. The qRT-PCR results validated the expression of thirteen genes (Fig 8). The differentially expression of one was not significant, while another was down-regulated. The overall trends of most of the genes were consistent, which indicated that the DGE results were reliable enough to interpret scientific problems we focused in this present study (Table 5). Note

Discussion
Resource trade-offs between immunity and reproduction may arise as a consequence of direct physiological conflicts between the two processes [10]. In the present study, the avirulent bacteria M. luteus was used to infect female L. migratoria, and the fat bodies and ovaries of the locusts were dissected at 4 h and 6 d after treatment. Female locusts develop to sexual maturity at~5 d after eclosion, and energy and resource accumulate during this period. Significant changes in fat bodies and in ovaries start at 7 to 10 d after eclosion, and varieties of ingredients in fat bodies and ovaries increased rapidly since 10 d [15]. Bacterial infection often elicits longterm changes in global host transcription. For example, flies that have survived from M. luteus infection remain chronically infected with~2 10 to 2 13 bacteria per fly for at least 5.5 d [16].
Considering that the early response includes both an aggressive initial immune response and injury-induced transcriptional regulation [17], infected insects were collected at 144 h (6 d) after treatment, with 4 h after treatment used as the control time point. GO analysis, which was used to explore the relationship between immunity and reproduction in locusts, indicated that GO terms related catalytic activity (n = 22) and response to oxidative stress (n = 18) were the most enriched. These biological functions are the basic response of the organism to external injury stimuli. Six GO terms associated with endosome metabolism were increased significantly. In eukaryotic cells, endosomes play a vital role in the regulation of fundamental processes, including nutrient uptake, immunity, signaling, adhesion, membrane turnover, and development, by regulating the re-utilization or degradation of membrane components. The up-regulation of functional groups associated with endosomes indicates an increase in metabolic waste emissions and cell activity in the fat bodies and ovaries [18]. The female locusts were at a stage where they were resisting bacterial infection and preparing for reproduction, which may have increased their metabolic activity. Lysosomes are single-membrane vesicles that release enzymes that degrade biological material [19] and, therefore, participate in waste removal. Four GO terms related to the JNK pathway, which regulates cell death in Drosophila and involved in the Toll pathway [20], were up-regulated, possibly indicating the occurrence of apoptosis in infected locust cells. GO terms for 'innate immune response,' 'defense response,' and 'immune response' were also up-regulated, thereby indicating that immune response of the infected locusts was heightened. Cellular immunity related GO terms, such as 'autophagosome,' 'autophagy,' and 'regulation of autophagy,' were significantly up-regulated, thereby indicating that the cells were fighting against pathogenic infection [21]. Three GO terms were involved in ovary development, including 'ovarian follicle cell development' (n = 23), 'ovarian follicle cell stalk formation' (n = 6), and 'ovarian nurse cell to oocyte transport' (n = 3). The MCM complex (n = 5) is a class of DNA replication genes. The JH-receptor complex has been reported to act on Mcm4 and Mcm7 to regulate DNA replication and polyploidy for vitellogenesis and oocyte maturation on migratory locusts [22]. The processes of immunity and reproduction are not separable from the nutrition metabolism of the body. Of the DEGs mapped to up-regulated GO terms, 158 and 38 were involved in immunity and reproduction, respectively, which suggested that more energy resources were allocated to bacterial defense.
The 1488 significantly altered transcripts were enriched in the KEGG pathway. The pathways related to immunity and reproduction was extracted to analyze trade-off strategies between the two processes. The top mapped pathways, namely 'Endocytosis' (n = 83), 'Phagosome' (n = 33), 'Lysosome' (n = 26), and 'Regulation of autophagy pathway' (n = 11), were involved in the elimination of infectious bacteria, viruses, and aging cells, as well as in the absorption of macromolecules in nutrient metabolism. Five DEGs were mapped to the Tolllike receptor signaling pathway, which is conserved among animals [23], and can be activated through gram-positive triggers [24]. In the present study, the M. luteus infection stimulated the immune response of the locusts, and the appropriate immune factors were up-regulated significantly. A total of 32 transcripts were mapped to 'FoxO signaling pathway'. In the red flour beetle, the FoxO transcription factor can control vitellogenesis, thereby, reproduction by negatively regulating insulin-like peptide signaling, and FoxO can also activate AMPs independently of immune pathways [25]. The significant up-regulation of 'Insulin resistance' (n = 25) pathways promoted the FoxO signaling pathway [26] and inhibited nutritional metabolism, thereby reducing or even preventing oviposition [27]. These demonstrate that, at the biological pathway level, locusts allocate more energy into immune responses when faced with a tradeoff between immunity and reproduction. Eighteen DEGs were mapped to the mTOR signaling pathway. It has been reported that the insulin-like peptide/TOR pathway plays a role in transducing nutritional information that regulates JH synthesis in mosquitoes and, thereby, Transcriptome analysis of immunity-and reproduction trade-off in Locusta migratoria determines the outcome of trade-offs between survival and reproduction [28]. The above analysis indicated that immune and reproductive processes are closely related to the nutritional status of individual locusts. 'Insulin signaling pathway' (n = 4) was up-regulated by infection. Combined with the sequence information provided by the NCBI and locust genome databases [29], VgA (LOCMI12128-m1) and VgB (LOCMI12129-m1) were found to be down-regulated (log 2 FC = -17.82 and -18.13, respectively). Vitellogenin is the precursor of vitelline which is the main egg-yolk protein in insects. It is mainly formed in fat bodies and is absorbed by oocytes through blood circulation [30]. Reduced vitellogenin expression in the fat body inhibits ovarian development and oocyte maturation. In the present study, Vg expression was significantly down-regulated, which suggests that locusts, when challenged by infection, may devote the majority of resources to maintaining humoral circulation and survival and reduce investment in reproduction. Insect defensins are effector components of the innate defense system. In locusts, there are four genes that code for defensins, namely LmDEF1, 3, 4, and 5 [31]. In the present study, defensin 3 (LmDEF3, LOCMI01321-m1) was significantly up-regulated in the infected M. luteus group (log 2 FC = inf). The prophenoloxidase-activating cascade is one of the key components of arthropod immunity. Wounding or infection triggers the release of PPO1 through a process that involves JNK activation [32]. In the present study, PPO1 (LOCMI17352-m1) was up-regulated (log 2 FC = 5.93) in the infected group, which suggests that locusts are able to adapt to immune challenge. Furthermore, JH acid methyltransferase (LOCMI17613-m1) was also up-regulated (log 2 FC = 7.25) in infected group. This enzyme is involved in final steps of JH biosynthesis in insects [33], and it's up-regulation directly increases JH expression. In turn, JH plays an important role in regulating Vg synthesis and oocyte maturation [22,34], as well as innate immunity. It has been reported in the mealworm  Transcriptome analysis of immunity-and reproduction trade-off in Locusta migratoria beetle, Tenebrio molitor that reduce of phenoloxidase, a major humoral immune effector, was mediated by JH [35] and that JH is a hormonal immuno-suppressor [36]. At 6 d after emergence, locusts begin to store energy for reproduction. During this period, JH secretion increases normally, but in the infected locusts, it fails to promote enough Vg secretion. The up-regulation of AMPs and PPO1, the markers of humoral immunity, indicates that the body Transcriptome analysis of immunity-and reproduction trade-off in Locusta migratoria has invested a lot of energy resources into infection resistance by M. luteus. It is inferred that locusts, in the process of balancing resource allocation between immunity and reproduction, devote more resources to eliminating infectious bacteria and maintaining their own survival  than to reproduction. The molecular mechanisms of locusts trade-off of energy allocation between immunity and reproduction were represented by differentially changed GO functional groups and biological pathways. The relationship of IIS, JH, the phenoloxidase cascade, and AMP secretion in locusts should be investigated further.

Conclusion
In this study, transcriptome analysis was used to identify M. luteus infection on the transcription of locust ovaries and fat bodies during the reproductive preparation period. The differentially expressed transcripts indicated that the expression of PPO1 and defensin 3 increased significantly, whereas the expression of Vgs decreased significantly, and that locusts, when faced with the trade-off of immune response and reproduction, allocate most resources to the physiological process of resistance to infection. These results provide a foundation for further studies of the molecular mechanisms underlying immune and reproductive trade-offs in locusts and provide insight for the biological control of locusts.