Transcriptomics reveals specific molecular mechanisms underlying transgenerational immunity in Manduca sexta

Abstract The traditional view of innate immunity in insects is that every exposure to a pathogen triggers an identical and appropriate immune response and that prior exposures to pathogens do not confer any protective (i.e., adaptive) effect against subsequent exposure to the same pathogen. This view has been challenged by experiments demonstrating that encounters with sublethal doses of a pathogen can prime the insect's immune system and, thus, have protective effects against future lethal doses. Immune priming has been reported across several insect species, including the red flour beetle, the honeycomb moth, the bumblebee, and the European honeybee, among others. Immune priming can also be transgenerational where the parent's pathogenic history influences the immune response of its offspring. Phenotypic evidence of transgenerational immune priming (TGIP) exists in the tobacco moth Manduca sexta where first‐instar progeny of mothers injected with the bacterium Serratia marcescens exhibited a significant increase of in vivo bacterial clearance. To identify the gene expression changes underlying TGIP in M. sexta, we performed transcriptome‐wide, transgenerational differential gene expression analysis on mothers and their offspring after mothers were exposed to S. marcescens. We are the first to perform transcriptome‐wide analysis of the gene expression changes associated with TGIP in this ecologically relevant model organism. We show that maternal exposure to both heat‐killed and live S. marcescens has strong and significant transgenerational impacts on gene expression patterns in their offspring, including upregulation of peptidoglycan recognition protein, toll‐like receptor 9, and the antimicrobial peptide cecropin.

One mechanism of transgenerational immunity which has been experimentally validated in insects involves the translocation of bacteria (Freitak et al., 2014) or bacterial cell wall components from mothers to their offspring (Salmela et al., 2015). In the honeycomb moth Galleria mellonella, Freitak et al. (2014) used fluorescently labeled, heat-killed Escherichia coli to trace the transfer of bacteria from the guts of mothers through the hemocoel and ovarioles into their developing eggs. They reported upregulation of prophenoloxidase, peptidoglycan recognition protein, glutathione-s-transferase, and lipopolysaccharide-binding protein in eggs of immune-challenged mothers. Salmela et al. (2015) identified vitellogenin protein as the carrier required for the transport of E. coli cell wall fragments into the developing eggs of the European honeybee Apis mellifera. In addition, several studies provide evidence of epigenetic transmission. Eggert, Kurtz, and Diddens-de Buhr (2014) reported that paternal transgenerational immunity in the beetle Tribolium castaneum exposed to Bacillus thuringiensis is passed via sperm, and to a lesser degree seminal fluid, suggesting that epigenetic modifications were involved.
A similar upregulation of fat body proteins has been documented in response to gram-negative Photorhabdus spp. include hemolin, immulectin-2, and peptidoglycan recognition protein (Eleftherianos, Millichap, Ffrench-Constant, & Reynolds, 2006). Ao, Ling, and Yu (2008) also reported an upregulation of toll-like receptor in Manduca exposed to gram-negative E. coli.
Evidence for transgenerational immunity in Manduca sexta has also been recently documented in the offspring of mothers exposed to heat-killed and live Serratia marcescens bacteria (Rosengaus et al., 2017). Rosengaus et al. (2017) demonstrated that maternal pathogen exposure significantly affected in vivo bacterial clearance by their offspring. Using an in vivo "clearance of infection" assay, they showed that first-instar larvae, offspring of Manduca females injected with either live or heat-killed S. marcescens had significantly lower microbial loads 24 and 48 hr after injection of live Serratia than offspring of females injected with a sterile saline control. Here, we used transcriptome-wide RNA sequencing to identify gene expression changes underlying the transgenerational phenotypic effects reported by Rosengaus et al. (2017). We compared gene expression patterns in fat body and ovariole tissues of mothers exposed to live and heat-killed S. marcescens and then compared gene expression patterns in their embryos to identify any transgenerational impacts on gene expression due to maternal pathogen exposure.
The gram-negative bacterium, S. marcescens, was chosen to elicit a maternal immune response because it is an ecologically relevant pathogen, commonly found on foliage and in soil (Sikorowski, Lawrence, & Inglis, 2001). Serratia marcescens is likely also encountered by developing larvae during herbivory and during the fifth and final larval stage when larvae wander and/or when they pupate subterraneously. In addition, our prior experiments on pathogen-induced maternal effects, which resulted in in vivo evidence of enhanced immune responsiveness across generations, were also carried out with the same strain of Serratia as this work (Rosengaus et al., 2017). It was important for us to maintain consistency with regard to the pathogenic strain so that we could transpose the current molecular data onto the in vivo clearing assay results of 2017.
Although previous work on Manduca included transcriptomic analyses of specific genes (Lee & Horodyski, 2006) in the context of starvation and mating, transcriptome-wide analyses for sex chromosome differences (Smith, Chen, Blissard, & Briscoe, 2014) and within-generation immune challenges (Van Munster et al., 2007), and phenotypic analyses of TGIP (Rosengaus et al., 2017) the present work, based on genome-wide transcriptomic analyses, provide more nuanced information with respect to gene expression changes involved in TGIP.

| ME THODS
Manduca sexta eggs were obtained from Carolina Biological Supply (Burlington, NC), then reared, and treated following the same protocol used by Rosengaus et al. (2017). Briefly summarizing, larvae were reared on a standard artificial diet (Bell & Joachim, 1976) at 25°C under a 16-hr:8-hr light/dark cycle. Two days prior to expected eclosion, female pupae were swabbed with 70% ethanol and injected with sterile saline (10 μl, n = 8), heat-killed Serratia (10 μl at 1 × 10 8 cells/ ml, n = 8), or live Serratia (10 μl at 4 × 10 5 CFU/ml, n = 8) using a 10 μl Hamilton syringe and sterile needle. After eclosion, twelve of the treated females (4 each from saline, heat-killed, and live exposures) were sacrificed and RNA was extracted from ovarioles and fat bodies to profile their transcriptomic response to pathogen exposure.
Total RNA was extracted from fat bodies and ovarioles of each adult female (n = 4 from the saline-treated mothers, n = 4 from the heatkilled mothers, and n = 4 from the live S. marcescens-injected mothers) using the Promega SV Total RNA Isolation Kit.

| Annotation
To facilitate KEGG pathway analysis, differentially expressed transcripts were mapped to KEGG ortholog IDs. The transcripts were aligned to Swiss-Prot (The Uniprot Consortium, 2017) using blastx (Camacho et al., 2009). Swiss-Prot hits were filtered using an e-value cutoff of 1e−5 and matched to KEGG orthologs using the KEGG API.
For each Manduca transcript, the KEGG ortholog corresponding to the lowest BLAST e-value was selected.

| Transcriptome-wide analyses
Although the independent PERMANOVAs did not identify signif-

| Gene-level analyses
Negative binomial GLMs were run separately for the two maternal tissues (fat body and ovarioles) as well as the embryos in order to identify genes that were differentially expressed due to heat-killed or live Serratia exposure compared to the saline control (Table 2).
Relative to the maternal saline treatment, fat bodies of heat-killed

| D ISCUSS I ON
Our results showed a strong upregulation of genes in the fat bodies, but not ovarioles of adult female moths in response to injections of heat-killed and live Serratia. Moreover, we observed a stronger upregulation of immune-related genes in embryos from heat-killed Serratia-injected mothers than that from embryos exposed to live bacteria. This may be a result of mothers dealing with an active (i.e.,

| Shared upregulated adult and offspring xenobiotic genes
There was very little overlap of shared DE xenobiotic genes detected in the adults and transgenerationally primed embryos. Only carbonyl reductase 3 (CBR3/Msex2.09445) was upregulated in adults treated with live Serratia and embryos whose mothers were exposed to heat-killed Serratia. Upregulation of carbonyl reductases has been reported in gypsy moth (Lymantria dispar) on protein-deficient diets (Lindroth, Barman, & Weisbrod, 1991). Upregulation of carbonyl reductase 3 has also been reported in human cancer cells in response to oxidative stress (Ebert, Kisiela, Malátková, El-Hawari, & Maser, 2010). To the best of our knowledge, this is the first report of upregulation in response to immune challenge in Lepidoptera.

| Upregulated adult-only immune and xenobiotic genes
The adult-only response was characterized by strong upregulation of xenobiotic genes in mothers injected with live bacteria. Cytochrome .13295) were upregulated in adults exposed to live Serratia.
CYP12 and CYP6 are members of the cytochrome P450 family involved in oxidation/reduction of organic chemicals including drugs, environmental toxins, and carcinogens in humans (Guengerich, Waterman, & Egli, 2016). Upregulation of cytochrome P450 has been reported in several invertebrates including silkworms (Bombyx mori) challenged with B. thuringiensis (Wu & Yi, 2018), flour beetles (T. castaneum) challenged with LPS (Altincicek et al., 2013), and abalone (Haliotis diversicolor) challenged with gram-negative and gram-positive bacteria (Wang, Ren, Xu, Cai, & Yang, 2008). The role upregulation of cytochrome P450 plays in the insect immune response is unclear, but Wu and Yi (2018) proposed that this may be a detoxification response to protect the host from intermediate F I G U R E 3 Differentially expressed genes by tissue (fat body, ovariole, nd embryo), treatment group (saline, heat-killed, and live), and direction (upregulated or downregulated). Fat body had the greatest number of differentially expressed genes, the largest number being upregulated in response to live Serratia exposure. Second to fat body in terms of number of differentially expressed genes was embryo, the largest number being upregulated in response to maternal treatment with heat-killed Serratia. Few genes were differentially expressed for ovariole alyzes the attachment of sugars to toxins to facilitate their excretion (Meech, Miners, Lewis, & Mackenzie, 2012). Upregulation of UGT in response to immune challenge has been reported in Drosophila exposed to E. coli (Johansson, Metzendorf, & Söderhäll, 2005), and perhaps its upregulation helps Manduca mothers detoxify the toxins produced by Serratia.
Only one immune gene was strongly upregulated in mothers injected with live bacteria-interferon gamma-inducible protein 30 (GILT). GILT is involved in MHC antigen processing in mammals (Jensen, 1993), and upregulation of GILT in response to immune challenge has been identified in several invertebrates including disk abalone exposed to gram-negative Vibrio alginolysticus (Zoysa & Lee, 2007), mosquito exposed to the malarial parasite Plasmodium falciparum (Schleicher et al., 2018), and fruit fly exposed to gram-negative E. coli (Kongton et al., 2014).

F I G U R E 4
Heatmap normalized gene expression of differentially expressed (DE) genes in the adult fat body with a log 2 -fold change (LFC) greater than 2 or less than −2. Exposure treatments (saline, live Serratia, and heat-killed Serratia) color coded across top of heatmap. Hierarchical clustering of genes shown along vertical axis with immune genes and xenobiotic genes coded in green and blue, respectively

| Downregulated adult-only immune and xenobiotic genes
Carboxylesterase 2 (CES2/Msex2.07504) and cathepsin B (CTSB/ Msex2.01721) were downregulated within fat body in mothers exposed to live Serratia. It is unclear why these two genes were downregulated as we find no comparative studies showing a similar pattern.
CES2 catalyzes the metabolism of ester and pyrethroid toxins (Wang et al., 2018). Cathepsins are proteases expressed in lysosomes. CTSB is highly expressed in the fat body of B. mori during the larval-pupal transformation, and it is involved in the programmed cell death of the fat body during metamorphosis (Lee et al., 2009). In addition to its role in metamorphosis, cathepsin B has been associated with the response to immune challenge. Wu et al. (2011) reported upregulation of cathepsins B and D in B. mori challenged with B. mori nuclear polyhedrosis virus (BmNPV). Our results showed a counter-intuitive pattern in M. sexta (downregulation) relative to results reported by Wu et al. (2011) in the closely related species B. mori, but this downregulation could be a host response to the known immunosuppressive effects of S. marcescens (Ishii et al., 2014).

F I G U R E 5
Heatmap normalized gene expression of differentially expressed (DE) genes in the embryos with a log 2 -fold change (LFC) greater than 2 or less than −2. Exposure treatments (saline, live Serratia, and heat-killed Serratia) color coded across top of heatmap. Hierarchical clustering of genes shown along vertical axis with immune genes and xenobiotic genes coded in green and blue, respectively  Msex2.01919, whose best BLAST hit was for mammalian TLR9, was upregulated in embryos of mothers exposed to heat-killed Serratia. In mammals, toll-like receptors are primarily involved in immune responses, while in insects, toll-like receptors contribute to both development and immune response roles (Imler & Zheng, 2004 (Ao et al., 2008).

| Signatures of TGIP in Embryos
Plasminogen activator inhibitor 1 (PAI-1) was upregulated in embryos of mothers exposed to heat-killed Serratia. PAI-1 is a serine protease inhibitor involved in hemostasis in mammals (Lijnen, 2005).
PAI-I is upregulated in humans during gram-negative sepsis caused by Burkholderia pseudomallei and protects the host by limiting bacterial growth, inflammation, and coagulation (Kager et al., 2011). We believe we are the first to report transgenerational PAI-1 upregulation in invertebrates, and it is possible that PAI-1 plays a similar protective anticoagulation role in Manduca.

| CON CLUS ION
Maternal exposure of M. sexta to both heat-killed and live S. marcescens had strong and significant transgenerational impacts on gene expression patterns of their offspring, and these patterns include upregulation of genes that could play a role in the transgenerational phenotypic effects reported by Rosengaus et al. (2017). This combination of phenotypic and transcriptomic transgenerational effects adds to the growing body of evidence for transgenerational immune priming in insects. Our results indicate that immune priming includes upregulation of genes associated with pathogen recognition, pathogen elimination, and modulation of downstream effects like coagulation. Further exploration is warranted to determine the mechanisms that drive these TGIP gene expression changes.

ACK N OWLED G M ENTS
We thank James Kemos, James MacArthur, Emily Moore, Shannon Rudolph, Matt Simhon, and Jamaica Siwak for their assistance in Manduca rearing and tissue sampling. This work was partially funded by a TIER1 award, Northeastern University.

CO N FLI C T O F I NTE R E S T
None declared. formal analysis (equal); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); supervision (lead); validation (equal); writing -review and editing (equal).