Recent studies addressing experimental host–parasite coevolution and transgenerational immune priming in insects provide evidence for heritable shifts in host resistance or parasitevirulence. These rapid reciprocal adaptations may thus be transferred to offspring generations by either genetic changes or mechanisms that do not involve changes in the germline DNA sequence. Epigenetic inheritance refers to changes in gene expression that are heritable across generations and mediated by epigenetic modifications passed from parents to offspring. Highlighting the role of epigenetics in host–parasite coevolution, this review discusses the involvement of DNA methylation, histone acetylation/deacetylation and microRNAs in the interactions between bacterial or fungal parasites and model host insects such as the greater wax moth Galleria mellonella and the red flour beetle Tribolium castaneum. These epigenetic mechanisms are thought to participate in generation-spanning transcriptional reprogramming in the host insect, often linking immunity with developmentally related gene expression and contributing to the heredity of acquired adaptations. It is proposed that the interactions during host–parasite coevolution can therefore be expanded beyond reciprocal genetic changes to include reciprocal epigenetic changes. Epigenetics is thus a promising and prospering field in the context of host–parasite coevolution.
Coevolution in an antagonistic context results in reciprocal adaptations that are reflected at the genetic level (Anderson and May, 1982, Woolhouse et al., 2002, Decaestecker et al., 2007, Paterson et al., 2009). The genetic toolbox that enables reciprocal adaptations during host–parasite coevolution encompasses the gain, loss or neofunctionalization of genes mediating host resistance or parasite virulence. The comparative analysis of insect genomes has provided insights into the evolutionary dynamism of the immune system (Sackton et al., 2007, Waterhouse et al., 2007). These studies indicate that immunity-related effectormolecules such as antimicrobial peptides (AMPs) and proteinase inhibitors display greater evolutionary dynamism than conserved components, such as the receptors that mediate non-self-recognition and immunity-related signaling pathways (Vilcinskas, 2013). The range and diversity of genes encoding AMPs differs remarkably among different insect species. At one extreme is the invasive harlequin ladybird beetle (Harmonia axyridis) with more than 50 AMPs, the highest number reported in any animal species thus far (Vilcinskas et al., 2013), whereas at the other is the pea aphid (Acyrthosiphon pisum) which has lost the entire repertoire of known AMPs targeting bacteria as well as components of the immune deficiency (IMD) pathway (Altincicek et al., 2008, Gerardo et al., 2010). The range and diversity of AMPs in different insect taxa is supposed to reflect adaptations imposed by pathogens, parasites or microbial symbionts (Mylonakis et al., 2016). The comparative analysis of host transcriptomes also provides a powerful tool to study how genetic changesthat occur during host–parasite coevolution have a direct impact on gene expression(Greenwood et al., 2016). There is increasing evidence that changes in host susceptibility and resistance to pathogens can reflect the transcriptional reprogramming of immunity-related genes (Haase et al., 2016; Yang et al., 2016). Pathogen-induced changes in immunity-related gene expression can be passed from parents to offspring (Freitak et al., 2014, Knorr et al., 2015). Such generation-spanning transcriptional reprogramming does not necessarily require genetic changes because epigenetic mechanisms such as DNAmethylation, histone acetylation and the expression of microRNAs can also modulate gene expression in a heritable manner (Bingsohn et al., 2016). Epigenetic inheritance refers to changes in gene expression that are heritable across generations but are not caused by changes in the DNA sequence. Environmental factors including pathogens and parasites are known to interfere with epigenetic mechanisms, but there is only a limited amount of data concerning their role in host–parasite coevolution (Gomez-Diaz et al., 2012).
Empirical studies to investigate the role of epigenetics in host–parasite coevolution require appropriate model organisms (Brockhurst and Koskella, 2013). The greater wax mothGalleria mellonella and the red flour beetle Tribolium castaneum have been used in the DFG Priority Programme 1399 (Host–parasite coevolution − rapid reciprocal adaptations and their genetic basis) to address three major issues: (i) whether shifts in host resistance during experimental coevolution with bacteria or fungal parasites are associated with inherited changes in the expression profiles of immunity-related genes; (ii) whether transgenerational immune priming is associated with inherited changes in the expression profiles of immunity-related genes; and (iii) whether such transcriptional reprograming is mediated by epigenetic mechanisms. The latter are postulated to benefit the host in line with the Red Queen hypothesis if they are involved in the heredity of adaptations imposed by parasites, particularly those influencing generation-spanning shifts in resistance by reprogramming of immunity-related genes.
2. Experimental coevolution can trigger shifts in host resistance
The greater wax moth has become established as model host to explore host–parasite coevolution, particularly in the context of its arms race with parasitic fungi (Vilcinskas and Götz, 1999, Vilcinskas, 2011a; Joop and Vilcinskas, 2016). Our laboratory has characterized the immunity-related genes of G. mellonella, including those encoding AMPs, in great detail (Vilcinskas, 2011b, Vogel et al., 2011). We therefore investigated whether differences in pathogen resistance exist among populations of this species. A darker (melanic) G. mellonella population has been discovered that is more resistant to parasitic fungi than the widely distributed non-melanic morph. Comparative anatomical, physiological and molecular analysis of the morphs revealed that the melanic form has a thicker cuticle(entomopathogenic fungi can infect insect hosts directly through the cuticle), a greater number of immune-competent hemocytes circulating in the hemocoel, a greater capacity to entrap fungal cells in multicellular aggregates, higher phenoloxidase activity, and higher AMP expression levels (Dubovskiy et al., 2013a). These results suggest that the differential expression of immunity-related genes among populations of the same species can mediate corresponding changes in resistance against fungal parasites. Interestingly, the higher investment of the darker morph in immunity-related traits was accompanied by reductions in longevity, fecundity and biomass, indicating a trade-off between the investments in immunity and fitness-associated parameters (Dubovskiy et al., 2013a).
Experimental coevolution is a prospering field that allows the analysis of rapid reciprocal adaptations. Recent coevolution experiments with insect hosts provide evidence for rapid changes in complex parameters such as pathogen virulence and host resistance within a few coevolved host generations (Dubovskiy et al., 2013b, Rafaluk et al., 2015). Because changes in the nucleotide sequence are not always associated with such changes, we postulated the existence of non-genetic mechanisms with a transgenerational impact that can determine pathogen virulence and/or host resistance.
To understand the basis of these rapid adaptations, we exposed G. mellonella to selective pressure imposed by the parasitic fungus Beauveria bassiana in a generation-spanning experiment. Larvae selected for resistance over 25 generations were compared with unselected control larvae to see whether resistance against the parasite was heritable. Accordingly, the selected larvae showed greatly enhanced resistance against B. bassiana, but not against an alternative fungal parasite, Metarhizium anisopliae (Dubovskiy et al., 2013b). By comparing the selected (resistant) and unselected (susceptible) G. mellonellalarvae, we observed differences in the tissue-specific expression of AMPs and other genes related to immunity or stress responses. Interestingly, constitutive immune responses were suppressed in the resistant larvae, but the expression of AMPs was enhanced in the cuticle and epidermis, representing the front line against invading fungi (Dubovskiy et al., 2013b). Our observations agree with previous studies in other lepidopteran species (Freitak et al., 2009) and in coleopteran species such as the meal worm beetle Tenebrio molitor (Moret, 2006) and the red flour beetle (Roth et al., 2009) suggesting that strain-specific resistance priming is widespread in insects. However, the mechanisms underlying the transgenerational manifestation of immunity-related resource reallocation (resulting in species-dependent resistance against parasites) had so far remained unclear.
3. Transgenerational immune priming
Parental investment in insects includes the preparation of offspring to deal with pathogens or parasites that only the parents have encountered, a phenomenon described as transgenerational immune priming (Little et al., 2003, Little et al., 2005, Sadd et al., 2005; Milutinovic et al., 2016). The provision of lepidopteran or coleopteran species with diets containing bacteria (to mimic natural infection) is sufficient to increase the survival of the same cohorts when they are later exposed to the same pathogen, but also to induce transcriptional changes in their offspring (Freitak et al., 2009, Roth et al., 2009, Milutinovic et al., 2014, Trauer-Kizilelma and Hilker, 2015). To determine whether this effect is pathogen-specific, we supplemented the diets of G. mellonella larvae with different bacteria such as Escherichia coli, Micrococcus luteus, Pseudomonas entomophila and Serratia entomophila.The contaminated diets resulted in pathogen-specific changes in the expression of immunity-related genes both in the midgut tissues of the exposed larvae and in the eggs of the resulting adults. Comparative proteomic analysis of eggs from parents which were fed with either contaminated or non-conditioned food provided additional evidence for pathogen-specific immune priming, because unique protein spots were identified in the eggs produced by the challenged larvae (Freitak et al., 2014). Similar pathogen-dependent changes in the relative expression levels of immunity and stress-related genes were also observed in T. castaneum eggs produced by adults that were fed during the larval phase with either non-supplemented flour or flour supplemented with lyophilized E. coli, M. luteus or P. entomophila (Knorr et al., 2015). Such oral priming with bacteria-supplemented diets and the subsequent transcriptional reprogramming was translated into increased survival of the T. castaneum beetles (Milutinovic et al., 2014,2016). Interestingly, even beetles in the subsequent generation, which were not kept on diets contaminated with bacteria, underwent transcriptional reprogramming that was specific for the bacteria fed to the parental generation. Fig. 1 shows a comparative analysis using heat maps to illustrate two-dimensional clustering of the top 50 differentially expressed genes modulated by the addition of particular bacteria to the diet of the parent generation. The specific transgenerational transcriptomic reprograming raised two questions, i.e. how is the presence of particular bacteria in the larval diet passed as information to the next generation, and how is this information translated into specific immune responses?
The specificity of transgenerational immune responses inspired us to explore what happens to the bacteria in the supplemented diet after ingestion. We therefore added fluorescent bacteria to the diets of G. mellonella and T. castaneum larvae and monitored the fate of the bacteria by fluorescence microscopy. Surprisingly, we discovered that the labeled bacteria were able to cross the midgut epithelium and induce cellular immune response in the hemocoel. Most remarkably, we also found that ingested bacteria accumulated within the ovary and were ultimately deposited in the eggs (Fig. 2). The translocation of bacteria from the gut to the eggs provides an intriguing mechanism potentially explaining the specificity of maternal transgenerational immune priming in insects (Freitak et al., 2014, Knorr et al., 2015). Furthermore, this mechanism may also indicate why female adults of another lepidopteran model host (the tobacco hornworm Manduca sexta) express immunity-related genes at higher levels if they originate from primed rather than non-primed parents (Trauer-Kizilelma and Hilker, 2015). A recent study suggests that this mechanism may also explain maternal transgenerational immune priming against fungal parasites in insects (Fisher and Hajek, 2015). Another recent study shows that the gut microbiota influences oral immune priming in T. castaneum (Futo et al., 2016). However, our discovery does not provide a basis for paternal immune priming in T. castaneum, which is less specific than the maternal mechanism (Roth et al., 2010, Eggert et al., 2014). The fact that even male parents can pass information about parasites to their offspring calls for a mechanism that does not require the transfer of microbes or microbial components, but which can be transferred with the sperm. Epigenetic mechanisms may provide an alternative explanation but this requires experimental validation (Youngson and Whitelaw, 2008).
4. Epigenetic mechanisms
4.1. DNA methylation
The heredity of acquired resistance against pathogens or parasites does not necessarily require changes in the host germline DNA sequence. Instead, epigenetic mechanisms have been proposed to mediate heritable changes in transcriptional reprogramming across generations (Meagher and Müssar, 2012, Mukherjee et al., 2015). Epigenetic modificationsimposed by the presence of pathogens or parasites may be passed from parents to offspring, causing hereditary changes in gene expression profiles. This could be achieved before transcription by the modification of chromatin, e.g. by the methylation of DNA or histones, or the acetylation and deacetylation of histones (Fig. 3), whereas post-transcriptional gene regulation has been attributed to small non-coding RNAs known as microRNAs (miRNAs) (Jaenisch and Bird, 2003, Gomez-Diaz et al., 2012).
Gene expression in eukaryotes can be regulated by DNA methylation, which acts before transcriptional initiation. In animals, the addition or removal of methyl groups on cytidineresidues in the sequence CpG results in the formation of 5-methylcytidine, which changes the way DNA interacts with proteins thus providing a mechanism for gene regulation (Fig. 3). The transfer of methyl groups is mediated by evolutionarily conserved enzymes called DNA methyltransferases. The degree of DNA methylation in genomes can be quantified using enzymatic assays, whereas the identification of specific methylated genes requires more expensive approaches such as bisulfite sequencing.
Feliciello et al. (2013) provided the first evidence of DNA methylation in T. castaneum and showed that changes in DNA methylation status affect heterochromatin structure as a response to heat stress. To determine whether DNA methylation also plays a role in transgenerational immune priming in this species, we compared the DNA methylation status of eggs from beetles that were fed during the larval phase either on contaminated or sterile flour, as described above (Fig. 1). We were unable to detect significant differences in total DNA methylation levels between these cohorts, suggesting that DNA methylation is unlikely to control paternal immune priming in this species (Knorr et al., 2015). The observed specific transgenerational changes in gene expression profiles in beetles whose parents were fed at the larval stage with bacteria instead suggest the existence of more specific regulatory mechanisms (Fig. 1).
In contrast to the above, DNA methylation is likely to play a role in the immune responses of insects against parasitic fungi because we found that three different strains of M. anisopliaecaused the differential expression of DNA methyltransferase genes in G. mellonella larvaeinfected in a natural manner (Fig. 4). Interestingly, the induction or suppression of these methyltransferases in infected larvae relative to uninfected controls was dependent on the impact of the fungal strains on the development of G. mellonella. Infection of the larvae with strain 43 and 97 delayed pupation while strain 79 accelerated the formation of pupae(Mukherjee et al., 2012). Accordingly, we also observed a fungal strain-specific impact on another epigenetic mechanism that regulates gene expression in eukaryotes before transcriptional initiation, i.e. the acetylation and deacetylation of histones (Mukherjee et al., 2012).
4.2. Histone acetylation and deacetylation
The acetylation of histones by histone acetyltransferases (HATs) generates a chromatin structure that increases the accessibility of DNA and promotes gene expression, whereas the removal of acetyl groups by histone deacetylases (HDACs) generates a condensed form of chromatin that silences gene expression. Two groups of enzymes with opposing activities thus modify chromatin structure and thereby control the ability of transcription factors to access DNA (Fig. 3).
We investigated the role of histone acetylation in host–parasite coevolution by screening our recently established comprehensive G. mellonella transcriptome database (Vogel et al., 2011) for the presence of genes encoding either HATs or HDACs. This allowed us to design real-time PCR primers for the quantification of HAT/HDAC gene expression during infection with bacteria or fungi. The topical infection of G. mellonella larvae with M. anisopliae conidiato experimentally mimic a natural infection induced HDACs more strongly than HATs, suggesting that the fungus suppresses host gene expression during the initial phase of pathogenesis (Mukherjee et al., 2012). This observation supports previous studies providing evidence that parasitic fungi suppress humoral immune responses in G. mellonella(Vilcinskas and Matha, 1997). In contrast, fungal cells or fungal cell wall preparations such as zymosan elicit strong immune responses when injected into G. mellonella (Gillespie et al., 2000). Taken together, these data suggest that parasitic fungi actively suppress immune responses in their insect hosts by interfering with the epigenetic regulation of immunity-related genes.
To determine whether bacterial pathogens can also interfere with host gene regulation by controlling histone acetylation, we injected G. mellonella with either non-pathogenic E. coli or virulent Listeria monocytogenes. The latter are food-borne human pathogens and we have established G. mellonella as an alternative and ethically more acceptable model host (compared to mammals) in which to study pathogenesis (Mukherjee et al., 2013). Both bacterial species induced the expression of HAT and HDAC genes, but only the virulent L. monocytogenes shifted the balance between these opposing enzymes in favor of histone deacetylases, suggesting that pathogenesis involves the active suppression of genes regulated by histone acetylation (Mukherjee et al., 2012).
Interestingly, the infection of G. mellonella with virulent L. monocytogenes or parasitic fungi delays larval development, whereas the injection of heat-killed fungal or bacterial cells accelerates the onset of metamorphosis (Vilcinskas and Matha, 1997, Mukherjee et al., 2012). Assuming that gene expression during metamorphosis is regulated by histone acetylation, we measured the expression of HAT and HDAC genes in the prepupae and pupae and found that these genes were induced during metamorphosis and the opposing activities were carefully balanced (Mukherjee et al., 2012). These observations inspired our hypothesis that parasites may suppress the expression of immunity-related genes that are regulated by histone acetylation, thereby also suppressing developmental genes controlled by the same epigenetic mechanism. To experimentally validate this hypothesis, we injected specific inhibitors of HATs and HDACs into G. mellonella larvae and observed striking diametric effects on the onset of metamorphosis and the expression of immunity-related genes. The administration of HDAC inhibitors accelerated larval development whereas HAT inhibitors delayed the onset of metamorphosis compared to untreated control larvae (Mukherjee et al., 2012). These data taken together indicate that histone acetylation/deacetylation co-regulates development and immunity. The injection of bacteria (or fungal elicitors of innate immune responses) into G. mellonella larvae causes the formation of open chromatin and facilitates the expression of immunity-related genes and developmental genes that accelerate metamorphosis. The manipulation of histone acetylation by parasites to suppress host immune responses thus causes a collateral delay in development (Mukherjee et al., 2012). The impact of parasites on histone acetylation may also influence other complex parameters, such as fecundity and longevity. For example, it has been shown in M. sexta that transgenerational immune priming benefits the larvae of the offspring, but inhibits reproduction in the adults (Trauer and Hilker, 2013).
We investigated whether histone acetylation plays a role in the transcriptional reprogramming associated with transgenerational immune priming in insects by feeding G. mellonella larvae with either E. coli or the entomopathogen S. entomophila. The latter tipped the balance between HDAC and HAT expression in the midgut of infected larvae in favor of HDACs, and strongly induced genes encoding HDACs or HATs in the eggs laid by parents fed with bacteria compared to eggs from naïve parents (Mukherjee et al., 2015). This indicates that histone acetylation does indeed play a role in the process of transgenerational immune priming in insects and solely contributes to transgenerational inherited chromatin structures (Meagher and Müssar, 2012).
MicroRNAs (miRNAs) are short non-coding RNAs (∼18–24 nucleotides in length) that regulate gene expression in eukaryotes at the post-transcriptional level by base-pairing with the 3′ untranslated regions (UTRs) of target messenger RNAs (mRNAs). The number of miRNAs based on computational predictions is greater than the number confirmed by experimental testing, but more than 30% of animal genes could be regulated by miRNAs and a single miRNA may control hundreds of different target genes (Bushati and Cohen, 2007, Sato et al., 2011). Recent evidence suggests that miRNAs are involved in the regulation of parasite virulence and host immunity (Harris et al., 2013). Our current knowledge of insect miRNAs was recently summarized by Asgari (2013).
The regulation of the mammalian immune system by miRNAs has been recognized (Xiao and Rajewsky, 2009), whereas little is known about the role of miRNAs in insect–pathogen interactions with the exception of antiviral responses (Hussain and Asgari, 2014). The potential role of miRNAs in our model host G. mellonella was therefore tested by infection with the parasitic fungus M. anisopliae or the entomopathogenic bacterium S. entomophilausing a microarray spotted with more than 2000 insect miRNA probe sequences (Mukherjee and Vilcinskas, 2014). Taking advantage of our comprehensive transcriptome database (Vogel et al., 2011), we identified 3′ untranslated regions that potentially form miRNA–mRNA duplexes by considering base pair complementarity, minimum free energy hybridization and the coexpression of selected miRNAs with their predicted target mRNAs. This approach enabled us, for the first time, to identify insect miRNAs that are differentially regulated during infection with a parasitic fungus. For example, dps-miR-200b was silenced during infection with M. anisopliae whereas its target genes were correspondingly upregulated (Mukherjee and Vilcinskas, 2014). Other miRNAs, such as api-miR-263a, were induced by infection with S. entomophila, providing evidence for the participation of miRNAs in the pathogen-specific regulation of target genes. Interestingly, we discovered that api-miR-263a was also more abundant in eggs from female imagoes which were fed as larvae with bacteria, suggesting that miRNAs also mediate transgenerational immune priming and host–parasite coevolution (Mukherjee and Vilcinskas, 2014).
We also used microarrays to quantify the differential expression of miRNAs in T. castaneumreared on diets supplemented with S. entomophila to mimic natural infection, or injected with peptidoglycans, revealing the experimental induction of both immunity-related genes and miRNAs (Freitak et al., 2012). The exposure of T. castaneum to infection, starvation or mild heat shock caused a striking stressor-dependent and sex-dependent response, with a much larger number of upregulated miRNAs in females. This suggests that miRNAs may play a key role in maternal transgenerational immune priming, which is more specific in terms of protection against particular species of pathogens and parasites (Freitak et al., 2012). On the other hand, the postulated role of miRNAs in the crosstalk between stress and immune responses may also explain why a temperature shock can mediate trans-generational immune priming in T. castaneum (Eggert et al., 2015). Taken together, our observations support the postulated role of miRNAs in the epigenetic regulation of gene expression during infections (Sato et al., 2011, Harris et al., 2013).
5. Concluding remarks
Recent studies provide evidence for the participation of epigenetic mechanisms such as DNA methylation, histone acetylation/deacetylation and microRNAs in generation-spanning transcriptional reprogramming associated with shifts in host resistance during host–parasite coevolution. Epigenetic inheritance in this antagonistic context is in line with Red Queen dynamics because it provides a complementary mechanism for rapid fixation and vertical transmission of adaptations and counter-adaptations if they are mediated via transcriptional reprogramming. Further studies are needed to determine how epigenetic mechanisms such as DNA methylation and miRNAs contribute to the arms race between parasites and their hosts as well as the generation-spanning fixation of shifts in complex parameters such as pathogen virulence or host resistance. I am convinced by the evidence summarized above that interactions during host–parasite coevolution can be expanded beyond reciprocal genetic changes to include reciprocal epigenetic changes.
The author acknowledges funding provided by the German Research Foundation for the project “The role of epigenetics in host–parasite coevolution” (VI 219/3-2) which is embedded within the DFG Priority Programme 1399 “Host–parasite coevolution − rapid reciprocal adaptations and its genetic basis”. He thanks the following colleagues for providing unpublished data: Dr. Dalial Freitak (University of Helsinki) for providing the heat map shown in Fig. 1, Dr. Henrike Schmidtberg (Justus-Liebig University of Giessen) for the microscopy images shown in Fig. 2, Dr. Krishnendu Mukherjee (Fraunhofer Institute for Molecular Biology and Applied Ecology) for Fig. 4, and Dr. Richard M. Twyman for editing the manuscript.
Read more:- https://www.microbiologyjournal.org