Oncopeltus-like gene expression patterns in Murgantia histrionica, a new hemipteran model system, suggest ancient regulatory network divergence.

Background
Much has been learned about basic biology from studies of insect model systems. The pre-eminent insect model system, Drosophila melanogaster, is a holometabolous insect with a derived mode of segment formation. While additional insect models have been pioneered in recent years, most of these fall within holometabolous lineages. In contrast, hemimetabolous insects have garnered less attention, although they include agricultural pests, vectors of human disease, and present numerous evolutionary novelties in form and function. The milkweed bug, Oncopeltus fasciatus (order: Hemiptera)-close outgroup to holometabolous insects-is an emerging model system. However, comparative studies within this order are limited as many phytophagous hemipterans are difficult to stably maintain in the lab due to their reliance on fresh plants, deposition of eggs within plant material, and long development time from embryo to adult.


Results
Here we present the harlequin bug, Murgantia histrionica, as a new hemipteran model species. Murgantia-a member of the stink bug family Pentatomidae which shares a common ancestor with Oncopeltus ~ 200 mya-is easy to rear in the lab, produces a large number of eggs, and is amenable to molecular genetic techniques. We use Murgantia to ask whether Pair-Rule Genes (PRGs) are deployed in ways similar to holometabolous insects or to Oncopeltus. Specifically, PRGs even-skipped, odd-skipped, paired and sloppy-paired are initially expressed in PR-stripes in Drosophila and a number of holometabolous insects but in segmental-stripes in Oncopeltus. We found that these genes are likewise expressed in segmental-stripes in Murgantia, while runt displays partial PR-character in both species. Also like Oncopeltus, E75A is expressed in a clear PR-pattern in blastoderm- and germband-stage Murgantia embryos, although it plays no role in segmentation in Drosophila. Thus, genes diagnostic of the split between holometabolous insects and Oncopeltus are expressed in an Oncopeltus-like fashion during Murgantia development.


Conclusions
The similarity in gene expression between Murgantia and Oncopeltus suggests that Oncopeltus is not a sole outlier species in failing to utilize orthologs of Drosophila PRGs for PR-patterning. Rather, strategies deployed for PR-patterning, including the use of E75A in the PRG-network, are likely conserved within Hemiptera, and possibly more broadly among hemimetabolous insects.


Introduction
Much of our understanding of the genetic and molecular mechanisms underlying animal development, morphology and physiology is based on studies of a handful of well-established model systems. Approaches in comparative biology have built upon discoveries from these model systems, revealing highly conserved genes and pathways shared across phyla. For example, based on work in the pre-eminent model organism Drosophila melanogaster, this approach revealed unexpected and conserved roles of Hox genes in embryonic development and Pax genes in eye development throughout Metazoa (reviewed in [1][2][3][4][5]). However, to discover novel genes and mechanisms, new model systems are required [6,7]. Insects are ideal model systems not only because of their short life cycles and embryos that can be collected in large numbers and readily manipulated, but also because insects represent the majority of terrestrial species on our planet, occupy diverse habitats on land and water, and display extensive diversity in morphology, behavior and more [8]. Most emerging insect model systems have focused on the group of holometabolous insects, to which Drosophila belongs (for examples, see [9][10][11][12][13]).
Hemiptera, a close outgroup to holometabolous insects, is a large order of insects, with > 85,000 species identified to date [14]. The piercing-sucking feeding mechanism seen within this clade enables many species to feed on plants, making many of them agricultural pests (e.g., aphids, scale insects, white flies, kudzu bug, stink bugs, psyllids), and even on humans in the case of kissing bugs, allowing them to vector human disease. The milkweed bug, Oncopeltus fasciatus, is easily reared in the lab on a simple diet of sunflower seeds and water, and has recently emerged as an excellent model system for this group of hemimetabolous insects. It was noted as early as the 1970s that " [t]here are very few basic biological or entomological problems for which the milkweed bug would not be a good experimental animal" [15]. In fact, prior to the explosion of Drosophila molecular genetics, Oncopeltus was used as a model for developmental studies [16]. More recently, the development of tools to examine and manipulate gene expression [17][18][19] and publication of transcriptomes and genome sequences [20][21][22] has moved Oncopeltus to "prime model status" [23]. Because of the practical importance of Hemiptera, the diversity within this order, its phylogenetic position, and because progress on Oncopeltus has revealed novelties in morphology and molecular mechanisms (see below), we are seeking to establish additional models within this clade and report here the development of the harlequin bug, Murgantia histrionica, as a new hemipteran model species. Unlike Oncopeltus, Murgantia is a major pest of cruciferous vegetables and thus the development of genetic tools for this species may be useful for pest control.
Research on early embryonic body plan establishment in Drosophila has revealed that sets of regulatory genes act in a hierarchy to sequentially subdivide the embryo into increasingly specified body regions, culminating in the establishment of repeated segmental units along the anterior-posterior axis [24]. Mutations in the pair-rule genes (PRGs) result in loss of alternate body segments and most PRGs are expressed in stripes in the primordia of every other segment (PR-stripes)-the alternate segmental units missing in corresponding mutants (e.g., [25][26][27]. Segment polarity genes, the next set of genes to act in the regulatory hierarchy, impact the development of equivalent portions of each body segment and are expressed in stripes in the primordia of every segment (segmentally). PRGs regulate the expression of segment polarity genes such as engrailed and wingless, with alternate segmental stripes missing in PRG mutants [28,29]; for example, in Drosophila, Fushi tarazu (Ftz) directly regulates the expression of alternate engrailed stripes [30,31].
Drosophila is a holometabolous insect with a derived mode of segmentation in which all segments are specified more or less simultaneously during the blastoderm stage. However, in most insect orders, segments are added sequentially to the germband, with only the most anterior segments specified at the blastoderm stage (sequential segmentation, reviewed in [18,32,33]). Given this major change in the process of segment formation between Drosophila and more basally branching insects, it seemed likely that the genetic mechanisms underlying segment formation would likewise differ. However, the finding that orthologs of many PRGs are expressed and function as typical PRGs-specifying alternate segmental units-in the sequentially specifying beetle Tribolium castaneum, demonstrated that PR-patterning per se is conserved in sequentially segmenting species. Further, many of the genes involved in this process play similar roles despite differences in the morphological steps involved in segment addition [34]. In particular, in beetles, orthologs of two PRGs (paired (prd) and sloppy-paired (slp)) function in the same way as their Drosophila orthologs, with knockdown leading to 'classical' PR-defects in which alternate segments are missing [34][35][36]. Other PRG orthologs have dual roles in sequential segmentation: three (even-skipped (eve), runt (run), and odd-skipped (odd)), which are expressed and function exclusively in PR-stripes in Drosophila, are expressed in the segment addition zone (SAZ) in sequentially segmenting holometabolous insects in addition to being expressed in PR-stripes in these species [34,37].
Strong knockdown of these genes results in truncated embryos, with extreme cases of 'head-only' embryos resulting after knockdown of eve in beetles, while weaker knockdown, in which the germband was able to elongate, revealed additional 'classical' PR-like roles for these genes [34,37]. Additionally, expression data from several other holometabolous insects indicate conserved roles for several orthologs of Drosophila PRGs in pair-rule patterning. For example, eve, run, and prd in the honey bee Apis mellifera; eve, odd, and run in the silk worm Bombyx mori and the parasitoid wasp Nasonia vitripennis, were also shown to have PR-expression patterns [10,[38][39][40][41][42]. The expression and/or function of other PRGs (ftz, hairy, odd-paired and ftz-f1) varies even within holometabolous insects [34,37,43,44]. In sum, within the Holometabola, orthologs of five PRGs-eve, odd, prd, slp, and run-are expressed and function in PR-stripes in the primordia of alternate segmental units in broadly divergent species within this clade.
In contrast to this, recent studies in the large milkweed bug, Oncopeltus fasciatus (Hemiptera: Lygaeidae), revealed that orthologs of Drosophila PRGs are expressed segmentally, rather than in a PR-like manner, more similar to segment polarity genes in Drosophila [45,46]. Yet PR-like patterning does occur in this species; Of-E75A is expressed in PR-stripes and knockdown resulted in PR-like defects [47]. Interestingly, E75A does not have PR-function in Drosophila [48,49]. Thus, PR-patterning appears to involve a different set of regulatory genes in Oncopeltus than in holometabolous insects. To ask whether these differences in the expression and function of PRG-orthologs in Oncopeltus compared to holometabolous insects is species-specific or a feature of other hemipterans, we used the harlequin bug, Murgantia histrionica, as a model system. With an estimated divergence time of ~ 200 million years from Oncopeltus [50][51][52], Murgantia and Oncopeltus are distant relatives both belonging to the order Hemiptera and infraorder Pentatomomorpha (Fig. 1a), a species-rich group with over 40,000 members [53,54]. Here we show that Murgantia can be adapted to laboratory conditions and maintained as a breeding colony for many generations. We focused on a set of genes that are diagnostic of differences between holometabolous insects and Oncopeltus (E75A,  [51]. (i) A condensed cladogram that includes some insect orders. Murgantia belongs to the order Hemiptera (colored orange) (ii) Shows a cladogram of Hemiptera, including some major suborders: Sternorrhyncha, Auchenorrhyncha and Heteroptera. Murgantia is a member of Heteroptera and the Pentatomomorpha superfamily (orange text). (iii) Murgantia' s relationship to Oncopeltus fasciatus (large milkweed bug) and Halyomorpha halys (brown marmorated stink bug). b Laboratory cage set up for Murgantia. Embryos are allowed to hatch and grow to the second instar in a small petri dish with a piece of wet cotton and kale. They are provided kale three times a week and kept in mesh cages (12 × 12 × 12 in) at 25 °C. A source of water can be provided (e.g. a wet cotton). c The life cycle of Murgantia. Our lab colony was observed to have five nymphal stages eve, odd, runt, prd, and slp) and found that Murgantia orthologs are expressed in an Oncopeltus-like fashion. This suggests that the roles of these genes are shared among Hemiptera and possibly other hemimetabolous insects. In addition to its utility for basic research, Murgantia is a pest of cruciferous vegetables and progress on molecular genetics approaches in this species can lead to novel pest control strategies.

Establishment of Murgantia as a lab model system
To serve as a viable model system for molecular genetic studies, organisms must be amenable to long term lab culturing, have a reasonably short life cycle, produce a large number of embryos, and molecular protocols must work well [6]. Before proceeding with genetic analyses, we assessed the feasibility of rearing Murgantia in the lab. We established lab colonies from field-collected Murgantia in summers of 2016 and 2017 and have maintained that colony continuously since that time. As described in the Methods, Murgantia were reared on a diet of fresh organic kale and cabbage leaves with wet cotton as a water source at 25 °C, 50% relative humidity, 16:8 light: dark cycle (Fig. 1b). Males can be readily distinguished from females by the lateral lobes of the genital capsule which are externally visible (arrows, Additional file 1: Fig. S1). Adults typically began mating 10-15 days after their last molt. In each cage, adults began laying eggs soon thereafter, first 1-2 clutches (12-24 eggs) per cage in 24 h and up to 7-8 clutches (84-96 eggs) in 24 h during their egg-laying peak, decreasing after 3 weeks. Embryos usually hatched 5-6 days after egg laying (AEL). A total of five nymphal stages were observed (Fig. 1c). The first instar lasted 4-5 days; the second, third and fourth instars each lasted 6-7 days; and the fifth and final nymphal instar lasted about 11-13 days. Thus, under our rearing conditions, development from egg to adult took ~ 38-45 days. While rearing on fresh food was effective, we have also tested whether Murgantia can be grown on a diet of seeds, similar to Oncopeltus. Preliminary results show that adults can survive on water and seeds (spider flower plus sunflower) though fecundity appears to be lower than those reared on kale/collard greens and cabbage (data not shown).
In order to understand embryogenesis in Murgantia, we performed nuclear staining of embryos carefully staged at 25 °C (Fig. 2). During the first 12 h AEL, nuclear division occurs and nuclei were observed at the periphery of the yolk. At the end of the first 24 h, invagination of the germband began at the posterior with cells aggregating at this site (Fig. 2a, arrowhead). For the next 24 h, the germband elongated along the ventral side of the embryo with segments added at the posterior end in the segment addition zone (SAZ) (Fig. 2b, asterisks). In early germbands, clear morphological segmentation was observed only in the thoracic region; as the germband elongated, abdominal segments were added at the posterior, later becoming clearly morphologically segmented. At 48-60 h AEL, a fully elongated germband with gnathal and thoracic appendages was observed (Fig. 2c).
To establish whole mount in situ hybridization and immunohistochemistry in Murgantia, we used engrailed (en), which is expressed in a highly conserved pattern at the border of each segment in developing arthropods and serves as a segmental marker across species [55][56][57][58]. The major challenge to isolating and fixing embryos for visualization of gene expression was that embryos are well-protected by a thick chorion that was difficult to penetrate. We found that immersion of embryos in hot water, followed by boiling and an extended time on ice facilitated physical removal of the pseudoperculum (anterior cap of the chorion, Additional file 2: Fig.  S2) which was necessary to allow fixative to access the embryo. In addition, incubation with xylene softened the chorion and facilitated dissection (see Methods for details). In late blastoderm-stage embryos, Mh-en was expressed in six stripes, suggesting that six segments are established in this species before germband elongation (Fig. 2di). Mh-en stripes were added as the germband elongated, presumably marking each segment (Fig. 2dii). Similarly, Mh-En was detected in segmental stripes by immunohistochemistry using monoclonal anti-Engrailed antibody 4D9, (Fig. 2diii). Overall, these experiments demonstrate that lab rearing and basic molecular genetic approaches to examine gene expression are effective in Murgantia. While no Murgantia histrionica genome has been sequenced yet, the USDA ARS has listed this species as a likely candidate for genome sequencing as part of its Ag100Pest initiative [59].

Murgantia E75A is expressed in PR-stripes
We chose a set of genes that are diagnostic of the difference between PRG-ortholog expression in Oncopeltus and in holometabolous insects to determine whether the genetic basis of segmentation in Murgantia is similar to one of these. E75A is expressed in PR-stripes in Oncopeltus but not in Drosophila and, to date, is the only PRG identified in Oncopeltus [46,47]. In Murgantia, Mh-E75A was first detected broadly in the anterior portion of early blastoderm-stage embryos (Fig. 3a). One clear Mh-E75A stripe appeared at the center of the embryo (Fig. 3b, arrowhead), as another stripe began to resolve from the anterior domain of expression (Fig. 3b). Another stripe appeared posterior to this center stripe (Fig. 3c, arrowhead). As germband invagination began, Mh-E75A was observed in three clear stripes (Fig. 3d). Based on the finding that en was expressed in six stripes at this stage ( Fig. 2di), we conclude that these domains of E75A expression correspond to the primordia of alternate segments and thus the pattern represents classic PR-expression. Mh-E75A continued to be expressed during germband elongation. In early germbands, Mh-E75A was expressed in two stripes anterior of the SAZ (Fig. 3e, f). In later germbands, one obvious stripe was observed with a second stripe resolving in the center of the SAZ (Fig. 3g, asterisks). Later, Mh-E75A continued to be expressed in the SAZ with older stripes disappearing anteriorly while two stripes remained in the center of the SAZ (Fig. 3h). The number and spacing of Mh-E75A stripes in the blastoderm, as it compares to Mh-en expression at the same stage, suggests a role in PR-patterning. The two-stripe cadence of E75A expression during germband elongation suggests that it has the potential to act as a PR regulator in abdominal segment patterning as well. In sum, Mh-E75A is expressed in a PR-like pattern, similar to that seen in Oncopeltus.

Mh-run is expressed in stripes in the blastoderm and broadly in the SAZ
In Oncopeltus, run was the only Drosophila PRG ortholog that showed any PR-like expression [46], making this Mh-E75A continues to be expressed broadly in the anterior with one stripe appearing in the middle of the embryo (arrow head); c Mh-E75A resolves into two stripes (arrowhead points to new stripe) with continued expression at the anterior part of the embryo; d In late blastoderm-stage embryos (24-36 h AEL), invagination of the embryo is beginning to occur (indicated with an arrowhead). At this point Mh-E75A has resolved into three stripes. e-h Mh-E75A expression during germband elongation. Mh-E75A is expressed at the segment addition zone. Asterisks highlight stripes; arrowhead indicates fading stripe. i Mh-run is expressed in the posterior half of an early blastoderm-stage embryo. j Mh-run resolves into one stripe with continued expression at the posterior. k Three stripes of Mh-run are observed in later blastoderm-stage embryos, with one stripe appearing to split. l Mh-run was observed anterior to broad SAZ expression in some early germband-stage embryos (arrowhead); m In an elongating germband, Mh-run continues to be expressed in the SAZ, and was also observed in dots of expression in the thoracic segments and the head lobes; n In later germbands, Mh-run was still observed broadly in the SAZ; o Occasionally stripes (arrowhead) of Mh-run were observed anterior to SAZ expression. Embryos are shown with the anterior to the left a particularly interesting gene to examine in another hemipteran. Mh-run was first detected broadly in the posterior half of the embryo (Fig. 3i). Expression then resolved into a broad stripe near the center of the embryo with continued expression at the posterior (Fig. 3j, arrowhead). As gastrulation began, three stripes were observed with continued expression in a posterior domain (Fig. 3k,  arrowhead). This expression pattern differs from that seen in blastoderm-stage Oncopeltus embryos, where two early broad stripes each appear to split [46]. While we never observed two earlier such stripes of Mh-run, we cannot rule out that a similar earlier expression pattern exists. It is important to note that Of-run expression was reported to be highly dynamic; given that Murgantia lay fewer embryos than Oncopeltus, this reduced sampling could mask similar variability in expression. Mh-run continued to be expressed during germband elongation, where the expression pattern also differs from that reported in Oncopeltus. Of-run was shown to be expressed in roughly two-segment-wide stripes in the SAZ [46]. In early germbands, Mh-run was observed broadly in the segment addition zone (SAZ); occasionally, a stripe of expression anterior to the SAZ was also observed (Fig. 3l, arrowhead). In older germbands, Mhrun was expressed in two dots in each segment, possibly in the central nervous system, with continued expression of a thick band in the SAZ, and additional expression in the head lobes ( Fig. 3m-o). This expression continued, with an occasional stripe seen anterior to the broad expression in the SAZ (Fig. 3o, arrowhead). In Oncopeltus, Of-run was observed in broad stripes in the anterior SAZ that appeared to split in two as segment primordia left the SAZ, suggesting the potential for PR modulation by Of-run. However, in Murgantia, such two-segmentwide stripes were never observed anterior to the SAZ. These Mh-run expression patterns in the blastoderm and during germband elongation are reminiscent of Of-run expression at these stages, though Of-run expression was more characteristic of classic PR patterning.

Mh-odd, Mh-eve and Mh-prd are expressed segmentally
Mh-odd was detected in a broad domain around the center of the early embryo, absent from the poles (Fig. 4a). This broad expression of Mh-odd resolved into five stripes in late blastoderm-stage embryos (Fig. 4b), in a manner similar to Of-odd [46]. Mh-odd continued to be expressed during early germband elongation and in fully elongated germbands in four stripes in the anterior SAZ with expression at the head lobes (Fig. 4c, d). This expression showed no PR-like character, similar to what was seen for Of-odd.
Mh-eve was first detected in a broad domain in early blastoderm-stage embryos, though absent from the anterior fourth of the embryo (Fig. 4e). In late blastoderm-stage embryos, Mh-eve resolved into five stripes (Fig. 4f ), and six stripes were observed as gastrulation began (Fig. 4g), similar to Mh-en (Fig. 2di). In early germbands, Mh-eve was expressed anterior to the SAZ with four visible stripes and a broader region of expression in the central SAZ (Fig. 4h). In older germbands, similar, closely spaced stripes in the anterior SAZ were also observed (Fig. 4i). Double staining with anti-Engrailed antibody was performed to compare the segmental expression of Mh-eve (blue) to En (brown) during germband elongation (Fig. 4j). This revealed that En stripes appear anterior to the SAZ where Mh-eve stripes were no longer detectable (arrowheads), showing no temporal overlap of Mh-eve and Mh-En. In sum, Mheve was expressed segmentally in blastoderm and later stage embryos, with no hint of PR-register. These Mh-eve expression patterns are nearly identical to those observed for Of-eve at each stage of embryonic development examined [46,58].
Mh-prd expression was first detected in late blastoderm-stage embryos in six stripes (Fig. 4k), presumably in the primordia of every segment. In early germbandstage embryos, Mh-prd was observed in two stripes in the anterior SAZ, as well as faintly in stripes in each more mature segment (Fig. 4l). In later germbands, Mhprd was seen in four stripes in the anterior SAZ (Fig. 4m), very similar to the expression patterns of Mh-eve and Mh-odd at this stage (Fig. 4i, d). However, unlike Mh-eve and like Mh-odd, Mh-prd was also expressed in dots in the head lobes (Fig. 4m, n).
In sum, Mh-odd, -eve and -prd were expressed in very similar patterns to their orthologs in Oncopeltus. For both species, these three genes appear to be expressed in the primordia of every segment of the blastoderm, and then at a register consistent with expression in the primordia of every segment during germband elongation.

Mh-slp is expressed in persistent segmental stripes
Mh-slp was first observed in the anterior third of early blastoderm-stage embryos (Fig. 5a). This expression resolved into one stripe (Fig. 5b); later, a second stripe was observed (Fig. 5c). Stripes continued to appear posteriorly until six stripes were visible (Fig. 5d). Comparison of the anterior stripe position of Mh-slp to that of Mhodd and Mh-eve (Fig. 4b, f ) suggests that the stripes of Mh-slp seen in early blastoderm-stage embryos (Fig. 5d) are much further anterior than those of eve or odd, and thus likely correspond to pre-mandibular segments. During germband elongation, Mh-slp continued to be expressed in every segment with no expression detected in the posterior SAZ (Fig. 5e, f ). In later germbands, Mh-slp was observed in every segment with expression concentrated along the midline in older segments. Additionally, expression was observed in the intercalary and antennal segments (Fig. 5g, h). This persistent segmental expression throughout the germband differs from other Mh-PRG orthologs and is similar to that seen for Of-slp.

Conclusions
We have established the harlequin bug, Murgantia histrionica, as a versatile insect model system. Murgantia are easy to maintain in the laboratory on a diet of fresh organic kale/collard greens and cabbage and can be bred en masse. Murgantia is a hemimetabolous insect with a reasonably short life cycle (~ 6 weeks from embryo to adult) and produces enough embryos to examine gene expression. As such, this species meets all the basic requirements to serve as a model for molecular genetic analysis. Our lab previously tried to rear another pentatomid, the invasive brown marmorated stink bug Halyomorpha halys, and observed a collapse of our lab colony after three generations, possibly due to infection by a microsporidian [60,61]. We therefore subjected Murgantia to a 'trial period' of lab rearing before beginning experiments to ensure that we could sustain our lab  Methods for examining gene expression in Murgantia were adapted from Oncopeltus protocols and can be performed readily with these modifications on a large number of samples. In addition to their utility as a model system, Murgantia is a pest of cruciferous vegetables (cabbage, kale, etc.). A native of Central America, it was first reported in the United States in 1864 [62]. After its introduction in Texas, the bug spread to other states making its way to most southern states by the late 1890s. Although it is most active during summer, it has been reported that Murgantia can remain active during mild winters. Over the last decade, warmer winters have led to increased survival and reproduction and larger populations in subsequent seasons [63,64]. Murgantia damage crops through their piercing-sucking mechanism of feeding, which leaves white blotches on leaves and can destroy crops if populations are large. Elucidation of the genes regulating Murgantia development can lead to non-chemical pest control methods, such as RNA interference.
Here, we used Murgantia as a model system to investigate conservation and variation in PR-patterning mechanisms in insects. We established methods for in situ hybridization and immunostaining; RNAi will be tested as a tool to assess gene function in the future, based on successful use of this technique in Oncopeltus. Previously, we and others had shown that orthologs of Drosophila PRGs are expressed segmentally, rather than in PR-stripes in the milkweed bug Oncopeltus [45,46,58]. Specifically, Of-eve, prd, odd, and slp were each found to be expressed in segmental stripes in the blastoderm with no PR-like expression observed at any stage, while Of-run showed some hint of PR-like expression [45,46,58]. In contrast, the Drosophila-like expression of PRGs in PRstripes, in the primordia of alternate segmental units, is shared among a number of holometabolous insects, particularly for the genes eve, prd, slp, run and odd (see Introduction). These findings led us to identify another hemipteran insect for comparative studies. We found that in Murgantia, Drosophila PRG-orthologs are expressed in segmental-stripes in blastoderm-stage embryos with varying expression patterns. Mh-odd, Mh-eve, Mh-prd and Mh-slp were found to be expressed in five or six stripes in the late blastoderm. During germband elongation, eve, odd, and prd were expressed in a series of stripes in the anterior SAZ showing no PR-register, with additional expression in the head lobes seen for odd and prd. Mh-slp was not observed in the posterior SAZ but was found in persistent segmental-stripes during elongation, marking every mature segment. These expression patterns bear remarkable resemblance to the expression patterns of the corresponding orthologs in Oncopeltus. This suggests that little change in the segmentation network-with regard to any role the Drosophila PRGorthologs might play in this network-occurred over the nearly 200 million years of each lineage's independent evolution.
The similarities in expression patterns between these two species extend to the timing of segmental stripe refinement. Stahi and Chipman [65] showed that expression of four genes-including eve and inv-in segmental stripes in the Oncopeltus blastoderm arise nearly simultaneously, similar to blastoderm segmentation in Drosophila. While we cannot be confident about the precise timing of segmental stripe refinement in Murgantia, we did not observe many intermediate striped expression patterns, suggesting that segmental stripes of Mh-en, -odd, -eve, and -prd arise nearly simultaneously in the blastoderm in this species as well. This pattern is in contrast to the more extreme short-germ mode of segmentation in which segmental stripes arise in an anterior to posterior order. Future studies of blastoderm segmentation in Murgantia could therefore provide further insight into the evolution of different modes of segmentation.
In Oncopeltus, Erezyilmaz et al. [47] showed and we confirmed [46] that E75A, which is not a PRG in Drosophila, is expressed in typical PR-stripes. This nuclear receptor E75A does not have PR-expression in Drosophila and is not known to function in segmentation in any holometabolous insects. Here, we show that Mh-E75A is expressed in a PR-like pattern during development, the first indication that this gene might be involved in regulating segmentation outside of Oncopeltus. Specifically, Mh-E75A is expressed in three stripes in the blastoderm, presumably in the primordia of every other segment. During germband elongation Mh-E75A is expressed in two stripes anterior to the SAZ that disappear in older segments, an expression pattern reminiscent of Of-E75A. Thus, it seems that there is a different set of genes acting as PRGs within Hemiptera. Whether or not this difference indicates a distinct rift in patterning mechanisms between holometabolous groups and more basally branching insects, or is simply a feature of hemipterans, remains to be investigated.
Here we have found a high level of conservation in expression patterns of PRG-orthologs and E75A in two distantly related hemipterans. While it is not surprising to find greater conservation of expression patterns between these two species than between either and Drosophila, the degree of conservation observed helps provide needed evolutionary context which is missing when network components are examined in highly diverged species in isolation. Thus far, it appears that genes responsible for PR-patterning are wholly different in Oncopeltus and Murgantia as compared to holometabolous insects. It will be of particular interest to examine other species in the hemipteran clade to determine if a broader hemipteran segmentation network exists. Future studies of the segmentation networks operating in even more basally branching species will help determine whether an Oncopeltus and Murgantia-like network is ancestral within insects or is Hemiptera-specific. These studies will provide insight into the timing of evolutionary transitions in core body patterning regulators and the mechanisms by which such changes have occurred.

Laboratory cultures, life cycle and embryo collection
Murgantia were collected from kale, cabbage, and spider flower (Cleome spp.) at local gardens. In one year, a garden plot was planted only with Cleome spinosa (Southern Exposure Seed Exchange) which was very effective at attracting Murgantia, suggesting its utility as a trap crop for this species. In the lab, Murgantia were reared in mesh cages (12 × 12 × 12 in) at 25 °C. A total of eight cages were kept at once, four adults cages (~ 70 adults per cage) and four growing nymph cages. Each cage was given one organic kale/collard leaf and one cabbage leaf three times a week. Cages were also supplied with wet cotton as a water source and were lightly misted every other day. Embryos were collected from adult cages to set up new cages. Embryos were kept in small petri dishes, allowed to hatch and grow to the second instar in the dish and later transferred to a clean mesh cage (Fig. 1b). This was done every 2-3 months to keep the colony going. Cages were generally kept for 4 months, after which their fecundity dramatically decreased. Used cages were washed with soapy water followed by bleach and rinsed well with tap water. To collect embryos, fresh food was placed in adult cages and embryos were collected from the leaves. For gene expression analysis embryos were collected during a 8-12 h window and fixed 12-17 h after collection. As we observed no egg laying during the nighttime all collections were done during daytime. Since females lay only 12 embryos at a time a collection window of at least 8 h was necessary to collect enough embryos for experiments. To determine relative ages of embryos, the invagination pore and extent of germband elongation were compared between embryos. Later experiments revealed that Kimwipes could be used to collect embryos by pinning them to the cages with pushpins. In order to count the number of nymphal instars for this species, individual hatchlings were kept in 2 oz. cups with small holes in the lids and cup. A piece of fresh kale was given every day along with a piece of wet cotton. Cups were observed every day and each molt recorded until adulthood. A total of five nymphal stages were observed (Fig. 1c).

cDNA, gene isolation and probe synthesis RNA extraction and cDNA preparation
Embryos were collected every 24 h for RNA extraction. Approximately 40-50 Murgantia embryos were collected within 24 h, homogenized in 150 ul TRIzol solution (Invitrogen), and stored at − 80 °C in preparation for RNA extraction. RNA extraction was done as described by the manufacturer. A total of 1 ug RNA template was used for cDNA synthesis. cDNA was prepared following NEB's standard protocol. Briefly, 1ug RNA template was combined with dNTPs (New England Biolabs), Random Primer Mix (New England Biolabs), and water. Tubes were incubated at 65 °C for 5 min to denature the RNA and immediately transferred to ice. Transcriptase buffer and RNAse inhibitor (New England Biolabs) were added and incubated at 37 °C for 2 min. Finally, M-MuLV Reverse Transcriptase (New England Biolabs) was added, and the reaction was incubated at 37 °C for 50 min followed by 15 min at 70 °C to inactivate the enzyme.

Probe synthesis
DNA templates for antisense RNA probes were generated by PCR using reverse primers containing the T7 polymerase promoter sequence. Digoxigenin-labeled RNA probes were synthesized from 200 ng of purified PCR product with digoxigenin labelling mix (Roche), RNase inhibitor, and T7 polymerase. RNA transcription reactions were incubated at 37 °C for 2 h, then RNA was precipitated with 100% cold ethanol and 6 M LiCl and incubated overnight at − 20 °C. The pellet was washed with 70% ethanol, dried, and dissolved in 20 μl of nuclease-free water. In order to ensure that the pellet was completely dissolved, tubes were incubated at 37 °C for 5-10 min.

Gene isolation
Mh-PRG orthologs were isolated using degenerate primers or primers matching orthologs from Halyomorpha halys (Hhal), based on sequences annotated in the i5k's publicly available draft genome [66]. Rapid Amplification of cDNA Ends (RACE) was performed when necessary. Mh-en, eve, E75A, and prd were isolated by degenerate PCR with Taq polymerase (see Additional file 3: Table S1 for primer sequences; degenerate primers used to isolate Mh-eve from [67]). 3′ and 5′ RACE were used to extend Mh-en, eve, E75, and prd using the FirstChoice RLM-RACE Kit (Invitrogen) or the SMARTer ® RACE 5′/3′ Kit (Takara Bio). Mh-slp, run, and odd were isolated with Hhal primers, and the SMARTer ® RACE 5′/3′ Kit (Takara Bio) was used to extend the 5′ and 3′ ends of Mh-odd. Orthology was determined by phylogenetic and sequence analysis using amino acid sequences of paralogs and orthologs of the target gene from various insect species (Additional file 4: Fig. S3; Additional file 5: Fig.  S4). Gene trees were constructed for each ortholog using the following homologs. en: en and invected; E75A: E75A and EcR; eve: eve, labial, and proboscipedia; runt: runt, lozenge, runxA, and runxB; odd: odd, brother of odd with entrails limited (bowl), and sister of odd and bowl (sob); prd: prd, gooseberry (gsb), and gooseberry-neuro (gsbn). Gene structure schematics show the regions of each gene that were isolated in this study (Additional file 6: Fig. S5). Note two isoforms of Mh-prd were found, as in Oncopeltus. These differ in their paired box with an insertion of 75 bp found in one isoform (Additional file 6: Fig. S5e). In some cases, the complete coding sequence was isolated as well as the 5′ or 3′ untranslated region (UTR). The sequences have been deposited to Gen-Bank with the following accession numbers

Embryo fixation
Murgantia embryos were fixed using the protocol for Oncopeltus embryo fixation [46,58,68] with some modifications. Appropriately aged embryos were carefully separated from each other under a dissecting scope using forceps and placed in 2 ml tubes (~ 35-40 embryos/tube). To begin fixation, 600 ul of boiling water was added to each tube, and tubes were submerged in boiling water for 3 min. Tubes of embryos were then immediately submerged in ice for 15-20 min. Water was removed and replaced with PBST (0.01 M phosphate buffered saline with 0.05% . Under a dissecting microscope, the pseudoperculum (or anterior "cap") of the chorion was removed using fine-tipped forceps (Additional file 2: Fig. S2). PBST was removed and replaced with 1.2 ml of 1:1 12% paraformaldehyde (PFA):heptane. Tubes were shaken at 200 rpm for 20 min. PFA (bottom layer of liquid) was removed and replaced with 600 ul of methanol. The heptane:methanol was removed, and embryos were washed three times with 100% methanol. 1.2 mL 1:1 xylene: methanol was added to each tube and tubes were rocked for 1 h at RT to soften the chorion for manual dissection. Xylene:methanol was removed, embryos were washed three times with methanol and stored in methanol at − 20 °C.
As dissecting embryos out of their eggshells is fairly time-consuming, this was usually done several days ahead of starting an in situ hybridization. Embryos were removed from storage at − 20º C and gradually washed from 100% methanol into PBST (3 min per wash in 100%, 75%, 50%, 25% methanol diluted in PBST), followed by one 3 min wash in PBST and three rinses with PBST. The chorion was manually removed under a dissecting microscope using forceps, and embryos were then washed gradually back into 100% methanol using the aforementioned series of washes in reverse order. Embryos were washed three times with 100% methanol, methanol was removed and replaced with 1.2 ml of 1:1 methanol:heptane in order to remove the vitelline membrane by inverting and manually shaking the embryos in the tubes for 30-45 s. The methanol:heptane mixture was removed, and embryos were washed two times with 100% methanol. The embryos were washed into PBST using the same methanol gradient as above. Embryos were washed three times with PBST and fixed for 1 h in 4% PFA. After fixation, embryos were washed three times with PBST, washed gradually back into 100% methanol, and stored at − 20 °C for later use.

In situ hybridization and SYTOX green nuclear staining
The in situ hybridization protocol described by Ben-David and Chipman [68] for Oncopeltus embryos was used with some modifications. Fixed embryos were gradually washed from 100% methanol into PBST (3 min per wash in 100%, 75%, 50%, 25% methanol diluted in PBST), followed by one 3 min wash in PBST and three rinses with PBST and incubated at 60º C with hybridization buffer for 3-4 h. Probes were heated for 3 min at 90 °C and placed in ice immediately. Probes were used at concentrations of 0.1-1 ng/μl. Embryos were incubated with probes at 60 °C overnight. Embryos were then incubated with pre-warmed hybridization buffer at 60 °C for 30 min, two times, followed by incubation with prewarmed 2× SSC (saline sodium citrate) at 60 °C, 2× SSC at room temperature (RT) and finally 0.2× SSC at RT for 30 min. Embryos were washed 3 times with PBST and incubated with 10% sheep serum in PBST for 2 h at RT. Sheep serum was removed and replaced with 1:1600 dilution of anti-digoxigenin-AP Fab fragments (Roche) in 10% sheep serum and incubated for 3-4 h at RT. Embryos were washed 3 times in PBST and kept in PBST overnight at 4º C. The following day, embryos were washed with PBST for 20 min five times. In preparation for staining, embryos were washed for 15 min in alkaline phosphatase (AP) staining buffer (100 mM NaCl, 50 mM MgCl 2 , 100 mM Tris-HCl, pH 9.5 and 0.1% TWEEN-20) and 15 min in AP-PVA staining buffer (100 mM NaCl, 50 mM MgCl 2 , 100 mM Tris-HCl, pH 9.5, 0.1% TWEEN-20 and 2% polyvinyl alcohol (PVA)). BCIP and NBT (Roche) in AP-PVA was used for color detection. SYTOX green staining was done using 1:1000 SYTOX Green (Invitrogen) in PBST for 1 h at RT after fixation of embryos with 4% PFA. Blastoderm state embryos were imaged in PBST. Germbands were hand dissected, removing as much yolk as possible, and mounted in 75% glycerol in PBST.

Immunohistochemistry
Embryos were fixed as described above, except that fixation after chorion removal was done for 30 min only. After fixation, the protocol described by Lu et al. [60] for Halyomorpha embryos was used except that post-fixation was not done after the staining reaction. If background staining was strong, embryos were washed in 50% methanol in PBST for 5 min and rinsed 3 times with PBST after the color reaction. The monoclonal anti-Engrailed antibody 4D9 (Developmental Studies Hybridoma Bank) was used at a 1:10 dilution. Double staining was performed by combining in situ hybridization and immunohistochemistry protocols. Protein staining was done first with some changes. First, instead of using PBST, freshly prepared PBTH (filtered 1X PBS, 0.1% Tween-20, 50 µg/ml heparin, and 250 µg/ml tRNA) was used for all steps. Additionally, RNase inhibitor was added to any incubation longer than 1 h. In situ hybridization followed antibody staining as described above. BCIP/NBT staining was done at 4 °C to decrease background staining and was monitored very closely under a dissection microscope.

Image processing
Germbands were imaged with an AxioCam MRc camera (Zeiss) mounted on a Zeiss Axio Imager M1. Picture of germbands were taken with a 20X objective requiring image merging with Adobe Photoshop. Blastoderms were imaged in PBST using an Axiocam 506 color camera mounted on a Zeiss Discovery.V12 SteREO dissecting microscope. SYTOX-stained embryos were imaged using an Axiocam mono camera mounted on Zeiss Discovery. V12 dissecting microscope in the dark.

Phylogenetic trees and gene structures
Protein sequences were subjected to phylogenetic analysis as described by Reding et al. [46]. Protein sequences of orthologs from various insects were collected and a multiple sequence alignment was generated using MUS-CLE. Alignments were then trimmed using AliView and exported as a FASTA file to TOPALi v2.5 where a phylogenetic tree was constructed using the Bayesian algorithm MrBayes. Trees were edited and formatted using MEGA7.

COI barcoding
To verify the species of our field-collected colony, part of the mitochondrial cytochrome c oxidase subunit I (COI) gene was sequenced and searched against the BLAST non-redundant nucleotide database. DNA was extracted from two clutches of eggs by pestling in 100 ul of DNA extraction buffer (10 mM Tris-Cl pH 8.2, 1 mM EDTA, 25 mM NaCl, and 200 μg/ml proteinase K), incubating at 37 °C for 30 min, then 95 °C for 2 min to inactive the proteinase K, then briefly spun down to pellet the yolk proteins and chorion. PCR was performed using 1 uL of the supernatant with the primers LCO1490 and HC02198 described in Folmer et al. [69] at an annealing temperature of 52 °C and an extension time of 15 s. The sequence of the PCR product was queried against NCBI's non-redundant nucleotide database and the Barcode of Life Database (BOLD; [70]). The NCBI database search revealed matches to two M. histrionica specimens whose sequences have also been deposited in the BOLD database (specimens 1 and 2 in Additional file 7: Fig. S6 with 16/658 and 4/658 mismatches relative to our sequence, respectively). The sequence with highest percent identity match to our own in the BOLD database had just 2/568 mismatches (specimen 4 in Additional file 7: Fig. S6). The COI sequence has been deposited to GenBank with accession number MT238119.