Ultrabithorax is essential for bacteriocyte developmentYu Matsuuraa,b,c, Yoshitomo Kikuchid, Toru Miurab, and Takema Fukatsua,c,1

aBioproduction Research Institute, National Institute of Advanced Industrial Science and Technology, Tsukuba 305-8566, Japan; bGraduate School ofEnvironmental Science, Hokkaido University, Sapporo 060-0810, Japan; cGraduate School of Life and Environmental Sciences, University of Tsukuba,Tsukuba 305-8572, Japan; and dBioproduction Research Institute, National Institute of Advanced Industrial Science and Technology, Hokkaido Center,Sapporo 062-8517, Japan

Edited by Nancy A. Moran, University of Texas at Austin, Austin, TX, and approved June 19, 2015 (received for review February 18, 2015)

Symbiosis often entails the emergence of novel adaptive traits inorganisms. Microbial symbionts are indispensable for diverse insectsvia provisioning of essential nutrients, wherein novel host cells andorgans for harboring the microbes, called bacteriocytes and bacter-iomes, have evolved repeatedly. Molecular and developmental mech-anisms underpinning the emergence of novel symbiotic cells andorgans comprise an unsolved question in evolutionary developmentalbiology. Here, we report that a conserved homeotic gene, Ultra-bithorax, plays a pivotal role in the bacteriocyte differentiation in ahemipteran insect Nysius plebeius. During embryonic development,six pairs of aggregated presumptive bacteriocytes appear on bothsides of six abdominal segments, incorporate the symbiotic bacteriaat the stage of germband retraction, and fuse into a pair of lateralbacteriomes at the stage of germband flip, where bacteriocyte-asso-ciated Ultrabithorax expression coincides with the symbiont infec-tion process. Suppression of Ultrabithorax expression by maternalRNA interference results in disappearance of the bacteriocytes andthe symbiont localization therein, suggesting that Ultrabithorax isinvolved in differentiation of the host cells for symbiosis. Suppressionof other homeotic genes abdominal-A and Antennapedia disturbsintegrity and positioning of the bacteriomes, affecting the configura-tion of the host organs for symbiosis. Our findings unveil the molec-ular and developmental mechanisms underlying the bacteriocytedifferentiation, which may have evolved either via cooption of thetranscription factors for inducing the novel symbiotic cells, or via re-vival of the developmental pathway for the bacteriocytes that hadexisted in the ancestral hemipterans.

bacteriocyte | homeotic gene | transcription factor | evolution | symbiosis

Symbiosis is the source of novel adaptive traits, thereby con-tributing to organismal evolution and diversification (1, 2). A

variety of insects are obligatorily dependent on microbial sym-bionts via provisioning of essential nutrients lacking in their diets(3, 4), wherein novel host cells and organs for harboring the mi-crobes, called bacteriocytes and bacteriomes, have evolved repeat-edly in such insect groups as hemipterans (aphids, whiteflies,mealybugs, leafhoppers, spittlebugs, etc.) (5–9), dipterans (tsetseflies, bat flies, etc.) (10, 11), coleopterans (weevils, etc.) (12, 13),and many others (14). Despite a considerable body of embryo-logical descriptions (14), molecular mechanisms underlying thebacteriocyte differentiation have been a long-lasting enigma inevolutionary developmental biology (15, 16). Although cellularand developmental aspects of the bacteriocyte formation havebeen best documented for the pea aphid, Acyrthosiphon pisum(5, 15, 17), the seed bug Nysius plebeius and allied heteropteranbugs of the superfamily Lygaeoidea have recently emerged as apromising model system for investigating the development, evo-lution, and origin of the bacteriocytes, on the grounds that(i) heteropteran bugs are generally associated with gut symbioticbacteria without bacteriocytes; (ii) thus, the bacteriocytes in theselygaeoid species are regarded as a novel trait whose evolution wasa relatively recent event (18, 19); and (iii) in N. plebeius and alliedlygaeoid species, RNA interference (RNAi) works efficiently,which enables functional analysis of genes involved in the bac-teriocyte formation (20, 21). Here we demonstrate that, by makinguse of the emerging model insect N. plebeius, conserved homeobox

transcription factors, in particularUltrabithorax (Ubx), are involvedin the development of the host cells and organs specialized forharboring the symbiotic bacteria.

Results and DiscussionBacteriocyte Development During Embryogenesis of N. plebeius. Weperformed a detailed description of the embryogenesis of N. ple-beius with special focus on the dynamics of the bacteriocyte-associated gammaproteobacterial symbiont “Candidatus Schneiderianysicola” (18) (Figs. 1 and 2, Table S1, and Movie S1). The sym-biotic bacteria were found as an aggregate at the anterior pole ofnewly laid eggs (Fig. 1 A and B). After blastoderm formation(12–24 h after oviposition; Fig. 2A) and germband elongation(24–33 h; Fig. 2 B–D), the symbionts were wrapped within ab-dominal segments of the germband (36–60 h; Figs. 1 I and J and 2E–G). After germband retraction (∼72 h; Fig. 2H), the symbiontsmigrated from the abdominal population to presumptive bacter-iocytes that appeared on both sides of abdominal segments A2–A7as six pairs of clusters (72–84 h; Figs. 1 K and L and 2 I and J).Then, during the process of drastic embryonic flip called katatrepsis,the six bacteriocyte clusters on each side fused into a coherentbacteriome located at abdominal segments A2–A4 (84–96 h;Figs. 1 M and N, and 2 K–M and Q–R). After the symbiont in-fection, the bacteriocytes accumulated red pigment and becameeasily recognizable (∼84 h; Figs. 1 C and 2 R–T). The red pig-mentation provided a visible marker useful for tracing thebacteriocytes, although it is unclear whether the red pigment wasderived from the symbiont or the host.

Significance

Among the most fundamental questions in developmental bi-ology is how novel cell types have emerged in the metazoanevolution. Among the most challenging questions in evolu-tionary biology is how sophisticated symbiotic associationshave evolved through less intimate interorganismal in-teractions. These fundamental biological issues are crystalizedin the evolution and development of insect’s bacteriocytesspecialized for harboring symbiotic bacteria. Here, we reportthat a conserved transcription factor Ultrabithorax is essentialfor bacteriocyte development in an insect, thereby uncoveringa molecular mechanism underlying the emergence of the novelhost cells for symbiosis. Our finding highlights the importanceof developmental cooption of preexisting transcription factorsand sheds new light on a long-lasting enigma in evolutionarydevelopmental biology.

Author contributions: Y.M. and T.F. designed research; Y.M., Y.K., and T.M. performedresearch; Y.M. analyzed data; and Y.M. and T.F. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

Data deposition: The nucleotide sequences determined in this study have been depositedin the DNA Data Bank of Japan database (accession nos. LC010622–LC010625).1To whom correspondence should be addressed. Email: t-fukatsu@aist.go.jp.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1503371112/-/DCSupplemental.

9376–9381 | PNAS | July 28, 2015 | vol. 112 | no. 30 www.pnas.org/cgi/doi/10.1073/pnas.1503371112

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

12, 2

020

Bacteriocyte-Associated Expression of Ubx in Embryogenesis ofN. plebeius. The homeobox genes encode transcription factorsthat assign segment identities and specify functional body partsin the development of insects and other animals (16, 22). In theembryogenesis of the pea aphid A. pisum, some homeobox geneproducts—including an appendage-patterning transcription fac-tor Distal-less (Dll), homeotic proteins Ubx or Abdominal-A(Abd-A), and a segment polarity protein Engrailed (En)—wereshown to localize to the bacteriocytes, although their functions inthe symbiotic cells have been elusive (15). We investigated thespatial expression patterns of some of these homeobox genesduring the development of N. plebeius (Fig. 3 A–F and Fig. S1).Dll was prominently expressed at the distal portion of append-ages in antennal, labial, and thoracic segments and faintly detectedin the maxillary segments, but not expressed in the areas of thepresumptive bacteriocytes (Fig. 3 A and A′), which agreed with theexpression patterns of Dll in a related lygaeid species withoutbacteriocytes, the milkweed bug Oncopeltus fasciatus (23, 24).Notably, Ubx exhibited remarkable expression patterns in theabdominal regions where the presumptive bacteriocytes weresupposed to differentiate: Although the strong expression at theabdominal segment A1 and the milder expression in the followingabdominal segments were observed during the germband elon-gation and retraction (Fig. S1 A–C), which are typical of the Ubxexpression in O. fasciatus and other insects (24, 25), Ubx was alsostrongly expressed as six pairs of clusters on both sides of

abdominal segments A2–A7 after the germband retraction (Fig. 3B–D and Fig. S1D), which agreed with the locations of the pre-sumptive bacteriocytes (Figs. 1 K and L and 2J). Meanwhile, abd-Awas expressed broadly across the abdominal segments withoutspecific association with the presumptive bacteriocytes (Fig. 3 Eand F and Fig. S1 H–J).

Disappearance of Bacteriocytes by Ubx Suppression. In heteropteranspecies including O. fasciatus, injection of double-stranded RNA(dsRNA) into mother insects was reported to efficiently suppresstheir offspring’s gene expression during embryogenesis (20, 26).Using the maternal RNAi technique, we suppressed the embry-onic expression of these genes and observed the development ofN. plebeius. Dll suppression affected neither the bacteriocyte for-mation nor the symbiont localization (Fig. 3 G and H). Strikingly,Ubx suppression resulted in disappearance of the red-pigmentedbacteriocytes and the symbiont localization associated with them(Fig. 3 I and J). Detailed observations of the Ubx-suppressed em-bryos revealed that, in contrast to the control embryos wherein thesymbionts aggregating in the abdomen were migrating to the bac-teriocytes (Fig. 3 M and M′) and strictly localized therein (Fig. 3O),the symbionts were dispersing within the embryonic body (Fig. 3 Nand N′) and lost subsequently (Fig. 3P). Meanwhile, abd-A sup-pression did not lead to disappearance of the bacteriocytes (Fig. 3K, L, Q, and R), although the spatial organization of the bacter-iocytes was affected as detailed later.

Fig. 1. Symbiont localization and bacteriocyte differentiation during the development of N. plebeius. (A and B) Newly laid eggs. (C and D) Embryos 5 d afteroviposition. (E and F) First-instar nymphs. (G and H) Adult insects. (I and J) Embryos ∼48 h after oviposition. (K and L) Embryos of 72–84 h. (M and N) Embryosof ∼96 h. (O and P) Embryos of ∼120 h. A, C, E, and G are light-microscopic images; I, K, M, and O are schematic illustrations of symbiont localization (red) inthe embryos; and B, D, F, H, J, L, N, and P are fluorescence microscopic images in which blue and green signals show the host nuclei and the symbiotic bacteria,respectively. Arrowheads depict bacteriocytes, and arrows indicate aggregated symbionts within the embryos. A1, A2, and A3, first, second, and thirdabdominal segments, respectively; ob, ovarial bacteriocytes in adult female; T3, third thoracic segment.

Matsuura et al. PNAS | July 28, 2015 | vol. 112 | no. 30 | 9377

EVOLU

TION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

12, 2

020

Although the majority of the control embryos hatched normallywith a pair of red-pigmented bacteriomes in the abdomen (Fig. 4A–C and Table S2), considerable proportions of the RNAi-treatedembryos failed to hatch with morphological abnormalities typicalof the homeobox gene defects (23, 25) (Table S2). Dll suppressionproduced the hatchlings without distal regions of the appendages,in which a pair of bacteriomes formed normally (Fig. 4 D–F, Fig.S2 A–D, and Table S2). Of 160 Ubx-suppressed hatchlings, all 42hatchlings with strong phenotypes, whose abdominal segment A1grew extra appendages, lacked the bacteriomes (Fig. 4 G–I andTable S2), whereas 48 hatchlings with moderate phenotypeslacked the symbiotic organs partially (Fig. S2 E and F and TableS2). Of 130 Ubx-suppressed unhatched embryos, 111 embryoslacked the bacteriomes (Fig. S2 G and H and Table S2). The ab-sence of the bacteriomes in the Ubx-suppressed hatchlings andembryos was evidently and significantly more frequent than in theDll-suppressed, abd-A–suppressed, and control hatchlings andembryos (Table S3).

Unfused Bacteriomes by abd-A Suppression. Notably, abd-A sup-pression affected the spatial organization of the bacteriocytes: Incontrast to a pair of fused bacteriomes in the control matureembryos (Figs. 1 O and P and 2 S and T), the abd-A–suppressed

mature embryos exhibited separate bacteriocyte clusters (Fig. 3K, L, Q, and R), as observed in the control younger embryos (Fig.3M). Whereas most of 329 abd-A–suppressed hatchlings and 68embryos retained the bacteriomes, all 23 hatchlings with strongphenotypes whose abdominal segments A2–A8 were transformedinto thorax-like identity, and also some embryos, exhibited un-fused bacteriomes, ranging from two to six in number, in the ab-domen (Fig. 4 J–L, Fig. S2 I–L, and Table S2). These resultssuggest that abd-Amay be involved in the integrity of the symbioticorgans via, presumably, homeotic regulation of segment identity inthe abdomen of N. plebeius.

Translocated Bacteriomes by Antp Suppression. Additionally, weperformed maternal RNAi of Antennapedia (Antp), which is ahomeotic gene expressed at thoracic and anterior abdominalsegments in O. fasciatus and other insects (24, 25). Antp sup-pression also affected the bacteriome localization. Of 211 Antp-suppressed hatchlings, all 13 hatchlings with strong phenotypes,whose six legs were deformed with swollen femurs and shortenedtibiae, exhibited the bacteriomes at the T3 thoracic segment andthe A1–A2 anterior abdominal segments. The bacteriomes werefound not only in the body trunk but also inside the hind legs (Fig.4 M–O, Fig. S2 M–P, and Table S2). These phenotypes look like

Fig. 2. Embryogenesis of N. plebeius. (A–P) Fluorescence microscopic images in which blue and green signals reflect the host nuclei and the symbiotic bacteria,respectively. Unless otherwise described, the embryos are oriented such that the anterior pole of the egg is to the left and the dorsal surface of the egg is upward.Arrows indicate the symbiont cells aggregating within the embryos, and arrowheads show the symbiont cells localized to the bacteriocytes/bacteriomes. (A) Theblastoderm stage (12–24 h after oviposition). (B) The early invagination stage (24–27 h). (C) Elongation of the germband (∼30 h). (D) Further elongation ofthe germband (∼33 h). (E) The late invagination stage (∼36 h) by which segmentation occurs so that embryonic head, thorax, and abdomen are recognizable.(F) The germband with growing appendages, whose tail is curving around the dorsal surface of the egg and wrapping the symbionts within the abdominalsegments (∼48 h). (G) The germband whose appendages are growing further (∼60 h). (H) The germband retraction stage (∼72 h), at which the embryonic bodythickens. (I) The retracted embryonic body with shrunk abdominal segments, on which all of the appendages are folded toward the longitudinal center (∼84 h).(J) An enlarged image of the thoracic and abdominal segments, where the symbiont infection to the embryonic body is taking place (∼84 h). The symbiontsaggregating on the abdominal segments (arrow) and the symbionts incorporated into six clusters of presumptive bacteriocytes on each side of the abdomen(arrowheads) are observed. (K) The katatrepsis, or embryonic flip, stage (84–96 h). During the katatrepsis, the embryo turns backward along the ventral surface ofthe egg, and the six clusters of the bacteriocytes on each abdominal side fuse into a coherent bacteriome. (L and M) Lateral and dorsal views of a postkatatrepsisembryo (∼96 h), which have flipped and escaped the yolk, and left serosal fold above its head. (N and O) Embryos after dorsal closure (108–120 h), in which thebacteriomes are oval in shape and located at both sides of the abdominal segments A2–A3 or A2–A4. (P) A mature embryo about to hatch (∼145 h). (Q–T) Light-microscopic images of developing embryos. (Q) An embryo at the germband retraction stage (∼72 h). The abdominal region where the symbionts are localizedbefore migrating to the presumptive bacteriocytes is recognizable by reddish hue (arrow). (R) An embryo in which symbiont migration to the presumptivebacteriocytes is taking place (∼84 h). The symbiont localizations before and after the migration are seen as red-colored abdominal regions (arrow and arrowhead,respectively). (S) An embryo after dorsal closure (∼108 h), whose bacteriomes are colored in deep red. (T) A mature embryo about to hatch (∼145 h). A1, A2, A3,and A4, abdominal segments 1, 2, 3, and 4, respectively; An, antenna; Gb, germband; Hl, head lobe; Lb, labium; Mn, mandible; Mx, maxilla; Sf, serosal fold; T1, T2,and T3, thoracic segments 1, 2, and 3, respectively. (Scale bars: 100 μm.)

9378 | www.pnas.org/cgi/doi/10.1073/pnas.1503371112 Matsuura et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

12, 2

020

invasion of posterior segmental identity into the thoracic segment,suggesting the possibility that, although speculative, Antp maynegatively regulate the bacteriocyte differentiation at the T3and A1 segments during the embryogenesis of N. plebeius.

ConclusionThe disappearance of the bacteriocytes by Ubx suppression, incombination with the bacteriocyte-associated Ubx expression,strongly suggests that Ubx plays a pivotal role in the bacteriocytedifferentiation in N. plebeius. In addition, abd-A and Antp mayaffect the bacteriome integrity and positioning. Our findings shedlight on the long-lasting enigma as to what molecular mechanismsunderlie the development of insect’s bacteriocytes and bacter-iomes, which unveil the importance of developmental cooption of

the preexisting and conservative homeobox transcription factorsfor acquisition of the novel cells and organs for symbiosis. Fig. 5schematically illustrates expression patterns of homeotic genesduring the embryogenesis of N. plebeius. The mechanism gov-erning the bacteriocyte-associated expression pattern of Ubx iscurrently unknown and deserves future study. As reported infruit flies (27), it seems plausible, although speculative, thatUbx might have acquired a novel cis-regulatory element for theunique regional expression in an ancestor of N. plebeius.

PerspectivesIt should be noted that the bacteriocyte-specific Ubx/Abd-A lo-calization was also observed in the aphid A. pisum (Sternor-rhyncha:Aphididae) (15), which is phylogenetically distinct from

Fig. 3. Expression patterns and RNAi phenotypes of Dll, Ubx, and abd-A during the embryogenesis of N. plebeius. (A and A′) Lateral and dorsal views of Dllexpression pattern (48–72 h after oviposition). (B and B′) Lateral and ventral views of Ubx expression pattern (48–72 h). (C and D) Dorsal and lateral views ofUbx expression pattern (∼84 h) just after the symbiont infection to the bacteriocytes. (E and E′) Lateral and dorsal views of abd-A expression pattern (48–72 h).(F) A dorsal view of abd-A expression pattern just after the symbiont infection to the bacteriocytes. In A–F, arrowheads indicate specific expressions of thehomeobox genes in the abdominal segments. Smaller arrowheads indicate bacteriocyte-associated expressions, where numbers 1–6 correspond to the regions ofthe abdominal segments A2–A7. Larger arrowheads show the other specific expressions in the abdominal region. An, antenna; Lb, labium; Pp, pleuropodium; T1,T2, and T3, thoracic segments 1, 2, and 3, respectively. (G and H) Lateral views of Dll-suppressed embryos just after katatrepsis (∼100 h). (I and J) Lateral views ofUbx-suppressed embryos just after katatrepsis (∼100 h). (K and L) Lateral views of abd-A–suppressed embryos just after katatrepsis (∼100 h). (M andM′) Dorsal andz-stacked images of a control embryo (∼84 h). (N and N′) Dorsal and z-stacked images of a Ubx-suppressed embryo (∼84 h). (O–Q) Dorsal views of control, Ubx-,and abd-A–suppressed embryos just after katatrepsis (∼100 h). (R) An enlarged lateral view of an abd-A–suppressed embryo (∼100 h). In G–R, G, I, and K are light-microscopic images in which the abdominal region for bacteriome formation is highlighted by a white circle, whereas the others are florescence microscopicimages in which blue and green signals indicate the host nuclei and the symbiotic bacteria, respectively. (Scale bars: 100 μm.)

Matsuura et al. PNAS | July 28, 2015 | vol. 112 | no. 30 | 9379

EVOLU

TION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

12, 2

020

N. plebeius (Heteroptera:Lygaeidae) in the Hemiptera. Consid-ering that the heteropterans without the bacteriocytes/bacter-iomes [with exceptions of bedbugs (Cimicidae) and some seedbugs (Lygaeoidea)] (18, 19, 28) constitute a monophyletic groupnested within the Hemiptera and all of the other hemipterangroups possess the bacteriocytes/bacteriomes (Fig. S3A) (29, 30),it is assumed that the bacteriocytes/bacteriomes were present in

the common ancestor of the Hemiptera, lost in the ancestor ofthe Heteroptera, and evolved again in the Lygaeoidea, to whichN. plebeius belongs (Fig. S3B) (14, 18, 31). Hence, the bacter-iocytes/bacteriomes of N. plebeius may have evolved either viacooption of the homeobox transcription factor for inducing thenovel cells for symbiosis or via revival of the developmentalpathway for the bacteriocytes/bacteriomes that had existed in theancestral hemipterans but been disrupted in the lineage of theheteropteran bugs. Whether the regain of the bacteriocytes/bacteriomes in association with the localized Ubx expression inthe stinkbug lineage is regarded as reversal, parallelism, or con-vergence is an intriguing issue in evolutionary developmentalbiology (32, 33), which should be addressed by comparativestudies on molecular mechanisms underlying the development ofsymbiotic cells and organs in A. pisum, N. plebeius, and otherhemipteran species. It is still an enigma what type of cells com-prises the developmental origin of the bacteriocytes. It is also ofinterest, but totally unknown, what mechanisms underlie thespecific targeting of the bacterial symbiont to the newly formedbacteriocytes at the stage of germband retraction in N. plebeius(Figs. 1 K and L and 3 M and M′). Although Ubx starts specificexpression in the presumptive bacteriocytes at this stage, it seemslikely that not Ubx itself but some factor(s) downstream of Ubxmust be responsible for the host–symbiont interplay, like flavo-noids in the legume-Rhizobium nitrogen-fixing symbiosis (34)and chitin oligosaccharides in the squid–Vibrio luminescentsymbiosis (35, 36). These evolutionary and developmental issues

Fig. 4. Strong RNAi phenotypes of Dll, Ubx, abd-A, and Antp in newborn nymphs of N. plebeius. (A–C) Control nymphs. (D–F) Dll-suppressed nymphs, whoseappendages are severely truncated. (G–I) Ubx-suppressed nymphs, whose abdominal segment A1 exhibits thorax-like pigmentation (arrow) and grows ap-pendage-like structures. (J–L) abd-A–suppressed nymphs, whose abdominal segments grow appendage-like structures on both sides. (M–O) Antp-suppressednymphs, whose legs are deformed and swollen (circle). A, D, G, J, J′, and M are light-microscopic images. B, E, H, H′, K, K′, and N are scanning electronmicroscopic images. C, F, I, L, and O are fluorescence microscopic images in which blue and green signals indicate the host nuclei and the symbiotic bacteria,respectively. Asterisks show cuticle-derived autofluorescence. Filled arrowheads indicate the bacteriomes, and open arrowheads highlight the absence of thebacteriomes. ap, appendage-like structure; vm, vitelline membrane. (Scale bars: 100 μm.) For mild and moderate RNAi phenotypes, see Fig. S2.

Fig. 5. Schematic illustration of expression patterns of homeotic genes inthe embryogenesis of N. plebeius. Note that we inspected the detailed ex-pression patterns of Ubx and abd-A only, and the expression patterns of theother genes are according to previous studies on a related lygaeid bugO. fasciatus (24, 25).

9380 | www.pnas.org/cgi/doi/10.1073/pnas.1503371112 Matsuura et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

12, 2

020

would be addressed by transcriptomics of the presumptive bac-teriocytes microdissected from embryos of N. plebeius or otherhemipteran insects.

Materials and MethodsInsect Material. N. plebeiuswas maintained at 25 °C under a long-day regime(16-h light, 8-h dark) on sunflower seeds, whole wheat, and distilled watersupplemented with 0.05% ascorbic acid.

In Situ Hybridization of Symbiont. Whole-mount fluorescence in situ hybrid-ization targeting bacterial 16S rRNAwas performed to visualize the symbiontlocalization in embryos and whole insects of N. plebeius as described (18)with a probe listed in Table S4.

Cloning of Homeobox Genes. Total RNA was extracted from embryos ofN. plebeius, reverse-transcribed to yield cDNA, and subjected to PCR amplifi-cation of the homeobox genes with degenerate primers listed in Table S4. ThePCR products were cloned and sequenced, and the complete or partial cDNAsequences were determined by rapid amplification cDNA end methods.

In Situ Hybridization of Host’s Homeobox Genes. Whole-mount in situ hybrid-ization of the homeobox gene transcripts was performed with digoxigenin-

labeled RNA probes, which were synthesized by T7 RNA polymerase (TaKaRa)and digoxigenin-11-UTP (Roche) using DNA templates amplified with primerslisted in Table S4. The dissected embryos were fixed with 4% paraformaldehyde,stored in absolute methanol, permeabilized, and hybridized with the probes at61 °C. Antidigoxigenin antibody conjugated with alkaline phosphatase was usedfor enzymatic visualization of the bound probes.

RNAi. Approximately 10 ng of dsRNA for each target gene was injected intoeach adult virgin female within a few days after eclosion. The eggs laid by theinjected mothers after 3 d and on were allowed to develop at 25 °C for 7–10 duntil hatching. Some of the embryos were subjected to phenotypic observa-tions, whereas other embryos were fixed and analyzed by scanning electronmicroscopy and in situ hybridization.

See SI Materials and Methods for complete details on the materialsand methods.

ACKNOWLEDGMENTS.We thank R. Futahashi, M. Moriyama, and T. Harumotofor technical advice; R. Koga, S. Koshikawa, and M. Matsunami for commentson the manuscript; and C. Ueda for embryo illustrations. This work was sup-ported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant25221107 (to T.F.). Y.M. was supported by the JSPS Fellowship for YoungScientists.

1. Margulis L, Fester R (1991) Symbiosis as a Source of Evolutionary Innovation (MITPress, Cambridge, MA), p 470.

2. McFall-Ngai M, et al. (2013) Animals in a bacterial world, a new imperative for the lifesciences. Proc Natl Acad Sci USA 110(9):3229–3236.

3. Moran NA, McCutcheon JP, Nakabachi A (2008) Genomics and evolution of heritablebacterial symbionts. Annu Rev Genet 42:165–190.

4. Douglas AE (2009) The microbial dimension in insect nutritional ecology. Funct Ecol23:38–47.

5. Koga R, Meng XY, Tsuchida T, Fukatsu T (2012) Cellular mechanism for selectivevertical transmission of an obligate insect symbiont at the bacteriocyte-embryo in-terface. Proc Natl Acad Sci USA 109(20):E1230–E1237.

6. Gottlieb Y, et al. (2008) Inherited intracellular ecosystem: Symbiotic bacteria sharebacteriocytes in whiteflies. FASEB J 22(7):2591–2599.

7. Koga R, Nikoh N, Matsuura Y, Meng XY, Fukatsu T (2013) Mealybugs with distinctendosymbiotic systems living on the same host plant. FEMS Microbiol Ecol 83(1):93–100.

8. Moran NA, Dale C, Dunbar H, Smith WA, Ochman H (2003) Intracellular symbionts ofsharpshooters (Insecta: Hemiptera: Cicadellinae) form a distinct clade with a smallgenome. Environ Microbiol 5(2):116–126.

9. Koga R, Bennett GM, Cryan JR, Moran NA (2013) Evolutionary replacement of obli-gate symbionts in an ancient and diverse insect lineage. Environ Microbiol 15(7):2073–2081.

10. Balmand S, Lohs C, Aksoy S, Heddi A (2013) Tissue distribution and transmission routesfor the tsetse fly endosymbionts. J Invertebr Pathol 112(Suppl):S116–S122.

11. Hosokawa T, et al. (2012) Reductive genome evolution, host-symbiont co-speciationand uterine transmission of endosymbiotic bacteria in bat flies. ISME J 6(3):577–587.

12. Login FH, et al. (2011) Antimicrobial peptides keep insect endosymbionts undercontrol. Science 334(6054):362–365.

13. Toju H, et al. (2010) “Candidatus Curculioniphilus buchneri,” a novel clade of bacterialendocellular symbionts from weevils of the genus Curculio. Appl Environ Microbiol76(1):275–282.

14. Buchner P (1965) Endosymbiosis of Animals with Plant Microorganisms (Interscience,New York), p 909.

15. Braendle C, et al. (2003) Developmental origin and evolution of bacteriocytes in theaphid-Buchnera symbiosis. PLoS Biol 1(1):E21.

16. Gilbert SF, Epel D (2009) Ecological Developmental Biology: Integrating Epigenetics,Medicine, and Evolution (Sinauer Associates, Sunderland, MA), p 459.

17. Miura T, et al. (2003) A comparison of parthenogenetic and sexual embryogenesis ofthe pea aphid Acyrthosiphon pisum (Hemiptera: Aphidoidea). J Exp Zoolog B Mol DevEvol 295(1):59–81.

18. Matsuura Y, et al. (2012) Evolution of symbiotic organs and endosymbionts in lygaeidstinkbugs. ISME J 6(2):397–409.

19. Kuechler SM, Renz P, Dettner K, Kehl S (2012) Diversity of symbiotic organs andbacterial endosymbionts of lygaeoid bugs of the families Blissidae and Lygaeidae(Hemiptera: Heteroptera: Lygaeoidea). Appl Environ Microbiol 78(8):2648–2659.

20. Liu PZ, Kaufman TC (2004) hunchback is required for suppression of abdominalidentity, and for proper germband growth and segmentation in the intermediategermband insect Oncopeltus fasciatus. Development 131(7):1515–1527.

21. Futahashi R, et al. (2011) Laccase2 is required for cuticular pigmentation in stinkbugs.Insect Biochem Mol Biol 41(3):191–196.

22. Carroll SB, Greiner JK, Weatherbee SD (2005) From DNA to Diversity: Molecular Ge-netics and the Evolution of Animal Design (Blackwell, Malden, MA), 2nd Ed, p 268.

23. Angelini DR, Kaufman TC (2004) Functional analyses in the hemipteran Oncopeltusfasciatus reveal conserved and derived aspects of appendage patterning in insects.Dev Biol 271(2):306–321.

24. Angelini DR, Kaufman TC (2005) Insect appendages and comparative ontogenetics.Dev Biol 286(1):57–77.

25. Angelini DR, Liu PZ, Hughes CL, Kaufman TC (2005) Hox gene function and interactionin the milkweed bug Oncopeltus fasciatus (Hemiptera). Dev Biol 287(2):440–455.

26. Khila A, Abouheif E, Rowe L (2009) Evolution of a novel appendage ground plan inwater striders is driven by changes in the Hox gene Ultrabithorax. PLoS Genet 5(7):e1000583.

27. Werner T, Koshikawa S, Williams TM, Carroll SB (2010) Generation of a novel wingcolour pattern by the Wingless morphogen. Nature 464(7292):1143–1148.

28. Hosokawa T, Koga R, Kikuchi Y, Meng XY, Fukatsu T (2010) Wolbachia as a bacter-iocyte-associated nutritional mutualist. Proc Natl Acad Sci USA 107(2):769–774.

29. Cryan JR, Urban JM (2012) Higher-level phylogeny of the insect order Hemiptera: IsAuchenorrhyncha really paraphyletic? Syst Entomol 37:7–21.

30. Cui Y, et al. (2013) Phylogenomics of Hemiptera (Insecta: Paraneoptera) based onmitochondrial genomes. Syst Entomol 38:233–245.

31. Kikuchi Y, Hosokawa T, Fukatsu T (2008) Diversity of bacterial symbiosis in stinkbugs.Microbial Ecology Research Trends, ed Dijk TV (Nova Science, New York), pp 39–63.

32. Collin R, Miglietta MP (2008) Reversing opinions on Dollo’s Law. Trends Ecol Evol23(11):602–609.

33. Wake DB, Wake MH, Specht CD (2011) Homoplasy: From detecting pattern to de-termining process and mechanism of evolution. Science 331(6020):1032–1035.

34. Subramanian S, Stacey G, Yu O (2007) Distinct, crucial roles of flavonoids during le-gume nodulation. Trends Plant Sci 12(7):282–285.

35. Mandel MJ, et al. (2012) Squid-derived chitin oligosaccharides are a chemotactic signalduring colonization by Vibrio fischeri. Appl Environ Microbiol 78(13):4620–4626.

36. Kremer N, et al. (2013) Initial symbiont contact orchestrates host-organ-wide tran-scriptional changes that prime tissue colonization. Cell Host Microbe 14(2):183–194.

37. Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ: 25 years of imageanalysis. Nat Methods 9(7):671–675.

38. Liu P, Kaufman TC (2009) Morphology and husbandry of the large milkweed bug,Oncopeltus fasciatus. Cold Spring Harb Protocols, 10.1101/pdb.prot5261.

39. Nagaso H, Murata T, Day N, Yokoyama KK (2001) Simultaneous detection of RNA andprotein by in situ hybridization and immunological staining. J Histochem Cytochem49(9):1177–1182.

40. Ogawa K, Ishikawa A, Kanbe T, Akimoto S, Miura T (2012) Male-specific flight ap-paratus development in Acyrthosiphon pisum (Aphididae, Hemiptera, Insecta): com-parison with female wing polyphenism. Zoomorphology 131:197–207.

41. Li M, Tian Y, Zhao Y, Bu W (2012) Higher level phylogeny and the first divergencetime estimation of Heteroptera (Insecta: Hemiptera) based on multiple genes. PLoSOne 7(2):e32152.

42. Shinmyo Y, et al. (2005) caudal is required for gnathal and thoracic patterning and forposterior elongation in the intermediate-germband cricket Gryllus bimaculatus.MechDev 122(2):231–239.

43. Tian X, et al. (2011) Phylogeny of pentatomomorphan bugs (Hemiptera-Heteroptera:Pentatomomorpha) based on six Hox gene fragments. Zootaxa 2888:57–68.

44. Futahashi R (2011) Whole-mount in situ hybridization of sectioned tissues of specieshybrids to detect cis-regulatory changes in gene expression pattern. MethodsMol Biol 772:319–328.

Matsuura et al. PNAS | July 28, 2015 | vol. 112 | no. 30 | 9381

EVOLU

TION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

12, 2

020