Ultrabithorax is essential for bacteriocyte development

Significance

Among the most fundamental questions in developmental biology is how novel cell types have emerged in the metazoan evolution. Among the most challenging questions in evolutionary biology is how sophisticated symbiotic associations have evolved through less intimate interorganismal interactions. These fundamental biological issues are crystalized in the evolution and development of insect’s bacteriocytes specialized for harboring symbiotic bacteria. Here, we report that a conserved transcription factor Ultrabithorax is essential for bacteriocyte development in an insect, thereby uncovering a molecular mechanism underlying the emergence of the novel host cells for symbiosis. Our finding highlights the importance of developmental cooption of preexisting transcription factors and sheds new light on a long-lasting enigma in evolutionary developmental biology.

Abstract

Symbiosis often entails the emergence of novel adaptive traits in organisms. Microbial symbionts are indispensable for diverse insects via provisioning of essential nutrients, wherein novel host cells and organs for harboring the microbes, called bacteriocytes and bacteriomes, have evolved repeatedly. Molecular and developmental mechanisms underpinning the emergence of novel symbiotic cells and organs comprise an unsolved question in evolutionary developmental biology. Here, we report that a conserved homeotic gene, Ultrabithorax, plays a pivotal role in the bacteriocyte differentiation in a hemipteran insect Nysius plebeius. During embryonic development, six pairs of aggregated presumptive bacteriocytes appear on both sides of six abdominal segments, incorporate the symbiotic bacteria at the stage of germband retraction, and fuse into a pair of lateral bacteriomes at the stage of germband flip, where bacteriocyte-associated Ultrabithorax expression coincides with the symbiont infection process. Suppression of Ultrabithorax expression by maternal RNA interference results in disappearance of the bacteriocytes and the symbiont localization therein, suggesting that Ultrabithorax is involved in differentiation of the host cells for symbiosis. Suppression of other homeotic genes abdominal-A and Antennapedia disturbs integrity and positioning of the bacteriomes, affecting the configuration of the host organs for symbiosis. Our findings unveil the molecular and developmental mechanisms underlying the bacteriocyte differentiation, which may have evolved either via cooption of the transcription factors for inducing the novel symbiotic cells, or via revival of the developmental pathway for the bacteriocytes that had existed in the ancestral hemipterans.

Symbiosis is the source of novel adaptive traits, thereby contributing to organismal evolution and diversification (1, 2). A variety of insects are obligatorily dependent on microbial symbionts via provisioning of essential nutrients lacking in their diets (3, 4), wherein novel host cells and organs for harboring the microbes, called bacteriocytes and bacteriomes, have evolved repeatedly in such insect groups as hemipterans (aphids, whiteflies, mealybugs, leafhoppers, spittlebugs, etc.) (5–9), dipterans (tsetse flies, bat flies, etc.) (10, 11), coleopterans (weevils, etc.) (12, 13), and many others (14). Despite a considerable body of embryological descriptions (14), molecular mechanisms underlying the bacteriocyte differentiation have been a long-lasting enigma in evolutionary developmental biology (15, 16). Although cellular and developmental aspects of the bacteriocyte formation have been best documented for the pea aphid, Acyrthosiphon pisum (5, 15, 17), the seed bug Nysius plebeius and allied heteropteran bugs of the superfamily Lygaeoidea have recently emerged as a promising model system for investigating the development, evolution, and origin of the bacteriocytes, on the grounds that (i) heteropteran bugs are generally associated with gut symbiotic bacteria without bacteriocytes; (ii) thus, the bacteriocytes in these lygaeoid species are regarded as a novel trait whose evolution was a relatively recent event (18, 19); and (iii) in N. plebeius and allied lygaeoid species, RNA interference (RNAi) works efficiently, which enables functional analysis of genes involved in the bacteriocyte formation (20, 21). Here we demonstrate that, by making use of the emerging model insect N. plebeius, conserved homeobox transcription factors, in particular Ultrabithorax (Ubx), are involved in the development of the host cells and organs specialized for harboring the symbiotic bacteria.

Results and Discussion

Bacteriocyte Development During Embryogenesis of N. plebeius.

We performed a detailed description of the embryogenesis of N. plebeius with special focus on the dynamics of the bacteriocyte-associated gammaproteobacterial symbiont “Candidatus Schneideria nysicola” (18) (Figs. 1 and 2, Table S1, and Movie S1). The symbiotic bacteria were found as an aggregate at the anterior pole of newly 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 abdominal segments of the germband (36–60 h; Figs. 1 I and J and 2 E–G). After germband retraction (∼72 h; Fig. 2H), the symbionts migrated from the abdominal population to presumptive bacteriocytes that appeared on both sides of abdominal segments A2–A7 as 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 coherent bacteriome located at abdominal segments A2–A4 (84–96 h; Figs. 1 M and N, and 2 K–M and Q–R). After the symbiont infection, the bacteriocytes accumulated red pigment and became easily recognizable (∼84 h; Figs. 1 C and 2 R–T). The red pigmentation provided a visible marker useful for tracing the bacteriocytes, although it is unclear whether the red pigment was derived from the symbiont or the host.

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 after oviposition. (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) Embryos of ∼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) in the 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 third abdominal segments, respectively; ob, ovarial bacteriocytes in adult female; T3, third thoracic segment.

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) The blastoderm stage (12–24 h after oviposition). (B) The early invagination stage (24–27 h). (C) Elongation of the germband (∼30 h). (D) Further elongation of the 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 abdominal segments (∼48 h). (G) The germband whose appendages are growing further (∼60 h). (H) The germband retraction stage (∼72 h), at which the embryonic body thickens. (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 symbionts aggregating 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 of the 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 postkatatrepsis embryo (∼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 the bacteriomes 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 localized before migrating to the presumptive bacteriocytes is recognizable by reddish hue (arrow). (R) An embryo in which symbiont migration to the presumptive bacteriocytes 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.)

Table S1.

Major developmental and symbiotic events during the embryogenesis of N. plebeius

Approx. time, h	Developmental event	Symbiotic event in N. plebeius
Before oviposition	Oogenesis and chorionization of embryo	Infection of symbiotic bacteria at the anterior pole of oocyte
0–12	Fertilization and nuclear division, followed by blastderm formation
12–24	After cellularization, invagination from a posterior–ventral region	Symbionts spreading over the anterior region of embryo in yolk
24–27	Anatrepsis: germband elongation
27–30
30–33	Anatrepsis: germband reaching the anterior end of egg	Gradual posterior movement of symbionts, coincident with arrival of the tip of germband
33–36	Anatrepsis: tip of germband rolling back to the posterior side	Symbionts trapped by abdominal segments of germband
36–48	Germband growing to form head lobes and appendages
48–60
60–72
72–84	Germband retraction with appearance of eye spots	Symbiont infection to bacteriocytes at distal rims of abdominal segments A2–A7, with bacteriocytes developing deep red color
84–96	Katatrepsis: germband turning backward, flipping upside down, and digesting extraembronic membrane	Fusion of bacteriocytes, forming a pair of bacteriomes on both sides of abdomen
96–108	Katatrepsis completion and onset of dorsal closure	Bacteriomes becoming oval and positioned at abdominal segments A2–A4
108–120	Embryo retraction
120–130
145–160	Epidermal cuticle pigmentation
168–192	Egg hatching

Bacteriocyte-Associated Expression of Ubx in Embryogenesis of N. plebeius.

The homeobox genes encode transcription factors that assign segment identities and specify functional body parts in the development of insects and other animals (16, 22). In the embryogenesis of the pea aphid A. pisum, some homeobox gene products—including an appendage-patterning transcription factor Distal-less (Dll), homeotic proteins Ubx or Abdominal-A (Abd-A), and a segment polarity protein Engrailed (En)—were shown to localize to the bacteriocytes, although their functions in the symbiotic cells have been elusive (15). We investigated the spatial expression patterns of some of these homeobox genes during the development of N. plebeius (Fig. 3 A–F and Fig. S1). Dll was prominently expressed at the distal portion of appendages in antennal, labial, and thoracic segments and faintly detected in the maxillary segments, but not expressed in the areas of the presumptive bacteriocytes (Fig. 3 A and A′), which agreed with the expression patterns of Dll in a related lygaeid species without bacteriocytes, the milkweed bug Oncopeltus fasciatus (23, 24). Notably, Ubx exhibited remarkable expression patterns in the abdominal regions where the presumptive bacteriocytes were supposed to differentiate: Although the strong expression at the abdominal segment A1 and the milder expression in the following abdominal segments were observed during the germband elongation and retraction (Fig. S1 A–C), which are typical of the Ubx expression in O. fasciatus and other insects (24, 25), Ubx was also strongly expressed as six pairs of clusters on both sides of abdominal segments A2–A7 after the germband retraction (Fig. 3 B–D and Fig. S1D), which agreed with the locations of the presumptive bacteriocytes (Figs. 1 K and L and 2J). Meanwhile, abd-A was expressed broadly across the abdominal segments without specific association with the presumptive bacteriocytes (Fig. 3 E and F and Fig. S1 H–J).

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 Dll expression 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 of Ubx 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 the homeobox genes in the abdominal segments. Smaller arrowheads indicate bacteriocyte-associated expressions, where numbers 1–6 correspond to the regions of the 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 of Ubx-suppressed embryos just after katatrepsis (∼100 h). (K and L) Lateral views of abd-A–suppressed embryos just after katatrepsis (∼100 h). (M and M′) Dorsal and z-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 microscopic images in which blue and green signals indicate the host nuclei and the symbiotic bacteria, respectively. (Scale bars: 100 μm.)

Fig. S1.

Expression patterns of Ubx and abd-A during embryogenesis of N. plebeius. Embryos carefully dissected from eggs were subjected to in situ hybridization of the homeobox genes. The embryos are oriented such that the anterior pole of the egg is to the left and the dorsal surface of the egg is to the top. (A) An early elongating germband (∼36 h), exhibiting strong Ubx signals at the abdominal segment 1 (A1) and weaker signals at the other abdominal segments. (B) An elongating germband (40–50 h) with Ubx expression in the abdominal segments 1–7 (A1–A7). (C) An elongating germband (50–60 h) with robust Ubx expression in A1 and weaker expression in A2–A7. In A–C, arrowheads point strong Ubx expression in A1, whereas line drawings encircle the other abdominal segments with moderate Ubx expression. (D) A retracting germband (72–84 h) with six pairs of Ubx expression foci in A2–A7, which agree with the locations of the presumptive bacteriocytes, in addition to the Ubx expression in A1. Note that the symbiont infection to the presumptive bacteriocytes occurs at this stage. (E–G) Negative control embryos hybridized with Ubx sense probe, exhibiting no specific signals. (H) An elongating germband (50–60 h) with abd-A expression in seven abdominal segments from A2 to A8. (I) Ventro-lateral view of a retracting germband (60–72 h) with abd-A signals in the same abdominal segments as above. (J) An embryo after germband retraction (∼84 h), in which the bacteriocyte formation and the symbiont infection are taking place, exhibiting faint abd-A signals in the abdominal segments. In H and J, line drawings indicate the abdominal regions with abd-A expression. (K–M) Negative control embryos hybridized with abd-A sense probe, exhibiting no specific signals. (Scale bars: 100 µm.)

Disappearance of Bacteriocytes by Ubx Suppression.

In heteropteran species including O. fasciatus, injection of double-stranded RNA (dsRNA) into mother insects was reported to efficiently suppress their offspring’s gene expression during embryogenesis (20, 26). Using the maternal RNAi technique, we suppressed the embryonic expression of these genes and observed the development of N. plebeius. Dll suppression affected neither the bacteriocyte formation nor the symbiont localization (Fig. 3 G and H). Strikingly, Ubx suppression resulted in disappearance of the red-pigmented bacteriocytes and the symbiont localization associated with them (Fig. 3 I and J). Detailed observations of the Ubx-suppressed embryos revealed that, in contrast to the control embryos wherein the symbionts aggregating in the abdomen were migrating to the bacteriocytes (Fig. 3 M and M′) and strictly localized therein (Fig. 3O), the symbionts were dispersing within the embryonic body (Fig. 3 N and N′) and lost subsequently (Fig. 3P). Meanwhile, abd-A suppression did not lead to disappearance of the bacteriocytes (Fig. 3 K, L, Q, and R), although the spatial organization of the bacteriocytes was affected as detailed later.

Although the majority of the control embryos hatched normally with a pair of red-pigmented bacteriomes in the abdomen (Fig. 4 A–C and Table S2), considerable proportions of the RNAi-treated embryos failed to hatch with morphological abnormalities typical of the homeobox gene defects (23, 25) (Table S2). Dll suppression produced 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 42 hatchlings with strong phenotypes, whose abdominal segment A1 grew extra appendages, lacked the bacteriomes (Fig. 4 G–I and Table S2), whereas 48 hatchlings with moderate phenotypes lacked the symbiotic organs partially (Fig. S2 E and F and Table S2). Of 130 Ubx-suppressed unhatched embryos, 111 embryos lacked the bacteriomes (Fig. S2 G and H and Table S2). The absence of the bacteriomes in the Ubx-suppressed hatchlings and embryos was evidently and significantly more frequent than in the Dll-suppressed, abd-A–suppressed, and control hatchlings and embryos (Table S3).

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, whose appendages are severely truncated. (G–I) Ubx-suppressed nymphs, whose abdominal segment A1 exhibits thorax-like pigmentation (arrow) and grows appendage-like structures. (J–L) abd-A–suppressed nymphs, whose abdominal segments grow appendage-like structures on both sides. (M–O) Antp-suppressed nymphs, 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 electron microscopic 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 the bacteriomes. ap, appendage-like structure; vm, vitelline membrane. (Scale bars: 100 μm.) For mild and moderate RNAi phenotypes, see Fig. S2.

Table S2.

Effects of maternal RNAi on nymphs and embryos of N. plebeius

Injected ds RNA (no. of mothers injected)	Hatched eggs					Unhatched eggs			Hatched nymphs in total eggs
	No. of visible effects	Specific phenotypes			Abnormal development	Hatching failure	Abnormal development	No germband formation
		Mild	Moderate	Strong
Dll (10)	0 (0.0) [0/0/0]	26 (18.7) [26/0/0]	4 (2.9) [4/0/0]	5 (3.6) [5/0/0]	7 (5.0) —	8 (5.8) [8/0/0]	63 (45.3) —	26 (18.7) —	42/139 (30.2) [43/0/0]
Ubx (21)	41 (12.2) [41/0/0]	23 (6.9) [23/0/0]	48 (14.3) [0/48/0]	42 (12.5) [0/0/42]	6 (1.8) —	130 (38.8) [0/19/111]	29 (8.7) —	16 (4.8) —	160/335 (47.8) [64/67/153]
abd-A (22)	147 (29.9) [147/0/0]	68 (13.8) [68/0/0]	64 (13.0) [64/0/0]	23 (4.7) [23/0/0]	27 (5.5) —	68 (13.8) [68/0/0]	41 (8.4) —	53 (10.8) —	329/491 (67.0) [370/0/0]
Antp (21)	101 (35.7) [101/0/0]	56 (19.8) [56/0/0]	25 (8.8) [25/0/0]	13 (4.6) [13/0/0]	16 (5.7) —	14 (4.9) [14/0/0]	27 (9.5) —	31 (11.0) —	211/283 (74.6) [209/0/0]
Water control (10)	90 (81.1) [90/0/0]	—	—	—	4 (3.6) —	6 (5.4) [6/0/0]	8 (7.2) —	3 (2.7) —	94/111 (84.7) [96/0/0]
β-lactamase control (10)	69 (60.5) [69/0/0]	—	—	—	33 (28.9) —	10 (8.8) [10/0/0]	1 (0.9) —	1 (0.9) —	102/114 (89.5) [79/0/0]

In parentheses is the percentage to the total number of eggs inspected. In brackets are presence/near absence/complete absence of the bacteriomes. Mild, moderate, and strong refer to mildly, moderately, and strongly affected individuals by RNAi experiments judged with their morphological defects specific to the treatment, respectively, whose morphologies are described in Fig. S3. Abnormal development refers to hatchlings or embryos that exhibited developmental arrest, abnormal morphologies, and/or morphological destruction due to cannibalism, in which development and integrity of the bacteriomes were difficult to evaluate.

Fig. S2.

RNAi phenotypes of Dll, Ubx, abd-A, and Antp in newborn nymphs and embryos of N. plebeius. (A) A Dll-suppressed nymph with mild phenotypes, whose tips of the appendages are reduced. (B) A Dll-suppressed nymph with moderate phenotypes, whose appendages are remarkably truncated. (C and D) Dll-suppressed unhatched embryos. In A–D, the bacteriomes are present despite the remarkable morphological defects (filled arrowheads). (E) A Ubx-suppressed nymph with mild phenotypes, in which the abdominal segment 1 (A1) grows tiny appendage-like structures and exhibits no thorax-like dorsal pigmentation. The bacteriomes look normal. (F) A Ubx-suppressed nymph with moderate phenotypes, in which thorax-like dorsal pigmentation occurs in A1 and appendage-like structures with several segments grow. (F’) The bacteriomes are reduced to small red spots. (G) A Ubx-suppressed unhatched embryo with appendage-like structures and spotty bacteriomes. (H) A Ubx-suppressed unhatched embryo with no bacteriomes. (I) An abd-A–suppressed nymph with mild phenotypes, in which A2 grows appendage-like structures and the following abdominal segments exhibit dark pigmentation on the ventral side. The bacteriomes look normal. (J) An abd-A–suppressed nymph with moderate phenotypes, in which appendage-like structures occur on A2–A8 in different numbers ranging from 2 to 4 each. The bacteriomes are either single or separated into a few pieces. (K) An abd-A–suppressed unhatched embryo with appendage-like structures and normal bacteriomes. (L) An abd-A–suppressed unhatched embryo with appendage-like structures and separate bacteriomes. (M) An Antp-suppressed nymph with mild phenotypes, in which distal portion of femurs of the legs from the thoracic segment 1 (T1) legs are swollen (M′). The bacteriomes look normal. (N) An Antp-suppressed nymph with moderate phenotypes, in which femurs of the legs are swollen and shortened. The bacteriomes are found at A1–A2 or in the coxae of the legs from T3. (O) An Antp-suppressed unhatched embryo with bacteriomes at A1–A2. (P) An Antp-suppressed unhatched embryo exhibiting abnormal development. Filled arrowheads indicate the locations of the bacteriomes, and open arrowhead highlights the absence of the bacteriomes. The locations of the bacteriomes were also verified by in situ hybridization. A1, abdominal segment 1; A2, abdominal segment 2; an, antenna; ap, appendage-like structure; lg, leg; sw, swollen part of femur; T3, thoracic segment 3. (Scale bars: 100 μm.)

Table S3.

Relationship between RNAi treatments and bacteriome development in N. plebeius

RNAi treatment	Bacteriome +† A	Bacteriome −‡ B	Bacteriome disappearance rate§ for B, %	Adjusted standardized residuals¶ for B	Bacteriome − or ±|| C	Bacteriome disappearance rate§ for C, %	Adjusted standardized residuals¶ for C
Dll	43 [35/8]	0 [0/0]	0.0 [0.0/0.0]	−2.824	0 [0/0]	0.0 [0.0%/0.0]	−3.306
Ubx	64 [64/0]	153 [42/111]	70.5* [39.6*/100*]	25.725	210 [90/130]	76.7* [58.4*/100*]	27.565
abd-A	370 [302/68]	0 [0/0]	0.0 [0.0/0.0]	−10.175	0 [0/0]	0.0 [0.0/0.0]	−11.742
Antp	209 [195/14]	0 [0/0]	0.0 [0.0/0.0]	−6.840	0 [0/0]	0.0 [0.0/0.0]	−7.958
Water control	96 [90/6]	0 [0/0]	0.0 [0.0/0.0]	−4.341	0 [0/0]	0.0 [0.0/0.0]	−5.071
β-lactamase control	79 [69/10]	0 [0/0]	0.0 [0.0/0.0]	−3.901	0 [0/0]	0.0 [0.0/0.0]	−4.561

^†Total number of newborn nymphs and full-grown unhatched embryos with bacteriomes [number of newborn nymphs with bacteriomes/number of full-grown unhatched embryos with bacteriomes].

^‡Total number of newborn nymphs and full-grown unhatched embryos without bacteriomes [number of newborn nymphs without bacteriomes/number of full-grown unhatched embryos without bacteriomes].

^§B or C/(A + B or C) x 100 [the rate for newborn nymphs/the rate for full-grown unhatched embryos]. *P < 0.01 (pairwise comparisons, Fisher’s test).

^¶After χ² test (χ² = 661.797, P = 8.9 x 10⁻¹⁴¹ for B; χ² = 759.825, P = 5.7 x 10⁻¹⁶² for C), the adjusted standardized residuals were calculated for evaluating how far the observed count is from the expected count. All values were statistically significant (P < 0.01).

^||Total number of newborn nymphs and full-grown unhatched embryos in the absence or near-absence of bacteriomes [number of newborn nymphs in the absence or near-absence of bacteriomes/number of full-grown unhatched embryos in the absence or near-absence of bacteriomes].

Unfused Bacteriomes by abd-A Suppression.

Notably, abd-A suppression affected the spatial organization of the bacteriocytes: In contrast to a pair of fused bacteriomes in the control mature embryos (Figs. 1 O and P and 2 S and T), the abd-A–suppressed mature embryos exhibited separate bacteriocyte clusters (Fig. 3 K, L, Q, and R), as observed in the control younger embryos (Fig. 3M). Whereas most of 329 abd-A–suppressed hatchlings and 68 embryos retained the bacteriomes, all 23 hatchlings with strong phenotypes whose abdominal segments A2–A8 were transformed into thorax-like identity, and also some embryos, exhibited unfused bacteriomes, ranging from two to six in number, in the abdomen (Fig. 4 J–L, Fig. S2 I–L, and Table S2). These results suggest that abd-A may be involved in the integrity of the symbiotic organs via, presumably, homeotic regulation of segment identity in the abdomen of N. plebeius.

Translocated Bacteriomes by Antp Suppression.

Additionally, we performed maternal RNAi of Antennapedia (Antp), which is a homeotic gene expressed at thoracic and anterior abdominal segments in O. fasciatus and other insects (24, 25). Antp suppression 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 shortened tibiae, exhibited the bacteriomes at the T3 thoracic segment and the A1–A2 anterior abdominal segments. The bacteriomes were found 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 invasion of posterior segmental identity into the thoracic segment, suggesting the possibility that, although speculative, Antp may negatively regulate the bacteriocyte differentiation at the T3 and A1 segments during the embryogenesis of N. plebeius.

Conclusion

The disappearance of the bacteriocytes by Ubx suppression, in combination with the bacteriocyte-associated Ubx expression, strongly suggests that Ubx plays a pivotal role in the bacteriocyte differentiation in N. plebeius. In addition, abd-A and Antp may affect the bacteriome integrity and positioning. Our findings shed light on the long-lasting enigma as to what molecular mechanisms underlie the development of insect’s bacteriocytes and bacteriomes, which unveil the importance of developmental cooption of the preexisting and conservative homeobox transcription factors for acquisition of the novel cells and organs for symbiosis. Fig. 5 schematically illustrates expression patterns of homeotic genes during the embryogenesis of N. plebeius. The mechanism governing the bacteriocyte-associated expression pattern of Ubx is currently unknown and deserves future study. As reported in fruit flies (27), it seems plausible, although speculative, that Ubx might have acquired a novel cis-regulatory element for the unique regional expression in an ancestor of N. plebeius.

Fig. 5.

Schematic illustration of expression patterns of homeotic genes in the embryogenesis of N. plebeius. Note that we inspected the detailed expression patterns of Ubx and abd-A only, and the expression patterns of the other genes are according to previous studies on a related lygaeid bug O. fasciatus (24, 25).

Perspectives

It should be noted that the bacteriocyte-specific Ubx/Abd-A localization was also observed in the aphid A. pisum (Sternorrhyncha:Aphididae) (15), which is phylogenetically distinct from N. plebeius (Heteroptera:Lygaeidae) in the Hemiptera. Considering that the heteropterans without the bacteriocytes/bacteriomes [with exceptions of bedbugs (Cimicidae) and some seed bugs (Lygaeoidea)] (18, 19, 28) constitute a monophyletic group nested within the Hemiptera and all of the other hemipteran groups 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 of the Heteroptera, and evolved again in the Lygaeoidea, to which N. plebeius belongs (Fig. S3B) (14, 18, 31). Hence, the bacteriocytes/bacteriomes of N. plebeius may have evolved either via cooption of the homeobox transcription factor for inducing the novel cells for symbiosis or via revival of the developmental pathway for the bacteriocytes/bacteriomes that had existed in the ancestral hemipterans but been disrupted in the lineage of the heteropteran bugs. Whether the regain of the bacteriocytes/bacteriomes in association with the localized Ubx expression in the stinkbug lineage is regarded as reversal, parallelism, or convergence is an intriguing issue in evolutionary developmental biology (32, 33), which should be addressed by comparative studies on molecular mechanisms underlying the development of symbiotic cells and organs in A. pisum, N. plebeius, and other hemipteran species. It is still an enigma what type of cells comprises the developmental origin of the bacteriocytes. It is also of interest, but totally unknown, what mechanisms underlie the specific targeting of the bacterial symbiont to the newly formed bacteriocytes at the stage of germband retraction in N. plebeius (Figs. 1 K and L and 3 M and M′). Although Ubx starts specific expression in the presumptive bacteriocytes at this stage, it seems likely that not Ubx itself but some factor(s) downstream of Ubx must be responsible for the host–symbiont interplay, like flavonoids in the legume-Rhizobium nitrogen-fixing symbiosis (34) and chitin oligosaccharides in the squid–Vibrio luminescent symbiosis (35, 36). These evolutionary and developmental issues would be addressed by transcriptomics of the presumptive bacteriocytes microdissected from embryos of N. plebeius or other hemipteran insects.

Fig. S3.

Evolution of symbiotic organs in the Hemiptera. (A) Phylogeny of higher taxa in the Hemiptera (29). Presence of the bacteriome is highly conserved among all of the taxa but the Heteroptara, suggesting a loss of the symbiotic organ in this group. (B) Phylogeny of higher taxa in the Heteroptera (41). Absence of the bacteriome is commonly observed across all of the taxa, with exceptions of some cimicids in the Cimicomorpha and a few lygaeoids in the Pentatomomorpha. +BM and −BM indicate presence and absence of the bacteriome, respectively.

Materials and Methods

Insect Material.

N. plebeius was maintained at 25 °C under a long-day regime (16-h light, 8-h dark) on sunflower seeds, whole wheat, and distilled water supplemented with 0.05% ascorbic acid.

In Situ Hybridization of Symbiont.

Whole-mount fluorescence in situ hybridization targeting bacterial 16S rRNA was performed to visualize the symbiont localization in embryos and whole insects of N. plebeius as described (18) with a probe listed in Table S4.

Table S4.

Oligonucleotides used in this study

Target	Name	Sequence,* 5′–3′	Orientation	Source
AlexaFluor647-labeled probe for in situ hybridization targeting bacterial 16S rRNA
16S rRNA	NpSch1274	TATACTTTTTGAGGTTCGCTTGCTC	Antisense	Ref. 18
Primers for PCR, cloning and sequencing of homeobox genes
Homeobox genes in general	F	GAGCTGGAGAAGGARTT	Forward	Ref. 42
	R	CCGCCGGTTCTGRAACCA	Reverse	Ref. 42
Dll	Dll_H0	TAYCCNTTYCCNMCNATGC	Forward	This study
	Dll_h2	GNGCNGCYTTCATCATYTT	Reverse	This study
	5F1	CTCCGCACCACGACCAGT	Forward	This study
	5R1	ACTGGTCGTGGTGCGGAG	Reverse	This study
	qF1	CCACAACAGCTATACAGGATACCA	Forward	This study
	qR1	GGCTCCTACATCCTCCAGAA	Reverse	This study
	3F2	CGTTAAACACCGGTTTCTGG	Forward	This study
Ubx	Ubx-F1	ATGAACTCTTATTTCGAGCAGACG	Forward	This study
	Ubx-R2	GGTCTGGTACCTCGTGTAGG	Reverse	This study
	Ubx-R3	GATGGCCACGAAGAGATGAC	Reverse	This study
abd-A	5′ end of abd-A	TCATCATCGATAGYATGYTNCCNAARTA	Forward	Ref. 43
	5′ end of abd-A	GATGGACATCCANGGRTANCKNGG	Reverse	Ref. 43
	abdA-F1	CTTCTGCGGTAAACTACTCC	Forward	This study
	abdA-R1	GTAGGGGTACATCCGTGAC	Reverse	This study
	abdAISH-F1	GGGCGGTTATAGTGTGTCTA	Forward	This study
	abdAISH-R1	GTGTACCGGCAGGACTT	Reverse	This study
Antp	AntpF0	GCTCTGCCTCACCGAG	Forward	This study
	AntpF1	CGGATGAAGTGGAAGAAGGA	Forward	This study
	AntpR1	CGTATAATACATTACGCACATTG	Reverse	This study
Primers for synthesis of dsRNA for RNAi
Dll	Dll-T7F	taatacgactcactatagggCTCACCACCCAAAGACGTTA	Forward	This study
	Dll-T7R	taatacgactcactatagggTATTTACTCCGCCGATTTTG	Reverse	This study
Ubx	Ubx-T7F	taatacgactcactatagggCTGGAGCTGGAGAAGGAG	Forward	This study
	Ubx-T7R	taatacgactcactatagggATAATAAACGGAAGGTGGAC	Reverse	This study
abd-A	abdA-T7F	taatacgactcactatagggACAGAATCCATGTACGG	Forward	This study
	abdA-T7R	taatacgactcactatagggTGAGGTAGTGGTTGAAGTG	Reverse	This study
Antp	Antp-T7F	taatacgactcactatagggCGGATGAAGTGGAAGAAGGA	Forward	This study
	Antp-T7R	taatacgactcactatagggTAGCCTTTGGTAGCAACACA	Reverse	This study
β-lactamase	amp-T7F	taatacgactcactatagggCTATGTGGCGCGGTATTAT	Forward	Ref. 21
	amp-T7R	taatacgactcactatagggCAGAAGTGGTCCTGCAACT	Reverse	Ref. 21
Primers for synthesizing probes for in situ hybridization from the plasmid pT7blue
Plasmid insert	T7-F	taatacgactcactatagggAGACTAGTCATATGGAT	Forward	Ref. 44
	T7-R	taatacgactcactatagggAGACCCGGGGATCCGAT	Reverse	Ref. 44
Plasmid insert	R20side	GACTAGTCATATGGAT	Forward	Ref. 44
	U19side	ACCCGGGGATCCGAT	Reverse	Ref. 44

^*Lowercase letters indicate T7 promoter sequence.

Cloning of Homeobox Genes.

Total RNA was extracted from embryos of N. plebeius, reverse-transcribed to yield cDNA, and subjected to PCR amplification of the homeobox genes with degenerate primers listed in Table S4. The PCR products were cloned and sequenced, and the complete or partial cDNA sequences were determined by rapid amplification cDNA end methods.

In Situ Hybridization of Host’s Homeobox Genes.

Whole-mount in situ hybridization 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 primers listed in Table S4. The dissected embryos were fixed with 4% paraformaldehyde, stored in absolute methanol, permeabilized, and hybridized with the probes at 61 °C. Antidigoxigenin antibody conjugated with alkaline phosphatase was used for enzymatic visualization of the bound probes.

RNAi.

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

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

SI Materials and Methods

Insect Material.

Several strains of N. plebeius were established from field-collected insects and maintained at 25 °C under a long-day regime (16-h light, 8-h dark) on sunflower seeds, whole wheat, and distilled water supplemented with 0.05% ascorbic acid.

Embryo Collection and Fixation.

Adult females were allowed to oviposit on a piece of cotton for 1–3 h. The cotton was briefly soaked in 70% (vol/vol) ethanol for sterilization, rinsed in tap water, and subjected to egg isolation. The eggs were carefully collected by using fine forceps and a paintbrush in water under a dissection microscope, transferred onto dry filter paper in Petri dishes, and allowed to develop at 25 °C. At ∼3- to 12-h intervals, dozens of developing embryos were collected and incubated at 65 °C in Carnoy’s solution (ethanol:chlroroform:acetate, 6:3:1) on a heat block for 5–10 min and then chilled at 4 °C overnight. After washing with 90% (vol/vol) ethanol twice, the fixed embryos were stored at −20 °C until use.

Embryo Dissection and Symbiont Localization.

The fixed embryos were dissected in 90% (vol/vol) ethanol to remove chorion and vitelline membrane by using a glass capillary and fine forceps in a glass dish. Whole-mount fluorescence in situ hybridization of “Ca. S. nysicola” was performed by using an oligonucleotide probe NpSch1274 (Table S4) specifically targeting 16S rRNA of the bacterial symbiont as described (18). Z-stack projection images were reconstructed by ImageJ (37). Developmental events during the embryogenesis of N. plebeius were compared with those reported for O. fasciatus (38).

Cloning and Sequencing of Homeobox Genes.

During the embryonic development of N. plebeius for a week, ∼10 embryos each day were collected and pooled, and total RNA was extracted by using RNAiso plus (TaKaRa). The RNA sample was subjected to cDNA synthesis by using SMARTScribe Reverse Transcriptase (Clontech). The cDNA sample was subjected to PCR amplification of several homeobox genes with degenerate primers (Table S4) under a temperature profile of 95 °C for 2 min followed by 35 cycles of 95 °C for 30 s, 53 °C for 1 min, and 72 °C for 1.5 min. The PCR products were cloned and sequenced as described (18). Based on the partial gene sequences, specific primers for rapid amplification of cDNA ends were designed (Table S4), with which the complete or partial cDNA sequences were determined by using SMARTer RACE cDNA Amplification Kit (Clontech).

In Situ Hybridization.

Fixation of embryos and in situ hybridization of insect genes were performed according to previous studies (38, 39) with modifications. Sense and antisense RNA probes were synthesized by T7 RNA polymerase (TaKaRa) and digoxigenin-11-UTP (Roche) using the templates amplified by PCR with the primers (Table S4). Appropriately staged embryos were collected in a plastic tube, fixed in boiled water for 45 s, and immediately placed on ice. After replacing the water with heptane by changing the solvent twice, the tube was vigorously shaken by hand for 2 min, and then heptane was replaced by methanol prechilled at −20 °C. The embryos were washed several times with a Pasteur pipette to crack chorion and stored in methanol at −20 °C for at least 12 h. Then, the embryos were dissected from chorion and vitelline membrane under a dissecting microscope by using a pair of fine forceps and a glass capillary. After washing with ethanol and incubation with ethanol–xylene (1:1) for 30 min at room temperature, the embryos were gradually hydrated through 100%, 75%, 50%, 25%, and 0% (vol/vol) ethanol in diethylpyrocarbonate (DEPC)-treated water for 10 min each, and preserved in 80% (vol/vol) acetone in DEPC-treated water prechilled at −20 °C. After incubation at −20 °C for 20 min, the embryos were washed with PBTwx [PBS (137 mM NaCl, 8.1 mM Na₂HPO₄, 2.7 mM KCl, 1.5 mM KH₂PO₄, pH 7.4) containing 0.1% Tween 20 and 0.1% Triton-X 100] several times and postfixed with 4% (wt/vol) formaldehyde in PBS for 20 min. After thorough washing with PBTwx twice, the embryos were equilibrated with a hybridization buffer [50% (vol/vol) formamide, 5× SSC (pH 7.0), 0.3% CHAPS, 0.1% Tween 20, 1× Denhardt’s reagent, 0.5 g/mL yeast tRNA, and 0.5 g/mL heparin in DEPC-treated water] for 2 h, and incubated with the hybridization buffer containing 100 ng/mL digoxigenin-labeled RNA probe for 16 h at 61 °C. After incubation with a washing hybridization buffer [50% (vol/vol) formamide, 5× SSC, 0.3% CHAPS, and 0.1% Tween 20 in DEPC-treated water] for 30 min at 61 °C, the embryos were washed with a diluted series [50%, 25%, 12.5%, and 6.25% (vol/vol)] of the washing hybridization buffer with 0.1% Tween 20 and 0.3% CHAPS in DEPC-treated water for 30 min each at 61 °C. After thorough washing with PBTw (PBS with 0.1% Tween 20), the embryos were blocked with 2% (wt/vol) BSA in PBTw for 30 min at room temperature, and then incubated with antidigoxigenin antibody conjugated with alkaline phosphatase (Roche) at a concentration of 1/2,000 in PBTw containing 2% (wt/vol) BSA at 4 °C overnight. Unbound antibody was washed with PBTw five times for 10 min each. Then, bound antibody was visualized by incubation with nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate substrate [100 mM NaCl, 50 mM MgCl₂, 100 mM Tris⋅HCl (pH 9.5), 0.1% Tween-20, 0.38 mg/mL nitroblue tetrazolium chloride, 0.25 mg/mL 5-bromo-4-chloro-3-indolyl-phosphate, and 4-toluidine salt] at room temperature for 40 min. Finally, the reaction was stopped by washing in cold PBTw three times and mounted onto the glass slides with glycerol containing medium for observation.

RNAi.

dsRNA for silencing Dll (280 bp), Ubx (263 bp), abd-A (343 bp), and Antp (295 bp), and ampicillin-resistant gene beta-lactamase (438 bp) from pT7Blue vector (Novagen) for a negative control, were designed and synthesized as follows. Each plasmid containing a clone of the transcript was amplified by PCR with specific oligonucleotides containing T7 promoter region (Table S4) to create a template for RNA polymerization. After purifying the product with QIAquick PCR purification kit (QIAGEN), both sense and antisense strands of the transcript were simultaneously synthesized by T7 RNA polymerase (TaKaRa) with ∼1 µg of template DNA at 37 °C for 2 h. The RNA product was denatured at 70 °C for 10 min, annealed at 20 °C for 20 min, and treated with DNase I to remove the template DNA. The dsRNA was purified by RNeasy Mini Kit column (QIAGEN), quantified by Nanodrop ND-1000 (Thermo Fisher Scientific), and checked by electrophoresis in 1.0% agarose gels. Adult females several days after eclosion were injected with 10 ng of dsRNA in 0.2 mL of water by a glass capillary using FemtoJet (Eppendorf). The treated females were allowed to mate with males for 3 d. Eggs were collected from the fourth day and on, because the eggs laid during the initial 3 d do not exhibit RNAi effects (20). Some embryos were fixed during embryogenesis in Carnoy’s solution for whole-mount fluorescence in situ hybridization, whereas other embryos were allowed to hatch for morphological inspection. Morphological effects of RNAi were observed under a dissecting microscope, photographed, and categorized with reference to previous studies on O. fasciatus (24, 25). The levels of bacteriome development were categorized into present, near-absent, or absent, and the levels of integration/fragmentation of the bacteriomes were also recorded.

Scanning Electron Microscopy.

Scanning electron microscopy was performed as described (40). Whole nymphs were fixed in Carnoy’s solution and washed twice by absolute ethanol. The samples were hardened by hexamethyldisilazane for 1 h and washed in absolute ethanol twice for 15 min each. Then, the samples were incubated in 100% t-butanol twice for 15 min each. After preservation at −20 °C for 2 h, the samples were dried by using a freeze dryer (Hitachi ES-2030) and coated with gold by using an ioncoater (Eiko IB-3). Scanning electron micrographs were obtained with JSM-5510LV system (JEOL).