Researchers are exploring whether these ubiquitous fluorinated molecules might worsen infections or hamper vaccine effectiveness.
Image credit: Shutterstock/Dmitry Naumov.
-
Edited by Kathryn V. Anderson, Sloan–Kettering Institute, New York, NY, and approved November 15, 2007 (received for review September 26, 2007)
Abstract
Higher cyclorrhaphan flies including Drosophila develop a single extraembryonic epithelium (amnioserosa), which closes the germband dorsally. In most other insects two extraembryonic epithelia, serosa and amnion, line the inner eggshell and the ventral germband, respectively. How the two extraembryonic epithelia evolved into one is unclear. Recent studies have shown that, in the flour beetle Tribolium and in the milkweed bug Oncopeltus, the homeobox gene zerknüllt (zen) controls the fusion of the amnion with the serosa before dorsal closure. To understand the origin of the amnioserosa in evolution, we examined the expression and function of zen in the extraembryonic tissue of lower Cyclorrhapha. We show that Megaselia abdita (Phoridae) and Episyrphus balteatus (Syrphidae) develop a serosa and a dorsal amnion, suggesting that a dorsal amnion preceded the origin of the amnioserosa in evolution. Using Krüppel (Kr) and pannier (pnr) homologues of Megaselia as markers for serosal and amniotic tissue, respectively, we show that after zen RNAi all extraembryonic tissue becomes indistinguishable from amniotic cells, like in Tribolium but unlike in Drosophila, in which zen controls all aspects of extraembryonic development. Compared with Megaselia and Episyrphus, zen expression in Drosophila is extended to cells that form the amnion in lower Cyclorrhapha and is down-regulated at the developmental stage, when serosa cells in lower Cyclorrhapha begin to expand. These expression differences between species with distinct extraembryonic tissue organizations and the conserved requirement of zen for serosa development suggest that the origin of an amnioserosa-like epithelium was accompanied by expression changes of zen.
The amnioserosa is a unique extraembryonic epithelium of higher flies (1). Because it has no direct equivalent in other insects, it provides a model to study the evolution of new morphology in connection with the underlying developmental gene network. Detailed comparisons of extraembryonic development between fly species with and without an amnioserosa would help to understand the mechanism by which this tissue evolved. In Drosophila, the amnioserosa develops from a small portion of the dorsal blastoderm into a squamous polyploid epithelial cell layer that closes the dorsal side of the gastrulating embryo (2). Later in development, epidermis replaces the amnioserosa, which disintegrates. Although the amnioserosa does not contribute embryonic tissue, it controls two vital morphogenetic movements of the embryo: germband retraction and dorsal closure. Germband retraction shortens the embryo and transforms the u-shaped germband into an essentially straight line of body segments. The amnioserosa mediates this process by signaling and physical interactions with the underlying yolk sac (3–5). Dorsal closure, a developmental process that follows germband retraction, seals the epidermis along the dorsal midline. The amnioserosa guides this process in conjunction with the yolk sac and the leading edge of the dorsal epidermis by providing contractile force (6–9). During dorsal closure, some of the amnioserosa cells segregate into the yolk, but the bulk of this tissue invaginates before it degrades, transiently forming a tube-shaped “dorsal organ” (5, 6, 10), similar to extraembryonic tissue in other insects (11). Unlike Drosophila, however, most insects develop two distinct extraembryonic epithelia, called amnion and serosa (11–15). Typically, these epithelia develop from an amnioserosal fold, which closes about the ventral side of the gastrulating embryo. The outer cell layer of the amnioserosal fold becomes the serosa. This epithelium detaches from the amnion and encloses the embryo. The inner cell layer of the amnioserosal fold becomes the amnion and retains continuity with the dorsal epidermis of the embryo. The distinct developmental trajectory of the amnioserosa prompts comparative developmental and genetic investigations that could reveal the mechanism that generated the amnioserosa in evolution.
Previous studies show that a wide range of insects require zen activity for extraembryonic development, but several variants in expression and function have been described (references in ref. 15). In Drosophila, zen controls all aspects of amnioserosa development; a second copy of zen (z2) is expressed in an identical pattern but is dispensable (16, 17). The cells of Zen-deficient Drosophila embryos either die or acquire an embryonic fate. Conversely, overexpression of zen causes an expansion of the amnioserosa (18). Zen-deficient Drosophila embryos also develop head defects, consistent with an expression domain of zen in the embryonic blastoderm (19, 20). Other insects express zen only in extraembryonic tissue (references in ref. 15). In the flour beetle Tribolium, zen controls the specification of serosal but not amniotic cells and later in development the fusion of the ventral amnion with the serosa, which precedes the dorsal contraction of the fused extraembryonic epithelium (14). In the milkweed bug Oncopeltus, zen activity may not be required for the specification of serosal blastoderm but controls the fusion of the amnion with the serosa, as in Tribolium (15). In this article we describe the expression and function of zen during extraembryonic development in two lower cyclorrhaphan taxa and propose a model for the evolutionary origin of the amnioserosa that integrates our findings and relevant published data from other species.
Results
Serosa and Dorsal Amnion in Lower Cyclorrhaphan Flies.
Previous work suggests that the amnioserosa evolved in the lineage of cyclorrhaphan flies (21). To map the origin of this tissue more precisely, we determined the occurrence of a serosa in four cyclorrhaphan dipterans (Lonchoptera, Megaselia, Episyrphus, and Themira) by nuclear staining (Fig. 1 A). In Themira, the only representative of higher Cyclorrhapha (Schizophora) in our sample, we identified extraembryonic tissue resembling the amnioserosa of Drosophila [supporting information (SI) Fig. 7]. In the lower cyclorrhaphan species Lonchoptera, Megaselia, and Episyrphus, which represent the two branches most closely related to Schizophora (22), we observed a complete serosa. To test whether lower Cyclorrhapha also develop an amnion, we labeled the cell membranes and nuclei of Megaselia and Episyrphus embryos (Lonchoptera embryos were not available in sufficient numbers) with a mixture of anti-phosphotyrosine antibodies and DNA stain, and we examined different stages under a confocal microscope. In both species we observed a serosa and an amnion (Fig. 1 and SI Fig. 8). Prospective serosa cells flatten and become polyploid during germband extension (Fig. 1 B). While the germband continues to extend, the serosa expands over the germband and fuses on the ventral side. At the posterior and lateral margin of the serosa, this process involves the formation of an amnioserosal fold, which disjoins at its edge (Fig. 1 C–E and SI Fig. 8A ). The cells of the outer layer of the amnioserosal fold (preserosa) contribute to the serosa (Fig. 1 F and G and SI Fig. 8 B and C ). Cells of the inner cell layer of the amnioserosal fold (preamnion) form the amnion. After disjoining from the edge of the preserosa, the preamnion develops into a dorsal cell layer (Fig. 1 H). Thus, throughout development the amniotic cells remain on the dorsal side. Removing the vitelline layer after completion of the serosa damages both the amnion and the serosa (SI Fig. 8 D and E ). This observation suggests that the amnion attaches to the serosa and that the serosa attaches to the vitelline layer. During dorsal closure, the amnion is replaced by dorsal epidermis. The serosa disappears during dorsal closure, but we did not observe the formation of a dorsal organ (contracted and invaginated serosa tissue), suggesting that the serosa disintegrates underneath the vitelline layer without contracting. In Episyrphus, unlike in Megaselia, the amnioserosal fold is shallow and the preserosa extends to the anterior pole (Fig. 1 I). Our findings in Megaselia and Episyrphus suggest that the evolution of a dorsal amnion preceded the evolutionary origin of the amnioserosa (see Discussion).
Fig. 1.
Phylogenetic position and extraembryonic development of Megaselia and Episyrphus. (A) Phylogenetic tree of species mentioned in the text. (B–H) Confocal Z-stacks of Megaselia embryos during germband extension (B–D) and germband retraction (F and H). Cartoons (E and G) depict serosa cells (red), amnion cells (blue), and the germband (green) at stages shown in D and F. Note the amnioserosal fold (curved arrows in B and C), the expansion of the preserosa (D–G), and the large amnion cells (H) of an embryo in which the serosa has been removed. (I) Episyrphus embryo showing the anterior edge of the preserosa (arrow). Embryos were labeled with anti-phosphotyrosine antibodies (green) and Topro-3 or DAPI (red in B, C, F, H, and I). Anterior is left, and dorsal is up.
Expression Differences of zen Between Lower Cyclorrhapha and Drosophila.
To assess whether the differences in zen expression could account for some of the differences in extraembryonic development, we compared the expression of zen in Megaselia, Episyrphus, and Drosophila. Previously we reported that Megaselia zen expression is strictly zygotic and marks developing extraembryonic tissue (23, 24). However, we inadvertently removed the expanded serosa together with the vitelline membrane and did not distinguish serosal and amniotic tissue. We therefore reexamined the expression of Megaselia zen. The first expression occurs at the beginning of blastoderm cellularization (Fig. 2 A and A′). As cell membranes grow inwards, the expression narrows and is subsequently restricted to the developing serosa, which unlike the prospective amnion does not invaginate with the proctodeum (Fig. 2 B, B′, and C). This expression persists during the expansion of the serosa (Fig. 2 D and D′). To determine which features of Megaselia zen expression are characteristic for lower Cyclorrhapha, we examined the expression of a newly identified zen homologue from Episyrphus (SI Fig. 9). As in Megaselia, maternal zen transcript could not be detected in ovaries or early embryos of Episyrphus (data not shown). Zygotic expression begins in the syncytial blastoderm (Fig. 2 E, E′, F, and F′). Later in development Episyrphus zen expression occurs in the developing serosa but not in the amnion or any embryonic tissue (Fig. 2 G, H, and H′). The comparison with Drosophila, in which transcript and protein patterns of zen closely match (17), reveals several expression differences. First, the strong and broad dorsal expression of zen during blastoderm formation in Drosophila (Fig. 2 I) has no equivalent in the expression patterns of the identified zen homologues of Megaselia and Episyrphus. Second, zen expression in Drosophila extends to the proctodeum (Fig. 2 J and K), unlike in Megaselia and Episyrphus, where the amnion separates zen expression from the proctodeum. Third, unlike in Megaselia and Episyrphus, zen expression in Drosophila is down-regulated during the late phase of germband extension [stage 8 (2)] (Fig. 2 L). In addition, Drosophila zen is expressed in the head and germ line (17, 20) but appears to be strictly extraembryonic in Megaselia and Episyrphus.
Fig. 2.
Comparison of zen expression among Megaselia, Episyrphus, and Drosophila. RNA in situ hybridizations of Megaselia (A–D), Episyrphus (E–H), and Drosophila (I–L) embryos are shown in lateral and dorsal orientation (A′–F′ and H′). Embryos are at the syncytial blastoderm stage (A, E, and I), early gastrulation (B, F, and J), germband extension (C, G, and K), or the extended germband stage (D, H, and L). Anterior is left. The proctodeal invagination is marked by an asterisk. In Megaselia and Episyrphus the amnioserosal fold (arrow) migrates during the progression of the fold to the posterior expression boundary of zen.
Phenotypic Effects of zen RNAi in Megaselia and Episyrphus.
We used RNAi to analyze the function of zen during embryogenesis in Megaselia and Episyrphus. We injected preblastoderm embryos with zen dsRNA and analyzed them at different time points (SI Fig. 10). Megaselia embryos injected with a 4.7 μM solution of dsRNA consistently developed a single extraembryonic epithelium on the dorsal side (Fig. 3 A and B). To test whether Megaselia zen is required for dorsal closure, we repeated the RNAi experiment and allowed the embryos to develop a cuticle. Approximately half of the developed embryos exhibited a “dorsal open” cuticle (Fig. 3 C), whereas the other half was indistinguishable from wild type (Fig. 3 D). The absence of dorsal closure defects in a large proportion of RNAi embryos could reflect residual zen activity. Alternatively, the dorsal open morphology could reflect a hypomorphic phenotype caused by an oversized dorsal epithelium that resulted from incomplete suppression of serosa development. At present we cannot distinguish between these possibilities. Although we observed protruding extraembryonic tissue in some of the open cuticles, lowering the concentration of dsRNA to 2.35 μM resulted in a moderate increase of embryonic viability rather than an increase in the proportion of RNAi embryos with a dorsal open phenotype (SI Fig. 10). In Episyrphus, all embryos developed a single extraembryonic epithelium on the dorsal side when injected with a 1.7 μM solution of Episyrphus zen dsRNA (Fig. 3 E–H). Other aspects of development, including germband retraction and dorsal closure, were not affected, except in two embryos in which incomplete dorsal closure was observed together with an oversized dorsal extraembryonic epithelium. We conclude that in both species zen is necessary for serosa development, whereas the serosa may not be essential for embryonic development including germband retraction and dorsal closure.
Fig. 3.
zen RNAi in Megaselia and Episyrphus causes the formation of a single extraembryonic epithelium on the dorsal side. (A) Single extraembryonic epithelium (arrow) of a Megaselia RNAi embryo during germband retraction in dorsolateral view. (B) Schematic representation of the RNAi phenotype in Megaselia depicting extraembryonic tissue (blue) and the germband (green). (C and D) RNAi cuticles of Megaselia larvae showing a dorsal open phenotype (arrow) in lateral view (C) and a “dorsal closed” phenotype in dorsolateral view (D). (E–G) Consecutive developmental stages of RNAi phenotypes in Episyrphus in lateral view. Midway through germband extension, extraembryonic cells are morphologically indistinguishable from embryonic cells, unlike in wild-type embryos (compare E with Fig. 1 I). After germband extension, a single layer of extraembryonic cells extends from the edge of the epidermis (arrows in F and G). Note ectopic anterior cells (broken arrow in G). (H) Schematic representation of the RNAi phenotype in Episyrphus. Anterior is left, and dorsal is up (unless specified otherwise). Embryos were stained with anti-phosphotyrosine antibodies and Topro-3 (A, F, and G) or DAPI (E).
To characterize the extraembryonic tissue that persists in RNAi embryos of lower cyclorrhaphan flies we searched for amnion and serosa markers in Megaselia. We cloned Megaselia homologues of C15, pnr, and Kr (SI Fig. 11). In Drosophila, the transcripts and proteins of all three genes are expressed throughout the developing amnioserosa, albeit in different time windows. Extraembryonic C15 expression starts in the blastoderm and continues until after germband retraction (25, 26). Extraembryonic pnr expression starts in the blastoderm and fades during the extended germband stage (stage 10) (27). Finally, Kr expression in the amnioserosa begins during the extended germband stage (stage 9) and persists until after germband retraction (28). C15 and pnr, but not Kr, are also expressed in the dorsal epidermis abutting the amnioserosa.
In Megaselia, C15 is first expressed during the formation of the syncytial blastoderm, spanning ≈60% of the trunk region (data not shown). This expression disappears during blastoderm cellularization. At the beginning of gastrulation, some cells of the serosa primordium weakly express C15 (Fig. 4 A), but this expression disappears shortly after the germband has started to extend. During germband extension, Megaselia C15 is activated in preamnion cells and the leading edge of the dorsal epidermis (Fig. 4 B and B′). In a few embryos we noticed C15 expression also in the posterior serosa primordium, suggesting that during early germband extension extraembryonic C15 expression is either dynamic or somewhat variable between individuals. In addition, we detected weak C15 expression in creases of the preserosa that transiently form during germband extension (data not shown). In the preamnion we observed an attenuation of C15 expression shortly before these cells begin to stretch (Fig. 4 C and C′).
Fig. 4.
Expression of C15, pnr, and Kr in Megaselia. (A–C) Megaselia C15. (D–F) Megaselia pnr. (G–I) Megaselia Kr. All embryos are shown in dorsal view (A, B′, C′, D, E′, F′, G, H′, and I′) or in lateral view with dorsal up. Anterior is left. Arrows point to the prospective amnion. The proctodeal invagination is marked by an asterisk. cf, cephalic furrow; tf1 and tf2, transverse folds in the preserosa.
Megaselia pnr expression starts in the mid-dorsal blastoderm but fades at the beginning of gastrulation in cells along the dorsal midline, leaving two parallel stripes that are joined at their posterior end (Fig. 4 D). During germband extension we consistently observed expression in the ectoderm and in the preamnion (Fig. 4 E, E′, F, and F′). In two embryos at the stage of early germband extension, we also detected expression in the posterior portion of the serosa primordium, suggesting that pnr expression in this tissue is initially either dynamic or variable between individuals, like the expression of Megaselia C15. In the preamnion expression levels of Megaselia pnr attenuate with the progression of the amnioserosal fold until high expression levels remain only in the leading edge of the dorsal epidermis (Fig. 4 F and F′). In summary, the expression patterns of pnr and C15 in germband-extending Megaselia embryos are similar, and high expression levels of both genes transiently mark the developing amnion.
Megaselia Kr expression in the blastoderm is similar to Kr expression in Drosophila except for a gap along the dorsal midline (data not shown). Here we consider only the extraembryonic expression of Megaselia Kr, which starts with gastrulation (Fig. 4 G). During germband extension we observed Kr expression in the expanding preserosa but not in the amnion (Fig. 4 H, H′, I, and I′). Thus, Megaselia Kr is a serosa marker.
We used the Megaselia homologues of pnr and Kr to examine the identity of extraembryonic cells after zen RNAi. As a time window, we chose germband extension stages, when pnr expression in Megaselia wild-type embryos is up-regulated in the prospective amnion and down-regulated in the prospective serosa. We performed double in situ hybridization experiments to examine simultaneously the expression of Kr and pnr (Fig. 5 A, A′, and A″). Megaselia zen RNAi embryos lacked extraembryonic but not embryonic Kr transcripts and expressed pnr evenly throughout the extraembryonic primordium and dorsal ectoderm (Fig. 5 B, B′, and B″). These results suggest that the extraembryonic epithelium of Megaselia zen RNAi embryos has an amniotic identity.
Fig. 5.
Extraembryonic primordium of Megaselia zen RNAi embryos expresses pnr but not Kr. (A) Wild type. (B) RNAi phenotype. Expression of pnr is shown in green or as a monochrome image of the green channel (A′ and B′), Kr in red, or as a monochrome image of the red channel (A″ and B″). In the RNAi embryo, note the suppression of extraembryonic Kr expression and the uniform expression of pnr in dorsal ectoderm including extraembryonic tissue (ext). am, preamnion; de, dorsal ectoderm; ser, preserosa; gb, posterior germband.
Discussion
Two Major Transitions in Extraembryonic Morphology in Dipteran Evolution.
Our survey of extraembryonic tissue organization in dipterans suggests three distinct trajectories of extraembryonic development. The transient formation of a serosa and a ventral amnion is common throughout lower (non-cylorrhaphan) Diptera and other insect orders and almost certainly reflects the primitive condition for extraembryonic development in Diptera (12, 29). The amnioserosa of schizophoran flies is therefore derived. Megaselia and Episyrphus develop a serosa and dorsal amnion without passing through a developmental stage with a ventral amnion. This developmental trajectory, hitherto unknown in dipterans, occurs at least in two paraphyletic taxa of the lower Cyclorrhapha (Aschiza): Megaselia (Phoroidea) and Episyrphus (Syrphoidea). Our findings therefore suggest that the last common ancestor of Megaselia, Episyrphus, and Schizophora shared this type. We conclude that the evolution of the amnioserosa was preceded by the evolution of a dorsal amnion and propose two major morphological transitions in the course of extraembryonic tissue evolution in flies. The first transition consisted in suppressing the formation of a ventral amnion, perhaps through a fusion of this tissue with the serosa before its ventral completion, and occurred most likely in the late Jurassic period after the cyclorrhaphan lineage had split from other extant dipterans. A similar transition occurred independently in the insect order Hymenoptera (30). The second transition consisted in the reorganization of the preserosa and the preamnion into a single dorsal epithelium, the amnioserosa, and occurred apparently toward the end of the Cretaceous period in the stem group of Schizophora.
Transition from a Serosa and Dorsal Amnion to an Amnioserosa.
The evolutionary transformation of the serosa and dorsal amnion into a single dorsal epithelium could have been achieved by suppressing the expansion of the preserosa and its disjunction from adjacent tissue. Similar to the amnioserosa of Drosophila, the preserosa of Megaselia and Episyrphus depends on the activity of zen and acquires its distinct cellular morphology during germband extension, ahead of amniotic cells. Importantly, the expansion of the preserosa in Megaselia and Episyrphus correlates with persisting zen expression in the developing serosa. In contrast, zen expression in the amnioserosa is down-regulated at about the same developmental stage when the presorsa of lower Cyclorrhapha begins to expand. This change in zen expression could have been critical for redirecting extraembryonic development in such a way that only a single dorsal epithelium forms. In other words, zen expression in the developing serosa could be essential for maintaining developmental properties of serosal and repressing properties of amniotic cells. Functional studies in other species are consistent with this hypothesis (Fig. 6).
Fig. 6.
Schematic comparison of extraembryonic zen expression between holometabolous insects. Successive developmental stages of Tribolium (A), Megaselia (B), and Drosophila (C) are shown with zen-expressing extraembryonic tissue depicted in red. Prospective extraembryonic tissue that does not express zen is depicted in blue, and embryonic tissue is in green. Note spatial differences of zen expression between species and the down-regulation of zen in Drosophila at about the same developmental stage that the preserosa of Megaselia begins to expand beyond the distal edge of the amnion. Anterior is left, and dorsal is up.
Across holometabolous insects, early zen expression is confined to the developing serosa and controls a developmental switch that allows the zen-expressing portion of the blastoderm to acquire serosal fate (Figs. 2–5) (14, 24, 31). In strong zen RNAi phenotypes of Tribolium and Megaselia, all extraembryonic cells are indistinguishable from amniotic cells, indicating that the specification of amniotic cells is independent of zen activity. In contrast, zen expression in Drosophila occurs in the entire extraembryonic primordium and is essential for the formation of all extraembryonic tissue (17). Together, the data from holometabolous insects suggests that in the Schizophora lineage zen expanded its expression domain to the entire amniotic primordium and became essential for all extraembryonic development (Fig. 6). Genetic changes upstream of zen must have occurred in the Drosophila lineage to allow these changes in zen expression and function (31). In addition, genetic changes immediately downstream of zen must have occurred to allow the coexpression of “serosa genes” (zen and Kr) together with C15, pnr, dorsoccross (doc), and tail-up (tup) (25, 32, 33), which have been described as amnion markers in other species (14, 31) (Fig. 4). The coexpression of zen and amnion genes could have resulted from the loss of a transcriptional repressor of amnion genes downstream of zen (31). However, despite the apparent derepression of amnion markers in the amnioserosa, their Drosophila homologues have either no or only a subtle role in the initial differentiation of the amnioserosa when zen is expressed. pnr activity in the amnioserosa appears altogether blocked (34). C15 null mutations do not interfere with larval hatching, indicating that their amnioserosa is essentially intact (35). Doc- and Tup-deficient embryos exhibit defects in the maintenance of amnioserosa cells only after stage 8 (25, 36). Thus, doc and tup unfold their function in the amnioserosa, with one exception (see below), after zen expression has been suppressed. Our model implies that these functions of doc and tup are comparable to the function of their homologues in the amnion of lower Diptera.
The amnioserosa of Doc-deficient embryos also fails to fold properly during germband extension, whereas zen is still expressed (25). This phenotype of a putative “amnion gene” in the early amnioserosa could suggest that some amnion properties of the amnioserosa unfold before zen expression is down-regulated. However, the folds in the amnioserosa of Drosophila occur likewise in the preserosa of Megaselia (compare to figure 2.8A in ref. 2) and probably constitute a preserosa rather than an amnion trait of the amnioserosa. Accordingly, we speculate that a doc homologue controls fold formation in the preserosa of Megaselia.
Finally, it has been suggested that only marginal cells of the late amnioserosa are equivalent of the amnion in less derived species (37). This idea, which is based on the observation that marginal amnioserosa cells are genetically distinct from other parts of this tissue, does not contradict our model because it refers to late properties of the amnioserosa. In addition, it does not rule out the possibility that central parts of the late amnioserosa are also amnion-like because the genetic differences between peripheral and central cells of the amnioserosa during dorsal closure could correspond to differences between peripheral and central cells of the dorsal amnion in lower cyclorrhaphan flies. More work in lower cyclorrhaphan species is needed to settle this question.
In summary, genetic modifications of extraembryonic development in the schizophoran lineage of Drosophila include changes upstream and downstream of zen as well as major changes in the expression pattern of zen itself both early and late in development. It is unlikely that all these changes happened at the same time. We propose that the early down-regulation of extraembryonic zen expression triggered the suppression of the expansion of the preserosa and the rift between this and adjacent tissue, thereby generating a single and strictly dorsal extraembryonic epithelium. At the same time, the early down-regulation of extraembryonic zen expression could have allowed preserosa tissue to acquire amniotic properties and to become essential for germband retraction and dorsal closure, which do not require the serosa (14) (Fig. 3). A more uniform amnioserosa anlage may have evolved gradually, perhaps to stabilize the developmental trajectory of the new morphology.
Loss of the Ventral Amnion in Evolution.
Currently, a mechanistic understanding of the evolutionary origin of the dorsal amnion in cyclorrhaphan flies is limited by the lack of functional data on extraembryonic development in lower dipterans. However, the comparison of our results in Megaselia and Episyrphus with data from Tribolium (14) and Oncopeltus (15) suggests a tentative model that is based on spatiotemporal changes in the expression of zen. In Tribolium, serosa and ventral amnion are completed during germband extension. Subsequently, both epithelia fuse and retract dorsally, such that a single extraembryonic epithelium closes the embryo. Initially, only the serosa expresses zen (Tc-zen1 and Tc-zen2), but, preceding the fusion of the amnion with the serosa, zen (Tc-zen2) is activated also in the amnion. RNAi against Tc-zen2 suppresses the fusion of the amnion with the serosa. Such embryos retain an intact amnion and close ventrally about the appendage buds. Likewise in Oncopeltus, the fusion of the completed amnion with the serosa is critical for normal dorsal closure and depends on the activity of zen, which is expressed throughout the serosa and at the site of fusion in the amnion. We speculate that zen-expressing amnion cells of Tribolium and Oncopeltus become indistinguishable from serosa cells and arrange to form a single epithelium, thereby causing the fusion of both tissues. To explain the transition from a ventral amnion to a dorsal amnion in dipteran evolution we propose a precocious expression of zen in the developing amnion. The earliest time point for zen expression in the developing amnion would be the blastoderm stage. Therefore, the boundaries of zen expression in the blastoderm relative to the boundaries of amnion-competent blastoderm might control the morphological difference between a ventral and a dorsal amnion perhaps not only in dipterans but also in hymenopterans, which modified the topology of the amnion in a similar manner (30).
Materials and Methods
Flies, Cloning Procedures, and RNAi.
Megaselia abdita Schmitz, Episyrphus balteatus Degeer, and Drosophila melanogaster (Oregon strain) were reared in the laboratory. Genomic fragments of Megaselia homologues of C15, pnr, and Kr were PCR-amplified from genomic DNA using the following degenerate primer pairs: 5′-ATHGGNCAYCCITAYCARAAY/5′-TTNACYTGIGCRTCNGTCATYTT for C15, 5′-RTNATGATGDSNWSNTGGMG/5′-CCRCANGCRTTRCANARRTARTG for pnr, and 5′-AARCAYGTRYTKMARAAYCAYGA/5′-YTTYARYTGRTTRSWRTCRSWRAA and 5′-GATCATCAYYTSAARACNCA/5′-TTMAGGTGRTTSGAGTCRSYRAA for Kr. A fragment of Episyrphus zen was obtained with degenerate primers for bicoid (38) because of mispriming of the lower primer. The cDNA amplification Kit SMART (Clontech) was used. cDNA was prepared from 0- to 5-h-old embryos (collected at room temperature). zen dsRNA from Megaselia and Episyrphus was prepared from cDNA nucleotides 34–760 and nucleotides 50–909, respectively (position 1 being the first nucleotide of the ORF; essentially as described in ref. 39). For microinjection, embryos were aligned on a glass slide along a 0.2-mm glass capillary, briefly desiccated, and covered with halocarbon oil [one part 27-oil (Sigma H773) and three parts 700-oil (Sigma H8898) for Episyrphus and 27-oil only for Megaselia]. Approximately 65 pl of dsRNA solution was injected per embryo.
In Situ Hybridization, Immunocytochemistry, and Microscopy.
In situ hybridization was done as described with minor modifications (40, 41). Protocols are available on request. RNA probes were labeled with fluorescein (zen and Megaselia pnr), biotin (Megaselia Kr), or digoxigenin (Megaselia C15, Megaselia zen, and Episyrphus zen). The Megaselia zen probe was complementary to nucleotides 120-1071 of the ORF. The Megaselia C15 probe was complementary to nucleotides 362–973 of the partial ORF and included 62 bases of 3′ UTR. The Megaselia pnr probe was complementary to the ORF and 111 nucleotides of 5′ UTR. The Megaselia Kr probe was complementary to the entire ORF, 44 nt of the 5′ UTR, and 79 nt of the 3′ UTR. Fab fragments against digoxigenin, fluorescein, and biotin (Roche) were used for probe detection. For fluorescent double in situ hybridizations, Tyramide Signal Amplification TSA (Molecular Probes) was used following the instructions of the manufacturer. To label cell membranes we used monoclonal mouse anti-phosphotyrosine (PY-plus mixture from Zymed) and anti-α-actin (Actin-4; gift of W. Sullivan, University of California, Santa Cruz, CA) primary antibodies and Alexa Fluor 488 or Cy3-conjugated secondary antibodies (Molecular Probes). The nuclei were stained with DAPI (Molecular Probes) or Topro-3 (Molecular Probes). Confocal scans were done with a Leica SP2 AOBS Spectral Confocal Microscope or a Zeiss Axiovert 200 Microscope. The 3D projection of image stacks was done with ImageJ software (Wayne Rasband, National Institute of Mental Health, Bethesda, MD).
Acknowledgments
Peter Katz kindly allowed us to perform initial experiments with Episyrphus at his company (Katz Biotech, Baruth, Germany). During this time, Hans Lehrrach (Max-Planck-Institut für Genetik, Berlin) and his collaborators Georgia Panopoulou and Albert J. Poustka provided important logistic support. Sofya Leonova helped with the cloning of Megaselia genes, and Annette Lemke helped with the fly cultures. Julia Bowsher in the laboratory of H. Frederik Nijhout (Duke University, Durham, NC) kindly fixed and sent eggs of Themira. We thank our colleagues, especially Chip Ferguson, Rick Fehon, and their respective laboratory members for discussions and sharing reagents. Irene Hsiao helped us to spot spelling and syntax errors in the final draft. S.L. was the recipient of a Boehringer Ingelheim fellowship. Funding was provided by the University of Chicago.
Footnotes
- §To whom correspondence should be addressed. E-mail: uschmid{at}uchicago.edu
-
Author contributions: A.M.R., S.L., and U.S.-O. designed research; A.M.R., S.L., and S.F. performed research; M.S. contributed new reagents/analytic tools; A.M.R., S.L., and U.S.-O. analyzed data; and A.M.R., S.L., and U.S.-O. wrote the paper.
-
↵ ‡Present address: Cancer Research UK, Developmental Genetics Laboratory, 44 Lincoln's Inn Fields, London WC2A 3PX, United Kingdom.
-
The authors declare no conflict of interest.
-
This article is a PNAS Direct Submission.
-
Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. DQ323932, EU287990, EU287991, and EU287992).
-
This article contains supporting information online at www.pnas.org/cgi/content/full/0709145105/DC1.
- © 2008 by The National Academy of Sciences of the USA
References
Evolutionary origin of the amnioserosa in cyclorrhaphan flies correlates with spatial and temporal expression changes of zen
Sign up for Article Alerts
Jump to section
You May Also be Interested in
Recent experiments and simulations are starting to answer some fundamental questions about how life came to be.
Image credit: Shutterstock/Radoslaw Lecyk.
Archaeologists have long tried to define the transition between the two time periods.
Image credit: Ceri Shipton.
Jonathon Yuly, David Beratan, and Peng Zhang investigate how electron bifurcation reactions work.
A study finds that giant pandas roll in horse manure to increase their cold tolerance.
Image credit: Fuwen Wei.