Introduction

The members of the Sp gene family encode evolutionarily conserved proteins, which are characterized by the presence of three zinc finger motifs (for nucleic acid binding) as well as an additional short conserved motif called Btd box (a transactivation domain), and are involved in a variety of developmental processes in both insects and vertebrates (reviewed in Zhao and Meng 2005). In Drosophila, two different Sp factor encoding genes, buttonhead (btd) and D-Sp1, have been shown to be expressed during leg development (Wimmer et al. 1996; Schöck et al. 1999; Estella et al. 2003). The btd gene is required for the specification and formation of the entire leg disc primordia (Estella et al. 2003; McKay et al. 2009). The gene is acting upstream of the well-characterized leg specification gene Distal-less (Dll) and seems to be capable of activating the entire leg developmental cascade when expressed ectopically (Estella et al. 2003). The role of the D-Sp1 gene during leg development is less clear, but seems to be partially redundant with btd (Schöck et al. 1999; Estella et al. 2003). The homolog of D-Sp1 in the red flour beetle Tribolium castaneum, Sp8, has been shown to be involved in the control of allometric growth of the leg segments, but the evidence for a role of Sp8 in the specification of the entire leg primordium is ambiguous (Beermann et al. 2004). Orthologous genes to D-Sp1 and Tc Sp8 have also been identified in the mouse, termed Sp8 and Sp9 (Bell et al. 2003; Treichel et al. 2003; Kawakami et al. 2004). Please note that the murine Sp8 gene has also been called inaccurately mBtd (Treichel et al. 2003), because it was named at a time when its orthology to either Drosophila btd or Drosophila D-Sp1 was unclear. Intriguingly, both murine genes are also involved in limb outgrowth.

Here we present the isolation of an Sp factor encoding gene from the milkweed bug O. fasciatus. Although our analysis of the Oncopeltus gene confidently places it as a homolog of the Drosophila D-Sp1, Tribolium Sp8, and the murine Sp8 and Sp9 genes, the exact orthology within this group is unclear and we therefore designate the Oncopeltus gene as Sp8/9. We have studied the role of the Oncopeltus Sp8/9 ortholog in order to investigate its evolutionary conservation in insect appendage development. We find that Sp8/9 is expressed in the appendages throughout development and the downregulation of Sp8/9 expression via RNAi leads to shortened legs, rostrum, and antennae. These data show that Sp8/9 is involved in the allometric growth of the appendages. However, we find no evidence for a more global role of Sp8/9 in appendage specification.

Materials and methods

Animal husbandry and embryology

Milkweed bugs were reared as described previously (Hughes and Kaufman 2000). Collected eggs were kept at 25°C. Embryos of all stages were fixed as reported in Liu and Kaufman (2004a). Oncopeltus embryo dissections before in situ staining were performed under a fluorescence stereomicroscope using SYTOX Green nucleic acid stain (Invitrogen) (Liu and Kaufman 2004b).

Isolation of Sp8/9 and sequence analysis

Oncopeltus embryos from 0 to 96 h were used for mRNA isolation using the MicroPoly(A)Purist kit (Ambion). This mRNA served as template for double-stranded (ds) cDNA synthesis (SMART PCR cDNA Synthesis kit, Clontech) and RACE template synthesis (SMART RACE cDNA Amplification Kit, Clontech). PCR with the primer pair Fw_GGC MGG GCI ACI TGY GAY TGY CCI AAY TG (GRATCDCPNC) and Rev_ARR TGR TCI SWI CKC ATR AAI CKY AA (LHDSRMFRK) resulted in a fragment of 311 bp. PCR fragments were cloned into the pCRII vector (Invitrogen). Additional sequence information was obtained by 5′ RACE PCR using the reverse primer CAG GTG AGC CTT GAG GTG CGA GGT C. Phylogenetic analysis of different Sp factor sequences was performed as described previously (Prpic et al. 2005). The Oncopeltus Sp8/9 sequence is available from the EMBL nucleotide database under the accession number FN396612.

In situ hybridization

The longest 5′ RACE fragment of Sp8/9 (1,078 bp comprising 181 bp 5′UTR and 897 bp ORF) served as template for the synthesis of digoxygenin-labeled RNA probes (Roche). In situ hybridization was performed as described previously (Liu and Kaufman 2004a).

Parental RNA interference

The template for dsRNA synthesis was prepared by PCR with T7 (GAA TTG TAA TAC GAC TCA CTA TAG G) and Sp6-T7 (TAA TAC GAC TCA CTA TAG GAT TTA GGT GAC ACT ATA GA) primers from the longest 5′ RACE fragment of Sp8/9 that has also been used for probe generation. Double-stranded RNA (dsRNA) was generated using the MEGAscript T7 Kit (Ambion) and resuspended in 1× injection buffer (1.4 mM NaCl, 0.07 mM Na2HPO4, 0.03 mM KH2PO4, 4 mM KCl) at a concentration of 4 µg/µl. RNA injections in adult virgin Oncopeltus females were performed as described previously (Liu and Kaufman 2004a). Injection of 1× injection buffer served as the negative control. To verify the RNAi phenotypes obtained with the full fragment, we repeated the parental RNAi with two shorter non-overlapping fragments of Of Sp8/9. A different 5′ RACE Of Sp8/9 fragment of 1,057 bp (133 bp 5′ UTR, 924 bp ORF) was cut with XhoI which resulted in two fragments of 421 and 636 bp. Parental RNAi experiments with dsRNA transcribed from these two fragments resulted in the same phenotype with a similar frequency as for dsRNA injections with the full longest 5′ RACE fragment (data not shown). As independent RNAi controls, we performed injections of dsRNA of EGFP and Of eve which resulted in no abnormal phenotype or the same phenotypes as previously published for Of eve, respectively (Liu and Kaufman 2005) (data not shown).

Results and discussion

Isolation of the Sp8/9 homolog of O. fasciatus

A short fragment of the Oncopeltus homolog of the Sp8/9 gene was isolated by PCR. More sequence information was obtained by 5′ RACE PCR. The available sequence comprises 181 bp of 5′ UTR followed by 1,008 bp of protein coding sequence also encoding the Btd box motif and the zinc finger motifs ZF 1, ZF 2, and a portion of ZF 3. An alignment of these conserved domains with the nine Sp factors from the mouse, as well as D-Sp1 and Btd from Drosophila and Tc Sp8 and Tc Btd (Schinko et al. 2008) from Tribolium, shows that the Btd box motif and the available sequences of the zinc finger motifs are almost identical in all Sp8/9 proteins, but clearly different from the homologous motifs in other Sp proteins (Fig. 1a). The linker regions between these motifs are also highly conserved in all Sp8/9 proteins, whereas differences in sequence and linker length exist in other Sp proteins (Fig. 1a).

Fig. 1
figure 1

Alignment and phylogenetic analysis of the Sp8/9 sequence from Oncopeltus. a Alignment of Sp proteins from Drosophila, Oncopeltus, Tribolium, and mouse. Shown is the protein region containing the Btd box motif (blue) and the three zinc finger motifs (red). The sequences of the Sp8/9 proteins including the D-Sp1 protein are almost identical and clearly different from other Sp proteins. b Phylogenetic analysis of all murine Sp factors and Sp factors from Tribolium, Drosophila, and Oncopeltus based on the alignment shown in (a). Shown is the unrooted majority rule consensus computed from 1,000 intermediate trees produced with the quartet puzzling method (Strimmer and von Haeseler 1996). The numbers at the tree edges indicate the reliability values. GenBank accession numbers: Dm_btd NP_511100, D-Sp1 NP_572579, Tc_btd NP_001107792, Tc_Sp8 NP_001034509, Mm_Sp1 NP_038700, Mm_Sp2 NP_084496, Mm_Sp3 NP_035580, Mm_Sp4 NP_033265, Mm_Sp5 NP_071880, Mm_Sp6 NP_112460, Mm_Sp7 NP_569725, Mm_Sp8 NP_796056, Mm_Sp9 NM_001005343. Species abbreviations: Mm Mus musculus, Tc Tribolium castaneum, Of Oncopeltus fasciatus, Dm Drosophila melanogaster

Full size image

In order to further corroborate the orthology of the Sp8/9 fragment isolated from Oncopeltus, we also performed a phylogenetic analysis (Fig. 1b) using the alignment in Fig. 1a in a maximum likelihood analysis with Tree Puzzle (Strimmer and von Haeseler 1996). Most edges in the phylogenetic tree are well supported with reliability values above 95. The Sp8/9 factors from Oncopeltus, Tribolium, and mouse cluster together in a group supported by the maximum reliability value of 100 and with very short edges, indicating that these genes are closely related. This grouping also includes D-Sp1 from Drosophila. These results further support the orthology of the Oncopeltus Sp8/9 gene with the other Sp8/9 genes and also give additional evidence to the previously published conclusion that the Drosophila D-Sp1 gene is actually the Drosophila Sp8/9 homolog (Beermann et al. 2004).

Sp8/9 is expressed during appendage development in O. fasciatus

Next we examined the expression profile of Sp8/9 in Oncopeltus embryos using whole mount in situ hybridization. During early stages of germband elongation, strong expression of Sp8/9 is seen in the limb buds of the thoracic legs and the antennae (Fig. 2a). Interestingly, the expression in the thoracic limb buds extends from there into the lateral tissue of the thoracic segments posterior to the limb buds, and thus is not restricted to the limb buds (Fig. 2a, arrows). A weaker expression is detected in the maxillary and labial segment, where distinct buds are not yet developed. A separate expression domain is present in the anterior head in the region where the labrum will develop. As the germband grows, a punctate expression in the ventral nervous system appears and the expression in the thorax becomes restricted to the leg limb buds (Fig. 2b). During later germband elongation stages, the punctate expression pattern in the ventral nervous system becomes stronger and some cells in the brain also express Sp8/9 (Fig. 2c). These brain cells are located at the rim of the head lobes in the area where the optic centers develop and also nearer to the ventral midline in the protocerebrum (Fig. 2c). All limb buds, including the emerging buds of the labrum, the mandibles, and the maxillae, now express Sp8/9. In fully elongated embryos, the expression pattern persists, but the expression in the thoracic legs becomes slightly non-homogeneous (Fig. 2d), indicating the emergence of iterated rings in the legs at later stages. At fully retracted germband stages, the expression pattern in the appendages is more restricted (Fig. 2e). In the antenna, Sp8/9 is restricted to a large ventral patch at the distal end and a smaller dot dorsal to it. The expression in the mandible has ceased and the expression in the maxilla is very weak. In the labium, Sp8/9 expression is largely confined to a faint ring near the middle of the appendage. In the thoracic legs, the expression is seen in three rings. These expression rings are located proximally adjacent to the constrictions that mark the presumptive joints between the leg segments (podomeres; note that not all future podomeres are yet developed at this stage of development) (Fig. 2f).

Fig. 2
figure 2

Expression of Sp8/9 in Oncopeltus embryos. a Early germband elongation stage. Arrows point to expression extending from the thoracic limb buds into the lateral tissue of the thoracic segments posterior to the limb buds. b Mid-elongation germband stage. c Embryo approaching the final stages of germband elongation. d Fully elongated germband stage. e Retracted germband stage. f Magnification of a leg at the retracted germband stage showing the expression rings of Sp8/9 abutting the presumptive joint constrictions (arrowheads). All embryos are oriented with the anterior end to the left. Abbreviations: lr labrum, an antenna, md mandible, mx maxilla, lb labium, t1 prothoracic leg, t2 mesothoracic leg, t3 metathoracic leg

Full size image

Functional analysis of Sp8/9 using RNAi

In order to establish the function of Sp8/9 during development of Oncopeltus, we have applied parental RNAi (Liu and Kaufman 2004a). The results of these injections are summarized in Table 1. In wild-type embryos, shortly before hatching, the appendages lie flat on the body and reach the posterior end of the abdomen (Fig. 3d, e). In Sp8/9 RNAi embryos, the appendages are severely shortened. They stick out (Fig. 3a) and do not reach further than the middle of the body (Fig. 3a, b).

Table 1 Parental RNAi with Sp8/9
Full size table
Fig. 3
figure 3

Parental RNAi with Sp8/9 in Oncopeltus. a, b Embryos after Sp8/9 RNAi shortly before hatching in lateral (a) and ventral (b) aspect. c Hatchling after Sp8/9 RNAi. d, e Wild-type embryos shortly before hatching in lateral (d) and ventral (e) aspect. f Wild-type hatchling. The arrows in (a) and (d) denote the tips of the legs. The arrows in (b) and (e) denote the tips of the antennae. All animals in (a)–(f) are oriented with the anterior end to the left. g–m Preparations of the appendages of wild-type hatchlings (g, i, k) and hatchlings after Sp8/9 RNAi (h, j, l, m). The wild-type rostrum (g) comprises the four-segmented labium, filiform maxillae and mandibles, and the pointed labrum. The rostrum of RNAi animals (h) is malformed with fused distal segments. The wild-type antenna (i) comprises an antennifer and four additional segments. The antenna of RNAi animals is severely shortened with fused segments (j). The arrow in (j) points to a small ectopic outgrowth on the antenna seen in about half of the hatchlings after Sp8/9 RNAi. The wild-type legs (k) comprise a coxa, trochanter, femur, tibia, and a two-segmented tarsus with two claws. The legs of RNAi animals are shortened with fused tarsal segments (l, m). The more proximal segments coxa, trochanter, and femur may be partially (l) or fully fused together (m). Abbreviations: lr labrum, an antenna, af antennifer, ro rostrum, md mandible, mx maxilla, lb labium, t1 prothoracic leg, t2 mesothoracic leg, t3 metathoracic leg, cx coxa, tr trochanter, fe femur, ti tibia, ta tarsus, cl claw

Full size image

In wild-type hatchlings the appendages are long and composed of several segments (podomeres) (Fig. 3f). The rostrum is a complex of four appendages (Fig. 3g). The labrum is thin and sharply pointed, the mandibles and the maxillae are long and thread-like, and the labium consists of four segments (Fig. 3g). The rostrum of the Sp8/9 RNAi animals is malformed (Fig. 3h). The labium is shortened; the distal segments are fused, bent, and enlarged at the tip. Therefore, the filiform mandibles and maxillae protrude from the labium at the distal end, while normally they are entirely ensheathed by it. The overall morphology of the mandibles, maxillae, and the labrum is normal in Sp8/9 RNAi animals, but they are shorter than in the wild type (Fig. 3h).

The wild-type antennae consist of a basal antennifer and four antennal segments (Fig. 3i). The antennae are severely affected in all Sp8/9 RNAi animals (Fig. 3c, j). The antennifer and the first segment of the antenna are roughly identical in size and shape to the wild type, but the three distal antennal segments are entirely fused and the antennal appendage as a whole is severely shortened. In about half of the cases, the fused distal antennal portion displays small ectopic outgrowths (arrow in Fig. 3j).

The wild-type thoracic legs consist of a short coxa, a trochanter which is closely attached to the femur, a tibia, and a two-segmented tarsus with two claws (Fig. 3k). The legs of Sp8/9 RNAi animals are much shorter than in the wild type (Fig. 3c, l, m). All leg segments are present but some of them are fused together. The tarsal segments are always fused, but they can still be distinguished because constrictions indicating the rudimentary joints are still present. The trochanter and femur are also always fused. In some specimens, the proximal podomeres are so severely malformed and fused that they cannot be distinguished anymore (Fig. 3m). In contrast to these joints, the joints between femur and tibia and between tibia and tarsus are always present.

Expression of Distal-less, dachshund, and homothorax in Sp8/9 RNAi embryos

Since the results of the RNAi experiments indicated a role of Sp8/9 in appendage development, we next studied the expression of leg developmental genes expressed at different positions along the proximal–distal leg axis.

The gene Distal-less (Dll) is expressed in the distal portion of all appendages except for the mandible (Fig. 4a, b) (Angelini and Kaufman 2004). In the head appendages, Dll is expressed in the tip of the labrum, the entire distal portion of the antenna, and the tips of maxilla and labium (Fig. 4a). The thoracic legs show two separate domains of Dll expression: a large distal domain (“sock”) and a ring of expression proximal to it (Fig. 4b). The antennae are shorter in Sp8/9 RNAi embryos, but they express Dll in their distal region (Fig. 4c). In addition, Dll is expressed in the distal tips of labrum, maxilla, and labium in Sp8/9 RNAi embryos (Fig. 4c). The thoracic legs are much shorter in Sp8/9 RNAi embryos than in the wild type, but they show a “ring and sock” pattern of Dll expression similar to the wild-type legs (Fig. 4d).

Fig. 4
figure 4

Expression of Distal-less, dachshund, and homothorax in Sp8/9 RNAi embryos. a, b Wild-type Dll expression in the head region (a) and the thoracic legs (b). c, d Expression of Dll in Sp8/9 RNAi embryos. Please note that the brownish staining in the head and the bases of all appendages (including the mandibles) in (c, d) is unspecific background sometimes produced during the staining procedure. Embryos in (a)–(d) are at the retracted germband stage. Note the expression in the tips of the maxilla and labium (arrowhead and arrow, respectively, in (a) and (c)). Also note the expression in a distal "sock" and a proximal "ring" in the thoracic legs (arrow and arrowhead, respectively, in (b) and (d)). e–g Expression of dac in wild-type (e) and Sp8/9 RNAi embryos (f, g). Please note that the embryo in (f) has been damaged during preparation (the head has come off and the t1 leg is damaged). The embryos in (e) and (f) are at the retracted germband stage and the embryo in (g) is at an even later retracted germband stage. h–i Expression of hth in wild-type (h) and Sp8/9 RNAi embryos (i). Both embryos are at the retracted germband stage. All animals are oriented with the anterior end to the left. Abbreviations: hl head lobe, lr labrum, an antenna, md mandible, mx maxilla, lb labium, t1 prothoracic leg, t2 mesothoracic leg, t3 metathoracic leg, pl pleuropodium

Full size image

The dachshund (dac) gene is in the wild-type expressed in a medial ring in the thoracic legs, and in a thin ring near the base in the labium and in the antenna (Fig. 4e) (Angelini and Kaufman 2004). The dac gene is also expressed very strongly in the mandible and in the maxilla (Fig. 4e). In the Sp8/9 RNAi animals, this pattern is not significantly altered. The expression ring in the antenna is thinner, but the expression in the mandible, maxilla, and labium is unchanged (Fig. 4f, g). The thoracic legs are significantly shortened, but the medial ring of dac expression is present as in the wild-type legs (Fig. 4f, g).

The gene homothorax (hth) is in the wild type expressed in the proximal area of all appendages (Fig. 4h), but the distal extension of expression is different in the different appendage types (Angelini and Kaufman 2004). Gene expression is restricted to the proximal third of the legs, but is expressed in the proximal two thirds of the appendage in antennae, mandible, and labium and fills the proximal half of the maxilla (Fig. 4h). This proportional pattern is identical in Sp8/9 RNAi animals although the legs are shorter than in the wild type (Fig. 4i).

A conserved role in appendage axis elongation

Previous work in Drosophila has shown that the D-Sp1 gene is expressed in the thoracic limb primordia in the embryo and in the leg imaginal discs in the larva (Wimmer et al. 1996; Schöck et al. 1999; Estella et al. 2003). In the leg discs, D-Sp1 is expressed in concentric rings that roughly correspond to the position of the future joints between the leg segments. This expression pattern is very similar to the late expression of Sp8/9 in the legs of the beetle T. castaneum. In this insect species, Sp8/9 is expressed in up to four segmental rings that lie at a similar location in the legs as the expression rings of the Serrate (Ser) gene, which encodes the ligand of the Notch (N) receptor (Beermann et al. 2004). RNAi experiments with Sp8/9 have shown that the podomeres are severely shortened and sometimes fused, and that the number of Ser rings is also reduced. Based on these data, Beermann et al. (2004) suggested that Sp8/9 is involved in the control of allometric growth of the individual leg segments, probably by interfering with the Notch pathway, which is known to control allometric podomere growth in Drosophila and other arthropod species (de Celis et al. 1998; Rauskolb and Irvine 1999; Bishop et al. 1999; Prpic and Damen 2009). It has been noted previously that the process of leg segment growth is tightly linked with the process of joint formation, because both processes are regulated by the Notch pathway and its targets (Milán and Cohen 2000), and this would then also explain the observed podomere fusions.

Our results in Oncopeltus support an evolutionarily conserved role of Sp8/9 in the control of allometric podomere growth. First, the Sp8/9 expression pattern in the legs after full germband elongation is very similar to the pattern in Tribolium and consists of several rings adjacent to the constrictions of the future leg joints. Second, after RNAi, all leg segments are still present but severely shortened compared to the wild-type legs. In addition, the podomere fusions observed in the legs and in the antennae are compatible with the notion that the Notch pathway is also affected, because similar podomere fusion phenotypes are observed in Drosophila when members of effectors of the Notch pathway are impaired (e.g., de Celis et al. 1998; Rauskolb and Irvine 1999; Bishop et al. 1999).

No evidence for a role of Sp8/9 in appendage specification

There is some evidence in Drosophila that D-Sp1 has a role in leg specification by activating Dll expression in the embryonic leg primordia together with Wg and Dpp signaling and Dll autoregulation (Estella et al. 2003; McKay et al. 2009). However, this evidence is not conclusive because the experiments were not able to discriminate between the effect of D-Sp1 and the effect of the neighboring gene btd. Beermann et al. (2004) suggest that Dll is a target gene of Sp8/9 in Tribolium, although Dll is still expressed in the legs after Sp8/9 RNAi. In summary, the evidence for a role of Sp8/9 in leg specification and Dll activation in the insects studied so far is inconclusive.

Our results with Sp8/9 RNAi in Oncopeltus provide no evidence for a role of Sp8/9 in appendage specification or Dll activation. All appendages are present in Sp8/9 RNAi animals and the legs consist of all podomeres including the distal claws. This suggests that the specification of the appendages and their overall proximal–distal patterning is not disrupted. This is further evidenced by the expression of the proximal–distal marker genes Dll, dac, and hth in Sp8/9 RNAi animals. Although the legs in these animals are much shorter than in the wild type, the expression of Dll, dac, and hth is proportionally identical to the wild-type expression, indicating that proximal, medial, and distal fates are present. These data also suggest that Dll, dac, and hth expression is not dependent on activation by Sp8/9. We note, however, that we found very weak residual expression of Sp8/9 after Sp8/9 RNAi (data not shown), and thus we cannot exclude the possibility that the phenotypes we obtained do not represent the loss-of-function (null) phenotype.