• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. May 26, 2009; 106(21): 8501–8506.
Published online May 11, 2009. doi:  10.1073/pnas.0809885106
PMCID: PMC2689000

The Drosophila DPP signal is produced by cleavage of its proprotein at evolutionary diversified furin-recognition sites


Maturation of bone morphogenetic proteins (BMPs) requires cleavage of their precursor proteins by furin-type proprotein convertases. Here, we find that cleavage sites of the BMP2/4/decapentaplegic (DPP) subfamily have been evolutionary diversified and can be categorized into 4 different types. Cnidaria BMP2/4/DPP is considered to be a prototype containing only 1 furin site. Bilateria BMP2/4/DPP acquired an additional cleavage site with either the combination of minimal–optimal or optimal–optimal furin sites. DPPs belonging to Diptera, such as Drosophila and mosquito, and Lepidoptera of silkworm contain a third cleavage site between the 2 optimal furin sites. We studied how the 3 furin sites (FSI–III) of Drosophila DPP coordinate maturation of ligands and contribute to signals in vivo. Combining mutational analysis of furin-recognition sites and RNAi experiments, we found that the Drosophila DPP precursor is initially cleaved at an upstream furin-recognition site (FSII), with consequent cleavages at 2 furin sites (FSI and FSIII). Both Dfurin1 and Dfurin2 are involved in the processing of DPP proproteins. Biochemical and genetic analyses using cleavage mutants of DPP suggest the first cleavage at FSII to be critical and sufficient for long-range DPP signaling. Our data suggest that the Drosophila DPP precursor is cleaved in a different manner from vertebrate BMP4 even though they are functional orthologs. This indicates that the furin-cleavage sites in BMP2/4/DPP precursors are tolerant to mutations acquired through evolution and have adapted to different systems in diversified species.

Keywords: bone morphogenetic protein (BMP), decapentaplegic, proprotein convertase

Bone morphogenetic proteins (BMPs), members of the transforming growth factor-β (TGF-β) superfamily, were originally identified as factors in demineralized bone extracts that could stimulate ectopic bone development. Since that time, BMPs have been found to modulate a variety of additional developmental processes, such as cell proliferation, apoptosis, differentiation, cell-fate determination, and morphogenesis (1). Moreover, molecular and embryological studies have demonstrated that BMPs function in specifying the body axis and in defining aspects of the embryonic pattern during development (2, 3).

The BMP2/4/decapentaplegic (DPP) subgroup of BMP type ligands has been proposed to play a conserved role patterning the dorsal-ventral axis of Bilateria including Vertebra, Cephalochordata, Hemichordata, Echinodermata, and Arthropoda (2, 4, 5). Recent studies demonstrate that BMP2/4/DPP orthologs are expressed in a localized fashion in the Phylum Cnidaria, such as coral Acropora millepora and sea anemone Nematostella vectensis, the closest outgroup to the Bilateria, indicating that these ligands are evolutionarily conserved in axis formation (6, 7).

In Drosophila, 3 BMP subgroup ligands, dpp, screw (scw), and glass bottom boat (gbb), have been identified, and each of these has a counterpart in mammals (8). Drosophila DPP appears to be a functional ortholog of vertebrate BMP4 because its recombinant proteins induce bone formation in mammalian cells and the human BMP4 genes are able to rescue dorsal embryonic pattern defects seen in Drosophila dpp mutants (9, 10). Furthermore, A. millepora BMP2/4 has been shown to be functional in the Drosophila embryo (6). Thus, it appears that the fundamental signaling mechanism used by BMPs during development is conserved throughout evolution.

In vertebrates, BMP4 is synthesized as an inactive precursor and is proteolytically activated by cleavage at the multibasic amino acid motif to yield a C-terminal mature protein. The combination of a potent protein inhibitor of furin and an in vitro digestion assay provided the evidence that furin and PC6 proteolytically activate BMP4 (11). Furthermore, the BMP4 precursor has been shown to be cleaved by furin in a sequential manner. Cleavage at an optimal furin site adjacent to the mature ligand domain allows for subsequent cleavage at an upstream minimal furin site within the prodomain. Further studies demonstrated that the pro- and mature domains of BMP4 remain noncovalently associated after optimal site cleavage, generating a complex that is targeted for rapid degradation. Subsequent cleavage at the minimal site liberates mature BMP4 from the prodomain, thereby stabilizing the protein (12, 13). These results indicate that the mature BMP4 ligand is produced as a single molecular form, and that the second cleavage site is functional for regulation of ligand secretion and diffusion. A recent study using mice carrying a point mutation that prevents processing of the minimal site within the prodomain of BMP4 showed severe loss of BMP4 activity in some tissues, such as testes and germ cells, suggesting that maturation and secretion of BMP4 type ligands may require different regulatory systems in different tissues (14).

Drosophila DPP protein is initially synthesized as an inactive 588-amino acid precursor protein. After dimerization and proteolytic cleavages, the active C-terminal mature forms are secreted from the cells. In contrast to the single mature form of BMP4, DPP proteins are produced as 2 different molecular forms, when tagged dpp is expressed in the cell culture and embryo (15). The N-terminal amino acid sequence of one of the mature forms has been identified to be cleaved after -RSIR456- (16), yet the additional cleavage site has not been determined. These facts raise the questions of how two different forms of DPP proteins are produced and how they contribute to sustain the signal in vivo?

Members of a proprotein convertase (PC) family are candidates for endogenous BMP convertases. PCs are proteolytic enzymes that activate precursor proteins into biologically active forms by limited proteolysis at 1 or multiple internal sites. In vertebrates, these enzymes constitute a family of 7 known basic amino acid-specific proteinases: furin, PC1/3, PC2, PC4, PACE4, PC5/6, and PC7 (17). In Drosophila, 3 members of this family have been identified: Dfurin1 (Dfur1), Dfurin2 (Dfur2), and amontillado (amon). The gene products of Dfur1 and Dfur2 were expressed in tissue culture cells and characterized as being PCs in vitro, but their mutants have not been analyzed yet (1820). Amon, in contrast, has been characterized as a PC2-type enzyme and amon mutants display partial embryonic lethality, defective larval growth, and arrest during the first to second instar larval molt (21, 22). However, there is no evidence of which enzymes are involved in the cleavage of DPP proproteins.

In this study, we identified 3 furin-recognition sites required for production of Drosophila DPP proteins. Mutational analysis of furin-recognition sites of DPP indicates that the upstream furin site is critical for ligand maturation and long-range signaling in wing development. Our results suggest that furin-cleavage sites in the BMP2/4/DPP prodomain have been diversified, even though the signaling mechanism is highly conserved; therefore, the cleavage sites are tolerant to the mutations acquired through evolution and have adapted to the system of different organisms.


Furin-Recognition Sites in the BMP2/4/DPP Proproteins.

In our previous studies, Drosophila DPP proteins were produced as 2 different molecular forms in Drosophila S2 cells and early embryos when HA-tagged dpp was expressed (15). To understand how the DPP precursor is cleaved into 2 different forms, despite vertebrate BMP4 being secreted as a single molecular form, we searched the sequences reported as BMP2/4/DPP-type ligands in the National Center for Biotechnology Information web site (www.ncbi.nlm.nih.gov/). Phylogenetic analyses of the C-terminal region of these proteins indicated that they belong to the BMP2/4/DPP subfamily of the TGF-β superfamily, however, how the furin-cleavage sites in the prodomains have been established had not been analyzed (6, 23, 24). Therefore, we compared the sequences of cleavage sites of BMP2/4/DPP proproteins. The sequence alignments suggest that the furin-cleavage sites have been diversified and can be categorized into 4 different types (Fig. 1 and Fig. S1). Cleavage type I is observed in Cnidaria BMP2/4/DPP and contains only 1 optimal furin site. Cleavage type II indicates 2 furin-recognition sites: an upstream minimal furin site (FSII: -RXXR-) and an optimal site adjacent to the mature ligand (FSI: -RXKR-). All BMP2/4 ligands belonging to Vertebrata and Echinodermata are placed in this category. Cleavage type III has optimal furin sites (-RXK/RR-) both upstream and adjacent to the ligand. This family contains BMP2/4/DPP of Cephalochordata, Urochordata, Hemichordata, and Arthropoda, excluding Diptera and Lepidoptera phyla. Cleavage type IV has only been found in Diptera such as Drosophila and mosquito, and Lepidoptera of silk worm: 2 conserved optimal furin-recognition sites are found both upstream and adjacent to the mature ligand and a minimal furin site lies between these two. These facts raise the question of why the furin-recognition sites of BMP2/4/DPP proproteins have been diversified. Are they suitable for supporting the multiple functions of BMP signaling?

Fig. 1.
The furin cleavage sites of the BMP2/4/DPP family are categorized into 4 different types. Type I contains 1 optimal furin site (RXK/RR) found in Cnidaria BMP2/4/DPP. Type II comprises a combination of minimal (RXXR) and optimal furin sites found in Vertebrata ...

Analysis of Furin-Recognition Sites in the Drosophila DPP Proproteins.

To understand how the furin-cleavage sites are regulated for signaling, we studied Drosophila DPP proteins. For that purpose, we mutated the 3 furin-cleavage sites; FSI (MFSI: RNKR to GNKG), FSII (MFSII: RLRR to GLRG), and FSIII (MFSIII: RSIR to GSIG) (Fig. 2A), and studied how they are produced in S2 cells (Fig. 2 B and C). S2 cells were transfected with HA-tagged, wild-type dpp, dppMFSI, dppMFSII, or dppMFSIII, and the protein products were analyzed by western blotting. In the cell lysates, DPP precursor proteins were detected as 90-kDa bands, and faint bands of 55-kDa fragments were observed (Fig. 2B). We normally observe equivalent levels of 90-kDa bands, indicating that proproteins of either wild-type or mutated DPP were produced at equivalent levels. Wild-type DPP was cleaved into 2 mature fragments, a 26-kDa form (DPP26) and a 23-kDa form (DPP23), and both forms were efficiently secreted into the culture medium (Fig. 2 B and C, lane 2). DPPMFSII-HA was cleaved much less efficiently than wild-type DPP-HA, even though the sizes of the mature ligands were the same as wild-type DPP-HA. The ratio of proteins in supernatants versus cell lysates of DPPMFSII-HA was equivalent to that of wild-type DPP-HA (Fig. 2 B and C, lanes 2 and 3). Taken together, we suppose that cleavage at the FSII site plays a critical role in DPP maturation. In contrast, DPPMFSI-HA was cleaved into DPP26, and these proteins were efficiently secreted into the culture medium (Fig. 2 B and C, lane 4). No DPP23 fragment was observed in either cell lysates or conditioned media. DPPMFSIII-HA was cleaved into an intermediate 30-kDa fragment (DPP30) and a DPP23 fragment, and both fragments were secreted, suggesting that processing of DPP in S2 cells is performed partially through FSII to FSI with low efficiency (Fig. 2 B and C, lane 5). These results support the hypothesis that the cleavage sites for producing the 2 different forms of DPP are FSI and FSIII. To confirm this, we mutated both FSI and FSIII. Drosophila S2 cells were transfected with HA-tagged dppMFSI/III. DPPMFSI/III-HA was cleaved into a larger form (DPP30) but not into the mature forms (DPP26 and DPP23) in the cell lysates; DPP30 was efficiently secreted into the media (Fig. 2 B and C, lane 6).

Fig. 2.
The Drosophila DPP precursor contains 3 furin-cleavage sites. (A) Schematic illustration of furin-cleavage sites and cleavage mutants of DPP proproteins. The MFSI, II, or III constructs carry 2 point mutations that make FSI, II, or III inactive, respectively. ...

These results indicate the possibility that the DPP precursor is cleaved at the FSII site to create the intermediate form and subsequently cleaved at the FSIII and FSI sites to produce mature ligands. Previous studies suggested that BMP4 is initially cleaved at the furin motif adjacent to the mature ligand (FSI), and that this allows for subsequent cleavage at an upstream motif (FSII) (12, 13). To confirm whether this is the case for Drosophila DPP processing or whether DPP cleavage is performed in a sequential manner from FSII to FSI/III, we produced the double mutants of MFSI/II and MFSII/III. When dppMFSI/II-HA was transfected into S2 cells, DPP26 was produced and secreted, but the protein levels were much lower than DPPMFSI-HA (Fig. 2 B and C, lane 8). Only faint levels of DPP23 were observed in cell lysates and in supernatants for DPPMFSII/III-HA (Fig. 2 B and C, lane 7), supporting that cleavage at the FSII site is critical for further cleavage at either FSIII or FSI, and is most likely a prerequisite for producing mature ligands.

Different Cleavage Forms of DPP Proteins Show the Same Signaling Intensities.

To examine how the different molecular forms of DPP ligands bind to their receptors, we used a cell-based signaling assay to measure BMP signaling (25). To compare the signaling intensities of each ligand, we first quantified protein levels of wild-type and cleavage mutants of DPP-HA by western blot analysis (Fig. S2). Using equivalent amounts of ligands, we measured the signaling activities. Wild-type DPP-HA, DPPMFSI-HA, DPPMFSIII-HA, and DPPMFSI/III-HA showed equivalent signaling intensities (Fig. 2D). These results suggest that, once they are secreted, differentially cleaved forms of DPP have the same affinity for binding to the receptors for signaling.

RNA Interference (RNAi) of Proprotein Convertases.

To understand how the cleavages at FSI-III are regulated to produce DPP26 and DPP23 fragments, for example, sequential cleavages by a single protease or cleavages by multiple proteases, we used an RNAi approach. Because furin-type proprotein convertases (PCs) are involved in BMP4 cleavage, we investigated how Drosophila PCs cleave DPP proproteins. We cotransfected dsRNAs against the 3 different PCs together with wild-type dpp-HA in S2 cells. We found that Dfur1 RNAi produced the intermediate form (DPP30), and Dfur2 RNAi reduced the production of the small mature form (DPP23) but increased large mature form (DPP26) productions (Fig. 3 A and B, lanes 4 and 5). However, we did not find any difference by knocking down amon (Fig. 3 A and B, lanes 9 and 10). Therefore, we suspected that PCs are partly redundant, and we coincubated 2 different dsRNAs to see whether this was true. The combination of both Dfur1 RNAi and Dfur2 RNAi drastically reduced the production of mature forms (Fig. 3 A and B, lane 6). To confirm that the reduction of mature DPP production comes from redundancies of DFur1 and DFur2, but not from the absolute dosage of PCs, we tested equal amounts of dsRNA against Dfur1 (2 μg/mL)/Dfur2 (2 μg/mL), Dfur1 (4 μg/mL), or Dfur2 (4 μg/mL). Only the combination of Dfur1/Dfur2 RNAi showed severe reductions of mature DPP proteins although the amounts of mature DPP proteins were slightly lower when increasing the dosage (4 μg/mL) of dsRNA in either Dfur1 or Dfur2 knockdown (Fig. 3 A and B, lanes 4–8). To further confirm that the RNAi effects were specific to the target genes, we measured the transcriptional and translational products of Dfur1 and Dfur2 after RNAi, showing that RNAi against Dfur1 and Dfur2 was highly specific and very effective (Fig. 3C and Fig. S3). Combining these data, we conclude that DFur1 and DFur2 cleave at the FSII and FSIII sites and that DFur2 is a primary enzyme for cleavage at the FSI site in S2 cells.

Fig. 3.
Both Dfur1 and Dfur2 are involved in the cleavage of DPP proproteins. (A and B) RNA interference of proprotein convertaseses in DPP cleavage. S2 cells were transfected with different combinations of dsRNA against amon, Dfur1, Dfur2, and gfp in a final ...

Functional Rescue in Wing Formation.

To investigate the physiological significance of the DPP cleavage, we studied the functions of DPP cleavage mutants and PCs in vivo. In Drosophila, dpp is expressed at the anterior–posterior boundary in the wing imaginal disc and forms an activity gradient via long-range diffusion (26). First, to confirm that the DPP precursor is processed during wing formation as it is in S2 cells, we compared the molecular forms of HA-tagged wild-type, MFSII, MFSI, and MFSI/III DPP expressed in the wing imaginal discs. Precursor 90-kDa fragments were observed in both wild-type and cleavage mutants (Fig. 4A). Both DPP26 and DPP23 were produced in wild type, and only DPP26 or DPP30 was produced in MFSI or MFSI/III. In contrast, only faint bands of mature ligands were observed in MFSII. To understand how DPP precursors are processed in the wing imaginal disc, we studied the expression patterns of the PCs. Both Dfur1 and Dfur2 are broadly expressed in the wing pouch, whereas amon is not (Fig. S4). We next tried to confirm the in vivo significance of Dfur1 and Dfur2. We obtained RNAi transgenic flies against Dfur1 and Dfur2 from the Vienna Drosophila RNAi Center. We tested two copies of Dfur1 or Dfur2 RNAi, or a combination of both Dfur1 and Dfur2 RNAi induced by the dpp-Gal4 driver and the MS1096-Gal4 driver, which is expressed ubiquitously in the wing disc, at 25 °C and 29 °C. However, we could not detect the phenotypes by using this method. We speculate that small fractions of Dfur1 and Dfur2, expressed after knockdown by RNAi, produce enough mature DPP to support wing development.

Fig. 4.
Functional rescue by cleavage mutants of dpp in wing formation. (A) A western blot analysis of wing disc extracts overexpressing wild-type or cleavage mutant dpps. Thirty wing discs of third instar larvae were collected and lysed by SDS/PAGE sample buffer. ...

To examine whether the cleavage mutants of DPP-HA can functionally complement DPP signaling during wing formation, we compared the phenotypes of wild-type DPP-HA or mutated DPP-HA expressed in a dppd6/dppd14 mutant background (Fig. 4C). Previous studies showed that the combination of 2 different hypomorphic alleles of dpp mutants left the flies viable with small wings, and these mutant phenotypes could be rescued by ectopic expression of dpp (27). Wild-type DPP-HA rescued the mutant phenotype as expected (Fig. 4D). DPPMFSI-HA and DPPMFSI/III-HA rescued the mutant flies, indicating that DPPMFSI and DPPMFSI/III are functional during wing development (Fig. 4 F and G). However, none of the transgenic lines expressing dppMFSII-HA that we tested rescued the mutant phenotype (Fig. 4E). To investigate how ectopically expressed proteins were produced and distributed, we stained DPP proteins with HA antibody in the wing imaginal disc by both conventional and extracellular staining. Both wild-type and cleavage mutants of DPP-HA were stably visualized in the anterior–posterior boundary in the wing disc by conventional staining (Fig. 4 H–K). To consider how DPP-cleavage mutant proteins are distributed in the wing imaginal disc, the extracellular localization of the ligands were visualized. Extracellular staining of wild-type, MFSI, and MFSI/III were observed in the wing pouch, and scanned images of them showed similar ranges of protein distribution, supporting the data of the rescue experiments (Fig. 4 H', J', and K'). In contrast, we could not detect extracellular staining of MFSII (Fig. 4I'). The heparan sulfate proteoglycan (HSPG), Dally, has been shown to be involved in DPP diffusion in wing disc; therefore it is important to identify how different forms of DPP proteins interact with HSPG. Previous work studied the interactions of truncated forms of DPP with heparin and demonstrated that deficiencies of the stretch of 5 basic amino acids of DPP ligands made the protein lose its affinity to heparin (28). This corresponds with the similarities observed in DPP diffusion pattern when comparing flies with the truncated form of DPP and dally mutants. We tested the interactions of various forms of DPP proteins with heparin and confirmed that DPPMFSI/III as well as wild-type DPP can bind heparin (Fig. S5). Because the FSI site is only 2 aa upstream of the basic amino acid core (HSPG binding site), the additional extension of N-terminal domain of DPP ligand does not have an impact on binding to HSPG. We conclude that DPP, once cleaved at the FSII site, is secreted and functional in vivo, supporting long-range signaling.


DPP Maturation Through Multistep Cleavages.

Here, we have shown that the maturation of Drosophila DPP requires multistep cleavages of proproteins: cleavage at the FSII site with sequential cleavage at the FSIII or FSI site for production of the large (DPP26) or small (DPP23) fragment of DPP. We also showed that the first cleavage at the FSII site is critical and sufficient for long-range DPP signaling in wing development. These results clearly indicate that Drosophila DPP precursor is cleaved in a different manner from vertebrate BMP4 even though they are functional orthologs.

We expect the sequential cleavages from FSII to either FSIII or FSI of Drosophila DPP to occur for the following reasons. (i) When knocking down both Dfur1 and Dfur2, the intermediate form DPP30 was produced and the production of mature ligands was drastically reduced (Fig. 3 A and B). We suppose that DPP30, which is normally processed into mature ligands promptly, is retained during the period of secretion if the amounts of PCs are very limited and the cleavage rate is extremely slow. (ii) DPPMFSI/III-HA proteins are expressed as the intermediate form and secreted as efficiently as wild-type DPP, suggesting that cleavage at FSII can occur without prerequirement of cleavage at either FSI or FSIII (Fig. 2 B and C). (iii) Even though Drosophila DPP and vertebrate BMP4 are functional orthologs, we find that their sequence motifs in the cleavage domains are not conserved (Fig. 1 and Fig. S1). Previous studies suggest that sequence motifs are critical for protein processing; mutation of the BMP4 FSII site from minimal to optimal changed the order of processing, in that the FSII site is cleaved rapidly, with equal efficiency to cleavage at the FSI site (12). Therefore, it is not surprising that the direction of DPP cleavage is different from that of BMP4.

Processing Enzymes for DPP Cleavage.

Based on the results of the RNAi analysis in S2 cells, we speculate that DFur1 prefers the cleavage of FSII and FSIII and that their cleavage is closely linked. However, DFur2 can cleave FSI, and is also capable of cleavage at FSII and FSIII. This suggests that 2 proteases may coordinate the maturation of DPP ligands. Even though DPP30 fragment was produced at low level after knocking down both Dfur1 and Dfur2, we suppose that the cleavage at the FSII site is redundant for DFur1 and DFur2 for the following reasons. There are only 3 furin-like proteases found in Drosophila. Among them, Amon can be excluded from DPP processing enzymes. Although the amounts of DFur1 and DFur2 have been significantly reduced by RNAi experiments, small fractions of ≈10% of the control mRNA levels are still being expressed (Fig. 3C). Therefore, we suspect that cleavage at the FSII site has been done by remaining DFur1 or DFur2, although we cannot deny the possibility that other unknown proteases perform this. In addition, if cleavage at the FSII site is done by other proteases than DFur1 or DFur2, we should see similar intensities of the DPP30 fragment to the ones observed in DPPMFSI/III. However, we normally see much weaker DPP30 bands than those of dppMFSI/III mutants when both Dfur1 and Dfur2 are knocked down. Therefore, we consider that the primary processing enzymes at the FSII site would be DFur1 and DFur2, but not other proteases. It is difficult to conclude in vivo functions of DFur1 and DFur2 at this point because mutant alleles for either Dfur1 or Dfur2 have yet to be characterized. However, our data provide the possibility that both DFur1 and DFur2 are proteases required for DPP maturation.

Evolutional Changes of Furin-Cleavage Sites in BMP2/4/DPP-Type Ligands.

We found that the furin-recognition sites of BMP2/4/DPP-type ligands have been diversified and can be categorized into 4 different types. Previous phylogenetic analyses have focused on the ligand domain starting with the first cysteine residue, demonstrating that this domain of the BMP2/4/DPP subfamily is highly conserved (e.g., Cnidaria coral BMP2/4 has an 80% amino acid sequence identity with Xenopus laevis BMP2) (6, 23, 24). These results suggest that the ligand domain is not tolerant to mutations through evolution. In contrast, furin-cleavage sites have been diversified and seem to be tolerant to mutations even though cleavage by furins is a critical step for producing biologically active ligands. These observations may be explained by the different stringencies required for ligand-receptor or protease-substrate interactions. The interactions of ligand-receptor require very high stringencies as the formation of rigid 3 dimensional structures is very critical to avoid misleading the signaling pathway (29). Furin-type proteases, however, are less stringent to their substrates and recognize a signal composed of multiple combinations of 4 to 6 amino acid residues called furin-recognition sites (30). In Drosophila, the FSII site of the DPP precursor is the most critical site for signaling, and the evolutionarily conserved FSI site may not be crucial anymore, even though the DPP precursor is cleaved at the FSI site in vivo (lost function). In vertebrates, the FSII site of BMP4 has been shown to be used for signal regulation after cleavage at the FSI site (acquired function).

In summary, our data suggest that BMP2/4/DPP-type ligands have acquired different types of furin-cleavage sites to be used for maturation in diversified species. Further analyses of type I and type III ligands will define how these furin-cleavage sites have adapted to different systems through evolution. Because the BMP2/4/DPP subfamily has been proposed to be one of the fundamental molecules for body axis formation, our findings provide insights for further studies of evolution and molecular diversities.

Materials and Methods

Constructs, RNAi, and RT-PCR.

dpp cleavage mutant cDNAs were constructed by using the QuikChange Site-Directed Mutagenesis kit (Stratagene). All constructions were confirmed by sequencing. The EcoRI fragments of these constructs were subcloned into pUAST for transgenic flies and pBRAcpA for cell culture. The PCR primers used for making dsRNA were designed as described in ref. 25. RNAi efficiency was measured by quantitative RT-PCR. RNA was isolated with TRIzol (Invitrogen). M-MLV reverse transcriptase (Promega) was used for cDNA synthesis and LightCycler 480 SYBR Green I Master (Roche) was used for RT-PCR. Detailed procedures are available on request.

Production of Recombinant Proteins and BMP Signaling Assay.

Drosophila S2 cells were transfected with wild-type or cleavage mutants of dpp-HA and incubated at 25 °C for 5 days. Conditioned medium was used for supernatants. The cells were lysed in lysis buffer (20 mM Tris·HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100) and the soluble fractions were used for cell lysates. Protein samples were analyzed with western blotting as described in ref. 25. Anti-HA 12CA5 (Roche), anti-β-tublin (Sigma), anti-flag M2 (Sigma) and anti-phosphoMad (courtesy of P. ten Dijke, Leiden University, Leiden, The Netherlands) antibodies were used as primary antibodies. BMP signaling assays were performed as described in ref. 25.

Fly Stocks and Rescue Experiments.

UAS-cleavage mutant dpp-HA flies were obtained by standard protocol. UAS-dpp-HA and A9-Gal4 flies are described in ref. 15. dpp-Gal4blk40C.6, dppd6, dppd14 flies were obtained from Bloomington Drosophila stock center. Flies were raised and crossed at 25 °C. For wing rescue experiments, flies of the genotypes w; dppd6/CyO; dpp-Gal4/Tm3Sb were crossed to w; dppd14, UAS-dpp-HA (wild-type or cleavage mutant)/CyO or dppd14/CyO; UAS-dpp-HA (wild-type or cleavage mutant)/Tm3Sb. To specify the molecular forms of DPP proteins in wing imaginal discs, wing disc extracts of A9-Gal4; UAS-dpp-HA (wild-type or cleavage mutants) flies were obtained and analyzed by western blotting.


Staining of wing discs was carried out following standard procedures. The wing discs were dissected from late third instar larvae in PBS and fixed with 4% formaldehyde/PBS for 20 min at room temperature. The fixed discs were stained with anti-HA 3F10 (Roche) as a primary antibody and Alexa 488 anti-rat (Molecular Probes) as a secondary antibody. To visualize extracellular proteins, tissues were incubated with the primary antibody before fixation at 4 °C. Intensity profiles for extracellular signals were generated by ImageJ 1.39.

Supplementary Material

Supporting Information:


We thank Mike O'Connor for his support during the initial phase of this work, Takuya Akiyama and Hiroshi Nakato for communication of results before publication, and Stuart Newfeld for thoughtful comments on the manuscript. We are grateful to Wim Van de Ven (University of Leuven, Leuven, Belgium) for anti-DFur1 and DFur2 antibodies. This work was supported by the Academy of Finland and the Sigrid Juselius foundation (O.S.). J.K. was supported by the Finnish Cultural Foundation and Viikki Graduate School in Biosciences.


The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at www.pnas.org/cgi/content/full/0809885106/DCSupplemental.


1. Hogan BL. Bone morphogenetic proteins in development. Curr Opin Genet Dev. 1996;6:432–438. [PubMed]
2. De Robertis EM. Evo-devo: Variations on ancestral themes. Cell. 2008;132:185–195. [PMC free article] [PubMed]
3. Whitman M. Smads and early developmental signaling by the TGFbeta superfamily. Genes Dev. 1998;12:2445–2462. [PubMed]
4. Lowe CJ, et al. Dorsoventral patterning in hemichordates: Insights into early chordate evolution. PLoS Biol. 2006;4:e291. [PMC free article] [PubMed]
5. Van der ZM, da Fonseca RN, Roth S. TGFbeta signaling in Tribolium: Vertebrate-like components in a beetle. Dev Genes Evol. 2008;218:203–213. [PubMed]
6. Hayward DC, et al. Localized expression of a dpp/BMP2/4 ortholog in a coral embryo. Proc Natl Acad Sci USA. 2002;99:8106–8111. [PMC free article] [PubMed]
7. Finnerty JR, Pang K, Burton P, Paulson D, Martindale MQ. Origins of bilateral symmetry: Hox and dpp expression in a sea anemone. Science. 2004;304:1335–1337. [PubMed]
8. Raftery LA, Sutherland DJ. TGF-β family signal transduction in Drosophila development: From Mad to Smads. Dev Biol. 1999;210:251–268. [PubMed]
9. Sampath TK, Rashka KE, Doctor JS, Tucker RF, Hoffmann FM. Drosophila transforming growth factor beta superfamily proteins induce endochondral bone formation in mammals. Proc Natl Acad Sci USA. 1993;90:6004–6008. [PMC free article] [PubMed]
10. Padgett RW, Wozney JM, Gelbart WM. Human BMP sequences can confer normal dorsal-ventral patterning in the Drosophila embryo. Proc Natl Acad Sci USA. 1993;90:2905–2909. [PMC free article] [PubMed]
11. Cui Y, Jean F, Thomas G, Christian JL. BMP-4 is proteolytically activated by furin and/or PC6 during vertebrate embryonic development. EMBO J. 1998;17:4735–4743. [PMC free article] [PubMed]
12. Cui Y, et al. The activity and signaling range of mature BMP-4 is regulated by sequential cleavage at two sites within the prodomain of the precursor. Genes Dev. 2001;15:2797–2802. [PMC free article] [PubMed]
13. Degnin C, Jean F, Thomas G, Christian JL. Cleavages within the prodomain direct intracellular trafficking and degradation of mature bone morphogenetic protein-4. Mol Biol Cell. 2004;15:5012–5020. [PMC free article] [PubMed]
14. Goldman DC, et al. Mutation of an upstream cleavage site in the BMP4 prodomain leads to tissue-specific loss of activity. Development. 2006;133:1933–1942. [PubMed]
15. Shimmi O, Umulis D, Othmer H, O'Connor MB. Facilitated transport of a Dpp/Scw heterodimer by Sog/Tsg leads to robust patterning of the Drosophila blastoderm embryo. Cell. 2005;120:873–886. [PubMed]
16. Panganiban GE, Rashka KE, Neitzel MD, Hoffmann FM. Biochemical characterization of the Drosophila dpp protein, a member of the transforming growth factor beta family of growth factors. Mol Cell Biol. 1990;10:2669–2677. [PMC free article] [PubMed]
17. Thomas G. Furin at the cutting edge: From protein traffic to embryogenesis and disease. Nat Rev Mol Cell Biol. 2002;3:753–766. [PMC free article] [PubMed]
18. Roebroek AJ, et al. Cloning and functional expression of Dfurin2, a subtilisin-like proprotein processing enzyme of Drosophila melanogaster with multiple repeats of a cysteine motif. J Biol Chem. 1992;267:17208–17215. [PubMed]
19. Roebroek AJ, Creemers JW, Pauli IG, Bogaert T, Van de Ven WJ. Generation of structural and functional diversity in furin-like proteins in Drosophila melanogaster by alternative splicing of the Dfur1 gene. EMBO J. 1993;12:1853–1870. [PMC free article] [PubMed]
20. De Bie I, et al. Processing specificity and biosynthesis of the Drosophila melanogaster convertases dfurin1, dfurin1-CRR, dfurin1-X, and dfurin2. J Biol Chem. 1995;270:1020–1028. [PubMed]
21. Rayburn LY, et al. Amontillado, the Drosophila homolog of the prohormone processing protease PC2, is required during embryogenesis and early larval development. Genetics. 2003;163:227–237. [PMC free article] [PubMed]
22. Siekhaus DE, Fuller RS. A role for amontillado, the Drosophila homolog of the neuropeptide precursor processing protease PC2, in triggering hatching behavior. J Neurosci. 1999;19:6942–6954. [PubMed]
23. Lapraz F, et al. RTK and TGF-beta signaling pathways genes in the sea urchin genome. Dev Biol. 2006;300:132–152. [PubMed]
24. Newfeld SJ, Wisotzkey RG, Kumar S. Molecular evolution of a developmental pathway: Phylogenetic analyses of transforming growth factor-beta family ligands, receptors and Smad signal transducers. Genetics. 1999;152:783–795. [PMC free article] [PubMed]
25. Shimmi O, O'Connor MB. Physical properties of Tld, Sog, Tsg, and Dpp protein interactions are predicted to help create a sharp boundary in Bmp signals during dorsoventral patterning of the Drosophila embryo. Development. 2003;130:4673–4682. [PubMed]
26. Affolter M, Basler K. The Decapentaplegic morphogen gradient: From pattern formation to growth regulation. Nat Rev Genet. 2007;8:663–674. [PubMed]
27. Teleman AA, Cohen SM. Dpp gradient formation in the Drosophila wing imaginal disc. Cell. 2000;103:971–980. [PubMed]
28. Akiyama T, et al. Molecular mechanisms of glypican co-receptor function: The role of Drosophila Dally in Dpp signaling. Dev Biol. 2008;313:408–419. [PMC free article] [PubMed]
29. Shi Y, Massague J. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell. 2003;113:685–700. [PubMed]
30. Rockwell NC, Krysan DJ, Komiyama T, Fuller RS. Precursor processing by kex2/furin proteases. Chem Rev. 2002;102:4525–4548. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...