The NDNF-like factor Nord is a Hedgehog-induced extracellular BMP modulator that regulates Drosophila wing patterning and growth

  1. Shu Yang
  2. Xuefeng Wu
  3. Euphrosyne I Daoutidou
  4. Ya Zhang
  5. MaryJane Shimell
  6. Kun-Han Chuang
  7. Aidan J Peterson
  8. Michael B O'Connor
  9. Xiaoyan Zheng  Is a corresponding author
  1. Department of Anatomy and Cell Biology and the GW Cancer Center, George Washington University School of Medicine and Health Sciences, United States
  2. Department of Genetics, Cell Biology & Development and the Developmental Biology Center, University of Minnesota, United States

Abstract

Hedgehog (Hh) and Bone Morphogenetic Proteins (BMPs) pattern the developing Drosophila wing by functioning as short- and long-range morphogens, respectively. Here, we show that a previously unknown Hh-dependent mechanism fine-tunes the activity of BMPs. Through genome-wide expression profiling of the Drosophila wing imaginal discs, we identify nord as a novel target gene of the Hh signaling pathway. Nord is related to the vertebrate Neuron-Derived Neurotrophic Factor (NDNF) involved in congenital hypogonadotropic hypogonadism and several types of cancer. Loss- and gain-of-function analyses implicate Nord in the regulation of wing growth and proper crossvein patterning. At the molecular level, we present biochemical evidence that Nord is a secreted BMP-binding protein and localizes to the extracellular matrix. Nord binds to Decapentaplegic (Dpp) or the heterodimer Dpp-Glass-bottom boat (Gbb) to modulate their release and activity. Furthermore, we demonstrate that Nord is a dosage-dependent BMP modulator, where low levels of Nord promote and high levels inhibit BMP signaling. Taken together, we propose that Hh-induced Nord expression fine-tunes both the range and strength of BMP signaling in the developing Drosophila wing.

Editor's evaluation

This manuscript is of broad interest to readers in the field of Bone Morphogenetic Proteins (BMPs) and Drosophila wing development. The identification of a Nord as an indirect regulator of BMP diffusion is a new contribution to our understanding of regulation of BMP function. A combination of genetics, transcriptomics, and biochemistry provide support for many of the conclusions in the paper.

https://doi.org/10.7554/eLife.73357.sa0

Introduction

Morphogens are conserved, secreted signaling molecules that pattern organs and tissues by eliciting graded responses from cells surrounding a localized source (Lawrence and Struhl, 1996; Wolpert, 1969). Secreted signaling proteins of the Epidermal Growth Factor (EGF), Fibroblast Growth Factor (FGF), Hedgehog (Hh), Notch, Transforming Growth Factor β (TGF-β)/Bone Morphogenetic Protein (BMP), and Wnt/Wingless (Wg) families have been shown to act as morphogens either at a short range or over a long distance. How these morphogens and their corresponding signaling pathways coordinate to control patterning and growth in various developmental systems is a central question in developmental biology.

The developing appendages of Drosophila provide a valuable model system to genetically and molecularly identify morphogens and members of their corresponding signaling pathways. They have also assisted in revealing the fundamental features and interplay between different morphogens during pattern formation. The Drosophila Hh and BMP signals function as short- and long-range morphogens, respectively, and together organize patterning along the anteroposterior (A/P) axis of the wing imaginal discs (Lecuit et al., 1996; Mullor et al., 1997; Nellen et al., 1996; Strigini and Cohen, 1997; Zecca et al., 1995).

In the case of Hh signaling, the quiescent state of the pathway is maintained by the Hh receptor Patched (Ptc)-dependent inhibition of Smoothened (Smo) (Ingham et al., 2011). This inhibition is lifted by Hh binding to Ptc and its co-receptors (Allen et al., 2011; Izzi et al., 2011; McLellan et al., 2008; Zheng et al., 2010), which releases Smo from repression and thus activates intracellular signal transduction (Beachy et al., 2010; Lum and Beachy, 2004; Varjosalo and Taipale, 2008). In the wing discs, Hh is secreted in the posterior (P) compartment and spreads toward the anterior (A) compartment (Basler and Struhl, 1994; Capdevila et al., 1994; Tabata and Kornberg, 1994). Hh signaling does not occur in P compartment cells because they do not express critical components of the Hh pathway, such as the major transcriptional effector Ci (Eaton and Kornberg, 1990). Conversely, A compartment cells can receive and respond to Hh but are unable to produce Hh. Hh synthesized by P cells diffuses into the A compartment adjacent to the A/P boundary, where it is sequestered by the Ptc-Ihog/Boi receptor complex (Chen and Struhl, 1996; Zheng et al., 2010), thus forming a short-range gradient that activates the localized expression ofseveral target genes, including ptc and decapentaplegic (dpp) (Basler and Struhl, 1994; Capdevila et al., 1994; Chen and Struhl, 1996; Ingham et al., 1991; Tabata and Kornberg, 1994).

Dpp is the homolog of vertebrate TGF-β superfamily ligands BMP2/4, which binds to the BMP type I receptor thickveins (Tkv) or Saxophone (Sax) and type II receptor Punt (Parker et al., 2004). Upon ligand binding, Tkv or Sax phosphorylates the transcription factor Mad (referred to as pMad; an indication of active Dpp signal transduction). Phosphorylated Mad then forms a trimeric complex with the co-Smad Medea and is translocated to the nucleus where it regulates downstream target genes (Moustakas and Heldin, 2009). In contrast to the Hh signal, Dpp diffuses into both the A and P compartments and acts as a long-range morphogen (Brook et al., 1996; Lawrence and Struhl, 1996). Dpp has multiple roles both in determining the wing size by promoting growth and epithelial morphogenesis during larval stages (Gibson and Perrimon, 2005; Schwank et al., 2008; Shen and Dahmann, 2005), and in patterning the longitudinal veins (LV) as well as the anterior crossvein (ACV) and posterior crossvein (PCV) during pupal development (de Celis et al., 1996; Ralston and Blair, 2005; Serpe et al., 2005; Shimmi et al., 2005; Sotillos and De Celis, 2005; Vilmos et al., 2005).

The diverse signaling outputs of Dpp are influenced by the formation of ligand heterodimers and by their binding to various extracellular factors. In Drosophila, two other characterized BMP family ligands, the BMP5/6/7/8 homolog Glass-bottom boat (Gbb), and the more distantly related Screw (Scw) (Arora et al., 1994; Padgett et al., 1987; Wharton et al., 1991), can form heterodimers with Dpp to augment the level and increase the range of BMP signaling in different cells and tissues. Other secreted or membrane-binding BMP-binding proteins, including Short gastrulation (Sog) (Ralston and Blair, 2005; Serpe et al., 2005), Twisted gastrulation (Tsg), Crossveinless (Cv) (Shimmi et al., 2005; Vilmos et al., 2005), Crossveinless-2 (Cv-2) (Conley et al., 2000; Serpe et al., 2008), Larval translucida (ltl) (Szuperák et al., 2011), Pentagone (Pent) (Vuilleumier et al., 2010), Kekkon5 (Kek5) (Evans et al., 2009), Dally and Dally-like protein (Dlp) (Akiyama et al., 2008; Belenkaya et al., 2004), have been identified as modulators that enhance or inhibit BMP signaling. Additionally, Tkv, the main type I Dpp receptor in Drosophila, critically affects Dpp tissue distribution through ligand trapping and internalization, and thus globally shapes the BMP activity gradient (Crickmore and Mann, 2006; Lecuit and Cohen, 1998; Tanimoto et al., 2000). Prominent among the many determinants that impact proper establishment and maintenance of the Dpp activity gradient is the action of the Hh signal, which simultaneously induces Dpp expression and lowers responsiveness to Dpp in the same cells at the A/P border by repressing transcription of the tkv receptor gene (Tanimoto et al., 2000). The coordinated activation and attenuation of BMP signaling activity by Hh illustrates the complex regulatory interactions of different morphogens in pattern and growth in developing systems.

Here, we show that the activity of BMPs is further fine-tuned by a previously unknown Hh-dependent mechanism involving induction of nord, which encodes a Neuron-Derived Neurotrophic Factor (NDNF)-like factor. Absence of nord results in reduced medial BMP signaling near the source of Dpp in the larval wing imaginal discs but increases long-range BMP activity surrounding the primordial PCV in the pupal wing. Thus, adult wings from nord mutant flies exhibit seemingly opposite phenotypes of reduced wing size and ectopic PCV. Expression and co-immunoprecipitation studies in cultured S2 cells demonstrate that Nord is secreted, associates with the extracellular matrix, and binds to Dpp or the heterodimer of Dpp-Gbb. We further show that Nord is a dosage-dependent BMP modulator, where low levels of Nord promote and high levels inhibit BMP signaling. Taken together, we propose that Hh-induced Nord expression modulates the distribution and level of BMP signaling in the developing Drosophila larval and pupal wing to obtain proper size and crossvein patterning.

Results

Genome-wide expression profiling in Drosophila wing imaginal discs identifies nord as a novel target gene of the Hh signaling pathway

The imaginal discs of Drosophila melanogaster, where most known Hh signaling target genes are expressed with a restricted pattern (Strigini and Cohen, 1997), offer an accessible model system for identifying novel targets of the Hh signaling pathway. In wing discs, cells near the A/P compartment boundary (B: ptc+) receive the highest level of Hh stimulation while A cells (A: hh-), located further from the border, receive lower levels of stimulation. In contrast, P cells (P: hh+) do not respond to Hh due to lack of the receptor Ptc and the transcription factor Ci (Eaton and Kornberg, 1990; Figure 1A, Figure 1—figure supplement 1). To identify target genes whose expression is controlled, directly or indirectly, by Hh signaling activity, we performed a systematic comparison of gene expression profiles among the three cell types. After genetic labeling, wing discs were dissected, cells dissociated, and then sorted using fluorescence-activated cell sorting (FACS). RNA was extracted from the sorted cell populations and subjected to microarray analysis (Figure 1A, Figure 1—figure supplement 1). Previously, by comparing genes differentially expressed in the A/P boundary adjacent cells (B: ptc+) and P cells (P: hh+), we identified an unknown Hh pathway target dTRAF1/TRAF4 and established that the Hh signal mediates JNK activity by regulating the expression of dTRAF1/TRAF4 in developmental organ size control (Willsey et al., 2016; Figure 1B). Here, we modified the gene expression analysis method by including additional transcriptome comparisons between the A/P boundary adjacent cells (B: ptc+) and A cells (A: hh-), and between A cells (A: hh-) and P cells (P: hh+). Genes whose expression is not only higher in A cells than P cells (FoldA/P > 1.2), but also higher in the A/P boundary adjacent cells than general A cells (FoldB/A > 1.5) were selected as potential Hh-induced target genes (Figure 1B and C). Hh-responsive genes known to be differentially expressed in the wing discs were found, including ptc and dpp, (Figure 1B–D, Supplementary file 1). We then focused on nord (FlyBase, CG30418), one of the top-ranking A/P boundary-enriched genes, with no previously characterized expression pattern or functional analysis (Figure 1B–D). We first verified the differential expression of nord across the Drosophila wing imaginal discs via quantitative reverse transcription PCR using RNA isolated from FACS sorted A, B, and P cells (Figure 1E). We also performed in situ hybridization to localize nord transcripts in the wing imaginal discs (Figure 1F). Like the known Hh signaling target gene ptc (Figure 1E and F), nord transcripts were absent from the P compartment and primarily detected in the A cells adjacent to the A/P compartment boundary (Figure 1E and F). Of note, unlike ptc and most other Hh pathway target genes, nord expression was not detected in the central wing pouch (Figure 1F). Collectively, these results suggested that nord is a potential target gene of the Drosophila Hh signaling pathway.

Figure 1 with 1 supplement see all
nord is a novel target gene of the Drosophila Hedgehog (Hh) signaling pathway.

(A) Schematic diagram of Drosophila wing imaginal disc: posterior compartment (P: hh+), anterior compartment (A: hh-), and anterior compartment cells adjacent to the A/P boundary (B: ptc+). A, P, and B cells from wing imaginal discs of third instar larvae carrying hh-Gal4 or ptc-Gal4-driven UAS-mCD8-GFP were dissociated and sorted by fluorescence-activated cell sorting (FACS). RNA was isolated and hybridized to microarrays. Differentially expressed genes were identified. (B) Dot plot shows all 14,448 annotated probe sets. Each dot indicates one probe set. The x-axis represents the log2 fold change of each gene in B vs. A cells, and the y-axis represents the log2 fold change of each gene in A vs. P cells. The blue dash lines indicated the threshold used to select differentially expressed genes in each group. Colored dots indicate known or novel Hh pathway targets dTraf1/Traf4 (orange), dpp (red), ptc (red), and nord (blue) in the graph. (C) Zoomed view of the bottom-right corner in panel (B) to show 59 differentially expressed genes, whose expression is significantly increased in B (ptc+) cells but decreased in P (hh+) cells compared with A (hh-) cells. (D) Heatmap shows the expression level of the 59 top-ranking differentially expressed genes in A, B, and P cells. The fold change in B vs. A cells was used to rank the order. (E) Fold changes of ci, hh, ptc, and nord mRNA expression, measured by quantitative reverse transcription PCR and normalized by the expression of the housekeeping gene pkg, in FACS-sorted A, B, or P cells. (F) In situ hybridization of ptc and nord in the third instar larval Drosophila wing discs. Scale bar, 50 μm.

Hh signaling regulates nord expression in the wing discs

To further analyze the expression pattern and investigate the function of Nord, we identified and characterized two Minos-Mediated Integration Cassette (MiMIC) lines from the Drosophila Gene Disruption Project (GDP) collection (Venken et al., 2011), one a gene-trap nord allele Mi{MIC}nordMI06414 and the second a protein-trap nord allele Mi{PT-GFSTF.2}nordMI06414-GFSTF.2. We also generated an additional protein-trap nord allele Mi{PT-RFPHA.2}nordMI06414-RFPHA.2. The nord gene-trap allele Mi{MIC}nordMI06414 contains a MiMIC consisting of a splice acceptor site followed by stop codons in all three reading frames. This transposon was inserted into the first coding intron of nord and thus interrupted transcription and translation of nord (Figure 2—figure supplement 1). This nord gene-trap allele was used in the functional analysis of Nord. The two protein-trap nord alleles Mi{PT-GFSTF.2}nordMI06414-GFSTF.2 and Mi{PT-RFPHA.2}nordMI06414-RFPHA.2 were derived from the Mi{MIC}nordMI06414 line by replacing the original gene-trap cassette with either an EGFP-FlAsH-StrepII-3xFLAG (GFSTF) or a TagRFP-T-3xHA (RFPHA) protein-trap cassette using Recombinase-Mediated Cassette Exchange strategies (Nagarkar-Jaiswal et al., 2015; Figure 2A). In these nord protein-trap alleles, the GFSTF or RFPHA tag was inserted in the appropriate orientation and reading frame of nord between amino acids 103R and 104F (Figure 2—figure supplement 2), which permitted visualization of the Nord protein localization in vivo. Immunostaining of wing imaginal discs from the nord protein-trap alleles showed that Nord is expressed along the A/P boundary within the hinge and notum but avoids the central wing pouch (Figure 2B and C), identical to the pattern of endogenous nord transcripts revealed by in situ hybridization (Figure 1F).

Figure 2 with 3 supplements see all
Hedgehog (Hh) signaling regulates nord expression in the Drosophila wing imaginal disc.

(A) Schematic diagram of the wild-type nord locus and the protein-trap alleles of nord. The EGFP-FlAsH-StrepII-3xFLAG (GFSTF) or TagRFP-T-3xHA (RFPHA) tag was inserted in the appropriate orientation and reading frame of nord, which permitted visualization of the Nord protein localization in vivo. (B) Wing imaginal discs from late third instar larvae carrying nord-GFP (Mi{PT-GFSTF.2}nordMI06414-GFSTF.2/+) were immunostained for GFP (green), Ptc (red), and Ci (blue). Maximum intensity z-projection and 3D reconstruction from a confocal image stack show nord expression in a representative wing imaginal disc. White brackets indicate the expression range of Ptc or Nord-GFP. (C, D) Wing imaginal discs from late third instar larvae carrying nord-RFP (Mi{PT-RFPHA.2}nordMI06414-RFPHA.2/+) and flip-out clones expressing the indicated UAS-transgenes were immunostained for HA (Nord-RFP, red), GFP (flip-out clones, green), and Ci (A compartment, blue). (D’–D”’) Zoomed view of the indicated area from panel (D). Note that ectopic nord-RFP is induced in UAS-SmoGlu-expressing clones located in the A compartment flanking the wing pouch (D’), but not in the P compartment (D”’). In the central wing pouch (D”), little (yellow star) or none (blue star) ectopic Nord-RFP was detected in SmoGlu-expressing flip-out clones. Dashed white lines indicate the clone boundary; dashed blue lines indicate the A/P compartment boundary, which is determined by the expression of endogenous Ci. Scale bar, 50 μm.

The selective upregulation of nord in the wing imaginal disc A cells adjacent to the A/P boundary indicated that nord expression may be controlled by Hh signaling activity.

We next examined this issue directly by following endogenous Nord expression in the protein-trap nord allele Mi{PT-GFSTF.2}nordMI06414-GFSTF.2 (hereafter referred to as nord-GFP). We used the heat shock (hs)-Gal4 driver (Halfon et al., 1997) either to activate Hh signaling by ectopically expressing UAS-Hh or inactivate Hh signaling activity by overexpressing UAS-Ptc. When compared to the wing discs carrying hs-Gal4 alone, Nord-GFP expression was expanded anteriorly following ectopic Hh expression by hs-Gal4 driving UAS-Hh (Figure 2—figure supplement 3A and B). Conversely, overexpression of Ptc, which keeps the Hh pathway silenced in the wing imaginal discs, dramatically reduced the Nord-GFP expression domain along the A/P boundary (Figure 2—figure supplement 3A and C).

To investigate whether Hh signaling activity can cell-autonomously induce the expression pattern of nord, we performed clonal analyses using the flip-out technique to ectopically activate the Hh signaling pathway (Ito et al., 1997). The expression of endogenous Nord was examined in the third instar wing discs carrying the protein-trap nord allele Mi{PT-RFPHA.2}nordMI06414-RFPHA.2 (hereafter referred to as nord-RFP) and flip-out clones expressing UAS-mCD8-GFP alone or in combination with a constitutively active form of Smo (UAS-SmoGlu) (Zhang et al., 2004; Figure 2C and D). We found that activation of Hh signaling via the constitutively active SmoGlu autonomously induced ectopic nord expression in the A compartment in clones that flanked the wing pouch (Figure 2D and D’). Consistent with the observation that endogenous nord expression is restricted to the A compartment and mostly absent from the center of the wing pouch (Figures 1F, 2B and C, Figure 2—figure supplement 3A), we noted that little or no ectopic Nord-RFP was detected in SmoGlu-expressing flip-out clones located in the central wing pouch region or those located in the P compartment, respectively (Figure 2D” and D”’). Furthermore, ectopic Hh, which is sufficient to activate the expression of the high-threshold target ptc, failed to induce nord expression in the center of the wing pouch (Figure 2—figure supplement 3B), indicating another mechanism excludes nord expression from this region. Nevertheless, these data demonstrated that nord expression in the hinge, the notum, and the edge of the wing pouch is regulated by Hh signaling activity. We thus identified nord as a novel target gene of the Drosophila Hh signaling pathway.

Nord belongs to an evolutionarily conserved family of secreted proteins

The Drosophila nord gene is associated with only one protein-coding transcript and one polypeptide containing 587 amino acids. It has a single homolog in the mouse genome, called NDNF (Kuang et al., 2010). Nord and NDNF belong to a family of evolutionarily conserved secreted proteins with a predicted signal peptide followed by one or two fibronectin type III-like repeats (FN3) and a domain of unknown function (pfam10179: DUF2369). Analyses of the genome and EST sequences from various organisms suggest that nearly all bilaterian animals have either single or multiple orthologous genes for Nord/Ndnf (Figure 3A). We cloned the full-length cDNAs from Drosophila Nord and NDNF from several different species into a mammalian expression vector and found that both Nord and various NDNF proteins were secreted into the medium after transient transfection into HEK 293T cells (Figure 3B). We immunostained unpermeabilized HEK 293T cells that were co-transfected with plasmids expressing Myc-tagged Nord or various NDNFs together with cytoplasmic-localized GFP proteins and detected a significant amount of Nord/NDNF proteins on the surface of the transfected cells (marked by cytoplasmic GFP expression), as well as the surrounding extracellular matrix within several cell diameters (Figure 3C, arrows). Consistent with the observations in cultured cells, when wing imaginal discs from third instar larvae carrying nord-GFP were immunostained in the absence of detergent, secreted Nord-GFP was noticed both flanking the A/P boundary and throughout the wing disc albeit at a much lower level (Figure 3D). Furthermore, when flip-out clones expressing an HA-tagged Nord were induced in the third instar wing imaginal discs, secreted Nord was detected outside of flip-out clones (Figure 3E). Together, these data demonstrated that Nord and its homolog NDNF belong to a family of secreted proteins, which likely exist in two spatially distinct pools: diffusible Nord/NDNF proteins that can reach a longer distance and membrane/matrix-associated Nord/NDNF proteins near the source cells.

Nord belongs to a family of secreted proteins.

(A) Phylogenetic analysis of Nord homologs from different species. The phylogenetic analysis of Nord homologs was performed with eggNOG (http://eggnogdb.embl.de/#/app/home). The phylogenetic tree is shown with domains predicted using Pfam (https://www.ebi.ac.uk/interpro/). The protein diagrams were then drawn proportional in length to the number of residues. Red stars (*) and red arrows (→) indicate the Ndnf/Nord proteins selected for further analysis. SP, signal peptide; FN3, fibronectin type 3 domain; DUF2369, domain of unknown function 2369. (B) Western blot analysis of HisMyc-tagged Nord or Ndnf protein in medium (M) and cell lysate (L) from transiently expressed in HEK-293 cells. Smeared band (*) indicates a portion of slow migrating Ndnf/Nord protein that may undergo extensive post-translational modification. Blue arrow (→) indicates Ndnf/Nord proteins that migrate as their predicted size. Hs, Homo sapiens; Dm, Drosophila melanogaster; Ce, Caenorhabditis elegans; Dr, Danio rerio; Xt, Xenopus tropicalis. (C) HEK-293 cells co-transfected with GFP and HisMyc-tagged Nord or Ndnf followed by cell-surface staining with anti-Myc antibody to reveal secreted Nord or Ndnf protein (yellow arrows). Note that secreted Nord-Myc and Ndnf-Myc (anti-Myc, red) were also detected several cell diameters away from the expressing cells (GFP, green). Scale bar, 10 μm. (D) Wing imaginal discs from third instar WT or nord-GFP (Mi{PT-GFSTF.2}nordMI06414-GFSTF.2/+) larvae were immunostained for GFP (red) without detergent treatment to label the secreted pool of Nord-GFP proteins. The total Nord-GFP proteins were detected by fluorescence from the GFP tag (green). Scale bar, 20 μm. (E) Wing imaginal discs from third instar larvae carrying flip-out clones expressing UAS-DsRed alone or in combination with UAS-Nord. The discs were immunostained in the absence of detergent to label the secreted pool of HA-tagged UAS-Nord (anti-HA, blue). Flip-out clones were detected by fluorescence from the UAS-DsRed transgene (red). Scale bar, 20 μm.

Nord is required for proper growth and crossvein patterning of the Drosophila wing

To investigate the function of Nord, we generated and characterized several nord loss-of-function alleles by using different genetic strategies. We first analyzed the phenotypes of two nord mutant alleles, nord3D and nord22A, created by the CRISPR/Cas9 strategy. Both alleles carry short deletions in the fourth coding exon that cause frameshifts and premature stop codons within the DUF2369 domain of Nord (Figure 4A, Figure 4—figure supplement 1). Animals homozygous or trans-heterozygous for nord22A and nord3D are viable and fertile. As homozygous or trans-heterozygotes, they show no obvious defects in the shape of the wing blade. However, a consistent and significant decrease in wing size was observed in both male and female flies (Figure 4—figure supplement 2). We next characterized the nord gene-trap allele Mi{MIC}nordMI06414 (hereafter referred to as nordMI06414) from the GDP collection (Venken et al., 2011), in which the transcription and translation of nord are interrupted after the first coding exon, resulting in a predicted truncated peptide containing only the first 103 amino acids without the conserved DUF2369 and FN3 domains (Figure 2—figure supplement 1, Figure 4A, Figure 4—figure supplement 1). Similar to the CRISPR-derived nord mutant alleles, the nordMI06414 homozygous flies are viable and fertile, and both male and female nord mutants showed a significant wing size reduction when compared to their wild-type counterpart (Figure 4C).

Figure 4 with 6 supplements see all
Nord is required for proper crossvein patterning and growth of the Drosophila wing.

(A) Upper panel: schematic diagram of the wild-type nord locus, the gene-trap, and the CRISPR alleles of nord. A Minos-Mediated Integration Cassette (MiMIC) cassette consisting of a splice acceptor site followed by stop codons in all three reading frames was inserted into the first coding intron of nord in the nord gene-trap allele nordMI0641. Detailed view of the deleted regions in the nord mutant alleles generated by the CRISPR/Cas9 system. Lower panel: schematic diagram of the human Neuron-Derived Neurotrophic Factor (NDNF) protein, Drosophila Nord protein, and the predicted polypeptide products from the indicated nord mutant alleles. (B) Adult wings obtained from male or female flies with indicated genotypes were pseudo-colored and overlapped to show the size difference. ACV, anterior crossvein; PCV, posterior crossvein; LV, longitudinal veins (L1–L5). (C) Quantification of wing size, distance between distal ends of LVs L2 and L5 (dL2–L5), L3 and L4 (dL3–L4), and the ratio of dL3–L4 /dL2–L5. Each bar shows the mean ± SD from n = 20 wings. All flies were grown at 25°C. One-way ANOVA followed by Sidak’s multiple comparison test or unpaired two-tailed t-test was used for statistical analysis. ***p<0.001, ****p<0.0001, ns, not significant; au, arbitrary units. (D) Adult wings of flies with the indicated genotypes. Yellow arrowhead indicates ectopic vein near posterior crossvein (PCV) or L5. (E) Quantification of the ectopic venation phenotype in adult wings from flies with the indicated genotypes. n > 50. Two-sided Fisher’s exact tests were used for statistical analysis. ***p<0.001, ****p<0.0001, ns, not significant. Scale bar, 500 μm.

Along with the growth defects in the wing, we also noted that 61% of females (59/97) and 21% (17/82) of males of the nord22A allele showed ectopic venation in the vicinity of the PCV of one or both wings (Figure 4D and E). The ectopic vein tissue was also found emanating from the PCV in the wing of adult flies homozygous for the nordMI06414 allele. Like nord22A, the ectopic venation displayed higher penetrance in females than in males, with 86% of mutant females and 36% of mutant males showing additional vein material at one or both wings (Figure 4D and E). This ectopic venation phenotype was not observed in nord3D homozygotes.

To further address whether the ectopic venation phenotype is caused by loss of nord, we performed genetic complementation tests between nordMI06414, nord3D, and nord22A with various deficiency lines covering the nord locus (Figure 4—figure supplement 3). All available deficiency lines in which the nord locus was entirely removed did not rescue the ectopic PCV phenotype of nordM106414, whereas adjacent deficiency lines with an intact nord locus fully rescued the crossvein defects in the nordMI06414 adult wings (Figure 4D and E, Supplementary file 2). We next tested Df(2R)BSC155 in trans to either nord22A and nord3D and once again found that nord22A exhibited high-frequency ectopic venation (Figure 4D and E; 69% females and 26% males), while nord3D showed a modest interaction (14% females and <1% males). Similar to the homozygotes, both nord22A and nord3D in trans to Df(2R)BSC155 also produced smaller wings compared to controls (Figure 4—figure supplement 4). To examine whether this size difference was the result of reduced cell proliferation or a smaller cell size, we examined wing trichome density as a proxy for cell size (Brummel et al., 1999; Dobzhansky, 1929; Martín-Castellanos and Edgar, 2002). We found no difference for nord MI06414/Df(2R)BSC155 or nord22A/Df(2R)BSC155 compared to heterozygous (nord MI06414/+ or nord22A/+) controls (Figure 4—figure supplements 5 and 6), suggesting that the alteration in wing size is caused by reduced cell proliferation. Together, these experiments indicate that Nord is required in wing imaginal discs for both proper growth and crossvein patterning and that the nord3D allele may retain some function.

Spatial-temporal overlapping expression of nord and dpp in the developing Drosophila wing

Both wing growth and crossvein patterning require precisely controlled BMP signaling activity (Affolter and Basler, 2007; De Celis, 2003). During wing development, BMP signaling activity is an output of the combined action of two BMP ligands, the Drosophila BMP2/4 homolog Dpp and the BMP5/6/7/8 homolog Gbb (Bangi and Wharton, 2006). Gbb is broadly and uniformly expressed in the larval and pupal wing, while Dpp, a well-known target of Hh signaling, is expressed in a stripe of cells in the anterior compartment along the A/P compartmental boundary of the larval wing imaginal disc. From this source, Dpp protein is thought to spread and form a concentration gradient to control the patterning and growth of the wing imaginal disc. In agreement with our finding that like dpp, nord is a target gene of the Hh signaling pathway in the wing imaginal discs (Figures 1 and 2), we observed a spatial-temporal correspondence between nord and dpp expression. Both Nord protein indicated by the Nord-GFP fusion derived from the nord-GFP protein-trap allele and Dpp precursor protein detected via an anti-Dpp prodomain antibody were present along the A/P boundary flanking the central wing pouch through the larval stage (Figure 5A). In the pupal wing, both Nord and Dpp expression does not change during the first 8 hr after pupation (AP) (Figure 5A). Subsequently, dpp disappears from the A/P boundary and commences expression in the differentiating LVs (deCelis, 1997; Yu et al., 1996); however, the Hh-dependent expression of nord along the A/P boundary remains for about 30 hr after pupation and diminishes after Dpp expression was detectable in both the LV and PCV regions (Figure 5A, Figure 5—figure supplement 1). Taken together, Nord proteins secreted from the A/P boundary stripe are expressed together or in close proximity with Dpp through the larval and early pupal (0–30 hr AP) stages.

Figure 5 with 4 supplements see all
Nord modulates Bone Morphogenetic Protein (BMP) signaling at or above the level of Mad phosphorylation.

(A) Expression of Nord in the wing discs through the third instar larval and early pupal stage. Upper panel: the third instar wing discs from nord-GFP (Mi{PT-GFSTF.2}nordMI06414-GFSTF.2/+) larvae were collected at indicated time points after egg laying (AEL) and were immunostained for GFP (Nord, green) and the pro-domain of Dpp (red). Bottom panel: pupal wings from pupae carrying both nord-GFP were collected at indicated time points after pupation (AP) and were immunostained for GFP (Nord, green) and the pro-domain of Dpp (red). Scale bar, 100 μm. (B) Wing imaginal discs from third instar wild-type and nord mutant (nordMI06414) larvae were immunostained for pMad (green) and Ptc (red). Scale bar, 50 μm. (C) Plotted pixel intensity of pMad or Ptc as a function of A/P position. Each point shows the mean ± SD. n = 15. au, arbitrary units. (D) Anti-pMad staining in wild-type and nordMI06414 pupal wings at indicated hours AP. Scale bar, 100 μm. (E) Quantification of the width (yellow dashed line in panel D) of pMad signal at primordial posterior crossvein (PCV) at indicated time points. Each bar shows the mean ± SD from n > 10 pupal wings. The unpaired two-tailed t-test was used for statistical analysis. ****p<0.0001.

Opposing effects of Nord in modulating BMP signaling activity during wing growth and PCV patterning

Both the wing growth and patterning defects observed in nord mutant animals and the overlapping expression patterns of Nord and the BMP ligand Dpp point to a possible role of Nord in mediating Dpp/BMP signal transduction (Figures 4 and 5A). We, therefore, asked whether elimination of Nord alters the level of phosphorylated Mad (pMad), the primary downstream signal transducer of BMP signaling in the wing disc. We quantified pMad signal intensity in nord mutant wing discs and compared it to that of wild-type controls. We found that the pMad intensity was slightly reduced in nord mutant wing discs, and this reduction of pMad levels was more evident in the A compartment where Nord is expressed (Figure 5B and C, arrow). In contrast, by measuring the expression of high-threshold target gene ptc and low-threshold gene dpp, we found no obvious difference in the Hh signaling activity in the wild-type and nord mutant discs (Figure 5B and C, Figure 5—figure supplement 2).

Of note, besides BMPs, Hh signaling also patterns the developing Drosophila wing. Hh short-range activity is responsible for patterning the central L3–L4 region and determining the distance between L3 and L4 LVs (Mullor et al., 1997; Strigini and Cohen, 1997). When quantifying the distance between distal ends of LVs L3 and L4 (dL3–L4) and that of L2 and L5 (dL2–L5), we found no specific reduction in the L3–L4 region indicated by comparable or higher dL3–L4/dL2–L5 ratio in the female and male nord mutant wings (Figure 4B and C), which is consistent with normal Hh signaling activity during the development of nord mutant wing imaginal discs (Figure 5B and C). Together, these findings suggested a positive role of endogenous Nord in augmenting BMP signaling activity to promote wing growth.

During pupal wing development, BMP signaling is activated in the prospective CV regions prior to the appearance of other known vein promoting signals (O’Connor et al., 2006; Ralston and Blair, 2005), and abnormal BMP signaling can selectively affect the PCV and leave the LVs largely or entirely intact (de Celis, 1997; Haerry et al., 1998; Khalsa et al., 1998; Nguyen et al., 1998; Ralston and Blair, 2005; Ray and Wharton, 2001; Wharton et al., 1999; Yu et al., 1996). Therefore, we assessed the possible role of Nord in Dpp/BMP signal transduction during crossvein patterning. It is known that pMad becomes gradually refined to a narrow strip of precursor cells that form the future PCV during the first 24–28 hr AP (Conley et al., 2000; Shimmi et al., 2005). To examine whether the ectopic PCV in nord mutants is a direct consequence of enhanced BMP signaling, we quantified pMad signal intensity at the presumptive PCV region in pupal wings at various time points after pupation. In contrast to the rather restricted pMad domain in wild-type pupal wings, we detected a broadened pMad domain around the presumptive PCV in nord mutants from 19 to 20 hr AP (Figure 5D and E). Gradually, the ectopic pMad accumulation became an expanded patch adjacent to the presumptive PCV, indicating that BMP signaling was abnormally elevated in this region (Figure 5D and E).

During early pupal stage, it is notable that Nord expression was seen neither in the LV nor primordia PCV (Figure 5A, Figure 5—figure supplement 1). In agreement, we did not notice any ectopic PCV in the adult flies when UAS-nord-RNAi was selectively expressed in the P compartment of the larval and pupal wing via the hh-Gal4 driver (Figure 5—figure supplements 3 and 4). Given that Nord is a secreted protein (Figure 3), Nord proteins secreted from the A/P boundary would likely play a role in reducing excessive BMP signaling in the L4–L5 intervein region to prevent ectopic venation. To test this possibility, we analyzed the ectopic venation phenotype in flies carrying ptc-Gal4-driven expression of UAS-nord-RNAi. Although ectopic venation phenotype was observed when nord was selectively knocked down in the Hh-responding cells (Figure 5—figure supplement 3), the frequency of flies carrying ectopic PCV was lower when compared to nord mutant flies (Figure 4D and E, Figure 5—figure supplement 4). This weaker phenotype likely resulted from residual Nord protein either due to incomplete ptc-Gal4>UAS-nord-RNAi-mediated nord knock down in the cells flanking the A/P boundary or due to possible Nord expression and secretion from other tissues. Nevertheless, opposite to the positive role of enhancing BMP signaling activity to promote growth of the larval wing discs, endogenous Nord also plays a negative role in inhibiting BMP signaling activity in the early pupal wing to prevent the formation of ectopic crossveins in the posterior compartment.

Nord is a biphasic modulator of BMP signaling during wing growth

To better understand the role of Nord in modulating BMP signaling in vivo, we generated transgenic flies carrying a Gal4-inducible UAS-Nord transgene. Under control of the ubiquitous wing blade driver nub-Gal4, ectopic Nord expression resulted in reduced range and level of the pMad gradient in the third instar wing imaginal discs (Figure 6A and B, 25°C), and accordingly decreased wing size in both adult males and females (Figure 6C and D, 25°C; Figure 6—figure supplement 1). In flies, minimal Gal4 activity is present at 16°C, while 29°C provides a balance between maximal Gal4 activity with minimal effects on fertility and viability due to growth at a high temperature (Duffy, 2002). Taking advantage of the temperature-dependent nature of Gal4 activity in Drosophila, we compared the dosage effect of Nord on pMad intensity in larvae raised at two different temperatures, 25 and 29°C. Indeed, in animals raised at 29°C expressing higher levels of exogenous Nord, we detected both a much reduced pMad gradient in the larval wing discs and more severely decreased wing size in the adult flies (Figure 6A–D, Figure 6—figure supplement 1). In contrast, Hh signaling activity is relatively normal in the wing discs expressing ectopic Nord. Although the Ptc-expressing domain became narrower, we did not notice any obvious decrease in the levels of Ptc expression. (Figure 6A and B). Consistently, ectopic Nord expression did not cause any specific reduction in the distance of L3 and L4 based on the dL3–L4/dL2–L5 ratio (Figure 6D), suggesting that the wing growth defect caused by ectopic Nord is unlikely due to inhibition of Hh signaling. Interestingly, partial L5 and PCV loss was also noticed in the wings with more dramatic size reduction (Figure 6C, arrowheads and arrows), and the frequency of disrupted L5 and PCV was more dramatic when the flies were raised at higher temperatures and expressed higher levels of ectopic Nord (Figure 6C). Therefore, our observations indicate that ectopic Nord attenuates BMP signaling, leading to inhibition of wing growth and vein patterning. Along with the positive role of endogenous Nord in enhancing BMP signaling to promote wing growth, we propose a model that Nord has both positive and negative effects in modulating BMP signaling activity, where low (endogenous) levels of Nord enhance and high (ectopic) levels of Nord inhibit BMP signaling.

Figure 6 with 1 supplement see all
Ectopic expression of Nord attenuates Bone Morphogenetic Protein (BMP) signaling in vivo.

(A) Larvae expressing the indicated transgenes driven by nub-Gal4 were raised at 25 or 29°C. Wing imaginal discs were collected from these larvae in the late third instar stage, and then immunostained for anti-pMad (red), anti-Ptc (blue), and anti-GFP (green). Scale bar, 20 μm. (B) Plotted pixel intensity of pMad or Ptc as a function of anteroposterior (A/P) position. Each point shows the mean ± SD. 25°C: n = 12. 29°C: n = 7. (C) Adult wings from male flies, which expressed the indicated transgenes driven by nub-Gal4 and were raised at 25 and 29°C. Scale bar, 500 μm. Arrows indicate reduced L5, and arrowheads indicate reduced crossveins. (D) Quantification of distance between distal ends of longitudinal veins L2 and L5 (dL2–L5), L3 and L4 (dL3–L4), and the ratio of dL3–L4 /dL2–L5. Each bar shows the mean ± SD from n > 14 wings. au, arbitrary units. The unpaired two-tailed t-test was used for statistical analysis. ****p<0.0001, ns, not significant.

Nord is a dosage-dependent modulator of BMP signaling during PCV patterning

To further examine the model that Nord is a dosage-dependent modulator of BMP signaling, we sought to manipulate Nord levels during pupal wing development in the posterior compartment and examine the effects on PCV formation since with this structure it is possible to assay both positive and negative roles by looking for an ectopic versus a loss of crossvein formation. Accordingly, we used both hh-Gal4 and en-Gal4 to drive different levels of ectopic Nord in the P compartment of the wing disc, including the PCV primordia where Nord is not normally expressed. To avoid the influence of prior larval stage Nord expression on the role of BMP signaling specifically during PCV pupal development, we used Gal4 together with tub-Gal80ts (a temperature-sensitive version of Gal80) to temporally control UAS-Nord expression. At a low temperature (18°C), Gal80ts represses the function of Gal4 bound to a UAS sequence but is unable to do so at the restrictive temperature (29°C) (McGuire et al., 2003). We performed temperature-shift experiments (from 18 to 29°C) to initiate ectopic Nord expression at different times during pupal development and characterized the impact on PCV patterning in the resulting adult wings. Consistent with a previous report (Roberts, 1998), the length of pupal period became shorter after the temperature was raised from 18 to 29°C due to temperature-dependent effects on the growth rate, and we found that shifting the temperature to 29°C right after pupation led to eclosion ~84–96 hr later (Figure 7A). We found that activation of UAS-Nord expression by temperature shifts at 78 hr or earlier before eclosion caused the most severe PCV phenotype, but activation at 66 hr before eclosion or later resulted in essentially normal PCV (Figure 7B and C, Figure 7—figure supplement 1). More importantly, we found that the abnormal PCV phenotypes relied on the levels of exogenous Nord: expression of a lower level of Nord (1XUAS-Nord) led to ectopic PCV, moderate levels (2XUAS-Nord) yielded a mixed phenotype with both ectopic and slightly reduced PCV, whereas high levels (3XUAS-Nord) gave rise to nearly complete loss of PCV (Figure 7B and C, Figure 7—figure supplement 1). Of note, consistent with the fact that Nord is a secreted protein, exogenous Nord expressed within the hh-Gal4 or en-Gal4-expressing domains was also able to influence crossvein patterning in the A compartment although with a lower frequency (Figure 7B, blue arrows and blue arrowheads, Figure 7—figure supplement 2).

Figure 7 with 2 supplements see all
Nord is a dosage-dependent modulator of the Bone Morphogenetic Protein (BMP) signaling.

(A) Upper row: schematic diagram of temporal UAS-Nord expression in the posterior compartment of wing discs using the temperature-sensitive Gal80 (Gal80ts) system. At the permissive temperature (18°C), Gal4 activity is blocked by Gal80ts. At the restrictive temperature (29°C), Gal80ts is unable to repress Gal4, which then induces expression of the UAS-transgenes. Middle and lower rows: timing of temperature shift. Embryos and larvae are grown at 18°C, and prepupae are transferred to 29°C at the indicated time points before eclosion (middle row) or before dissection (lower row). (B) Representative adult wings from flies that carry the indicated transgenes under the control of tub-Gal80ts together with hh-Gal4 or en-Gal4. The animals were grown at 18°C till prepupa stage, and then transferred to 29°C at 78 hr before eclosion. Yellow arrowhead indicates ectopic posterior crossvein (PCV); yellow arrow indicates reduced PCV; blue arrowhead indicates ectopic anterior crossvein (ACV); blue arrow indicates reduced ACV. Scale bar, 500 μm. (C) Quantification of PCV phenotypes in adult wings of female and male flies with indicated genotypes. The animals were grown at 18°C till prepupa stage, and then transferred to 29°C around 78, 72, or 66 (±1) hr before eclosion. n > 30 wings for each genotype at a given temperature shift time point. (D) Representative pupal wing from larvae that carry the indicated transgenes under the control of tub-Gal80ts together with en-Gal4. The animals were grown at 18°C till prepupa stage, and then transferred to 29°C at 12 hr before dissection. The collected pupal wings were immunostained for anti-pMad (red), anti-Ptc (blue), and anti-GFP (Nord-GFP, green). Yellow arrowhead indicates ectopic pMad around the primordial PCV; yellow arrow indicates reduced pMad signal around the primordial PCV. Scale bar, 100 μm.

The critical time window in which the primordial PCV responds to exogenously expressed Nord (66–78 hr before eclosion at 29°C) coincides with the stage during which BMP signaling induces PCV formation. We next tested whether the PCV phenotypes caused by ectopic Nord were correlated with alterations of BMP signaling activity. As shown in Figure 7D, adult PCV defects and pMad patterns in pupal wings showed a correlation with the level of ectopic Nord, where the temperature shift occurred 12 hr after the start of pupal development. In pupal wings expressing a low level of exogenous Nord (1XUAS-Nord), ectopic pMad and ectopic crossveins were detected, while moderate overexpression (2XUAS-Nord) led to a mixed phenotype of partial pMad and crossvein vein loss or wider pMad and ectopic veins. When high Nord levels (3XUAS-Nord) were expressed in the pupal wings, pMad and the PCV were largely absent. Although the temporal resolution is somewhat limited, the results clearly indicate that the level of Nord influences the outcome of PCV patterning during the early pupal development where a lower level of exogenous Nord resulted in enhanced BMP signaling and ectopic PCV, while higher levels of exogenous Nord inhibited BMP signaling and caused disrupted or depleted PCVs. Taken together, these results demonstrated that Nord is a dosage-dependent modulator of BMP signaling both in wing growth and crossvein patterning.

Nord binds to Dpp and interferes with BMP signaling in vitro

The activity of BMPs is modulated by a large variety of binding proteins that can either enhance or inhibit their signaling in a context-dependent manner (Chang, 2016; Umulis et al., 2009). Given the spatial-temporal overlapping expression of nord and dpp and the dosage-dependent modulation of Nord on BMP signaling in wing growth and crossvein patterning, we assessed whether Nord modulates BMP signaling via binding to either of the two BMP ligands, Dpp and Gbb, that are expressed in the developing Drosophila wing. We turned to an in vitro model to examine possible interactions of Nord with Dpp and Gbb by carrying out co-immunoprecipitation assays. We added a GFP tag to the C-terminus of Nord and expressed the fusion protein in Drosophila S2 cells. The conditioned medium from Nord-GFP-expressing cells was collected and mixed with medium from cells transfected with FLAG-tagged Dpp and HA-tagged Gbb alone or in combination. The mixed-media were then incubated with anti-FLAG or anti-HA antibody-coupled beads to precipitate the BMP ligands. We found that Nord co-precipitated with Dpp and, to a lesser extent, with Gbb (Figure 8A). Additionally, we observed an increased level of Nord proteins co-precipitated with Dpp or Gbb when Dpp and Gbb were co-expressed (Figure 8A), indicating that Nord may have a higher affinity for Dpp-Gbb heterodimers formed in cells co-expressing Dpp and Gbb.

Nord binds to Decapentaplegic (Dpp) and attenuates Bone Morphogenetic Protein (BMP) signaling in vitro.

(A) Co-immunoprecipitation of Nord with the BMP ligands Dpp and Glass-bottom boat (Gbb). Medium from S2 cells transfected for expression of GFP-tagged Nord were mixed with medium from cells expressing FLAG-tagged Dpp and HA-tagged Gbb alone or in combination, followed by incubation with anti-FLAG or anti-HA antibody-coupled beads overnight at 4°C. Precipitated proteins were analyzed by Western blotting with indicated antibodies. Nord was immunoprecipitated with Dpp and, to a lesser extent, with Gbb. The amount of immunoprecipitated Nord was increased when Dpp and Gbb were co-transfected. (B) S2 cells were transfected for expression of FLAG-tagged Dpp or Gbb with or without GFP-tagged Nord (source cell). Both cell lysate and conditioned medium from the source cells were collected and followed by Western blot analysis. Loading was controlled by probing the blot for tubulin. The amount of Dpp or Dpp-Gbb ligands released into the medium was reduced when Nord was co-expressed in the source cells. (C) Comparison of BMP signaling activities of conditioned media in a cell-based signaling assay. After incubating the conditioned media collected in (B) with S2 cells stably expressing the FLAG-Mad transgene (Mad-S2, responding cell) for 1 hr at room temperature, the responding cells were washed and lysed. The lysates were probed with anti-pMad and anti-FLAG antibodies to detect both the phosphorylated Mad and total Mad protein, respectively. (D, E) Quantification of the Western blot data in panels (B) and (C). The levels of the secreted BMP ligands from the medium (anti-FLAG in panel B), phosphorylated Mad (anti-pMad in panel C), and total Mad (anti-FLAG in panel C) were measured based on the band intensity. The signaling activity from an equal volume of conditioned medium was determined by the ratio of pMad and the corresponding total Mad (pMad/Mad) in panel (C), which was then normalized to the condition without the exogenous Nord-GFP to calculate a relative signaling activity (D), or normalized to the secreted ligand amount to calculate the relative ligand activity ([pMad/Mad]/BMPs) (E). (F) Mad-S2 cells were treated with a recombinant Dpp peptide (rDpp) in the absence or presence of conditioned medium containing raising levels of Nord for 1 hr at room temperature, the responding cells were washed and lysed. The lysates were probed with anti-pMad and anti-FLAG antibodies to detect both the phosphorylated Mad and total Mad protein, respectively. (G) Quantification of the Western blot data in panel (F). The phosphorylated Mad (anti-pMad) and total Mad (anti-FLAG) levels were measured based on the band intensity. The signaling activity from each conditioned medium was determined by the ratio of pMad and the corresponding total Mad (pMad/Mad), which was then normalized to the condition with 2 nM Dpp but no additional Nord. Panel (F) is representative of n = 3 independent experiments. Each point shows the mean ± SD. One-way ANOVA test with Tukey’s multiple comparison was used for statistical analysis, and a significant difference was considered by *p<0.05. (H) Mad-S2 cells were transiently transfected for expression of GFP or Nord-GFP. 48 hr after transfection, the cells were treated with recombinant Dpp peptides for 1 hr. Upon treatment, the cells were washed, fixed, and stained by anti-FLAG to detect total Mad and anti-pMad to detect phosphorylated Mad. The average pMad levels were measured and normalized to the total Mad levels, and then plotted against different Dpp concentrations (0–5 nM). Each point shows the mean ± SD, n > 10. au, arbitrary units. The unpaired two-tailed t-test was used for statistical analysis. ***p<0.001, ****p<0.0001. (I) Representative images of Mad-S2 cells treated by 0.625 nM Dpp. Scale bar, 10 μm.

We next collected conditioned medium from S2 cells expressing Dpp or Dpp-Gbb with or without co-expressed Nord (source cell) and compared the levels of processed Dpp or Dpp-Gbb within the different conditioned media. We found that far less cleaved Dpp protein was released into the conditioned medium when Nord was co-expressed (Figure 8B, compare lane M3 with M2). Likewise, the same negative effect of Nord was observed when the medium was collected from cells expressing both Dpp and Gbb (Figure 8B, compare lane M5 with M4), albeit the increased total amount of ligands likely reflects the release of Dpp-Gbb heterodimers (Figure 8B, compare lane M4 with M2). Thus, the presence of Nord affected the release into the media of both Dpp and Dpp-Gbb, likely via binding to the BMP ligand. We then determined the activity of the collected conditioned media in an S2 cell-based signaling assay. Because endogenous levels of Mad protein in S2 cells are low, we established Mad-S2 responding cells that stably express a FLAG epitope-tagged Mad transgene (FLAG-Mad) (Ross et al., 2001). Upon incubation of the Mad-S2 responding cells with conditioned medium collected from the source cells, BMP signaling activity was monitored by measuring the pMad signal intensity. Conditioned medium containing either Dpp or Dpp-Gbb, but not that containing Nord alone, was able to induce Mad phosphorylation (Figure 8C). In agreement with the dramatically reduced amount of Dpp or Dpp-Gbb ligands (Figure 8B), the conditioned medium collected from source cells co-expressing Nord showed lower pMad signal intensity compared to that collected from source cells lacking Nord co-expression (Figure 8C and D). Of note, Nord is not a membrane-tethered protein (Figure 3A), but we observed noticeable amounts of Nord deposited on the surface of source cells, as well as the surrounding extracellular matrix (Figure 3C and D). Thus, the matrix-associated Nord may sequester Dpp and Dpp-Gbb ligands to the source cells and thereby reduce ligand level in the media, which in turn leads to decreased BMP signaling activity in the responding cells (Figure 8C and D).

Dosage-dependent modulation of BMP signaling by Nord in vitro

In vivo, our loss- and gain-of-function analyses suggested that Nord is a dosage-dependent modulator of BMP signaling in wing growth and crossvein patterning. In the in vitro signaling assay, ectopically expressing Nord in the source cells led to reduced BMP ligand release and thus decreased the signaling activity (pMad/Mad) from an equal volume of conditioned medium (Figure 8D). However, when the signaling activity was further normalized to the ligand amount (Figure 8B), the relative ligand activity ([pMad/Mad]/BMPs) from the medium without Nord is lower than that with Nord (Figure 8E), suggesting that while ectopic Nord expressed in the source cells reduced the levels of released ligand, that soluble ligand appears to have a higher signaling activity perhaps as a result of an association with Nord (Figure 8E). To further assess whether Nord directly modulates the signaling activity of released BMP ligands in a dosage-dependent manner, we treated Mad-S2 cells with a recombinant Dpp peptide (rDpp) in the absence or presence of conditioned medium containing increasing amounts of Nord (Figure 8F). Consistent with our in vivo analyses, in vitro, we also observed a dosage-dependent signaling profile of rDpp in the presence of increasing concentrations of Nord where the lowest tested level of Nord enhanced signaling while higher Nord levels reduced Dpp signaling activity (Figure 8F and G).

Notably, the highest level of Nord supplied from conditional medium failed to completely inhibit the free rDpp-induced BMP signaling. We next transiently transfected Mad-S2 cells for expression of Nord-GFP. Similar to Nord supplied from conditioned medium, we found that transient expression of Nord-GFP in the Mad-S2 cells partially inhibited the average Mad phosphorylation induced by exogenous rDpp (Figure 8H). Remarkably, immunofluorescent staining revealed that cells expressing Nord-GFP exhibited much lower pMad when compared to the surrounding Mad-S2 cells lacking Nord-GFP expression (Figure 8I). In contrast, cells expressing GFP did not cause a noticeable reduction in pMad level in rDpp-treated Mad-S2 cells (Figure 8I). The much stronger cell-autonomous inhibition of Nord on BMP signaling is likely due to a much higher level of ectopic Nord when overexpressed in the Mad-S2 cells (Figure 8I). Taken together, both our in vivo and in vitro assays demonstrated that Nord not only can sequester BMP ligands, thereby impeding their release from the source cells, but may also directly modulate the activity of released BMP ligands in a dosage-dependent manner, where low levels promote and high doses attenuate BMP signaling.

Discussion

In Drosophila, the short-range morphogen Hh and the long-range morphogen BMP function together to organize wing patterning (Lecuit et al., 1996; Mullor et al., 1997; Nellen et al., 1996; Strigini and Cohen, 1997; Zecca et al., 1995). It has been previously shown that the Hh signal shapes the activity gradient of BMP by both inducing the expression of Dpp and simultaneously downregulating the Dpp receptor Tkv, resulting in lower responsiveness to Dpp in cells at the A/P compartment border (Tanimoto et al., 2000). In this study, we showed that the activity of BMP is further fine-tuned by another previously unknown Hh-dependent mechanism. Using a genome-wide expression profiling of the Drosophila wing imaginal discs, we identified nord as a novel target gene of the Hh signaling pathway (Figures 1 and 2). Nord and its homolog NDNF belong to a family of secreted proteins that can exist in two distinct pools: diffusible Nord/NDNF proteins that can reach a longer distance and membrane/matrix-associated Nord/NDNF proteins spreading within a short distance from the source cells (Figure 3). During larval and early pupal wing development, Nord is expressed together or in close proximity with the BMP ligand Dpp along the A/P compartment boundary (Figure 5A, Figure 5—figure supplement 1). Elimination of nord caused a reduction of overall wing size and resulted in ectopic PCV formation. Both of these phenotypes are attributable to alterations of BMP signaling activity as monitored by the level of Mad phosphorylation, yet in opposite directions: loss of nord led to decreased pMad in larval wing discs, whereas ectopic pMad surrounded the primordial PCV in nord mutant pupal wings (Figures 4 and 5). Moreover, expressing exogenous Nord at different levels and during different developmental stages and contexts showed that Nord is a dosage-dependent modulator of BMP signaling both in wing growth and crossvein patterning (Figures 6 and 7). At the molecular level, we further demonstrated that Nord is a BMP-binding protein that directly enhances or inhibits BMP signaling in cultured S2 cells (Figure 8).

Combining the genetic and biochemical evidence, we propose that Nord mediates BMP signaling activity through binding of the BMP ligands Dpp and Dpp-Gbb (Figure 9). Depending on the levels of Nord proteins and the source/types of BMP ligands, Nord-mediated binding of Dpp and Dpp-Gbb may promote or repress BMP signaling activity. Additionally, the existence of two spatially distinct pools of diffusible and membrane/matrix-associated Nord proteins may introduce further complications in Nord-mediated BMP signaling regulation. In the wild-type wing discs, expressed in a subset of Dpp-secreting cells along the A/P boundary, Nord binds and enhances the local BMP signaling activity by augmenting ligand concentration near the Nord/Dpp-secreting cells. Meanwhile, Nord also impedes the mobilization of Dpp, especially the long-range BMP signaling mediator Dpp-Gbb heterodimer (Figure 9A). Loss of nord simultaneously led to reduced local BMP and increased long-range BMP activities, and therefore gave rise to the seemingly opposite phenotypes of reduced wing size and ectopic PCV (Figure 9B). In contrast, low levels of ectopic Nord in the P compartment autonomously increased BMP signaling activity (Figure 9C), whereas high levels of Nord, either in the P compartment or throughout the wing pouch, inhibited BMP signaling activity likely through interfering with the normal BMP reception (Figure 9D and E). Taken together, we propose that Hh-induced Nord expression provides an exquisite regulation of the strength and range of BMP signaling in the developing Drosophila wing.

A model for dosage-dependent modulation of Bone Morphogenetic Protein (BMP) signaling by the Hedgehog (Hh) target gene nord.

Illustration of potential concentration- and location-dependent mechanism of Nord as a dosage-dependent BMP signaling modulator. The source cells of Nord (and Decapentaplegic [Dpp]) are labeled in blue, and the responding cells in focus are indicated by the dashed line. Note that Nord could be either diffusible or membrane/matrix-associated. (A, A’) In the wild-type wing discs, nord is induced by the Hh signal along the anteroposterior (A/P) compartment boundary flanking the central wing pouch, and thus is expressed in a subset of Dpp-secreting cells. We propose that membrane/matrix-associated Nord-mediated binding of Dpp and Dpp-Gbb both enhances the local BMP signaling activity by augmenting ligand concentration near the Nord/Dpp-secreting cells and impedes the mobilization of Dpp, especially the long-range BMP signaling mediator Dpp-Gbb heterodimer. The diffusible pool of Nord is more likely to interfere with BMP reception. (B, B’) Loss of nord simultaneously leads to reduced local BMP and increased long-range BMP activities, and therefore gives rise to the seemingly opposite phenotypes of reduced wing size and ectopic posterior crossvein (PCV), both of which are attributable to alteration of BMP activity at the level of Mad phosphorylation (pMad). (C–E, C’-E’) Nord misexpression experiments. Low levels of ectopic Nord in the P compartment increase BMP signaling activity (C, C’), whereas high levels of Nord, either in the P compartment (D, D’) or throughout the wing pouch (E, E’), inhibit BMP signaling activity. The altered BMP signaling activities are reflected by the pMad levels.

Nord is a novel, multifunctional BMP-binding protein

The activity of TGF-β type factors, including the BMP subfamily, is modulated by a large variety of binding proteins that can either enhance or inhibit their signaling in a context-dependent manner (Chang, 2016; Umulis et al., 2009). These modulator proteins vary broadly in structure, location, and mechanism of action. Well-known extracellular and freely diffusible proteins include Noggin, Tsg, Follistatin, the CR (cysteine-rich) domain containing proteins such as Chordin/Sog, and the Can family named after two founding members, Dan and Cerberus (Chang, 2016). With the exception of Tsg and Tsg/Sog or Tsg/Chordin complexes that in some cases can promote BMP signaling, all of these factors behave as antagonists, where BMP binding prevents association of the ligand with the receptor complex.

The other broad category of BMP-binding proteins includes membrane-bound or matrix-associated proteins and, in contrast to the highly diffusible class of BMP-binding factors, these proteins often act as either agonists or antagonists depending on context. These proteins are also structurally diverse, but to date, none contain FN3 or DUF2369 domains that are characteristic of Nord and NDNF, its vertebrate counterpart. From a mechanistic point of view, perhaps the two most instructive Drosophila members of this class of modulators are the heparan sulfate proteoglycan (HSPG) Dally and the CR-containing protein Cv-2. HSPGs are well characterized as modulators of growth factor signaling (Nakato and Li, 2016). In the case of FGFs, HSPGs act as true co-receptors in which they form a tripartite complex with ligand and FGFR, the signaling receptor (DiGabriele et al., 1998; Schlessinger et al., 2000). However, they can also mediate signaling in other ways. Analysis of dally loss-of-function clones in imaginal discs demonstrates that it has both cell-autonomous and non-autonomous effects with respect to BMP signaling (Akiyama et al., 2008; Belenkaya et al., 2004; Fujise et al., 2003). In general, low levels tend to promote signaling while high doses attenuate signaling. Many models have been put forth to explain these opposing effects and often come down to balancing ligand sequestration and diffusion properties. For instance, in the absence of HSPGs, Dpp may more freely diffuse away from the disc epithelial cell surface. In this case, HSPG acts to enhance signaling by keeping Dpp tethered to the cell surface where it can engage its signaling receptors. On the other hand, a high level of HSPG may compete with signaling receptors for BMP binding and thereby reduce signal (Nakato and Li, 2016).

The situation with respect to signal modulation becomes even more complex for factors such as Nord that bind both HSPGs (Akiyama et al., 2021) and BMPs (this report). An instructive example to consider is Cv-2, a secreted factor that, like Nord, binds both to HSPGs and BMPs and is also induced by BMP signaling (Serpe et al., 2008). Like Dally, Cv-2 also has dose-dependent effects on signaling in wing imaginal discs, where low levels enhance while high levels inhibit BMP signaling. By virtue of being bound to HSPGs, it may simply function as an additional tethering molecule that keeps BMPs localized near the cell surface. However, Cv-2 has the unique property that it is also able to bind Tkv, a Drosophila BMPR type I receptor (Serpe et al., 2008). This has led to speculation that it could act as an exchange factor that aids in handing off a BMP ligand from the HSPG pool to the type I receptor. Mathematical modeling showed that this mechanism can produce a biphasic signal depending on affinities of the various BMP-binding proteins involved and their concentrations (Serpe et al., 2008).

In the case of Nord, its mechanism of action is likely compatible with a variety of these and/or alternative models. While we have shown that Nord is a BMP-binding protein and Akiyama et al., 2021 have shown that it also binds HSPGs, it is not clear whether the BMP and HSPG-binding sites overlap or are distinct and where they are positioned relative to the FN3 and DUF2369 domains. This is an important issue to consider with respect to the two CRISPR mutants that we generated that truncate Nord within the DUF2369 domain. Interestingly, the nord3D allele appears to retain some function since it does not generate ectopic crossveins as do the nordMI06414 or nord22A alleles, yet nord3D still produces small wings in transheterozygous combination with a deficiency or nord22A, consistent with having lost the BMP growth-promoting ability. The discrepancy in crossvein patterning between the different nord alleles may be explained by a difference in residual function of the various truncated Nord protein products (Figure 4). Because the nordMI06414 allele yields a much shorter predicted Nord peptide compared to the two CRISPR alleles, it is likely to behave as a protein null with a stronger phenotype. The two nord CRISPR alleles, although similar in the sequence deleted from the C-terminus, differ in how many non-nord encoded amino acids occur between the frameshift and the stop codon. The nord22A allele has additional 14 amino acids relative to nord3D. Perhaps this extension of the truncated fragment destabilizes or interferes with residual function found in the nord3D allele. Additional biochemical studies defining the BMP and HSPG-binding sites, the stability of truncated Nord fragments, and whether Nord can also associate with either the type I or II receptors will aid in formulating a more precise mechanistic model.

Is Nord structure and function conserved across species?

Nord shows some sequence similarity to the NDNF family of proteins (Figure 3A). Based on a very recent study, like many other neurotrophic factors, NDNF arose in the ancestor of bilaterians or even later (Heger et al., 2020). In agreement, by analyzing the genome and EST sequences from various organisms, we found that nearly all bilaterian animals have either single or multiple orthologous genes for Nord/Ndnf (Figure 3A). Of note, we did not identify any Ndnf homologs in the flatworm Planarian, but these factors are highly conserved across vertebrates (Kuang et al., 2010). All vertebrate family members contain a signal peptide, two FN3-like repeats, and a domain of unknown function (DUF2369) that is now referred to as the NDNF domain. The NDNF domain partially overlaps with the first FN3 but shows some additional conservation that extends between the two FN3 domains (Figure 3A). The FN3 module is quite diverse in sequence but is thought to exhibit a common fold that is used as an interaction surface or spacer (Campbell and Spitzfaden, 1994; Koide et al., 1998). The function of the NDNF domain is not clear, but it may also provide a protein interaction surface.

Although the vertebrate NDNFs are highly conserved throughout the entire protein length, the Caenorhabditis elegans and Drosophila relatives are quite divergent in primary sequence and show little conservation beyond a few key residues that define the second FN3 and NDNF domains (Kuang et al., 2010). Notably, the Drosophila protein is missing the first FN3 domain, and therefore it is not clear the extent to which Nord and the vertebrate NDNFs may exhibit functional conservation. Ironically, the original human NDNF clone was identified on the basis of domain structure conservation with Drosophila Nord, which was identified via enhancer trapping to be a gene expressed in mushroom bodies and whose loss leads to defects in olfactory learning and memory (Dubnau et al., 2003). Unfortunately, that particular LacZ enhancer trap line that disrupted the nord locus is no longer available. The use of our new alleles should prove helpful for either confirming or eliminating the involvement of Nord as a modulator of learning and memory and/or other neuronal functions in larva and adult Drosophila.

In the mouse, NDNF is highly expressed in many neurons of the brain and spinal cord (Boyle et al., 2011; Kuang et al., 2010; Schuman et al., 2019). Studies using cultured mouse hippocampal neurons revealed that it promotes neuron migration and neurite outgrowth, hence its name (Kuang et al., 2010). In later studies, NDNF was also found to be upregulated in mouse endothelial cells in response to hindlimb ischemia, where it promotes endothelial cell and cardiomyocyte survival through integrin-mediated activation of AKT/endothelial NOS signaling (Ohashi et al., 2014; Joki et al., 2015). Additionally, recent studies have shown that NDNF expression is significantly downregulated in human lung adenocarcinoma (LUAD) and renal cell carcinoma (RCC), indicating that NDNF may also provide a beneficial function as a tumor suppressor (Xia et al., 2019; Zhang et al., 2019).

Taken together, these studies have suggested some possible functions for vertebrate NDNF. However, they have primarily relied on in vitro cell culture models, and only recently have in vivo loss-of-function studies been reported (Messina et al., 2020). Remarkably, NDNF mutants were discovered in the genomes of several probands with congenital hypogonadotropic hypogonadism (CHH), a rare genetic disorder that is characterized by absence of puberty, infertility, and anosmia (loss of smell) (Boehm et al., 2015; Lima Amato et al., 2017). This phenotype is very similar to that produced by loss of the anos1, which also encodes an FN3 superfamily member and is responsible for Kallmann syndrome, a condition that similarly presents with CHH and anosmia due to lack of proper GnRH and olfactory neuron migration (Stamou and Georgopoulos, 2018). Although in vitro studies indicated that NDNF modulates FGFR1 signaling after FGF8 stimulation, the in vivo molecular mechanism responsible for the neuronal migration defects is not clear (Messina et al., 2020). The results of our study on the function of Drosophila Nord raise the issue of whether any of the ascribed vertebrate NDNF functions could involve alterations in BMP signaling. In the case of angiogenesis and EMT, BMPs, as well as other TGF-β family members, participate at many levels (Goumans et al., 2018; Kahata et al., 2018). At present, however, no involvement of BMP or TGF-β signaling has been implicated in migration of the GnRH neurons, although BMP signaling does define neurogenic permissive areas in which the olfactory placode forms (Forni and Wray, 2015). A clear objective for the future is to determine if the vertebrate NDNF factors bind BMPs and/or HSPG proteins such as Dally-like glypicans to modulate BMP signaling activity. On the Drosophila side, additional non-BMP-modulating roles for Nord should also be examined.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Genetic reagent (Drosophila melanogaster)hs-FLPGolic and Lindquist, 1989FBti0000785Chr X
Genetic reagent (D. melanogaster)Actin>y+>Gal4Ito et al., 1997FBti0009983Chr 2
Genetic reagent (D. melanogaster)ptc-Gal4Hinz et al., 1994FBal0287777Chr 2
Genetic reagent (D. melanogaster)hh-Gal4BDSCRRID:BDSC_67046Chr 3
Genetic reagent (D. melanogaster)en-Gal4BDSCRRID:BDSC_30564Chr 2
Genetic reagent (D. melanogaster)nub-Gal4BDSCRRID:BDSC_25754Chr 2
Genetic reagent (D. melanogaster)MS1096-Gal4BDSCRRID:BDSC_8860Chr X
Genetic reagent (D. melanogaster)MS1096-Gal4; UAS-Dcr-2BDSCRRID:BDSC_25706Chr X; Chr 2
Genetic reagent (D. melanogaster)A9-Gal4BDSCRRID:BDSC_8761Chr X
Genetic reagent (D. melanogaster)hs-Gal4Halfon et al., 1997FBtp0065595Chr 2
Genetic reagent (D. melanogaster)tub-Gal80tsBDSCRRID:BDSC_7019Chr 2
Genetic reagent (D. melanogaster)tub-Gal80tsBDSCRRID:BDSC_7018Chr 3
Genetic reagent (D. melanogaster)UAS-Dicer-2BDSCRRID:BDSC_24650Chr 2
Genetic reagent (D. melanogaster)UAS-Dicer-2BDSCRRID:BDSC_24651Chr 3
Genetic reagent (D. melanogaster)UAS-GFPBDSCRRID:BDSC_1521Chr 2
Genetic reagent (D. melanogaster)UAS-GFPBDSCRRID:BDSC_1522Chr 3
Genetic reagent (D. melanogaster)UAS-mCD8-GFPBDSCRRID:BDSC_5137Chr 2
Genetic reagent (D. melanogaster)UAS-DsRedBDSCRRID:BDSC_6282Chr 3
Genetic reagent (D. melanogaster)UAS-PtcJohnson et al., 1995Chr 3
Genetic reagent (D. melanogaster)UAS-HhLee et al., 1992Chr 3
Genetic reagent (D. melanogaster)UAS-SmoGluZhang et al., 2004Chr 3
Genetic reagent (D. melanogaster)UAS-nord-RNAi-1VDRCRRID:VDRC_v38151Chr 2
Genetic reagent (D. melanogaster)UAS-nord-RNAi-2VDRCRRID:VDRC_v38152Chr 2
Genetic reagent (D. melanogaster)w[1,118]; Df(2R)BSC770/SM6aBDSCRRID:BDSC_26867Chr 2
Genetic reagent (D. melanogaster)[1,118]; Df(2 R)BSC356/SM6aBDSCRRID:BDSC_24380Chr 2
Genetic reagent (D. melanogaster)y(1) w[1,118]; Df(2R)BSC155/CyO-Df(2R)B80, y[+]BDSCRRID:BDSC_9691Chr 2
Genetic reagent (D. melanogaster)w[1,118]; Df(2R)BSC780/SM6aBDSCRRID:BDSC_27352Chr 2
Genetic reagent (D. melanogaster)w[1,118]; Df(2R)BSC603/SM6aBDSCRRID:BDSC_25436Chr 2
Genetic reagent (D. melanogaster)w[1,118]; Df(2R)ED4061,
P{w[+ mW.Scer\FRT.hs3] = 3'.
RS5 + 3.3'}ED4061/SM6a
BDSCRRID:BDSC_9068Chr 2
Genetic reagent (D. melanogaster)w1118; PBac{y[+ mDint2]=vas-Cas9}VK00027BDSCRRID:BDSC_51324Chr 3
Genetic reagent (D. melanogaster)y wBDSCRRID:BDSC_1495Chr X
Genetic reagent (D. melanogaster)y w; Mi{MIC}nordMI06414BDSCRRID:BDSC_42389Chr 2
Genetic reagent (D. melanogaster)y w; Mi{PT-GFSTF.2}nordMI06414-GFSTF.2/CyOBDSCRRID:BDSC_60250Chr 2
Genetic reagent (D. melanogaster)w1118; PBac{y[+ mDint2]=vas-Cas9}VK00027BDSCRRID:BDSC_51324Chr 3
Genetic reagent (D. melanogaster)y w; Mi{PT-RFPHA.2}nordMI06414-RFPHA.2This paperChr 2
Genetic reagent (D. melanogaster)UAS-Nord-HA-GFP (Chr.2)This paperChr 2
Genetic reagent (D. melanogaster)UAS-Nord-HA-GFP (Chr.3)This paperChr 3
Genetic reagent (D. melanogaster)w; nord22AThis paperChr 2
Genetic reagent (D. melanogaster)w; nord3DThis paperChr 2
AntibodyAnti-Ci (rat monoclonal)DSHBCat# 2A1; RRID:AB_2109711IF: (1:50)
AntibodyAnti-Ptc (mouse monoclonal)DSHBCat# Apa 1; RRID:AB_528441IF: (1:50)
AntibodyAnti-Dpp prodomain (rabbit polyclonal)Akiyama and Gibson, 2015A gift from M. GibsonIF: (1:100)
AntibodyAnti-GFP (rabbit polyclonal)Molecular ProbesCat# A-11122, RRID:AB_221569IF: (1:2000)
AntibodyAnti-GFP (chicken polyclonal)AbcamCat# ab13970, RRID:AB_300798IF: (1:2000)
AntibodyAnti-beta tubulin (mouse monoclonal)DSHBCat#E7; RRID:AB_2315513WB: (1:5000)
AntibodyAnti-HA.11 (mouse monoclonal, 16B12)CovanceCat# MMS-101P-1000; RRID: AB_291259IF: (1:1000)
AntibodyAnti-HA (rabbit polyclonal, SG77)Thermo Fisher ScientificCat# 71-5500; RRID:AB_2533988WB: (1:1000)
AntibodyAnti-FLAG (mouse monoclonal, M2)Sigma-AldrichCat# F3165; RRID:AB_259529WB: (1:200)
AntibodyAnti-FLAG (rabbit polyclonal)Sigma-AldrichCat# F7425; RRID:AB_439687WB: (1:200)
AntibodyAnti-Myc (mouse monoclonal)Santa Cruz BiotechnologyCat# sc-40; RRID:AB_627268WB: (1:200)
AntibodyAnti-anti-phospho-Smad1/5
(rabbit polyclonal, Ser463/465)
Cell Signaling TechnologyCat# 9516; RRID:AB_491015IF: (1:100)
AntibodyAnti-FLAG
(mouse monoclonal, M2) Magnetic Beads
Sigma-AldrichCat# M8823; RRID:AB_2637089IP
AntibodyAnti-HA (rat monoclonal, 3F10) Affinity MatrixRocheCat# 11815016001;RRID:AB_390914IP
AntibodyFluorophore-conjugated secondary antibodiesJackson ImmunoResearch LabNAIF: (1:500)
AntibodyHRP-conjugated secondary antibodiesJackson ImmunoResearch LabNAWB: (1:10,000)
OtherDAPIMilliporeSigmaCat# D9542(1 µg/mL)
OtherFetal bovine serumOmega ScientificCat# FB-02
OtherSchneider’s Drosophila MediaInvitrogenCat# 21720
OtherDulbecco’s Modification of Eagle’s Medium (DMEM)CorningCat# 10-013CM
OtherBlasticidin S HClThermo Fisher ScientificCat# R21001
OtherPenicillin-Streptomycin-Glutamine (100×)Thermo Fisher ScientificCat# 10378016
OtherAntifade mounting mediaVECTASHIELDCat# H-1000
OtherFuGENE HD transfection reagentPromegaCat# E2311
Other16% paraformaldehyde aqueous solutionElectron Microscopy SciencesCat# 15710
OtherDissociation bufferSigmaCat# C-1544
OtherRecombinant DppR&D SystemsCat# 59-DP-020
OtherElastaseSigmaCat# E-0258
OtherPropidium iodideInvitrogenCat# P3566
OtherDNAse IThermo Fisher ScientificCat# AM2222
OtherMaxima Reverse TranscriptaseThermo Fisher ScientificCat# EP0742
OtherSYBR Green SupermixBio-RadCat# 1708880
OtherRNeasy MiniQIAGENCat# 74104
Recombinant DNA reagentpAcSV-Nord-GFPSupplementary file 3Nord coding sequence (NM_138056) was fused in frame with
C-terminal GFP tag and cloned into pAcSV vector
Recombinant DNA reagentpcDNA3.1-Nord-HisMycSupplementary file 3Nord coding sequence (NM_138056) was fused in frame with
C-terminal HisMyc tag and cloned into pcDNA3.1 vector
Recombinant DNA reagentpcDNA3.1-HsNDNF-HisMycSupplementary file 3HsNDNF coding sequence (NM_024574) was fused in frame with
C-terminal HisMyc tag and cloned into pcDNA3.1 vector
Recombinant DNA reagentpcDNA3.1-CeNdnf-HisMycSupplementary file 3CeNdnf coding sequence (NM_067881) was fused in frame with
C-terminal HisMyc tag and cloned into pcDNA3.1 vector
Recombinant DNA reagentpcDNA3.1-DrNdnf-HisMycSupplementary file 3DrNdnf coding sequence (XM_684842) was fused in frame with
C-terminal HisMyc tag and cloned into pcDNA3.1 vector
Recombinant DNA reagentpcDNA3.1-XtNdnf-HisMycSupplementary file 3XtNdnf coding sequence (NM_001122800) was fused in frame
with C-terminal HisMyc tag and cloned into pcDNA3.1 vector
Recombinant DNA reagentpBRAcpA- Dpp-FLAGSupplementary file 3
Recombinant DNA reagentpBRAcpA- Gbb-FLAGSupplementary file 3
Recombinant DNA reagentpBRAcpA-Gbb-HASupplementary file 3
Recombinant DNA reagentpBRAcpA-FLAG-MadSupplementary file 3
Sequence-based reagentPrimer: pkgForward:
GTCCCAAGACCCGTGAGCTCTTCGC
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: pkgReverse:
TCCGTGTTCCACTTGGCGCAGCAAG
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: ciForward:
CGACCACCAGGAGGAAGTAT
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: ciReverse:
AATCGGAATAAGGCGATGAC
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: hhForward:
GGATTCGATTGGGTCTCCTA
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: hhReverse:
GAATCTGACTTGACGGAGCA
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: ptcForward:
CGATGTGGTCGATGAGAAAT
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: ptcReverse:
CGAGGTGGGACTGGAATACT
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: nordForward:
CACCGCAAAAGTGTCCTTCG
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: nordReverse:
CAGGTTCAGCACAAATCGCT
This paperPrimers used in Figure 1 for
quantitative reverse transcription PCR
Sequence-based reagentPrimer: ptcForward:
atggaccgcgacagcctccca
Hsia et al., 2017Anti-sense probe (600 bp) for ptcPrimers used in Figure 1 for
generating in situ probes
Sequence-based reagentPrimer: ptcReverse:
TAATACGACTCACTATAGGG
gaggtggcgcaggatctgctc
Hsia et al., 2017Anti-sense probe (600 bp) for ptcPrimers used in Figure 1 for
generating in situ probes
Sequence-based reagentPrimer: nordForward:
gaaatccgggtgaagctgctacg
This paperAnti-sense probe (666 bp) for nordPrimers used in Figure 1 for
generating in situ probes
Sequence-based reagentPrimer: nordReverse:
TAATACGACTCACTATAGGG
atgcagcgaagctttgggtatgg
This paperAnti-sense probe (666 bp) for nordPrimers used in Figure 1 for
generating in situ probes
Sequence-based reagentP1: nord exon 1Forward:
GCAAGTGGCAAGAGCTGAAC
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentP2: nord exon 1/2Reverse:
GTGTTCTGCGGTTTTGCCTG
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentP3: nord exon 2Forward:
CGCACTCAGAGGTTGTTTCA
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentP4: nord exon 4/5Reverse:
GCTCCTTTCCCACTTGACGA
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentP5: nord exon 6Forward:
AGGCTCTGTTCCGGGATTTG
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentP6: nord exon 8Reverse:
AAATGCAGCGAAGCTTTGGG
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentGapdhForward:
GCCACCTATGACGAAATCAAGGCTA
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagentGapdhReverse:
GGAGTAACCGAACTCGTTGTCGTAC
This paperPrimers used in Figure 2—figure supplement 1 for confirming the
predicted transcripts from different Nord alleles
Sequence-based reagent5′-GGACCTGTTCGGAATCCACC-3′This paperGuide RNA sequence used to generate different Nord
alleles using CRISPR/Cas9 system
Sequence-based reagent5′-GGGTGAGGTTCTGTCTACCC-3This paperGuide RNA sequence used to generate different Nord
alleles using CRISPR/Cas9 system
Cell line (D. melanogaster)S2DGRCS2-DGRC
Cell line (Homo sapiens)embryonic kidney cell line HEK 293ATCCCRL-1573
Software, algorithmFijiNIHRRID:SCR_002285
Software, algorithmGraphPad PrismGraphPad softwareRRID:SCR_002798

Drosophila maintenance

Request a detailed protocol

Animals were grown on standard food containing molasses at room temperature unless otherwise indicated. The hs-FLP and actin>y+>Gal4 (Ito et al., 1997) driver was used to generate random flip-out clones expressing various UAS-transgenes. The hs-Gal4 was used to induce random ectopic expression of UAS-Hh or UAS-Ptc (Figure 2—figure supplement 3). Larvae of the corresponding genotypes were incubated at 37°C for 30–60 min during the second instar larval stage (Figure 2, Figure 2—figure supplement 3) or 15–20 min in the mid-third larval stage (Figure 3) to induce flip-out clones. Wing imaginal discs were dissected from the larvae containing flip-out clones or hs-Gal4-expressing clones at the wondering larva stage. The hh-Gal4 or en-Gal4 driver together with tub-Gal80ts (McGuire et al., 2003) was used for transient expression of transgenic constructs. Fly crosses, embryos, and larvae were maintained at 18°C, and the Gal80ts repressor was inactivated for the indicated number of hours at restrictive temperature (29°C) before adult fly eclosion or dissection (see Figure 7 legend for details). The genotypes of larvae, pupae, or adult flies used in each figure are listed in Supplementary file 3. Drosophila stocks used in this study are listed in the Key resources table.

Dissociation and sorting of imaginal disc cells

Request a detailed protocol

Wing imaginal discs were dissected from wandering third instar larvae of the genotypes hh-Gal4; UAS-mCD8-GFP or ptc-Gal4; UAS-mCD8-GFP. Discs were stored in Schneider’s Drosophila Media (21720, Invitrogen) plus 10% fetal bovine serum (FBS) (10438, Invitrogen) on ice for less than 2 hr prior to cell dissociation. Discs were washed twice with 1 mL cell dissociation buffer (Sigma, C-1544). Elastase (Sigma, E-0258) was diluted to 0.4 mg/mL in fresh cell dissociation buffer once discs were ready. Discs were incubated for 20 min at room temperature in 0.4 mg/mL elastase with stirring by a magnetic micro-stir bar. Undissociated tissue was spun out, cell viability was measured (>80%), and cells were immediately isolated using the BD FACSAria system. Dead cells labeled with propidium iodide (P3566, Invitrogen) were excluded during FACS, and purity of sorted cells was greater than 99% by post-sorting FACS analysis. Total RNA was extracted from sorted cells (RNeasy, QIAGEN) and stored at −80°C. Quality was assessed with the Agilent Bioanalyzer 2100 (RIN > 7.0).

Identification of target genes of the Hh signaling pathway

Request a detailed protocol

As described in Willsey et al., 2016, using total RNA extracted from sorted A (hh-), B (ptc+), and P (hh+) cells (see details in ‘Dissociation and sorting of imaginal disc cells’), we acquired three primary transcriptome datasets via the Affymetrix D. mel GeneChip Genome 2.0 microarrays. The raw microarray data were deposited to the Gene Expression Omnibus public repository (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE180120; Gene Expression Omnibus series no. GSE180120). In our previous analysis (Willsey et al., 2016), we focused on two datasets and searched for genes differentially expressed in the A/P boundary adjacent cells (B: ptc+) and P cells (P: hh+). Here, we modified the gene expression analysis method by including the third dataset from A cells (hh-) and additional transcriptome comparisons between the A/P boundary adjacent B cells (ptc+) and A cells (hh-), and between A cells (hh-) and P cells (hh+).

Briefly, all analyses were conducted in R version 4.0.2. Expression values were determined using the affy package (Gautier et al., 2004), available from BioConductor (http://bioconductor.org). The Drosophila 2.0 CDF environment was utilized. Probe-level data from the CEL files were imported using the function ReadAffy and converted to expression values using rma with default settings. This method implemented the multi-array average (RMA) for background correction followed by quantile normalization. Expression values were log2 transformed. Probe sets were mapped to genes using the Drosophila_2.na30.annot.xml annotation file, available from the Affymetrix website. 14,448 of 18,952 (76.2%) probe sets map to gene isoforms—13,016 (90.1%) of which correspond to unique genes (some genes are mapped by ≥1 probe set). Probe sets mapping to the same gene were not combined to minimize technical artifacts. Genes (probe sets) whose expression is not only higher in A cells than P cells (FoldA/P > 1.2), but also higher in the A/P boundary adjacent cells than general A cells (FoldB/A > 1.5), were selected as potential Hh-induced target genes (Figure 1). A total of 61 probe sets (59 unique genes) were identified as potential Hh signaling target genes (Supplementary file 1). The heatmap was generated in R with the ggplot2 package, and the genes were ordered by the FoldB/A change.

Quantitative reverse transcription PCR

Request a detailed protocol

Total RNA was extracted from FACS-sorted A, B, and P cells (see details in ‘Dissociation and sorting of imaginal disc cells’). Possible contamination of genomic DNA was excluded by treatments of DNAse I (AM2222, Thermo Fisher Scientific). RNA was reverse-transcribed to cDNA using Maxima Reverse Transcriptase (EP0742, Thermo Fisher Scientific) with random hexamers. All samples within an experiment were reverse-transcribed at the same time; the resulting cDNA was stored in aliquots at –80°C until used. cDNA was PCR-amplified using SYBR Green Supermix (1708880; Bio-Rad). qPCR was carried out with an ABI PRISM Sequence Detection System (Applied Biosystems). Reactions were run in triplicate in three independent experiments. Expression data were normalized to the geometric mean of the housekeeping gene pkg and were analyzed using the 2–ΔΔCT method. The primer sequences are provided in the Key resources table.

In situ hybridization

Request a detailed protocol

In situ hybridization of wing discs was performed as previously described (Hsia et al., 2017). Briefly, RNA probes were created from in vitro transcription of PCR products carrying the T7 RNA polymerase recognition sequence at one end and synthesized by using a digoxigenin (Dig)-labeling kit (Roche). Wing discs of L3 larvae were hybridized with probes overnight at 56°C using standard procedures and visualized using anti-Dig-AP (1:1000; Roche). Primers used for generating PCR templates are listed in the Key resources table.

Generation of Nord Crispr alleles

Request a detailed protocol

The nord3D and nord22A alleles were generated using the Crispr/Cas9 system. The following guides 5′-GGACCTGTTCGGAATCCACC-3′ and 5′-GGGTGAGGTTCTGTCTACCC-3 were separately cloned into the BbsI site of pU6-BbsI-chiRNA plasmid (obtained from Addgene) and both were simultaneously injected by Best Gene into w1118; PBac{y[+ mDint2]=vas-Cas9}VK00027 on chromosome 3 (Bloomington Stock Center #51324). G0 flies were crossed to a balancer stock (w; Pin/CyoStar) and then individual males were crossed to w; Gla, Bc/CyO{GFP} to establish stocks. DNA from homozygous adults was amplified by PCR using primers that flanked the two Crispr target sites and sequenced. The nord22A allele was a 5 bp deletion generated at guide sequence 2 site, while the nord3D allele was an 11 bp deletion generated at the guide 1 site.

Cell culture and transfection

Request a detailed protocol

Drosophila S2 cells (S2- DGRC) were obtained directly from the Drosophila Genomics Resource Center (DGRC) and cultured in Drosophila Schneider’s medium supplemented with 10% of FBS (Omega Scientific) and 1% Penicillin-Streptomycin-Glutamine (Thermo Fisher) at 25°C in a humidified incubator. S2 cells stably expressing the FLAG-Mad transgene (Mad-S2) cells were generated by co-transfecting pBRAcpA-FLAG-Mad (Jensen et al., 2009) and pCoBlast, and then followed by selection with 12.5 μg/mL blasticidin. HEK 293 cells were obtained directly from ATCC and cultured in Dulbecco’s Minimal Essential Medium with 10% FBS (Omega Scientific) and 1% Penicillin-Streptomycin-Glutamine (Thermo Fisher) at 37°C in a humidified incubator with 5% CO2. Transfection was performed with FuGENE 6 transfection reagent (Promega). All the cell lines were regularly confirmed to be free of contamination (e.g., mycoplasma) through PCR-based tests as recommended by the NIH.

Antibodies

Antibodies and dilutions used were mouse anti-Ptc antibody 1:50 (DSHB, Apa1; 1/50); rat anti-Ci antibody 1:50 (DSHB, 2A1; 1/50); rabbit anti-Dpp prodomain (Akiyama and Gibson, 2015; 1/100); rabbit anti-GFP (Invitrogen, A-11122; 1/1000); chicken anti-GFP (Abcam Cat# ab13970; 1/2000); rabbit anti-phospho-Smad1/5 (Ser463/465) (Cell Signaling Technology, 9516; 1/100); mouse anti-HA.11 (Covance, Cat#101P; 1/1000); rabbit anti-HA (Thermo Fisher Scientific, Cat# 71-5500; 1/1000); mouse anti-β-tubulin (DSHB, E7; 1/5000); mouse anti-β-galactosidase (Promega, Z378A; 1/100); mouse anti-Myc (Santa Cruz Biotechnology, 9E10; 1/200); mouse anti-FLAG (Sigma-Aldrich, M2; 1/200); rabbit anti-FLAG (Sigma-Aldrich, Cat# F7425; 1/200); HRP-conjugated and fluorophore-conjugated secondary antibodies were from Jackson ImmunoResearch Lab and Thermo Fisher. The antibody information is also listed in the Key resources table.

Constructs

Request a detailed protocol

The coding sequence of Nord or Ndnf from various species was fused in frame with C-terminal GFP tag or HisMyc tag, and cloned into the pAcSV vector for expression in the Drosophila S2 cells, or into the pcDNA3.1 vector for expression in the HEK 293 cells. The coding sequence of Nord was fused with a C-terminal HA-GFP tag and cloned into the pUAST vector to generate the transgenic UAS-Nord line. Constructs expressing Dpp-FLAG, Gbb-FLAG, Gbb-HA, and FLAG-Mad were generated using PCR methods to tag and amplify the gene of interest from a full-length cDNA and then cloned into the S2 cell constitutive expression vector pBRAcpA (Cherbas et al., 1994). The sequences are provided in Supplementary file 4.

Imaginal discs and pupal wing immunostaining and imaging

Request a detailed protocol

Wing discs from third instar larvae were dissected, fixed in 4% formaldehyde in PBS, blocked and permeabilized by 5% normal goat serum (NGS) and 0.3% Triton X-100 in PBS, incubated with primary antibody in PBS containing 5% NGS and 0.3% Triton X-100 overnight at 4°C, washed three times with 0.3% Triton X-100/PBS, incubated with secondary antibody, and washed with 0.3% Triton X-100/PBS. To selectively stain the secreted Nord (Figure 3), the above immunostaining procedure was carried out in the absence of Triton X-100 (PBS alone for blocking and antibody incubation buffer; 0.01% Tween-20/in PBS for washing buffer). Pupal wings were collected and pre-fixed as previously described (Classen et al., 2008), then followed by the procedure described above for immunostaining of the larval wing discs. The stained larval wing discs or pupal wings were mounted and imaged with a Zeiss spinning disc confocal microscope.

Image collection and quantification of fluorescence intensity

Request a detailed protocol

To compare the expression profile of pMad, Dpp, or Ptc in different genotypes, we used wing imaginal discs at the same developmental stage, which were dissected from wandering larvae, corresponding to 1–6 hr before the entry into the pupal stage. Larvae from the control and corresponding experimental group were raised at the same temperature and density. Wing discs were dissected, fixed, immunostained, and mounted by following the same protocol. All images were taken using the same confocal microscope settings. The pixel intensities of pMad, Dpp, or Ptc were obtained within a fixed rectangular region across both ventral and dorsal compartments using the Plot Profile function of Fiji. Then, the average pixel intensities from multiple discs were plotted using the GraphPad Prism software. The number of wing imaginal discs used in each experiment is provided in the corresponding figure legend.

Cell immunostaining

Request a detailed protocol

48 hr after transfection, NIH 293 cells were washed twice with PBS, fixed in 4% formaldehyde in PBS, blocked by 1.5% NGS in PBS, incubated with the primary antibody in PBS containing 1.5% NGS for overnight at 4° (to stain surface Nord or Ndnf), washed with 0.01% Tween-20/1× PBS, incubated with secondary antibody, and washed with 0.01% Tween-20/PBS. Mad-S2 cells immunostaining was carried out through similar procedures but in the presence of 0.1% Triton X-100 during blocking, antibody incubation, and washing steps. The stained cells were mounted and imaged with a Zeiss spinning disc confocal microscope.

Immunoprecipitation assay

Request a detailed protocol

S2 cells were separately transfected to express Nord with a C-terminal GFP tag, or the FLAG-tagged Dpp, HA-tagged Gbb alone, or Dpp-FLAG/Gbb-HA in combination. >72 hr after transfection, conditioned medium from transfected cells were collected. The medium containing Nord or the BMP ligands were mixed, followed by incubation with anti-FLAG or anti-HA antibody-coupled beads (anti-M2 Affinity Matrix from Sigma; anti-HA Affinity Matrix from Roche) overnight at 4°C. Precipitated proteins were analyzed by Western blotting using anti-GFP, anti-HA, and anti-FLAG antibodies. Beads were washed five times with washing buffer (50 mM Tris-HCl at pH 6.8, 150 mM NaCl, and 1% NP40). Proteins bound to the beads were recovered in the SDS-PAGE sample buffer. Procedures from the medium collection were carried out at 4°C or on ice. Proteins samples were resolved by SDS-PAGE and transferred to PVDF membranes (Millipore) for Western blot analysis.

S2 cell-based BMP signaling assay

Request a detailed protocol

The S2 cell-based BMP signaling assay was adopted from assays as previously described (Shimmi and O’Connor, 2003). Recombinant Dpp (159-DP-020, R&D Systems) was diluted in the culture medium according to the manufacturer’s recommendations. To prepare BMP ligands secreting source cells, briefly, Drosophila S2 cells were transfected with plasmids to express Dpp-FLAG or/and Gbb-FLAG with or without Nord-HA-GFP. >72 hr after transfection, the conditioned medium was collected, and the cells were lysed in lysis buffer (50 mM Tris-HCl at pH 6.8, 150 mM NaCl, 1% NP40, and protease inhibitors). S2 cells stably expressing the FLAG-Mad transgene (Mad-S2) cells were generated and used as BMP responding cells. Alternatively, we also transiently transfected S2 cells or Mad-S2 cells with plasmids to express GFP or Nord-GFP as described in the figure legends. >48 hr after transfection, the responding cells were incubated with the conditioned medium collected from the source cells for 1 hr. After incubation, the responding cells were washed and then lysed in the lysis buffer. Both the conditioned medium and the lysates were clarified by centrifugation, and proteins were recovered directly in the SDS-PAGE sample buffer. Proteins were separated by SDS-PAGE under reducing conditions and then transferred onto PVDF membranes (Millipore). The membranes were blocked and immunostained with primary antibodies and HRP-conjugated secondary antibodies. Blots were developed using Immobilon Forte Western HRP substrate (MilliporeSigma, WBLUF0100) with GeneSys Imaging System (Syngene). In Figure 8, Western blot quantification was performed using the Fiji software. Following background subtraction, the intensity of each band was measured by the function of plot lanes (GelAnalyzer). Then, the band intensity was used to calculate the signaling activity from an equal volume of conditioned medium (pMad/Mad) or the relative ligand activity ([pMad/Mad]/BMPs) as indicated in the figure legend.

Wing size and wing trichome measurements

Request a detailed protocol

Adult wings were dissected from animals in 100% ethanol and mounted in 1:1 Wintergreen oil-Canadian balsam medium. The wings were imaged using a ×4 objective and area was measured using Fiji (ImageJ). Measurements were taken from the end of the costa on the anterior portion of the wing hinge to the end of the alula on the posterior. Area in pixels squared was converted to millimeters squared with a calibration value determined using a hemocytometer under the ×4 objective. Wing trichomes from the dorsal wing surface were imaged with a ×40 objective in the region between veins 4 and 5 just distal to the PCV (Figure 4—figure supplement 6). Trichomes were counted manually within the imaged area (37,500 µm2).

Statistical analysis

Request a detailed protocol

All data in column graphs are shown as mean values with SD and plotted using GraphPad Prism software. As described in the figure legends, one-way ANOVA followed by Sidak’s multiple comparison test, unpaired two-tailed t-test, or two-sided Fisher’s exact test was used for statistical analysis. The sample sizes were set based on the variability of each assay and are listed in the figure legends. Independent experiments were performed as indicated to guarantee reproducibility of findings. Differences were considered statistically significant when p<0.01.

Data availability

The raw microarray data were deposited to the Gene Expression Omnibus public repository (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE180120; Gene Expression Omnibus series no. GSE180120).

The following data sets were generated
    1. Zheng X
    (2021) NCBI Gene Expression Omnibus
    ID GSE180120. Genome-wide expression profiling in Drosophila wing imaginal discs.

References

    1. Capdevila J
    2. Estrada MP
    3. Sánchez-Herrero E
    4. Guerrero I
    (1994)
    The Drosophila segment polarity gene patched interacts with decapentaplegic in wing development
    The EMBO Journal 13:71–82.
    1. de Celis JF
    (1997)
    Expression and function of decapentaplegic and thick veins during the differentiation of the veins in the Drosophila wing
    Development 124:1007–1018.
    1. Ito K
    2. Awano W
    3. Suzuki K
    4. Hiromi Y
    5. Yamamoto D
    (1997)
    The Drosophila mushroom body is a quadruple structure of clonal units each of which contains a virtually identical set of neurones and glial cells
    Development 124:761–771.
    1. Martín-Castellanos C
    2. Edgar BA
    (2002)
    A characterization of the effects of Dpp signaling on cell growth and proliferation in the Drosophila wing
    Development 129:1003–1013.
    1. Ray RP
    2. Wharton KA
    (2001)
    Context-dependent relationships between the BMPs gbb and dpp during development of the Drosophila wing imaginal disk
    Development 128:3913–3925.
    1. Strigini M
    2. Cohen SM
    (1997)
    A Hedgehog activity gradient contributes to AP axial patterning of the Drosophila wing
    Development 124:4697–4705.

Decision letter

  1. Erika A Bach
    Reviewing Editor; New York University School of Medicine, United States
  2. Utpal Banerjee
    Senior Editor; University of California, Los Angeles, United States
  3. Erika A Bach
    Reviewer; New York University School of Medicine, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for choosing to send your work, "The NDNF-like factor Nord is a Hedgehog-induced BMP modulator that regulates Drosophila wing patterning and growth", for consideration at eLife. Your submission has been assessed by a Senior Editor in consultation with a member of the Board of Reviewing Editors. Although the work is of interest, we regret to inform you that the findings at this stage are too preliminary for further consideration at eLife.

Specifically, the reviewers felt that many experiments would be needed to better characterize the nord mutant phenotype, particularly its role in BMP signaling. They wanted the authors to:

1. Measure Dpp levels in nord mutants. This could be accomplished using the anti-pro-Dpp, or enhancer traps that are more accurate than BS3.0 (which not only lack the shortvein enhancers but is overly sensitive to anterior en).

2. Measure Ptc levels in nord mutants.

3. Show a cell autonomous effect of Hh signal reception on Nord expression.

4. Examine pMad at earlier stages of PCV development (20 hr AP) before EGFR and Notch signaling are active in this region.

5. Test whether the posterior defects they see are caused by loss of anterior Nord, or whether they might be caused by loss of low-level Nord expression in the posterior. In other words, what is the result when Nord is depleted from anterior or from posterior cells. Additionally, the authors should determine how far secreted Nord can spread in the disc.

7. Test how much of the PCV loss in en>Nord is caused by loss of the adjacent L5.

8. Add missing controls such as a nub-Gal4/+ 29°C control wing and a control disc stained with anti-GFP.

9. improve quantification for pMad analysis, particularly in Figure 4E, 4F and 6.

The reviewers also suggested extensive editorial changes that would make the manuscript easier for the reader to follow.

Reviewer #1:

The authors used cell sorting and genome-wide expression profiling to identify new Hedgehog (Hh) target genes in the developing Drosophila wing disc, which has been a powerful system to study Hh signaling. They identify nord, which encodes a secreted factor that associates with cell membranes and the extracellular matrix. Null mutants of nord are adult viable. Paradoxically, these animals have slightly smaller wings and ectopic posterior cross vein (PCV) material. Both phenotypes can be caused by alterations in BMP signaling. They demonstrate that recombinant Nord binds to Dpp and Gbb and restricts their release from cells and diffusion in the media. They use the Gal4/UAS mis-expression to test the model that Nord has both positive and negative effects on BMP signaling depending on where and when Nord is expressed.

The strengths of this paper are the appropriate use of techniques such as cell sorting, transcriptomics, Drosophila loss of function genetics, cell culture. These approaches support their model that Nord regulates availability of Dpp, Gbb and/or Dpp-Gbb dimers. However, the use of mis-expression of wild-type Nord in the pouch or posterior compartment lead similar phenotypes as global nord loss-of-function. This is a weakness because it is difficult for the reader to reconcile logically these paradoxical results. This work may impact the mammalian fields because Nord is conserved through evolution, but Nord homologues have not been implicated in BMP signaling.

a) While I appreciate that they authors have done a lot of work and the data are of high quality, I don't follow the reasoning of how similarity in loss- and gain-of-function nord phenotypes supports their model (or Nord having both positive and negative effects on BMP signaling depending on where and when Nord is expressed). Could the authors use available Gbb and Dpp antibodies to determine the expression patterns of these proteins in nord LOF and GOF?

Related to this, the authors do not control for the number of UAS transgenes in Figure 7. If you want to compare 3X UAS-Nord to 1X UAS-Nord, the 1X UAS-Nord sample should have 2X UAS-LacZ, GFP or another neutral protein. Being able to prove that they have not titrated Gal4 is critical to their model about low vs high Nord. These experiments should be performed again with all animals have the same number of UAS transgenes.

b) The authors do not provide a description of micro-array analysis. Additionally, there is no mention of the deposition of these data into public repository like NCBI GEO, to which the reviewers should be provided access. Related to this, there is no information on fold change of nord among the various sorted populations. They discuss top-ranked genes, but none of these data are presented. All of these need to be addressed.

c) We need more information about the Halfon hs-Gal4 technique that they used. Additionally, there are too many "clones" in 1H and we don't know where the "clones" are in Figure 1G. Could the authors use FLP/FRT flip-out approaches to generate a few discrete clones that mis-express Hh or Ptc?

d) The reader needs more information about the nord homologs that were examined in cell culture in Figure 2. For example, what is DrNdnf?

e) Figure 2C, the labels should be M and L, not M and C.

f) The font size is too small in Figure 2A. B, Figure 3D, Figure 3F, Figure 3M. Please enlarge.

Reviewer #2:

Wu et al. identified nord as a potential Hh-target gene in the Drosophila wing imaginal disc. Nord shares sequence similarity with a family of secreted proteins containing a domain of unknown function (DUF2369) and one or two fibronectin type II-like repeats which contains the neuron-derived neurotrophic factor (NDNF) which has been shown to promote neuronal migration and neurite outgrowth in vertebrates, such as the gonadotropin-releasing hormone producing neurons critical in the regulation of puberty and reproduction. The authors explored the function of nord in wing development. Hh is known to be secreted from cells in the posterior compartment of the wing disc and act on cells across the A/P boundary where an adjacent stripe of anterior cells express the Hh receptor ptc. Interestingly, nord is expressed in a subset of these 'stripe' cells consistent with it being a Hh-target. The dpp gene is a well known Hh-target expressed throughout the 'stripe' like ptc. dpp encodes a BMP2/4 ligand which is known, with the Drosophila BMP5/6/7 ligand, Gbb, to play multiple roles in wing growth and patterning. Wu et al. find while flies homozygous for LOF mutations in nord are viable and fertile, their wings are slightly smaller and they contain ectopic vein material at the posterior cross vein (PCV). A reduction in BMP signaling results in smaller wings but an increase in BMP signaling is associated with PCV spurs. Based on these phenotypes, Wu et al. investigate the potential relationship between nord function and BMP signaling. BMP signaling in the larva and its roles in growth and patterning and its later roles in the pupal wing disc in laying down vein material are temporally separable events. Furthermore, both the Blair and O'Connor labs have nicely delineated a mechanistic explanation for BMP involvement in the formation of the PCV. How nord is involved in each of these processes is not completely clear, especially given nord's expression pattern. Must the spatial and temporal requirements for BMPs in these processes be re-evaluated?

Is there a connection between how NDNF influences the migration neurons and its extracellular role in modulating molecules such as BMPs? Does it provide a link with extracellular matrix and other extracellular molecules that also appear to be modulators of BMP activity? Could Nord impact BMP output by influencing wing disc cellular projections?

Strengths:

– Identification of a Hh-target gene that appears to modify BMP signaling output.

– Ectopic vein material near the posterior cross vein (PCV) is attributed to loss of nord.

– A loss of function nord mutants exhibit a reduction in wing size.

– Nord-GFP is shown to bind to Dpp, and less so to Gbb, by co-immunoprecipitations in S2 cells.

Weaknesses:

– The manuscript would benefit from stating at the start that the researcher's findings suggest that nord may both facilitate and antagonize BMP signaling. It will be easier for the reader to follow.

– Better clarification of the different functions for BMP signaling in wing growth, patterning, and development in the larval versus pupal period would strengthen the manuscript.

– The authors can then follow up on these points as to how nord impacts these different functions

– For the most part nord is not expressed in either wing pouch cells or vein primordia. Furthermore, based on expression in cell culture, Nord is not found far from the source cell. How is Nord though to influence pMad, for example, in cells in the anterior wing pouch?

– It is import to reconcile the findings of others that the pMad gradient is not required for wing disc growth.

General points:

It would be helpful to the reader if the authors stated at the outset that they are suggesting that the impact of nord on BMP signaling can be both as a facilitator and an antagonist, and then go on to present data and the contexts in which they see different behaviors.

In the wing disc, pMad represent the output of BMP signaling, both the action of Dpp and Gbb. In most instances it is impossible to separate the effects of the two ligands, whether they act as a heterodimer or as homodimers, therefore, in nearly all places in the text "Dpp signaling" should be replaced with "BMP signaling", unless the authors know for certain they are examining signaling induced solely by Dpp.

Methods:

All Drosophila strains should be listed under "Drosophila strains".

Method used for overexpression of Hh or Ptc should be presented under some other subheading than Drosophila strains.

Third instar larval period is 48hrs at 22°C. If animals were heat shocked in early third and then dissected after two days they should have pupated. More accurate documentation of developmental time should be provided, or the authors should indicate that the larvae were anywhere between X and Y hrs or days old.

Epitomics 1880-1 was raised against phospho-Smad3, not pMad.

Constructs – the nucleotide sequences of the Nord and Ndnf – GFP and 3xMyc fusions should be shown.

Dpp-FLAG, Gbb-FLAG, Gbb-HA, and Mad-FLAG are not described in Serpe et al. 2005 – a description and characterization of activity should be provided

Section "RT-qPCR" needs to be renamed to indicate it contains the description of imaginal disc cell isolation. Were these the cells and RNA used for microarray analysis referred to in the text?

Imaginal discs and gene expression patterns change significantly during the 3rd larval instar. At what time were wing imaginal discs harvested for RNA isolation and immunohistochemistry?

No description of procedure for FACS or microarray was provided.

Results

Figure 1 – when first describing nord gene expression on page 7 Figure 1D & E, its absence from the wing pouch should be noted and that it is expressed anterior to the A/P boundary in the hinge progenitors.

Description of Nord-GFP in discs following over expression of Hh or Ptc is confusing (bottom of page 8) as the tissues are derived from larvae not flies.

hs>hh leads to expansion of Nord-GFP in hinge and wing periphery

hs>ptc breadth of Nord-GFP domain narrowed, only remaining is at A/P boundary.

Figure 1 H – do the author's have an explanation for the uneven expression of ptc and the corresponding response of Nord-GFP expression? Nord-GFP appears to be lacking in cells where Ptc is present in Figure 1G but Nord-GFP expression and Ptc are present in quite a number of cells in the hinge region of a wild type wing disc Figure 1F.

Figure legend: error? – "White brackets indicate the expression range of Ptc (E) or No-GFP (E-H)".

Figure S2 panel A – include a scale in bp.

panel E – why is the expected 732bp band seen when P3 + P4 are used to amplify MI06414-GFSTF.2?

S2 figure legend lacks reference to panels D-F.

Figure S3 – sequencing described?

Residue numbers should be included in figure.

Figure 2 – should note that melanogaster was used – provided species name for other genus' clone names should be listed in figure legend, ie which species the Ndnf gene was derived from;

panel B – no size markers.

lanes in figure B marked "M" and "C" but in figure legend noted as "medium (M)" and "cell lysate (L)".

Indicate that the resulting Ndnf proteins were tagged with Myc – in figure legend.

Explanation for difference in size of DmNord found in media versus cell lysate?

What are the predicted sizes for Nord/Ndnf proteins?

Ideas about the differential abundance of Ndnf in media from lysate from different species.

An important control is needed – WT disc with Anti-GFP.

Is it known that MI0641 produces a truncated peptide, or is it predicted?

Does higher prevalence of PCV spur in females suggest a second site enhancer? Dosage compensation?

What is the correlation between ectopic vein formation in the pupal wing and expression in the larval disc?

Authors should be careful when describing protein versus RNA expression patterns, as well as loss of gene function versus loss of gene expression. For example on page 12, neither deCelis 1997 or Yu et al. 1996 examined Dpp (protein) expression They report in situs (endogenous RNA) or lacZ enhancer trap expression patterns.

In the sentence "We asked whether elimination of Nord alters level of pMad" – should be rephrased " We asked if the knock down (or reduction) in nord function …"

pMad is a downstream signal transducer, not a transcription transducer

Harmansa et al. 2015 proposed that dpp spreading is required for medial but not lateral wing growth. How would this finding apply to the propose function of nord with respect to wing growth?

Harmansa 2017 showed that eGFP-Dpp dispersal in basolateral is required for patterning & growth. Does Nord have subcellular localization?

An antibody against the Dpp prodomain is used in Figure 4A. BMP proproteins are cleaved separating the prodomain from the ligand domain proper, with the ligand sometimes secreted with its prodomain and other times, not. Do the authors feel this pattern reflects that of the secreted Dpp ligand?

Nord-GFP is in a subset of Dpp prodomain expression in larval wing disc (Figure 4A) and shows even less overlap in pupal wing disc (Figure 4B) – this needs to be more clearly addressed in the text.

Data not shown for Nord-GFP in later pupal wing when dpp-lacZ transitions to longitudinal veins.

"While dpp is also known to be expressed in the developing longitudinal veins (deCelis, 1997), we see no Nord expression in either the longitudinal vein or crossvein despite the fact that Nord loss of function leads to ectopic crossvein formation." The data referred to in the underlined text top page 13 is not shown. (?)

Figure 4E – Is quantification plot an average of multiple discs? how many discs were measured? Where was the measurement taken? Is it a profile from a line across the disc or a compilation of all signal in the disc? Would the authors expect pMad to be altered in the wing pouch?

Figure 4F – not at all clear what is meant by "maximum pMad intensity in the A and P compartment" in the figure legend. Is this the highest intensity measurement in the entire disc?

Is the proposal that the slight reduction in pMad in the anterior affects only growth and patterning of the position of the wing veins?

If Nord is required to facilitate "Dpp" movement, then why is signaling activation (as measured by pMad) in the majority of the anterior compartment beyond the peak unchanged. If in the absence of Nord, Dpp does not move, then one would expect all signaling in the anterior to be affected.

Instead of facilitating movement, could Nord prevent Dpp-Gbb turnover, thus in its absence Dpp-Gbb is less stable and a drop in signaling (pMad) is observed. Difficult to understand how Nord has an effect on ligands when its not present in wing pouch cells

Top half of pg 16 – authors suggest that media from cells co-expressing Nord and Dpp/Gbb elicit less signaling (pMad) than media produced by cells without Nord -if the same amount of media is added to cells in Figure 5C then it seems that M4 media is less effective at inducing pMad compared to the lower amount of Dpp/Gbb in M5 which elicits nearly the same level of pMad.

What does a normalized plot show?

Using a recombinant Dpp peptide in Figure 5D, the authors find that in the presence of Nord-GFP, equivalent amounts of rDpp induce less pMad but this is the opposite of what is observed in vivo in the disc where signaling is reduced in nord mutants. Perhaps the authors can expand on this observation. Is this result related to the fact that a recombinant peptide was used in rather than the entire Dpp proprotein or prodomain + ligand domain?

Figure 5E in WT – seems odd that one cannot see increased concentration of Mad-Flag in the nucleus.

Suggest that Nord-GFP could prevent release of ligand from cell surface and inhibit ligand from media to act on responding cell – what about ligand-receptor binding?

Westerns in Figure 5 are cropped and it is not possible to see the other cleaved forms of ligands. Does Nord show differential interaction with different ligand forms, as has been shown by Anderson 2017 to be the case with different forms of Gbb and receptors?

Figure 6

How were pMad profiles measured? Superimposing profiles in D-F would be helpful.

Authors state ptc levels have not changed – provide measurement of ptc levels.

Loss of L5 and PCV are typical of gbb LOF phenotypes and not seen in dpp LOF genotypes, therefore, it is hard to reconcile these phenotypes with if Nord is thought to preferentially affect dpp.

In Figure 6H – nub/+ label could easily be confused with a nub mutant heterozygote.

A nub-Gal4/+ 29°C control wing should be shown.

In some places nord is indicated as a secreted protein and in others membrane-associated – a little bit confusing.

Akiyama 2015 showed that the pMad gradient in the wing pouch, as well as Dpp were not required for wing growth. How do the authors think changes in the pMad profile may change growth? It is curious that changes in the pMad gradient seen in Figure 6F has no effect on longitudinal vein development in Figure 6K.

The Discussion contains a considerable amount of background information that is not particularly relevant to the findings presented in manuscript. Focusing the discussion a bit more would be helpful to the reader.

To the reader it is not at all clear how expression in the posterior compartment during larval period influences PCV formation in the pupal wing when ligands are thought to be transported from the longitudinal veins.

Does uniform expression in posterior using enGal4 interfer with the normal mechanism?

Suppl Figure 12 – orientation of high mag of trichomes is incorrect.

Overall, please check labelling of figure for units, number of samples (n=?), and accuracy of figure legend description. In some places in the manuscript this information is robust but in other places it is lacking.

Reviewer #3:

The study focuses on a possible role for the secreted Nord protein in BMP signaling in Drosophila. The authors demonstrate Nord binding to Dpp and Gbb in vitro, and suggest that this explains the in vivo phenotypes observed. The in vitro work is quite suggestive, but some caution may be in order as other other extracellular molecules can affect multiple pathways, and some of these other pathways can have have similar effects on growth and vein formation. While this is a worthwhile publication with strong in vitro data, I have a few points below that I think would strengthen the in vivo data.

1) Nord as a Hh target- The expression of Nord in the imaginal and pupal wing tissues largely mirrors that of Hh targets, being heightened anterior to the anterior-posterior compartments boundary. Its expression is however excluded from the distal wing blade, something that the authors do not further investigate. The authors can widen or narrow Nord expression by widening or reducing Hh signaling, but this will also widen BMP expression and signaling, so it would have been helpful to have one experiment showing a cell autonomous effect of Hh signal reception on Nord expression.

2) Nord mutants- The authors generate and examine Nord mutants. The effects in the wing are fairly mild, consisting of very slight undergrowth mirrored by slightly reduced BMP signaling (pMad) in the disc, and a few ectopic veins, suggestive of gains in BMP signaling. The only other growth-inducing pathways the authors examine are Hh, looking at a single target, Ptc. They do not however quantify Ptc staining in the same way they did to detect the weak effects on pMad, nor do they look at other Hh targets with different sensitivities.

The ectopic veins in mutants are consistent with defects seen with mild increases in BMP signaling, but there are many other ways of inducing similar ectopic venation, especially through EGF gains or reductions in Notch or its ligands. The authors concentrate almost exclusively on one defect, a vein extension distal to the PCV. This is a common defect seen in many mutant fly stocks as it is something of a “hot-spot” for ectopic venation. The authors show that pMad is high in this ectopic vein, but do this staining at 26hr AP, a fairly late stage of PCV development. By this time LVs and cells near the ACV have all been expressing dpp for some time and have BMP signaling, while dpp expression in the PCV is initiating. I would wager this ectopic vein is expressing dpp, and will also have rhomboid and thus EGFR activity and high Δ and Serrate. The authors could greatly strengthen their case by examining pMad at earlier stages of PCV development (20 hr AP) before EGFR and Notch signaling are active in this region.

The authors also do not test whether the posterior defects they see are caused by loss of anterior Nord, or whether they might be caused by loss of low-level Nord expression in the posterior. Since they have a Nord-RNAi line, what happens with anterior knockdown? Posterior?

3) Nord overexpression and biphasic effects on BMP signaling – Increased Nord expression causes undergrowth and reduced pMad but ectopic venation. In that sense, Nord has biphasic effects, with similar mixed effects (both BMP losses and gains) caused by either too much or too little Nord. The authors extend this by showing that further increases in Nord overexpression change the ectopic PCV extension into PCV loss, the latter often caused by loss of BMP signaling. However, this loss may not always be due to a direct effect on the PCV. While the nub-gal4 example looks ok, the en-gal4 example has lost L5 adjacent to the PCV, likely due to loss of the distal L5 primordium at the disc stages (Ray and Wharton, Bangi and Wharton). That will in turn block or reduce dpp expression in L5 and thus Dpp’s availability for signaling in the early PCV.

4) In vitro studies- These provide the strongest proof that Nord can affect BMP signaling, although in this case the observed effects are all inhibitory. Secreted Nord can co-IP with secreted Gbb or Dpp, and co-expression with Nord reduces Gbb or Dpp secretion and thus its ability to signal. Nord can also reduce the BMP signaling induced by addition of exogenous Dpp. The only shortcoming here is that the authors have not apparently tested binding to any of the extracellular modulators of BMP signaling.

I need to see whether there is ectopic pMad in the nord mutant PCV at a stage before dpp, rhomboid and Δ/Serrate expression in the region. Otherwise I am not convinced that the ectopic vein is initiated by ectopic BMP signaling.

Quantify Ptc or some other Hh marker (Dpp would be especially helpful) in Nord mutants.

How much of the PCV loss in en>Nord might be caused by loss of the adjacent L5?

Cell-autonomous effects of Hh signaling on Nord expression would be helpful, as would the effects of anterior vs. posterior Nord RNAi.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for submitting your article “The NDNF-like factor Nord is a Hedgehog-induced extracellular BMP modulator that regulates Drosophila wing patterning and growth” for consideration by eLife. Your article has been reviewed by 2 peer reviewers, including Erika A Bach as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Utpal Banerjee as the Senior Editor.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) Soften their conclusions about their in vitro work (Figure 8) in the results and discussion. Add a statement to the effect that without additional replicates in Figure 8 other conclusions are possible.

2) Provide additional data on the Nord in the posterior compartment using ptc-gal4, UAS-nord-RNAi.

3) Add information about the cross vein phenotype of MS-1096 to the manuscript.

4) Make changes to the Figure fonts.

5) Correct the typos and awkward sentences.

Reviewer #1:

The authors have made strong efforts to respond to the essential revisions and the reviewers’ critiques. Overall, the manuscript is much improved. However, I have concerns about the conclusions regarding Figure 8, the in vitro work, which the authors use conclude that Nord in vitro also has biphasic function on BMPs in vitro. The authors present results from a single experiment and then normalize these single data points to draw conclusions about Nord function in vitro. Specifically, the authors’ new analyses in Figure 8 panels D,E,G show relative pMad levels normalized to Mad (8D,8G) and to Mad and BMP (8E). My concerns are as follows. First, the authors need statistical analyses. Second, BMP released from cells appear to be equally active in the absence (M4) or presence (ME) of Nord, and these data only become supportive of their argument when “normalized” to the amount of ligand. Third, there are additional bands consistent with other Dpp/Gbb isoform in the source blot for Figure 8A. These may be bioactive molecules but are not accounted for in the manuscript. The pMad bands in M4 and M5 also appear to be saturated, again making it difficult to determine differences between the lanes. In panel C, the WB: Flag row is "smiling" with some lanes compacted and others not, again making it difficult to normalize. Without more replicates, the authors statement that “soluble ligand appears to have a higher signaling activity perhaps as a result of an association with Nord” (p.21) is not sufficiently supported by the data. This same issue holds for panel F, with only one replicate. In its current form, the data are not conclusive and other interpretations are also possible. The authors should soften their conclusions about their in vitro work in the results and discussion. Additionally, they need add a statement to the effect that without additional replicates in Figure 8 other conclusions are possible.

They are other issues that need to be addressed:

1. The figure in the response to reviewers was too small. In the future, please upload a larger image.

2. P. 7. The authors write “In wing discs, cells near the A/P compartment boundary (B: ptc+) receive the highest level of Hh stimulation while A cells (A: hh-), located further from the border, receive lower levels of stimulation. In contrast, P cells (P: hH+) do not respond to Hh due to lack of the receptor Ptc and the transcription factor Ci (Tabata et al., 1992; Eaton and Kornberg, 1990).” It is very important that the authors add “Figure 1” when discussing B, A and P cells.

3. P. 11 and Figure 3C – please clarity what tissue is being used for the flip-out clones.

4. Top of P. 15. Please rephrase “which is consistent with the obviously reduced BMP, but not Hh, signaling activity during wing imaginal disc development (Figure 5B, C)” because it is not clear what is being discussed. I suggest “which is consistent with normal Hh signaling activity during the development of nord mutant wing imaginal discs (Figure 5B, C).”

5. P. 18 The authors cite Robert 1998 but there is no such citation in the references cited section. Please add this reference or remove it from the manuscript.

6. P. 21. As noted, please soften your conclusions about the in vitro studies and add a statement to the effect that without additional replicates in Figure 8 other conclusions are possible.

7. P. 25, line 9 “it’s” should be “its”.

8. P. 25, line 13 “Akiyama et al. 2021 accompanying manuscript” is not in the references cited.

9. Add to the Materials and methods the procedure you used for normalization of the data in Figure 8E,D,G.

10. Figure 1B – increase the font size of the colored/highlighted genes. Currently we can’t read them.

11. Figure 2A – increase the font size of the RMCE vectors. Currently we can’t read them.

12. Figure 2D, E, F, G, H – increase the width of the lines outlining regions of the wing disc and clones. Currently we can’t see them.

13. Figure 2C-H, it would be helpful to add C’ and C” etc to the unlabeled panels.

14. Figure 4E – The graphs are too small – I cannot read anything. Please increase the size of the graphs and the font of the labels.

15. Figure 5D – increase the width of the lines of the boxes and the enlargements so that they can be seen on a print out.

16. Figure 7B – increase the width of the lines of the boxes and the enlargements so that they can be seen on a print out.

17. Figure 2, figure supplement 3. Increase the size of the brackets in the Nord-GFP monochrome panels. Currently, it is very hard to see them.

18. Figure 5—figure supplement 2B – label the blue and the red lines.

Reviewer #2:

The authors have added some valuable data that have improved the manuscript, enough for acceptance. I do however have some points I would like the authors to seriously consider. (1) and (2) concern data that are problematic, but could be removed from the manuscript without damaging its conclusions.

1) Hh signaling and Nord expression- The autonomous gains in Nord expression caused by Smo overexpression are convincing. The loss caused by Ptc overexpression, shown for a single flpout clone Figure 3D, is not. To be a valid result, the dotted blue line drawn by the authors must accurately follow the A/P. However, outside the Ptc clone at the bottom of the figure, the dotted line does not at all correspond to the line shown by Nord expression, even though the Nord outside the clone should define the A/P. Nor does the line correspond to the A/P defined by Ci staining. Instead, the true A/P boundary is further left, and thus it is not at all clear that the Ptc clone is in the anterior compartment. If the authors want to keep Ptc flpout data in the manuscript, they need a more convincing example.

2) MS1096-gal4- The authors use MS1096 for some of the nord-RNAi data. This is a Gal4 enhancer trap insertion into the X-linked Bx (dLMO) locus, and in hemizygous males has a mild Bx wing mutant phenotype with variable penetrance that includes disruption of the crossveins (see Figure 6 in Milan, Diaz-Benjumea, and Cohen 1998). The crossvein phenotype tends to disappear in stocks, but often reappears when outcrossing to other lines. It is therefore not an appropriate driver to use when assessing crossvein phenotypes in males. If the authors want to use it, they have to explain how they dealt with the crossvein phenotypes that can be caused by MS1096 on its own.

And while MS1096 is expressed in the wing disc, it is not clear that MS1096 is “wing-specific” as the authors state. Bx (dLMO) is expressed in many tissues, including the CNS. If MS1096 follows the expression of Bx, the authors cannot say that this experiment “further validates that the observed phenotypes in the nord mutants are not indirect effects associated with loss-of-nord in tissues other than the wing discs.”

3) Is Nord in the posterior compartment? The authors make several absolute statements about the absence of Nord expression from the posterior of the developing wing, but do not provide any staining of controls that would allow one to distinguish signal from background. While Nord is certainly much higher anterior to the A/P boundary, the data does not rule out low-level Nord expression in the posterior. The authors have addressed this in their revision, not by changing any of their absolute statements about expression, but by added a new experiment showing that “we did not notice any ectopic PCV in the adult flies when UAS-nord-RNAi was selectively expressed in the P compartment of the larval and pupal wing via the hh-Gal4 driver.” But this negative result could simply reflect stronger RNAi expression with MS1096-gal4 or A9-gal4. It would be preferable if they had tested the effects of nord-RNAi when it was limited to the anterior compartment, something that their model predicts should affect the posterior crossvein. This could easily be done with ptc-gal4 or ci-gal4.

The authors also model the action of Nord as if it only diffuses over a short distance, and thus is only affecting BMP movement from neighboring cells, such as those”in L4 in pupal wings, and is never present in the PCV. However, the extracellular staining of Nord-GFP in Figure 3 looks substantially above background throughout the wing disc.

4) Biphasic or multiphasic effects- The authors have added new data that clarify some of the effects of Nord in the PCV, but complicate their story about biphasic effects. In growth Nord appears truly biphasic, promoting growth only at normal levels, and inhibiting growth and BMP signaling with either decreased or increased Nord expression. But in PCV formation things are more complex. Both loss of endogenous Nord and mild gains induce ectopic PCV phenotypes. Since normal levels do not alter the PCV, that also fits with a biphasic effect, albeit the opposite of the growth effect (ectopic signaling with both losses and mild gains). Overexpression also has a biphasic effect, as strong Nord gains do the opposite of weak gains and cause loss of the PCV and signaling. However, if one puts all of this together, the PCV effect is more multiphasic than biphasic, because of the normal signaling observed at wild type. Nord loss increases signaling, wild type is normal, mild Nord overexpression increases, strong Nord decreases.

The only detailed explanation of these effects is in the legend of Figure 9, which is not very complete and leaves some puzzles. If loss of Nord increases long-range BMP diffusion into the PCV, why isn’t the same thing happening in a third instar wing pouch? And there are ideas hidden in the figure itself that are not explained in the legend or the Discussion, especially the switch from BMP presentation to BMP sequestration with a change from low local to high local Nord concentrations. A fuller discussion in the text would be helpful.

5) Biphasic in vitro- The in vitro work remains strong, especially for Nord’s inhibitory effects on BMP signaling. But the only evidence for a positive and thus biphasic effect of adding Nord relies on a single lane in a single gel: the lowest Nord concentration in Figure 8F, then quantified in 8G. It would be helpful if the authors could state how many times they have repeated this result.

6) Given the long discussion of NDNF, the authors should mention the mechanisms that have been proposed for NDNF activity: by binding integrins, or by modulation of FGF signaling by binding HSPGs.

https://doi.org/10.7554/eLife.73357.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Specifically, the reviewers felt that many experiments would be needed to better characterize the nord mutant phenotype, particularly its role in BMP signaling. They wanted the authors to:

1. Measure Dpp levels in nord mutants. This could be accomplished using the anti-pro-Dpp, or enhancer traps that are more accurate than BS3.0 (which not only lack the shortvein enhancers but is overly sensitive to anterior en).

The suggested experiment was performed, in which no noticeable change of the Dpp (anti-Dpp prodomain) expression levels were detected in the nord mutant wing discs. These data are presented in Figure 5—figure supplement 2 in the revised manuscript.

2. Measure Ptc levels in nord mutants.

The suggested experiment was performed and included in Figure 5C in the revised manuscript.

To further differentiate the effect of loss-of-nord on Hh vs. BMP signaling pathways, we measured the dL3-L4 /dL2-L5 ratio in the adult wing (Figure 4C), and these results were described on pages14-15 (below).

“Of note, beside BMPs, Hh signaling also patterns the developing Drosophila wing. Hh shortrange activity is responsible for patterning the central L3-L4 region and determining the distance between L3 and L4 longitudinal veins (Mullor et al., 1997; Strigini and Cohen, 1997). When quantifying the distance between distal ends of longitudinal veins L3 and L4 (dL3-L4) and that of L2 and L5 (dL2-L5), we found no specific reduction in the L3-L4 region indicated by comparable or higher dL3-L4 /dL2-L5 ratio in the female and male nord mutant wings (Figure 4B, C), which is consistent with the obviously reduced BMP, but not Hh, signaling activity during wing imaginal disc development (Figure 5B, C). Together, these findings suggested a positive role of endogenous Nord in augmenting BMP signaling activity to promote wing growth.”

3. Show a cell autonomous effect of Hh signal reception on Nord expression.

The suggested experiment was performed and included in Figure 2C-H in the revised manuscript. Briefly, we found that Nord expression is cell-autonomously induced/inhibited by Hh signaling within the anterior compartment in the wing discs’ hinge and outer wing pouch region.

4. Examine pMad at earlier stages of PCV development (20 hr AP) before EGFR and Notch signaling are active in this region.

The suggested experiment was performed and included in Figure 5D-E in the revised manuscript. Expansion of pMad range around the primordial PCV was noticed in the nord mutants starting at 19-20 hr after pupation.

5. Test whether the posterior defects they see are caused by loss of anterior Nord, or whether they might be caused by loss of low-level Nord expression in the posterior. In other words, what is the result when Nord is depleted from anterior or from posterior cells.

Our findings show that nord, like ptc and dpp, is a target gene of the Hh signaling pathway and not expressed in the posterior compartment of the wing discs (Figures 1, 2, 5A; Figure 2—figure supplement 3; Figure 5—figure supplement 1). Consistently, when we knocked down nord in the posterior compartment using hh-Gal4, we did not observe the ectopic PCV phenotype, which is frequently caused by knocking down nord throughout the wing discs via MS1096-Gal4 or A9- Gal4. These data are presented in Figure 5—figure supplements 3 and 4 in the revised manuscript.

6. Additionally, the authors should determine how far secreted Nord can spread in the disc.

We expressed UAS-Nord in small clones in the wing discs and found that Nord was mainly detected on the expressing cells' surface or nearby extracellular matrix (Figure 3D). We also detected very low levels of secreted Nord-GFP in the wing discs carrying the nord-GFP protein trap allele (Figure 3E). Additionally, we observed that ectopic Nord driven by hh-Gal4 or en-Gal4 could induce ectopic crossvein formation in the anterior compartment close to the A/P boundary, indicating secreted Nord can spread and act several cell diameters away. These data are presented in Figure 7B and Figure 7—figure supplement 2 in the revised manuscript. Based on these observations, we think a portion of Nord can spread at least several cell diameters away from the source cells.

7. Test how much of the PCV loss in en>Nord is caused by loss of the adjacent L5.

High levels of ectopic Nord driven by en-Gal4 (or hh-Gal4) in the larval stage can sequester BMP ligands and inhibit BMP signaling, thus lead to loss of L5 (original Figure 7). To avoid affecting L5 development, in the revised version, we restricted ectopic Nord expression to the pupal wing and then studied its effect on PCV patterning. The temporal control of en-Gal4 (or hh-Gal4) driven expression of UAS-Nord was achieved by including the tub-Gal80ts allele. In this new setting, in the presence of normal L5, ectopic Nord caused PCV phenotypes in a dosage-dependent manner. These data are presented in a new Figure 7 in the revised manuscript.

8. Add missing controls such as a nub-Gal4/+ 29°ting and genome-wide expression proC control wing and a control disc stained with anti-GFP.

The suggested experiment was performed and included in Figure 6 in the revised manuscript.

9. Improve quantification for pMad analysis, particularly in Figure 4E, 4F and 6.

The suggested experiment was performed and included in Figure 5 (original Figure 4) and Figure 6 in the revised manuscript.

10. The reviewers also suggested extensive editorial changes that would make the manuscript easier for the reader to follow.

Based on the reviewers' suggestions, we reorganized the Results section to better communicate the idea that Nord is a dosage-dependent biphasic modulator of BMP signaling. We also shortened the discussion by mainly focusing on the mechanistic aspects of Nord/Ndnf-mediated regulation of BMP signaling.

Reviewer #1:

The authors used cell sorting and genome-wide expression profiling to identify new Hedgehog (Hh) target genes in the developing Drosophila wing disc, which has been a powerful system to study Hh signaling. They identify nord, which encodes a secreted factor that associates with cell membranes and the extracellular matrix. Null mutants of nord are adult viable. Paradoxically, these animals have slightly smaller wings and ectopic posterior cross vein (PCV) material. Both phenotypes can be caused by alterations in BMP signaling. They demonstrate that recombinant Nord binds to Dpp and Gbb and restricts their release from cells and diffusion in the media. They use the Gal4/UAS mis-expression to test the model that Nord has both positive and negative effects on BMP signaling depending on where and when Nord is expressed.

The strengths of this paper are the appropriate use of techniques such as cell sorting, transcriptomics, Drosophila loss of function genetics, cell culture. These approaches support their model that Nord regulates availability of Dpp, Gbb and/or Dpp-Gbb dimers. However, the use of mis-expression of wild-type Nord in the pouch or posterior compartment lead similar phenotypes as global nord loss-of-function. This is a weakness because it is difficult for the reader to reconcile logically these paradoxical results. This work may impact the mammalian fields because Nord is conserved through evolution, but Nord homologues have not been implicated in BMP signaling.

We want to thank the reviewer for the time and effort invested in reviewing our manuscript, as well as the instructive comments. In the revised manuscript, we have provided additional experimental evidence to support our model that Nord is a biphasic regulator of BMP signaling.

In addition, extensive editorial changes were made to help the reader to reconcile logically the paradoxical results associated with the biphasic nature of Nord in regulating BMP signaling.

a) While I appreciate that they authors have done a lot of work and the data are of high quality, I don't follow the reasoning of how similarity in loss- and gain-of-function nord phenotypes supports their model (or Nord having both positive and negative effects on BMP signaling depending on where and when Nord is expressed). Could the authors use available Gbb and Dpp antibodies to determine the expression patterns of these proteins in nord LOF and GOF?

We have provided additional experimental results to support the model that Nord is a dosage depend biphasic BMP modulator, where low levels of Nord promote, and high levels inhibit BMP signaling. The biphasic feature of Nord was observed both in the larval wing discs during wing growth and in the pupal wing in the process of PCV patterning. Additionally, we show the biphasic activity of Nord on BMP signaling in cultured S2 cells. These data are presented in Figures 6-8 in the revised manuscript.

The suggested experiment was performed, in which no noticeable change of the Dpp (anti-Dpp prodomain) expression levels were detected in the nord mutant wing discs. These data are presented in Figure 5—figure supplement 2 in the revised manuscript.

Related to this, the authors do not control for the number of UAS transgenes in Figure 7. If you want to compare 3X UAS-Nord to 1X UAS-Nord, the 1X UAS-Nord sample should have 2X UAS-LacZ, GFP or another neutral protein. Being able to prove that they have not titrated Gal4 is critical to their model about low vs high Nord. These experiments should be performed again with all animals have the same number of UAS transgenes.

The suggested experiment was performed and is included in Figure 7 in the revised manuscript.

b) The authors do not provide a description of micro-array analysis. Additionally, there is no mention of the deposition of these data into public repository like NCBI GEO, to which the reviewers should be provided access. Related to this, there is no information on fold change of nord among the various sorted populations. They discuss top-ranked genes, but none of these data are presented. All of these need to be addressed.

In the method section of the revised manuscript, we now have provided the GEO information related to the array. See Page 30.

We also added new figures (Figure 1; Supplementary file 1) to provide detailed information about how nord was selected from the array analysis. See Page 7.

c) We need more information about the Halfon hs-Gal4 technique that they used. Additionally, there are too many "clones" in 1H and we don't know where the "clones" are in Figure 1G. Could the authors use FLP/FRT flip-out approaches to generate a few discrete clones that mis-express Hh or Ptc?

The suggested experiment was performed and included in Figure 2C-H in the revised manuscript. Briefly, we found that Nord expression is cell-autonomously induced/inhibited by Hh signaling within the anterior compartment in the wing discs’ hinge and outer wing pouch region. See Pages 9-10.

d) The reader needs more information about the nord homologs that were examined in cell culture in Figure 2. For example, what is DrNdnf?

As suggested, the definition of Dm, Dr, Hs, Ce, and Xt were added into the legend of Figure 3 (original Figure 2). See Pages 39-40.

e) Figure 2C, the labels should be M and L, not M and C

Suggested changes were made in Figure 3 (original Figure 2) of the revised manuscript.

f) The font size is too small in Figure 2A. B, Figure 3D, Figure 3F, Figure 3M. Please enlarge.

Suggested font size change was made in Figures 3 and 4 (original Figures 2 and 3) in the revised version.

Reviewer #2:

Wu et al. identified nord as a potential Hh-target gene in the Drosophila wing imaginal disc. Nord shares sequence similarity with a family of secreted proteins containing a domain of unknown function (DUF2369) and one or two fibronectin type II-like repeats which contains the neuron-derived neurotrophic factor (NDNF) which has been shown to promote neuronal migration and neurite outgrowth in vertebrates, such as the gonadotropin-releasing hormone producing neurons critical in the regulation of puberty and reproduction. The authors explored the function of nord in wing development. Hh is known to be secreted from cells in the posterior compartment of the wing disc and act on cells across the A/P boundary where an adjacent stripe of anterior cells express the Hh receptor ptc. Interestingly, nord is expressed in a subset of these 'stripe' cells consistent with it being a Hh-target. The dpp gene is a well known Hh-target expressed throughout the 'stripe' like ptc. dpp encodes a BMP2/4 ligand which is known, with the Drosophila BMP5/6/7 ligand, Gbb, to play multiple roles in wing growth and patterning. Wu et al. find while flies homozygous for LOF mutations in nord are viable and fertile, their wings are slightly smaller and they contain ectopic vein material at the posterior cross vein (PCV). A reduction in BMP signaling results in smaller wings but an increase in BMP signaling is associated with PCV spurs. Based on these phenotypes, Wu et al. investigate the potential relationship between nord function and BMP signaling. BMP signaling in the larva and its roles in growth and patterning and its later roles in the pupal wing disc in laying down vein material are temporally separable events. Furthermore, both the Blair and O'Connor labs have nicely delineated a mechanistic explanation for BMP involvement in the formation of the PCV. How nord is involved in each of these processes is not completely clear, especially given nord's expression pattern. Must the spatial and temporal requirements for BMPs in these processes be re-evaluated?

Is there a connection between how NDNF influences the migration neurons and its extracellular role in modulating molecules such as BMPs? Does it provide a link with extracellular matrix and other extracellular molecules that also appear to be modulators of BMP activity? Could Nord impact BMP output by influencing wing disc cellular projections?

Strengths:

– Identification of a Hh-target gene that appears to modify BMP signaling output.

– Ectopic vein material near the posterior cross vein (PCV) is attributed to loss of nord.

– A loss of function nord mutants exhibit a reduction in wing size.

– Nord-GFP is shown to bind to Dpp, and less so to Gbb, by co-immunoprecipitations in S2 cells.

Weaknesses:

– The manuscript would benefit from stating at the start that the researcher's findings suggest that nord may both facilitate and antagonize BMP signaling. It will be easier for the reader to follow.

– Better clarification of the different functions for BMP signaling in wing growth, patterning, and development in the larval versus pupal period would strengthen the manuscript.

– The authors can then follow up on these points as to how nord impacts these different functions

– For the most part nord is not expressed in either wing pouch cells or vein primordia. Furthermore, based on expression in cell culture, Nord is not found far from the source cell. How is Nord though to influence pMad, for example, in cells in the anterior wing pouch?

– It is import to reconcile the findings of others that the pMad gradient is not required for wing disc growth.

We want to thank the reviewer for the very detailed and constructive comments towards improving the quality and presentation of our manuscript. In the revised manuscript, we have provided additional experimental evidence, with better spatial (Figure 5, compared to original Figure 4) and temporal (Figure 7 compared to original Figure 7) resolution, to support our model that Nord can both facilitate and antagonize BMP signaling in a dosage-dependent manner. In addition, as suggested, the abstract and majority of the result section were re-written, and several figures were re-formatted to help the readers follow. The specific changes were explained below in the point-by-point response.

As for the three major questions raised by the reviewer, we commented on the second question “Does Nord provide a link with extracellular matrix and other extracellular molecules that also appear to be modulators of BMP activity?”, which is included in the Discussion section under “Nord is a novel, multi-functional BMP binding protein.” However, at this point, we would not be able to comment on the other two, which will be explored in future studies.

General points:

It would be helpful to the reader if the authors stated at the outset that they are suggesting that the impact of nord on BMP signaling can be both as a facilitator and an antagonist, and then go on to present data and the contexts in which they see different behaviors.

In the revised manuscript, the abstract and majority of the result section were re-written by following reviewer 2’s suggestion. See Pages 2, 5, 14-22.

In the wing disc, pMad represent the output of BMP signaling, both the action of Dpp and Gbb. In most instances it is impossible to separate the effects of the two ligands, whether they act as a heterodimer or as homodimers, therefore, in nearly all places in the text "Dpp signaling" should be replaced with "BMP signaling", unless the authors know for certain they are examining signaling induced solely by Dpp.

Changes were made throughout the text as suggested.

Methods:

All Drosophila strains should be listed under "Drosophila strains"

Method used for overexpression of Hh or Ptc should be presented under some other subheading than Drosophila strains.

In the revised manuscript, we have listed all the used Drosophila strains in the Key Resources Table. The original “Drosophila strains” section was renamed as “Drosophila maintenance.”

Third instar larval period is 48hrs at 22°C. If animals were heat shocked in early third and then dissected after two days they should have pupated. More accurate documentation of developmental time should be provided, or the authors should indicate that the larvae were anywhere between X and Y hrs or days old.

Changes were made as suggested under the method section “Drosophila maintenance.”

Epitomics 1880-1 was raised against phospho-Smad3, not pMad.

The description of antibody used to recognize pMad was corrected under Method section under “Antibodies.”

Constructs – the nucleotide sequences of the Nord and Ndnf – GFP and 3xMyc fusions should be shown.

Information about Nord/Ndnf expression constructs is provided in the Key Resources Table and Supplementary file 3.

Dpp-FLAG, Gbb-FLAG, Gbb-HA, and Mad-FLAG are not described in Serpe et al. 2005 – a description and characterization of activity should be provided

Information about the constructs used to express tagged Dpp, Gbb and Mad is provided in the Key Resources Table and Supplementary file 3. We also corrected the related citation in the Method section under “Constructs.”

Section "RT-qPCR" needs to be renamed to indicate it contains the description of imaginal disc cell isolation. Were these the cells and RNA used for microarray analysis referred to in the text? Imaginal discs and gene expression patterns change significantly during the 3rd larval instar. At what time were wing imaginal discs harvested for RNA isolation and immunohistochemistry?

As suggested, we have separated the “Dissociation and sorting of imaginal disc cells” section from the “Quantitative reverse transcription PCR” section. The cells were dissociated from wondering third instar larvae, which is added into the revised method section. The procedure for FACS are included under “Dissociation and sorting of imaginal disc cells” and microarray were carried out under standard Affymetrix labeling and hybridization protocol, which was provided when depositing the GEO raw data.

Results

Figure 1 – when first describing nord gene expression on page 7 Figure 1D & E, its absence from the wing pouch should be noted and that it is expressed anterior to the A/P boundary in the hinge progenitors.

The suggested change was made in the revised manuscript. See Page 6 (first paragraph).

No description of procedure for FACS or microarray was provided.

Description of Nord-GFP in discs following over expression of Hh or Ptc is confusing (bottom of page 8) as the tissues are derived from larvae not flies.

hs>hh leads to expansion of Nord-GFP in hinge and wing periphery

hs>ptc breadth of Nord-GFP domain narrowed, only remaining is at A/P boundary.

The suggested change was made in the revised manuscript. See Page 9 (second paragraph).

Figure 1 H – do the author's have an explanation for the uneven expression of ptc and the corresponding response of Nord-GFP expression? Nord-GFP appears to be lacking in cells where Ptc is present in Figure 1G but Nord-GFP expression and Ptc are present in quite a number of cells in the hinge region of a wild type wing disc Figure 1F.

The uneven expression (staining) of Ptc is due to the fact that anti-Ptc detected both endogenous Ptc and exogenous Ptc from UAS-Ptc driven by the Hs-Gal4. In the revised manuscript, we generated a Nord-RFP protein trap allele and examined Nord expression in Flipout clones with increased/reduced Hh signaling activity. In this experiment, the clones were labeled with the expression of mCD8-GFP, which allowed to avoid the use of Ptc antibody. We showed that Nord expression was cell-autonomously induced/inhibited by Hh signaling within the anterior compartment in the wing discs’ hinge and outer wing pouch region. These new data were included in Figure 2C-H in the revised manuscript.

Figure legend: error? – "White brackets indicate the expression range of Ptc (E) or No-GFP (E-H)".

The figure legend error was corrected in the revised manuscript.

Figure S2 panel A – include a scale in bp.

The scale in bp was provided in the revised manuscript in Figure 2—figure supplement 2.

Fig S2 panel E – why is the expected 732bp band seen when P3 + P4 are used to amplify MI06414-GFSTF.2?

It is likely due to alternative splicing by skipping the inserted cassette. Alternatively, it could due to a small portion of heterozygous flies was used during RNA extraction.

S2 figure legend lacks reference to panels D-F.

The figure legend error was corrected in the revised manuscript.

Figure S3 – sequencing described?

Residue numbers should be included in figure.

The scale in bp was provided in the revised manuscript in Figure 4—figure supplement 1.

Figure 2 – should note that melanogaster was used – provided species name for other genus' clone names should be listed in figure legend, ie which species the Ndnf gene was derived from;

The species name was added into the legend of Figure 3 (original Figure 2) in the revised manuscript.

panel B – no size markers.

The size marker was added into Figure 3B (original Figure 2B).

Lanes in figure B marked "M" and "C" but in figure legend noted as "medium (M)" and "cell lysate (L)".

The figure label error was corrected in the revised manuscript in Figure 3B (original Figure 2B).

Indicate that the resulting Ndnf proteins were tagged with Myc – in figure legend.

The suggested information was added into the revised legend of Figure 3 (original Figure 2)

Explanation for difference in size of DmNord found in media versus cell lysate?

Secreted Ndnf/nord may undergo additional posttranslational modification, which lead to the slower migration of Ndnf/Nord in the Media compared to that in the cell lysate. Alternatively, the salt concentration in the media vs. lysate may also contribute to the different migration speed.

What are the predicted sizes for Nord/Ndnf proteins?

The predicted size for Nord (587 aa) /NDNF (568 aa) proteins is around 65-70 kDa, however we observed that these proteins migrate a slower than the predicted size, especially the secreted potion, which is likely due to post translational modification.

Ideas about the differential abundance of Ndnf in media from lysate from different species.

We are not sure about why there is differential abundance of Ndnf in the media from lysate from different species.

An important control is needed – WT disc with Anti-GFP.

A WT disc with Anti-GFP surface staining was added as the control in to Figure 3E (original Figure 2D).

Is it known that MI0641 produces a truncated peptide, or is it predicted?

We have modified the text to clarify that the truncate peptide is predicted.

Does higher prevalence of PCV spur in females suggest a second site enhancer? Dosage compensation?

Higher prevalence of PCV in females was also observed in man1 mutants (Wagner et al., 2010 Dev Bio). It was thought to related to the differences in wing size. We are not entirely clear about the mechanisms underlying the PCV+ associated female preference.

What is the correlation between ectopic vein formation in the pupal wing and expression in the larval disc?

We propose the ectopic vein formation in the pupal wing is related to Nord expression in the early pupal wing discs (Figure 5A lower row; Figure 5—figure supplement 1), not that in the larval wing disc.

Authors should be careful when describing protein versus RNA expression patterns, as well as loss of gene function versus loss of gene expression. For example on page 12, neither deCelis 1997 or Yu et al. 1996 examined Dpp (protein) expression They report in situs (endogenous RNA) or lacZ enhancer trap expression patterns.

We changed “Dpp” to “dpp” in the revised manuscript. See Page 13 (bottom line).

In the sentence "We asked whether elimination of Nord alters level of pMad" – should be rephrased " We asked if the knock down (or reduction) in nord function …"

pMad is a downstream signal transducer, not a transcription transducer.

The suggested clarification was made in the revised manuscript. See Page 14 (second paragraph).

Harmansa et al. 2015 proposed that dpp spreading is required for medial but not lateral wing growth. How would this finding apply to the proposed function of nord with respect to wing growth?

We measured L2-L5 distance (dL2-L5) in the revised manuscript and found that dL2-L5 is significantly reduced in the nord mutant wings (Figure 4C). The reduced dL2-L5 indicates Nord-Dpp interaction is likely required for the medial wing growth. See Figure 4C.

Harmansa 2017 showed that eGFP-Dpp dispersal in basolateral is required for patterning & growth. Does Nord have subcellular localization?

Nord localized to both apical and basolateral side of the wing epithelium.

An antibody against the Dpp prodomain is used in Figure 4A. BMP proproteins are cleaved separating the prodomain from the ligand domain proper, with the ligand sometimes secreted with its prodomain and other times, not. Do the authors feel this pattern reflects that of the secreted Dpp ligand?

We could not detect any specific signal when we used anti-Dpp prodomain antibody in the absent of detergent (surface staining protocol). The Dpp signal in the original Figure 4 (Figure 5 in the revised version) should come from the Dpp prodomain that stayed inside of the producing cells.

Nord-GFP is in a subset of Dpp prodomain expression in larval wing disc (Figure 4A) and shows even less overlap in pupal wing disc (Figure 4B) – this needs to be more clearly addressed in the text.

The suggested clarification was made in the revised manuscript. See Page 13-14.

Data not shown for Nord-GFP in later pupal wing when dpp-lacZ transitions to longitudinal veins.

Nord-GFP (co-stained with anti-Dpp prodomain antibody) expression in later pupal wing is provided in Figure 5A (lower row) in the revised manuscript.

"While dpp is also known to be expressed in the developing longitudinal veins (deCelis, 1997), we see no Nord expression in either the longitudinal vein or crossvein despite the fact that Nord loss of function leads to ectopic crossvein formation." The data referred to in the underlined text top page 13 is not shown. (?)

The related Nord expression data is presented in the lower row of Figure 5A of the revised manuscript.

Figure 4E – Is quantification plot an average of multiple discs? how many discs were measured? Where was the measurement taken? Is it a profile from a line across the disc or a compilation of all signal in the disc? Would the authors expect pMad to be altered in the wing pouch?

Figure 4F – not at all clear what is meant by "maximum pMad intensity in the A and P compartment" in the figure legend. Is this the highest intensity measurement in the entire disc?

Is the proposal that the slight reduction in pMad in the anterior affects only growth and patterning of the position of the wing veins?

The quantification procedure was described with more detail in “Image collection and quantification of fluorescence intensity”. In the new Figure 5C, we present pMad intensity profile plot as the mean ± SD from n = 15 wing imaginal discs. In the nord mutants, reduced pMad signal was observed, which we thought is due to increased spreading of Dpp when Nord is absent. The increased spreading will likely cause reduced local Dpp concentration, thus reduce BMP signaling in the medial wing discs (see the proposed model in Figure 9).

If Nord is required to facilitate "Dpp" movement, then why is signaling activation (as measured by pMad) in the majority of the anterior compartment beyond the peak unchanged. If in the absence of Nord, Dpp does not move, then one would expect all signaling in the anterior to be affected.

We propose Nord-mediated binding of Dpp (and/or Dpp-Gbb) interferes with (not facilitate) Dpp movement. This is because Nord is associated with the matrix, and does not spread too far from the producing cells (Figure 3D, Figure 7D). Also, we detected much less Dpp/Dpp-Gbb ligands in the medium when Nord is co-expressed (Figure 8B). We propose that in the absence of endogenous Nord (low level), Dpp spreading will increase, which in turn lead to the reduced medial BMP signaling and associated with decreased wing growth.

Instead of facilitating movement, could Nord prevent Dpp-Gbb turnover, thus in its absence Dpp-Gbb is less stable and a drop in signaling (pMad) is observed. Difficult to understand how Nord has an effect on ligands when its not present in wing pouch cells.

Instead of “impeding (not facilitating)” Dpp movement, we agree endogenous Nord may also prevent Dpp-Gbb internalization and turnover, thus enhancing BMP signaling. However, these models need to be tested in future experiments.

Secreted Nord, although it does not travel too far, may explain its potential effect in the center wing pouch.

Top half of pg 16 – authors suggest that media from cells co-expressing Nord and Dpp/Gbb elicit less signaling (pMad) than media produced by cells without Nord -if the same amount of media is added to cells in Figure 5C then it seems that M4 media is less effective at inducing pMad compared to the lower amount of Dpp/Gbb in M5 which elicits nearly the same level of pMad.

What does a normalized plot show?

We thank the reviewer for pointing out an additional interpretation of Figure 8C (original Figure 5C). In the revised manuscript, we added panel D to show the relative signaling activity of an equal volume of conditioned medium (pMad/Mad). The medium signaling activity was then normalized to the ligand amount (band intensity in panel B) to calculate the normalized ligand activity [(pMad/Mad)/ligand amount], which was found higher in the conditioned medium with Nord (Figure 8E), suggesting released Nord may enhance Dpp/Gbb activity. We expanded this observation and showed that Nord could modulate recombinant Dpp activity in a biphasic manner in vitro (Figure 8F-I). These in vitro results are consistent with what we found in vivo (Figures 4, 6, and 7).

Using a recombinant Dpp peptide in Figure 5D, the authors find that in the presence of Nord-GFP, equivalent amounts of rDpp induce less pMad but this is the opposite of what is observed in vivo in the disc where signaling is reduced in nord mutants. Perhaps the authors can expand on this observation. Is this result related to the fact that a recombinant peptide was used in rather than the entire Dpp proprotein or prodomain + ligand domain?

In Figure 8H (original Figure 5D), high levels of exogenous Nord were expressed under the actin promoter (pAcSV plasmids) in transiently transfected S2 cells, therefore, only inhibitory effects of Nord were observed here, which is similar to the inhibitory effects of high levels of Nord produced by en>3XUAS-Nord or nub>UAS-Nord at 29°C.

Figure 5E in WT – seems odd that one cannot see increased concentration of Mad-Flag in the nucleus.

This is likely due to (1) FLAG-Mad is expressed at a high level in the stable S2 cells, and (2) very low dose of Dpp (0.625 nM) was used. Thus, only a small portion (for example <5%) of total FLAG-Mad is phosphorylated, and its subcellular localization change will not be evident when visualized with the FLAG antibody.

Suggest that Nord-GFP could prevent release of ligand from cell surface and inhibit ligand from media to act on responding cell – what about ligand-receptor binding?

It is very likely that Nord functions via additional BMP interacting proteins, including the receptors. We propose that low levels of Nord enhance and high levels inhibit BMP signaling likely by promoting or interfering ligand-receptor binding either directly or indirectly, which will be tested in future studies.

Westerns in Figure 5 are cropped and it is not possible to see the other cleaved forms of ligands. Does Nord show differential interaction with different ligand forms, as has been shown by Anderson 2017 to be the case with different forms of Gbb and receptors?

In the medium used in the IP experiment, majority of the ligands, as showing in Figure 8A, are smaller than 25 kDa (see the uncropped blot in Figure 8-source data 1). We also detected a smaller fraction of other isoforms of Dpp /Gbb (*), and further studies are required to figure out the interacting domain on Dpp/Gbb and Nord.

Figure 6

How were pMad profiles measured? Superimposing profiles in D-F would be helpful.

Authors state ptc levels have not changed – provide measurement of ptc levels.

Suggested changes were made in the revised Figure 6. The procedures used to measure pMad and Ptc was described with more detail in “Image collection and quantification of fluorescence intensity”.

Loss of L5 and PCV are typical of gbb LOF phenotypes and not seen in dpp LOF genotypes, therefore, it is hard to reconcile these phenotypes with if Nord is thought to preferentially affect dpp.

Based on Posakony, et al., 1991 and Ray and Wharton 2001, dpp is required for the distal half of L5. In Figure 6 and the original Figure 7 (replaced by a new Figure 7), high levels of ectopic Nord (driven by nub-Gal4 or en-Gal4) indeed only caused loss of the distal half of L5. Thus, these observations are consistent with the model that ectopic Nord impedes Dpp spreading, thus affect L5 growth due to reduced Dpp ligands available in the distal wing discs.

In Figure 6H – nub/+ label could easily be confused with a nub mutant heterozygote.

A nub-Gal4/+ 29°C control wing should be shown.

The suggested changes were made and the suggested control (UAS-GFP) were added in the revised Figure 6.

In some places nord is indicated as a secreted protein and in others membrane-associated – a little bit confusing.

Nord is a secreted protein, but associated with the extracellular matrix and not freely diffusible. To avoid confusion, we now changed “membrane-associated” to “matrix-associated” in the revised version.

Akiyama 2015 showed that the pMad gradient in the wing pouch, as well as Dpp were not required for wing growth. How do the authors think changes in the pMad profile may change growth?

In Akiyama 2015, their conclusion is “These results indicate that during the latter half of larval development, the Dpp morphogen gradient emanating from the anterior–posterior compartment boundary is not directly required for wing disc growth.” There are also publications (Barrio and Milan, 2017 eLife) demonstrated that “Boundary Dpp promotes growth of medial and lateral regions of the Drosophila wing”. We propose that ectopic Nord (driven by nub-Gal4 throughout larval development) impedes the spreading and inhibits the activity of Dpp (reflected at the level of pMad intensity), which in turn inhibits the growth of the wing.

It is curious that changes in the pMad gradient seen in Figure 6F has no effect on longitudinal vein development in Figure 6K.

Based on Posakony, et al., 1991 and Ray and Wharton 2001, dpp is required for the distal half of L5. In the original Figures 6 and 7 (replaced by a new Figure 7 to void the L5 phenotype), high levels of ectopic Nord (driven by nub-Gal4 or en-Gal4) indeed only caused loss of the distal half of L5. Thus, these observations are consistent with the model that ectopic Nord impedes Dpp spreading, thus affecting L5 growth due to reduced Dpp ligands available in the distal wing discs.

The Discussion contains a considerable amount of background information that is not particularly relevant to the findings presented in manuscript. Focusing the discussion a bit more would be helpful to the reader.

As suggested, we shortened the discussion.

To the reader it is not at all clear how expression in the posterior compartment during larval period influences PCV formation in the pupal wing when ligands are thought to be transported from the longitudinal veins.

We propose the ectopic vein formation in the pupal wing is related to Nord expression in the early pupal wing discs (Figure 5A lower row; Figure 5—figure supplement 1), not that in the larval wing disc. The text was extensively reorganized to better communicate this message. See Pages 13-22.

Does uniform expression in posterior using enGal4 interfer with the normal mechanism?

In the original Figure 7, we showed en-Gal4 driven expression of Nord during the larval stage and pupal stage caused reduced wing size, especially the posterior compartment size.

However, en-Gal4 driven expression of Nord during the pupal stage has no noticeable effect on wing size (Figure 7 of the revised version).

Suppl Figure 12 – orientation of high mag of trichomes is incorrect

The orientation of the zoomed area was corrected in Supplementary file 4 of the revised version.

Overall, please check labelling of figure for units, number of samples (n=?), and accuracy of figure legend description. In some places in the manuscript this information is robust but in other places it is lacking.

Suggested changes were made in the revised version.

Reviewer #3:

The study focuses on a possible role for the secreted Nord protein in BMP signaling in Drosophila. The authors demonstrate Nord binding to Dpp and Gbb in vitro, and suggest that this explains the in vivo phenotypes observed. The in vitro work is quite suggestive, but some caution may be in order as other other extracellular molecules can affect multiple pathways, and some of these other pathways can have have similar effects on growth and vein formation. While this is a worthwhile publication with strong in vitro data, I have a few points below that I think would strengthen the in vivo data.

We thank the reviewer for the instructive comments and suggestions. As suggested, we have performed additional experiments and analysis to strengthen the in vivo data and clarify the model of how Nord functions.

1) Nord as a Hh target- The expression of Nord in the imaginal and pupal wing tissues largely mirrors that of Hh targets, being heightened anterior to the anterior-posterior compartments boundary. Its expression is however excluded from the distal wing blade, something that the authors do not further investigate. The authors can widen or narrow Nord expression by widening or reducing Hh signaling, but this will also widen BMP expression and signaling, so it would have been helpful to have one experiment showing a cell autonomous effect of Hh signal reception on Nord expression.

The suggested experiment was performed and included in Figure 2C-H in the revised

manuscript. Briefly, we found that Nord expression is cell-autonomously induced/inhibited by Hh signaling within the anterior compartment in the wing discs’ hinge and outer wing pouch region.

2) Nord mutants- The authors generate and examine Nord mutants. The effects in the wing are fairly mild, consisting of very slight undergrowth mirrored by slightly reduced BMP signaling (pMad) in the disc, and a few ectopic veins, suggestive of gains in BMP signaling. The only other growth-inducing pathways the authors examine are Hh, looking at a single target, Ptc. They do not however quantify Ptc staining in the same way they did to detect the weak effects on pMad, nor do they look at other Hh targets with different sensitivities.

In the revised manuscript, we measured the Ptc staining profile and examined the expression of a low-threshold Hh signaling target gene dpp (anti-Dpp Prodomain) in the nord mutant wing discs. No noticeable change of the Ptc or Dpp (anti-Dpp prodomain) expression levels were detected in the nord mutant wing discs. These data are presented in Figure 5 and Figure 5-figure supplement 2 in the revised manuscript.

Additionally, we measured the distance between distal ends of longitudinal veins L3 and L4 (dL3-L4) and that of L2 and L5 (dL2-L5), we found no specific reduction in the L3-L4 region indicated by comparable or higher dL3-L4 /dL2-L5 ratio in the female and male nord mutant wings (Figure 4B, C), which is consistent with the obviously reduced BMP, but not Hh, signaling activity during wing imaginal disc development (Figure 5B, C). Together, these findings suggested a positive role of endogenous Nord in augmenting BMP signaling activity to promote wing growth. See Page 14-15.

The ectopic veins in mutants are consistent with defects seen with mild increases in BMP signaling, but there are many other ways of inducing similar ectopic venation, especially through EGF gains or reductions in Notch or its ligands. The authors concentrate almost exclusively on one defect, a vein extension distal to the PCV. This is a common defect seen in many mutant fly stocks as it is something of a "hot-spot" for ectopic venation. The authors show that pMad is high in this ectopic vein, but do this staining at 26hr AP, a fairly late stage of PCV development. By this time LVs and cells near the ACV have all been expressing dpp for some time and have BMP signaling, while dpp expression in the PCV is initiating. I would wager this ectopic vein is expressing dpp, and will also have rhomboid and thus EGFR activity and high Δ and Serrate. The authors could greatly strengthen their case by examining pMad at earlier stages of PCV development (20 hr AP) before EGFR and Notch signaling are active in this region.

The suggested experiment was performed and included in Figure 5D-E in the revised manuscript. Expansion of pMad range around the primordial PCV was noticed in the nord mutants starting from 19-20 hr AP.

The authors also do not test whether the posterior defects they see are caused by loss of anterior Nord, or whether they might be caused by loss of low-level Nord expression in the posterior. Since they have a Nord-RNAi line, what happens with anterior knockdown? Posterior?

Our findings show that nord, like ptc and dpp, is a target gene of the Hh signaling pathway and not expressed in the posterior compartment of the wing discs (Figures 1, 2; Figure 2-figure supplement 3). Consistently, when we knocked down nord in the posterior compartment using hh-Gal4, we did not observe the ectopic PCV phenotype, which is frequently caused by knocking down nord throughout the wing discs via MS1096-Gal4 or A9-Gal4. These data are presented in Figure 5—figure supplements 3 and 4 in the revised manuscript.

3) Nord overexpression and biphasic effects on BMP signaling – Increased Nord expression causes undergrowth and reduced pMad but ectopic venation. In that sense, Nord has biphasic effects, with similar mixed effects (both BMP losses and gains) caused by either too much or too little Nord. The authors extend this by showing that further increases in Nord overexpression change the ectopic PCV extension into PCV loss, the latter often caused by loss of BMP signaling. However, this loss may not always be due to a direct effect on the PCV. While the nub-gal4 example looks ok, the en-gal4 example has lost L5 adjacent to the PCV, likely due to loss of the distal L5 primordium at the disc stages (Ray and Wharton, Bangi and Wharton). That will in turn block or reduce dpp expression in L5 and thus Dpp's availability for signaling in the early PCV.

High levels of ectopic Nord driven by nub-Gal4, en-Gal4 or hh-Gal4 in the larval stage can sequester BMP ligands and inhibit BMP signaling, thus leading to loss of L5 (Figure 6; original Figure 7). To avoid affecting L5 development, we restricted ectopic Nord expression to the pupal wing and then studied its effect on PCV patterning. The temporal control of en-Gal4/hh-Gal4>UAS-Nord was achieved by adding the tub-Gal80ts allele. In this new setting, in the presence of normal L5, ectopic Nord caused PCV phenotypes in a dosage-dependent manner.

These data are presented in a new Figure 7 in the revised manuscript.

4) in vitro studies- These provide the strongest proof that Nord can affect BMP signaling, although in this case the observed effects are all inhibitory. Secreted Nord can co-IP with secreted Gbb or Dpp, and co-expression with Nord reduces Gbb or Dpp secretion and thus its ability to signal. Nord can also reduce the BMP signaling induced by addition of exogenous Dpp. The only shortcoming here is that the authors have not apparently tested binding to any of the extracellular modulators of BMP signaling.

Interaction of Nord and the BMP binding protein Dally was presented in a manuscript co-submitting with ours. We discussed the possibility that Nord can bind both BMPs and Dally in regulating wing growth and vein patterning in the Discussion section under “Nord is a novel, multi-functional BMP binding protein.”

I need to see whether there is ectopic pMad in the nord mutant PCV at a stage before dpp, rhomboid and Δ/Serrate expression in the region. Otherwise I am not convinced that the ectopic vein is initiated by ectopic BMP signaling.

The suggested experiment was performed and included in Figure 5D-E in the revised manuscript. Expansion of pMad range around the primordial PCV was noticed in the nord mutants starting from 19-20 hr AP.

Quantify Ptc or some other Hh marker (Dpp would be especially helpful) in Nord mutants.

In the revised manuscript, we have measured the Ptc staining profile and examined the expression of a low-threshold Hh signaling target gene dpp (anti-Dpp Prodomain) in the nord mutant wing discs. No noticeable change of the Ptc or Dpp (anti-Dpp prodomain) expression levels were detected in the nord mutant wing discs. These data are presented in Figure 5 and Figure 5—figure supplement 2 in the revised manuscript.

How much of the PCV loss in en>Nord might be caused by loss of the adjacent L5?

High levels of ectopic Nord driven by en-Gal4 or hh-Gal4 in the larval stage can sequester BMP ligands and inhibit BMP signaling, thus lead to loss of L5 (original Figure 7). To avoid affecting L5 development, we restricted ectopic Nord expression to the pupal wing and then studied its effect on PCV patterning. The temporal control of en-Gal4/hh-Gal4>UAS-Nord was achieved by adding the tub-Gal80ts allele. In this new setting, despite essentially normal L5, ectopic Nord caused PCV phenotypes in a dosage-dependent manner. These data are presented in Figure 7 in the revised manuscript.

Cell-autonomous effects of Hh signaling on Nord expression would be helpful, as would the effects of anterior vs. posterior Nord RNAi.

The suggested experiment was performed and included in Figure 2C-H in the revised manuscript. Briefly, we found that Nord expression is cell-autonomously induced/inhibited by Hh signaling within the anterior compartment in the wing discs’ hinge and outer wing pouch region.

[Editors’ note: what follows is the authors’ response to the second round of review.]

Essential revisions:

1) Soften their conclusions about their in vitro work (Figure 8) in the results and discussion. Add a statement to the effect that without additional replicates in Figure 8 other conclusions are possible.

We include additional replicates and statistical analysis in Figure 8G. See changes in the Figure 8 legend on Page 45. We also softened the conclusion on Page 22 of the revised version.

2) Provide additional data on the Nord in the posterior compartment using ptc-gal4, UAS-nord-RNAi.

Suggested experiments were performed and included in Figure 5—figure supplements 3 and 4.

3) Add information about the cross vein phenotype of MS-1096 to the manuscript.

As suggested by the reviewer, we removed MS1096-Gal4 related data in Figure 4.

4) Make changes to the Figure fonts.

Suggested changes (see below) to the figure founts were made in Figures 1, 2, 4, and 7.

5) Correct the typos and awkward sentences.

Suggested changes (see below) were made.

Reviewer #1:

The authors have made strong efforts to respond to the essential revisions and the reviewers' critiques. Overall, the manuscript is much improved. However, I have concerns about the conclusions regarding Figure 8, the in vitro work, which the authors use conclude that Nord in vitro also has biphasic function on BMPs in vitro. The authors present results from a single experiment and then normalize these single data points to draw conclusions about Nord function in vitro. Specifically, the authors' new analyses in Figure 8 panels D,E,G show relative pMad levels normalized to Mad (8D,8G) and to Mad and BMP (8E). My concerns are as follows. First, the authors need statistical analyses. Second, BMP released from cells appear to be equally active in the absence (M4) or presence (ME) of Nord, and these data only become supportive of their argument when "normalized" to the amount of ligand. Third, there are additional bands consistent with other Dpp/Gbb isoform in the source blot for Figure 8A. These may be bioactive molecules but are not accounted for in the manuscript. The pMad bands in M4 and M5 also appear to be saturated, again making it difficult to determine differences between the lanes. In panel C, the WB: Flag row is "smiling" with some lanes compacted and others not, again making it difficult to normalize. Without more replicates, the authors statement that "soluble ligand appears to have a higher signaling activity perhaps as a result of an association with Nord" (p.21) is not sufficiently supported by the data. This same issue holds for panel F, with only one replicate. In its current form, the data are not conclusive and other interpretations are also possible. The authors should soften their conclusions about their in vitro work in the results and discussion. Additionally, they need add a statement to the effect that without additional replicates in Figure 8 other conclusions are possible.

We want to thank the reviewer for carefully reviewing our manuscript. We agree with the concerns regarding panels B-E, and therefore added the experiments in F and G, where rDpp was used. In the revised version, we include additional replicates and statistical analysis in Figure 8G. See changes in the Figure 8 legend on Page 45. We also softened the conclusion on Page 22.

They are other issues that need to be addressed:

1. The figure in the response to reviewers was too small. In the future, please upload a larger image.

Suggestions will be followed in the future.

2. P. 7. The authors write "In wing discs, cells near the A/P compartment boundary (B: ptc+) receive the highest level of Hh stimulation while A cells (A: hh-), located further from the border, receive lower levels of stimulation. In contrast, P cells (P: hH+) do not respond to Hh due to lack of the receptor Ptc and the transcription factor Ci (Tabata et al., 1992; Eaton and Kornberg, 1990)." It is very important that the authors add "Figure 1" when discussing B, A and P cells.

Suggested changes were made on Page 7 of the revised manuscript.

3. P. 11 and Figure 3C – please clarity what tissue is being used for the flip-out clones.

Suggested changes were made on Page 11 (Figure 3D) of the revised manuscript.

4. Top of P. 15. Please rephrase "which is consistent with the obviously reduced BMP, but not Hh, signaling activity during wing imaginal disc development (Figure 5B, C)" because it is not clear what is being discussed. I suggest "which is consistent with normal Hh signaling activity during the development of nord mutant wing imaginal discs (Figure 5B, C)."

Suggested changes were made on Page 15 of the revised manuscript.

5. P. 18 The authors cite Robert 1998 but there is no such citation in the references cited section. Please add this reference or remove it from the manuscript.

The missing citation was added to the reference list.

6. P. 21. As noted, please soften your conclusions about the in vitro studies and add a statement to the effect that without additional replicates in Figure 8 other conclusions are possible.

We include additional replicates and statistical analysis in Figure 8G. See changes in the Figure 8 legend on Page 45. We also softened the conclusion on Page 22 of the revised version.

7. P. 25, line 9 "it's" should be "its".

Suggested changes were made on Page 25 of the revised manuscript.

8. P. 25, line 13 "Akiyama et al. 2021 accompanying manuscript" is not in the references cited.

The missing citation was added to the reference list.

9. Add to the Materials and methods the procedure you used for normalization of the data in Figure 8E,D,G.

The procedure used for normalization of the data in Figure 8E, D, G was added to the figure legend (Page 45-46) and Materials and methods (Page 37).

10. Figure 1B – increase the font size of the colored/highlighted genes. Currently we can't read them.

The suggested change was made in Figure 1B.

11. Figure 2A – increase the font size of the RMCE vectors. Currently we can't read them.

12. Figure 2D, E, F, G, H – increase the width of the lines outlining regions of the wing disc and clones. Currently we can't see them.

13. Figure 2C-H, it would be helpful to add C' and C" etc to the unlabeled panels.

The suggested changes were made in Figure 2, as well as the related figure legend.

14. Figure 4E – The graphs are too small – I cannot read anything. Please increase the size of the graphs and the font of the labels.

The suggested change was made in Figure 4E.

15. Figure 5D – increase the width of the lines of the boxes and the enlargements so that they can be seen on a print out.

The suggested change was made in Figure 5D.

16. Figure 7B – increase the width of the lines of the boxes and the enlargements so that they can be seen on a print out.

The suggested change was made in Figure 7B.

17. Figure 2, figure supplement 3. Increase the size of the brackets in the Nord-GFP monochrome panels. Currently, it is very hard to see them.

The suggested change was made in Figure 2—figure supplement 3.

18. Figure 5—figure supplement 2B – label the blue and the red lines.

The suggested change was made in Figure 5—figure supplement 2B.

Reviewer #2:

The authors have added some valuable data that have improved the manuscript, enough for acceptance. I do however have some points I would like the authors to seriously consider. (1) and (2) concern data that are problematic, but could be removed from the manuscript without damaging its conclusions.

We want to thank the reviewer for carefully reviewing our manuscript. We have seriously considered and addressed each concern. Below is a point-by-point response.

1) Hh signaling and Nord expression- The autonomous gains in Nord expression caused by Smo overexpression are convincing. The loss caused by Ptc overexpression, shown for a single flpout clone Figure 3D, is not. To be a valid result, the dotted blue line drawn by the authors must accurately follow the A/P. However, outside the Ptc clone at the bottom of the figure, the dotted line does not at all correspond to the line shown by Nord expression, even though the Nord outside the clone should define the A/P. Nor does the line correspond to the A/P defined by Ci staining. Instead, the true A/P boundary is further left, and thus it is not at all clear that the Ptc clone is in the anterior compartment. If the authors want to keep Ptc flpout data in the manuscript, they need a more convincing example.

We agree with the reviewer’s concern and have removed the related data from Figure 4 and Figure 5—figure supplementary 3. We also removed related text, including the statement "further validates that the observed phenotypes in the nord mutants are not indirect effects associated with loss-of-nord in tissues other than the wing discs" from Page 13 and 15.

2) MS1096-gal4- The authors use MS1096 for some of the nord-RNAi data. This is a Gal4 enhancer trap insertion into the X-linked Bx (dLMO) locus, and in hemizygous males has a mild Bx wing mutant phenotype with variable penetrance that includes disruption of the crossveins (see Figure 6 in Milan, Diaz-Benjumea, and Cohen 1998). The crossvein phenotype tends to disappear in stocks, but often reappears when outcrossing to other lines. It is therefore not an appropriate driver to use when assessing crossvein phenotypes in males. If the authors want to use it, they have to explain how they dealt with the crossvein phenotypes that can be caused by MS1096 on its own.

And while MS1096 is expressed in the wing disc, it is not clear that MS1096 is "wing-specific" as the authors state. Bx (dLMO) is expressed in many tissues, including the CNS. If MS1096 follows the expression of Bx, the authors cannot say that this experiment "further validates that the observed phenotypes in the nord mutants are not indirect effects associated with loss-of-nord in tissues other than the wing discs."

We agree with the reviewer’s concern and have removed the related data from Figure 4 and Figure 5—figure supplementary 3. We also removed related text, including the statement "further validates that the observed phenotypes in the nord mutants are not indirect effects associated with loss-of-nord in tissues other than the wing discs" from Page 13 and 15.

3) Is Nord in the posterior compartment? The authors make several absolute statements about the absence of Nord expression from the posterior of the developing wing, but do not provide any staining of controls that would allow one to distinguish signal from background. While Nord is certainly much higher anterior to the A/P boundary, the data does not rule out low-level Nord expression in the posterior. The authors have addressed this in their revision, not by changing any of their absolute statements about expression, but by added a new experiment showing that "we did not notice any ectopic PCV in the adult flies when UAS-nord-RNAi was selectively expressed in the P compartment of the larval and pupal wing via the hh-Gal4 driver." But this negative result could simply reflect stronger RNAi expression with MS1096-gal4 or A9-gal4. It would be preferable if they had tested the effects of nord-RNAi when it was limited to the anterior compartment, something that their model predicts should affect the posterior crossvein. This could easily be done with ptc-gal4 or ci-gal4.

As suggested, we analyzed the ectopic venation phenotype in flies carrying ptc-Gal4 driven expression of UAS-nord-RNAi (Figure 5—figure supplements 3 and 4). See the related changes on Page 15-16 of the revised manuscript.

The authors also model the action of Nord as if it only diffuses over a short distance, and thus is only affecting BMP movement from neighboring cells, such as those in L4 in pupal wings, and is never present in the PCV. However, the extracellular staining of Nord-GFP in Figure 3 looks substantially above background throughout the wing disc.

We thank the reviewer for pointing out our incomplete interpretation regarding how far Nord may diffuse. We propose that the secreted Nord/NDNF proteins likely exit in two spatially distinct pools: one pool with more freely diffusible proteins that can reach a longer distance, whereas another pool containing proteins mainly associated with source cells or nearby extracellular matrix. See Pages 23-34 and 47 of the revised manuscript for related changes.

4) Biphasic or multiphasic effects- The authors have added new data that clarify some of the effects of Nord in the PCV, but complicate their story about biphasic effects. In growth Nord appears truly biphasic, promoting growth only at normal levels, and inhibiting growth and BMP signaling with either decreased or increased Nord expression. But in PCV formation things are more complex. Both loss of endogenous Nord and mild gains induce ectopic PCV phenotypes. Since normal levels do not alter the PCV, that also fits with a biphasic effect, albeit the opposite of the growth effect (ectopic signaling with both losses and mild gains). Overexpression also has a biphasic effect, as strong Nord gains do the opposite of weak gains and cause loss of the PCV and signaling. However, if one puts all of this together, the PCV effect is more multiphasic than biphasic, because of the normal signaling observed at wild type. Nord loss increases signaling, wild type is normal, mild Nord overexpression increases, strong Nord decreases.

We agree with the reviewer that we need to be cautious about using the word “biphasic” when describing how Nord modulates BMP signaling in PCV patterning. Therefore, we have replaced “biphasic” with “dosage-dependent” when describing the activity of Nord in the context of PCV patterning.

The only detailed explanation of these effects is in the legend of Figure 9, which is not very complete and leaves some puzzles. If loss of Nord increases long-range BMP diffusion into the PCV, why isn't the same thing happening in a third instar wing pouch?

Loss of nord likely also leads to increased Dpp spreading in the 3rd instar wing pouch. For instance, a recent study reported that nord mutation rescued the L5 defects in dally mutants (Akiyama et al., 2021). However, in nord single mutant, the Dpp spreading increase associated gain of BMP activity is not strong enough to overturn the reduced BMP signaling associated with medial Dpp reduction during wing growth regulation.

And there are ideas hidden in the figure itself that are not explained in the legend or the Discussion, especially the switch from BMP presentation to BMP sequestration with a change from low local to high local Nord concentrations. A fuller discussion in the text would be helpful.

The suggested changes were made on Pages 23-34 and 47 of the revised manuscript.

5) Biphasic in vitro- The in vitro work remains strong, especially for Nord's inhibitory effects on BMP signaling. But the only evidence for a positive and thus biphasic effect of adding Nord relies on a single lane in a single gel: the lowest Nord concentration in Figure 8F, then quantified in 8G. It would be helpful if the authors could state how many times they have repeated this result.

We include additional replicates and statistical analysis in Figure 8G. See Page 45 for the changes in the legend of Figure 8. We also softened the conclusion on Page 22.

6) Given the long discussion of NDNF, the authors should mention the mechanisms that have been proposed for NDNF activity: by binding integrins, or by modulation of FGF signaling by binding HSPGs.

The suggested changes were made on Page 28.

https://doi.org/10.7554/eLife.73357.sa2

Article and author information

Author details

  1. Shu Yang

    Department of Anatomy and Cell Biology and the GW Cancer Center, George Washington University School of Medicine and Health Sciences, Washington, United States
    Contribution
    Data curation, Formal analysis, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review and editing
    Contributed equally with
    Xuefeng Wu
    Competing interests
    No competing interests declared
  2. Xuefeng Wu

    Department of Anatomy and Cell Biology and the GW Cancer Center, George Washington University School of Medicine and Health Sciences, Washington, United States
    Contribution
    Data curation, Formal analysis, Visualization, Writing – original draft, Writing – review and editing
    Contributed equally with
    Shu Yang
    Competing interests
    No competing interests declared
  3. Euphrosyne I Daoutidou

    Department of Genetics, Cell Biology & Development and the Developmental Biology Center, University of Minnesota, Minneapolis, United States
    Contribution
    Data curation, Formal analysis, Visualization
    Competing interests
    No competing interests declared
  4. Ya Zhang

    Department of Anatomy and Cell Biology and the GW Cancer Center, George Washington University School of Medicine and Health Sciences, Washington, United States
    Contribution
    Data curation, Formal analysis, Visualization
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9060-3777
  5. MaryJane Shimell

    Department of Genetics, Cell Biology & Development and the Developmental Biology Center, University of Minnesota, Minneapolis, United States
    Contribution
    Data curation, Visualization
    Competing interests
    No competing interests declared
  6. Kun-Han Chuang

    Department of Anatomy and Cell Biology and the GW Cancer Center, George Washington University School of Medicine and Health Sciences, Washington, United States
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3241-8586
  7. Aidan J Peterson

    Department of Genetics, Cell Biology & Development and the Developmental Biology Center, University of Minnesota, Minneapolis, United States
    Contribution
    Data curation, Validation, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6801-3364
  8. Michael B O'Connor

    Department of Genetics, Cell Biology & Development and the Developmental Biology Center, University of Minnesota, Minneapolis, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3067-5506
  9. Xiaoyan Zheng

    Department of Anatomy and Cell Biology and the GW Cancer Center, George Washington University School of Medicine and Health Sciences, Washington, United States
    Contribution
    Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing
    For correspondence
    xzheng@gwu.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-4983-5503

Funding

National Institute of General Medical Sciences (R01GM117440)

  • Xiaoyan Zheng

National Institute of General Medical Sciences (R35GM118029)

  • Michael B O'Connor

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Gibson, MC, the Bloomington Drosophila Stock Center, the Vienna Drosophila RNAi Center, Kyoto Stock Center, Addgene, and the Developmental Studies Hybridoma Bank for fly strains and reagents. MBO thanks Pierre Leopold in whose lab some of this work was carried out while he was on sabbatical. This work was supported by the National Institutes of Health grants R01GM117440 to XZ and R35GM118029 to MBO.

Senior Editor

  1. Utpal Banerjee, University of California, Los Angeles, United States

Reviewing Editor

  1. Erika A Bach, New York University School of Medicine, United States

Reviewer

  1. Erika A Bach, New York University School of Medicine, United States

Publication history

  1. Received: August 25, 2021
  2. Preprint posted: September 6, 2021 (view preprint)
  3. Accepted: January 15, 2022
  4. Accepted Manuscript published: January 17, 2022 (version 1)
  5. Version of Record published: February 18, 2022 (version 2)

Copyright

© 2022, Yang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 487
    Page views
  • 100
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Shu Yang
  2. Xuefeng Wu
  3. Euphrosyne I Daoutidou
  4. Ya Zhang
  5. MaryJane Shimell
  6. Kun-Han Chuang
  7. Aidan J Peterson
  8. Michael B O'Connor
  9. Xiaoyan Zheng
(2022)
The NDNF-like factor Nord is a Hedgehog-induced extracellular BMP modulator that regulates Drosophila wing patterning and growth
eLife 11:e73357.
https://doi.org/10.7554/eLife.73357

Further reading

    1. Cell Biology
    Désirée Schatton et al.
    Research Article Updated

    Proliferating cells undergo metabolic changes in synchrony with cell cycle progression and cell division. Mitochondria provide fuel, metabolites, and ATP during different phases of the cell cycle, however it is not completely understood how mitochondrial function and the cell cycle are coordinated. CLUH (clustered mitochondria homolog) is a post-transcriptional regulator of mRNAs encoding mitochondrial proteins involved in oxidative phosphorylation and several metabolic pathways. Here, we show a role of CLUH in regulating the expression of astrin, which is involved in metaphase to anaphase progression, centrosome integrity, and mTORC1 inhibition. We find that CLUH binds both the SPAG5 mRNA and its product astrin, and controls the synthesis and the stability of the full-length astrin-1 isoform. We show that CLUH interacts with astrin-1 specifically during interphase. Astrin-depleted cells show mTORC1 hyperactivation and enhanced anabolism. On the other hand, cells lacking CLUH show decreased astrin levels and increased mTORC1 signaling, but cannot sustain anaplerotic and anabolic pathways. In absence of CLUH, cells fail to grow during G1, and progress faster through the cell cycle, indicating dysregulated matching of growth, metabolism, and cell cycling. Our data reveal a role of CLUH in coupling growth signaling pathways and mitochondrial metabolism with cell cycle progression.

    1. Cell Biology
    Robert J Tower et al.
    Research Article

    De novo limb regeneration after amputation is restricted in mammals to the distal digit tip. Central to this regenerative process is the blastema, a heterogeneous population of lineage-restricted, dedifferentiated cells that ultimately orchestrates regeneration of the amputated bone and surrounding soft tissue. To investigate skeletal regeneration, we made use of spatial transcriptomics to characterize the transcriptional profile specifically within the blastema. Using this technique, we generated a gene signature with high specificity for the blastema in both our spatial data, as well as other previously published single-cell RNA-sequencing transcriptomic studies. To elucidate potential mechanisms distinguishing regenerative from non-regenerative healing, we applied spatial transcriptomics to an aging model. Consistent with other forms of repair, our digit amputation mouse model showed a significant impairment in regeneration in aged mice. Contrasting young and aged mice, spatial analysis revealed a metabolic shift in aged blastema associated with an increased bioenergetic requirement. This enhanced metabolic turnover was associated with increased hypoxia and angiogenic signaling, leading to excessive vascularization and altered regenerated bone architecture in aged mice. Administration of the metabolite oxaloacetate decreased the oxygen consumption rate of the aged blastema and increased WNT signaling, leading to enhanced in vivo bone regeneration. Thus, targeting cell metabolism may be a promising strategy to mitigate aging-induced declines in tissue regeneration.