Hemocyte-secreted papilin bearing mucin-type O-glycans regulates peripodial stalk formation via epidermal JAK/STAT signaling in Drosophila

Summary

Protein glycosylation is an essential post-translational modification. In evolutionarily conserved mucin-type O-glycosylation, the most common O-glycan, T antigen, is synthesized by core 1 β1,3-galactosyltransferase 1 (C1GalT1). Loss of C1GalT1 leads to developmental defects across organisms. We previously found that Drosophila C1GalT1 mutants exhibit malformed legs, but the underlying mechanism was unclear. Here, we identify a glycan-mediated inter-tissue signaling mechanism wherein embryonic hemocytes regulate leg morphogenesis. We show that T antigen-modified Papilin (Ppn), an extracellular matrix (ECM) protein secreted by embryonic hemocytes, suppresses JAK/STAT signaling in the epidermis surrounding Keilin’s organ. This repression is essential for proper tubulogenesis of the peripodial stalk anchoring the leg disc and ensuring its correct positioning during development. Disrupted mucin-type O-glycosylation impairs Ppn secretion and causes mislocalized leg discs and morphogenetic defects. These findings identify Ppn carrying mucin-type O-glycan as long-range modulators of epithelial signaling and underscore the role of immune-like cells in coordinating organogenesis via ECM.

Graphical abstract

Highlights

• •Loss of dC1GalT1 leads to mispositioned leg discs due to the failure of stalk formation
• •Embryonic hemocyte-derived T antigen suppresses the epidermal JAK/STAT pathway
• •Papilin is a candidate core-protein carrying T antigen in hemocytes
• •Blood cell-secreted Papilin drives stalk tubulogenesis

Biochemistry; Glycobiology; Cell biology; Developmental biology

Introduction

Protein glycosylation is a widespread post-translational modification involved in diverse biological processes. O-Glycosylation is characterized by the addition of a monosaccharide to the hydroxyl group of a serine (Ser) or threonine (Thr) residue. Among various O-glycan structures, mucin-type O-glycans form a major class and are evolutionarily conserved across species. They are typically found in mucin-like domains—namely, protein regions enriched in Ser and Thr residues. Mucin-type O-glycosylation is initiated by the transfer of N-acetylgalactosamine (GalNAc) in an α1-linkage to the Ser or Thr residue of the protein by polypeptide N-acetylgalactosaminyl-transferases (GALNTs).^1,2 The resulting terminal GalNAc glycan structure is called Tn antigen (GalNAcα1-Ser/Thr; Figure 1A). The core 1 β1,3-galactosyltransferase 1 (C1GalT1, also known as T-synthase) subsequently transfers galactose (Gal) in a β1-3 linkage to the GalNAc residue to produce T antigen—namely, core 1 structure (Galβ1-3GalNAcα1-Ser/Thr)—the most common structure found in mucin-type O-glycans.^{3,4,5,6,7,8,9,10,11}

Figure 1.

dC1GalT1 expression in hemocytes is required for proper leg formation

(A) Structure of mucin-type O-glycans in Drosophila, including Tn antigen (GalNAcα1-Ser/Thr), T antigen (Galβ1-3GalNAcα1-Ser/Thr), and glucuronylated T antigen (GlcAβ1-3Galβ1-3GalNAcα1-Ser/Thr). T antigen is synthesized by dC1GalT1, which adds galactose (Gal) to N-acetyl galactosamine (GalNAc) of Tn antigen. Glucuronylated T antigen is synthesized by dGlcAT-P, which adds glucuronic acid (GlcA) to Gal of T antigen.

(B–E) Micrographs of a third leg in WT (B) and dC1GalT1^EY/EY mutant (C–E) flies. Arrowheads indicate malformed parts of the leg.

(F–H) Groin areas of legs in WT (F) and dC1GalT1^EY/EY mutant (G and H) flies, in which all third legs are left at the base and separated from there to the tip. Arrowheads in G show the separated basal portions of the left and right legs.

(I) Pieces of a third leg burrowed into the basal portion, indicated by the arrowhead in H.

(J–O) Micrographs of a third leg in WT (J) and dC1GalT1^EY/EY mutant (K) flies, and in mutant flies rescued using Act5C-Gal4 (L), arm-Gal4 (M), Hml-Gal4 (N), and He-Gal4 (O). See also Videos S1 and S2, and Tables S1 and S2.

In mammals, C1GalT1 requires a specific molecular chaperone, Cosmc (also known as C1GalT1C1), for enzymatic activity.¹² In Drosophila, however, C1GalT1 does not require a chaperone, and an equivalent protein to Cosmc has not been identified.⁵ In mammals, T antigen can be further modified by ST3Gal-I to synthesize sialylated T antigen (Siaα2-3Galβ1-3GalNAcα1-Ser/Thr).¹ In contrast, Drosophila uses β1,3-glucuronyltransferase-P (dGlcAT-P)¹⁰ to synthesize glucuronylated T antigen (GlcAβ1-3Galβ1-3GalNAcα1-Ser/Thr) (Figure 1A)^10,13,14,15 in place of sialylated T antigen. Both GlcA and Sia are negatively charged monosaccharides; therefore, glucuronylated T antigen is thought to functionally correspond to its sialylated mammalian counterpart.

Aberrant expression of Tn antigen, due to the reduced expression of C1GalT1 or Cosmc, is associated with several human pathologies, including Tn syndrome,^16,17 IgA nephropathy,^18,19 and various cancers.^20,21,22 The disruption of T antigen synthesis in model organisms also results in a broad spectrum of phenotypes. In mice, for example, loss of C1galt1 or Cosmc causes vascular abnormalities,^23,24,25,26 thrombocytopenia,^26,27 spontaneous colitis and colorectal tumors,^28,29,30 as well as glomerular damage including proteinuria and sclerosis.^31,32 Similarly, loss of function of Drosophila C1GalT1 (dC1GalT1) leads to developmental and physiological abnormalities, such as reduced numbers of circulating larval hemocytes,⁶ excess differentiation of prohemocytes in lymph glands,⁷ impaired tissue invasion of embryonic hemocytes,³³ elongation of the larval ventral nerve cord (VNC),³⁴ and defects in larval neuromuscular junctions and muscles.⁸ These findings underscore the essential roles of dC1GalT1-synthesized O-glycans across multiple pathways. Furthermore, we previously showed that the third legs of dC1GalT1 mutant flies are malformed⁶; however, the molecular mechanism underlying leg malformation remains unknown.

In Drosophila, the adult legs are derived from three pairs of leg imaginal discs—that is, sac-like structures located within the larval body cavity. The leg discs themselves originate from the ventral embryonic epidermis, where the three pairs of leg disc primordia (small clusters of cells) are specified.³⁵ Each primordium is closely associated with a peripheral sensory organ known as Keilin’s organ (KO), which is located at the surface of the larval epidermis. As the larval stages progress, the primordia invaginate into the body cavity and develop into sac-like leg discs, which are connected to the larval epidermis by a thin epithelial stalk composed of peripodial cells.³⁶ The disc cells proliferate slowly at the second instar larval stage and then rapidly expand to form mature leg discs by the late third-instar larval stage.

After pupal formation (APF), the stalk lumen opens and expands, allowing the leg discs to evert and push through the stalk opening to the space between the epidermis and puparium.^36,37 The peripheral regions of the leg discs then spread and fuse with adjacent discs to form the continuous adult epidermis. Subsequent contraction of abdominal muscles drives head eversion and leg inflation, leading to rapid leg extension. The final adult leg shape is sculpted through a reduction in leg width. Thus, leg morphogenesis in Drosophila involves dynamic transformation of the leg disc, which is coordinated with stalk-mediated positioning and tissue remodeling.

Several Drosophila epithelial organs, such as the trachea, salivary gland, and hindgut, have served as models of epithelial tube formation.^38,39,40,41 In all cases, the tubes are typically formed by the invagination of polarized epithelial primordia, followed by elongation through cell migration and/or cell shape changes. In the hindgut, tube elongation also involves an increase in cell size and convergent extension in a process driven by the JAK/STAT signaling.⁴² Although the peripodial stalk that connects the leg disc to the epidermis resembles an epithelial tube, the molecular mechanism underlying its formation remains unexplained.

In this study, we have investigated how mucin-type O-glycans synthesized by dC1GalT1 regulate epithelial morphogenesis during leg development. We demonstrate that the adult leg defects observed in dC1GalT1 mutants arise from the mislocalization of the third leg discs due to a failure of stalk formation. Rescue experiments pinpoint embryonic hemocytes as the source of functional dC1GalT1 activity required for this process. Mechanistically, we show that the absence of T antigen in dC1GalT1 mutants leads to the ectopic activation of JAK/STAT signaling in the adjacent epidermis, which disrupts stalk elongation. Defective stalk formation was partially rescued by mild inactivation of JAK/STAT signaling. Moreover, we identify Papilin (Ppn), a mucin-type O-glycosylated extracellular matrix (ECM) protein secreted by hemocytes, as a key effector in this signaling axis. Loss of Ppn phenocopies dC1GalT1 mutants, resulting in shortened stalks and an upregulation of epidermal JAK/STAT signaling. Our biochemical data confirm that Ppn carries the T antigen, supporting its role as a glycosylated instructive cue. Furthermore, mucin-type O-glycosylation is essential for Ppn secretion from hemocyte-derived cells. Collectively, our findings establish an inter-tissue communication mechanism whereby embryonic hemocytes sculpt the local epithelial signaling environment via glycan-modified ECM proteins to direct tube formation and organ positioning.

Results

Mutation of dC1GalT1 leads to the malformation of adult third legs

We previously reported that dC1GalT1 mutant flies display malformations in the third (metathoracic) pair of legs (Figures 1B–1E),⁶ while the first (prothoracic) and second (mesothoracic) pairs are unaffected. To confirm that dC1GalT1 is the gene responsible, we first analyzed three alleles with a P element insertion (dC1GalT1^EY13370, dC1GalT1^c01812, and dC1GalT1^KG02976) and a chromosome deficiency line (Df(2L)Exel7040) in which the entire dC1GalT1 locus is deleted. Both homozygous and transheterozygous mutated flies exhibited a significantly higher frequency of third leg defects as compared with wild-type (WT; Table S1), confirming that mutations in dC1GalT1 lead to the malformation of the third legs.

Next, we examined whether dC1GalT1 mutations affect specific leg segments. The Drosophila adult leg consists of nine segments (coxa, trochanter, femur, tibia, and five tarsal segments),⁴³ and mutations in segmentation genes often lead to segment-specific defects.⁴⁴ Here, the malformations in dC1GalT1 mutants were observed in various leg segments in a non-specific pattern (Figures 1C–1E, arrowheads), suggesting that dC1GalT1 is not directly involved in leg segmentation. We observed, however, that the basal portions of the third legs were mispositioned in dC1GalT1 mutants. In WT, the left and right legs were symmetrically positioned (Figure 1F); in mutants, by contrast, they were abnormally separated from each other (Figure 1G, arrowheads). In some individuals, only one-third leg was visible while the other remained embedded at the base (Figures 1H and 1I, arrowhead), indicating a defect in both basal positioning and leg morphology.

To determine whether the defects stem from aberrant leg extension during metamorphosis, we performed live imaging of pupal leg development using GFPπRas driven by Dll-Gal4 to visualize the legs. In controls, the three pairs of legs aligned at the ventral midline by ∼6 h APF (Video S1), and head eversion and full leg extension occurred normally by ∼11 h APF. Subsequently, the shape of legs was slimly arranged. In dC1GalT1 mutants, leg alignment failed at 6 h APF, resulting in misdirected extension during head eversion (Video S2). These observations suggest that adult leg malformation arises from misaligned leg extension during pupal development.

Video S1. Normal leg formation during metamorphosis, related to Figure 1

Live imaging of leg formation during the pupal stage. The developing legs were visualized with GFPπRas driven by Dll-Gal4 (Dll>GFPπRas; control). Extension of the legs starts at the timing of head eversion. Immediately after extension, the legs have the thick shape of an extended sac, and the shape of the adult legs is formed over time.

Video S2. Abnormal leg formation in dC1GalT1 mutants during metamorphosis, related to Figure 1

Live imaging of leg formation during the pupal stage in dC1GalT1^EY/EY mutant (Dll>GFPπRas, dC1GalT1^EY/EY). The developing legs were visualized with GFPπRas driven by Dll-Gal4. Some legs are seen to be distorted in an unnatural direction as they extend.

Hemocyte-expressed dC1GalT1 is required for normal third leg formation

To identify the tissue responsible for the observed leg phenotype, we conducted rescue experiments. Ubiquitous overexpression of dC1GalT1 under the Gal4 driver (Act5C-Gal4 or arm-Gal4) in the dC1GalT1 mutant background rescued the leg defects (Figures 1J–1M and Table S2). However, overexpression using leg-specific drivers (C855a-Gal4, Ubx-Gal4, 1151-Gal4, tsh-Gal4, or Gug^AGiR-Gal4) failed to rescue the phenotype (Table S2), suggesting that dC1GalT1 expression in the leg tissue is not required.

Next, to clarify which other tissues expressing dC1GalT1 are involved in leg formation, we rescued dC1GalT1 expression using hemocyte-specific (Hml-Gal4 and He-Gal4), neuron-specific (elav-Gal4), or salivary gland-specific (sd-Gal4) Gal4 drivers (Table S2). Notably, the leg phenotype was restored to WT only when hemocyte-specific Gal4 drivers were used (Figures 1N and 1O). These results demonstrate that dC1GalT1 activity in hemocytes is critical for proper third leg development.

Loss of dC1GalT1 disrupts third leg disc positioning due to stalk formation failure

To further investigate the origin of leg malformation in dC1GalT1 mutants, we examined larval leg disc positioning. In WT, the third leg discs were localized adjacent to muscle 33 (ventral intersegmental 5; Figures 2A and Aʹ). In dC1GalT1 mutants, by contrast, the third leg discs were mislocalized beneath the ventral muscle (Figures 2B and Bʹ, arrowheads). To clarify the positional relationship between the third leg discs and ventral muscles, we observed a cross-section of the larval body using transmission electron microscopy. Whereas the third leg disc was localized on the dorsal side of the ventral muscles in WT (Figures 2C and Cʹ), it was localized between the muscles and cuticulated epithelium and compressed physically in dC1GalT1 mutants (Figures 2D and Dʹ). These observations indicate that mislocalization of the third leg discs in dC1GalT1 mutants causes the malformation of adult third legs.

Figure 2.

Short stalk phenotype in dC1GalT1 mutant larvae

(A–B′) Micrographs and schematic show ventral muscles and imaginal disc positions in WT (A and A′) and dC1GalT1^EY/EY mutant (B and B′) larvae. Dashed line indicates the ventral midline. Arrowheads indicate mislocalized third leg discs (Leg 3) beneath the ventral muscles.

(C–D′) Transmission electron micrographs of larval cross-section in WT (C and C′) and dC1GalT1^EY/EY mutant (D and D′) larvae. High magnification views of the area bordered by the rectangle in C and D are shown in C′ and D′ with third leg discs indicated in magenta. m, muscle; c.e., cuticulated epithelium.

(E–F′) Confocal micrographs of imaginal discs and epidermal cells in WT (E and E′) and dC1GalT1^EY/EY mutant (F and F′) larvae. High magnification views of E and F are shown in E′ and F'. The cell membrane is labeled with anti-Fas III antibody. Closed arrowheads in E and E′ indicate the thin peripodial stalk that anchors the third leg disc to the epidermis. Open arrowhead in F and F′ indicates the lack of a stalk.

(G–I″) High magnification views of the edge of the third leg disc in WT (G–G″) and dC1GalT1^EY/EY mutant (H–I″) larvae. Cell membrane and nucleus are labeled with anti-Fas III antibody (magenta) and TOTO-1 (green), respectively. Arrowheads in G–G″ indicate the stalk cells in WT. Arrows in H–H″ indicate a few stalk cells. Open arrowheads in I–I″ indicate the lack of stalk cells.

(J) Stalk phenotypes are divided into three categories: normal, modest, and strong. Semi-translucent square shows the position of the ventral muscles. In the normal stalk phenotype, the stalk is sufficiently extended, and the third leg disc is not covered by the ventral muscles. In the modest stalk phenotype, the stalk length is shorter, and the leg disc is partly covered by the ventral muscles. In the strong stalk phenotype, the third leg disc is completely covered by the ventral muscles due to the lack of a stalk.

(K) Percentages of stalk phenotypes in WT, dC1GalT1^EY/EY, and dC1GalT1^KG/KG mutants at 25 °C. Note that the phenotype was stronger at 29 °C; however, the number of individuals that died before reaching adulthood increased. Shading in the graphs represents the categories of stalk phenotypes shown in J. The number of observations is indicated on the right. See also Figures S1 and S2.

To visualize the peripodial stalk that anchors each disc to the epidermis, we labeled the cell membranes using anti-fasciclin III (Fas III) antibody. In WT, a long and thin stalk connected the edge of each third leg disc to epidermis (Figures 2E–Eʹ and 2G–Gʺ, arrowheads). In dC1GalT1 mutants, by contrast, the stalk cells were reduced or absent (Figures 2F–Fʹ and 2H–Iʺ). Thus, we named this phenotype “short stalk”. This defect was significantly enriched in dC1GalT1^EY/EY mutants (Figures 2J and 2K) and, in addition to malformed third legs, was rescued by hemocyte-specific dC1GalT1 expression (Figure S1A). The stalks anchoring other imaginal discs—eye discs, first and second leg discs, haltere discs, wing discs, and genital discs—were also shorter in dC1GalT1 mutants, although the positioning of these discs remained normal (data not shown).

Next, using a temperature-sensitive Gal80 (Gal80^ts) system, we performed stage-specific rescue to identify the point at which dC1GalT1 is required. Rescuing dC1GalT1 expression in hemocytes from the egg-laying stage to the early-second instar larval stage (Figure S1B), to the mid-first instar larval stage (Figure S1C), or only to the end of the embryonic stage (Figure S1D) led to the restoration of the short stalk to a normal phenotype in each case. By contrast, the short stalk phenotype was observed if dC1GalT1 was not expressed at any point in the whole stage (Figure S1E). These results indicate that hemocyte-derived dC1GalT1 expression during the embryonic stages is essential for both stalk formation and disc positioning.

Convergent extension contributes to stalk formation

Next, we examined whether the failure of stalk formation in dC1GalT1 mutants is due to defects in convergent extension driven by the contraction of actomyosin—namely, the actin and myosin complex—along the cell-cell junction.^45,46 To disrupt convergent extension, we performed knockdown of the spaghetti squash (sqh) gene encoding the myosin regulatory light chain using an esg-Gal4 construct that is expressed not only in the epidermal and disc cells but also in the stalk cells (Figures S2A and S2B). Knockdown of sqh led to an increase in stalk width (Figures S2C–S2E), suggesting that convergent extension driven by actomyosin contractility is critical for stalk morphogenesis.

Normal development of third leg discs at the late-second instar larval stage

To investigate the early developmental process of third leg disc formation, we analyzed the positioning and behavior of epidermal cells surrounding the KO and disc cells beneath it during the late second-instar larval stage of WT flies (about 36–48 h after egg hatching; AEH). We visualized cell morphology using two fluorescent markers: sqh-GFP, which labels the non-muscle myosin regulatory light chain⁴⁷; and canoe (cno)-GFP, the Drosophila ortholog of mammalian Afadin, which localizes to the adherens junctions of epithelial cells.⁴⁸ Based on sequential cellular changes, we defined eight developmental stages of early third leg disc formation (Figure S3A). At stage 1, in optical sections of epidermis, both sqh-GFP and cno-GFP were expressed in epidermal cells, while smaller cells expressing cno-GFP but not sqh-GFP began to appear around the KO. Simultaneously, in the body cavity, disc cells beneath the KO expressed sqh-GFP, but not cno-GFP. From stages 1–5, the sqh-GFP-negative area around the KO gradually expanded, and the number of cno-GFP-positive, smaller cells increased. In parallel, the underlying disc cell cluster also expanded as these cells proliferated.

At stage 6, both the sqh-GFP-negative area and the number of cno-GFP-positive smaller cells started to decrease (Figure S3A). The smaller cells became visible not only in the epidermis but also within the body cavity, suggesting that they had migrated from the epidermis to the disc cell cluster. Simultaneously, the disc cluster began to migrate laterally while continuing to proliferate. By stage 8, the sqh-GFP-negative area had disappeared completely, and the region occupied by cno-GFP-positive cells was nearly identical in size to the disc cell cluster, indicating the formation of a sac-like structure lined with epithelial cells. Importantly, these early events were not significantly altered in dC1GalT1 mutants (data not shown), suggesting that initial leg disc development proceeds normally in dC1GalT1 mutant flies and that the defects subsequently emerge during the third-instar stage, when the stalk forms and disc positioning becomes critical.

Collectively, our findings enable us to present a schematic model outlining the process of early disc formation in late-second instar larvae (Figure S3B). From stages 1–5, disc primordia and KO-associated epithelial cells expand. At stage 6, KO-associated epithelial cells begin migrating into the underlying disc cell cluster. At stage 8, the disc cell cluster just beneath the KO becomes constricted, and the sac-like structure lined by cno-GFP-positive cells is formed. The constricted disc cell cluster is presumed to extend and form the long, thin stalk during the third-instar larval stage by convergent extension.

dC1GalT1 mutants exhibit upregulated JAK/STAT signaling in larval epidermis

To determine whether the dysregulation of a specific signaling pathway underlies the stalk formation defects observed in dC1GalT1 mutants, we systematically screened the activity of several major pathways involved in epithelial remodeling—namely, Notch, epidermal growth factor receptor (EGFR), Hedgehog, and JAK/STAT—using transcriptional reporters and pathway-specific markers. Among them, we identified the aberrant activation of the JAK/STAT signaling specifically in epidermal cells of dC1GalT1 mutants. In mutant larvae, reporter activity was notably elevated in epidermal cells surrounding the third leg disc (Figures 3A, 3B, 3D, and 3E), whereas the activity in disc cells themselves remained comparable to that in controls (Figures 3Dʹ and 3Eʹ). This epidermal upregulation was suppressed when dC1GalT1 expression was restored in the hemocytes (Figures 3C and 3F), indicating a non-cell-autonomous effect from blood cells. Morphologically, the disc cell cluster beneath the KO failed to constrict in dC1GalT1 mutants, appearing significantly wider than in controls (Figures 3Dʹ and 3Eʹ, double-headed arrow). This defect was rescued by hemocyte-specific dC1GalT1 expression (Figure 3Fʹ). Together, these findings suggest that dC1GalT1 expression in hemocytes is essential for maintaining appropriate levels of epidermal JAK/STAT signaling, which in turn is required for disc constriction and proper stalk morphogenesis.

Figure 3.

Upregulation of JAK/STAT signaling in dC1GalT1 mutant epidermis

(A–C) JAK/STAT reporter activity in epidermal cells of control (A), dC1GalT1^EY/EY mutant (B), and Hml-Gal4 rescued mutant (C) larvae during late second instar larval stages (38–48 h AEH). The activity of JAK/STAT signaling was visualized with 10×STAT-dGFP. An optical section of epidermis around the KO is shown. KO, Keilin’s organ; L, lateral; V, ventral; A, anterior; P, posterior. Scale bars, 50 μm.

(D–F′) JAK/STAT activity in epidermal and disc cells of control (D and D′), dC1GalT1^EY/EY mutant (E and E′), and Hml-Gal4 rescued mutant (F and F′) larvae at 48 h AEH (stage 8). Optical sections of epidermis are shown in D, E, and F; optical sections of disc cells (in the body cavity) are shown in D′, E′, and F'. Double-headed arrow in D′, E′, and F′ indicates the width of the disc cell cluster beneath the KO. Scale bars, 50 μm. See also Figure S3.

Moderate activation of epidermal JAK/STAT signaling is required for proper stalk formation

Next, we explored whether the ectopic activation of the JAK/STAT signaling is sufficient to disrupt stalk formation. We overexpressed domeless (dome), the pathway’s transmembrane receptor, in the epidermis by using tsh-Gal4 combined with tub-Gal80^ts for stage-specific control. The resulting stalk phenotypes were classified into four categories—namely, normal, weak, modest, and strong—and quantified (Figures 4A–4F). Overexpression of dome from the early first-instar to late third-instar larval stages induced a strong short stalk phenotype (Figure 4B). Similarly, dome induction from the mid-second-instar to late third-instar larval stages caused severe defects (Figure 4C). In contrast, activation limited to the third-instar larval stage or from the mid-to late-second instar larval stage did not replicate the strong short stalk phenotype (Figures 4D and 4E). We confirmed that the short stalk phenotype was not induced if dome was not overexpressed during any of the stages (Figure 4F). Together, these results indicate that moderate JAK/STAT activity must be precisely maintained from the mid-second-instar to late third-instar larval stages for proper stalk formation.

Figure 4.

Moderate JAK/STAT activation is required for stalk formation

(A) Stalk phenotypes are divided into four categories: normal (score 1), weak (score 2), modest (scores 3 and 4), and strong (score 5). Box indicates the position of the ventral muscles. In the normal phenotype, the stalk is sufficiently extended, and the third leg disc is not covered by the ventral muscles. In the weak phenotype, the stalk length is shorter than that in the normal phenotype, but the third leg disc is not covered by the ventral muscles. In the modest phenotype, the stalk length is shorter than that in the weak phenotype, and the third leg disc is partly covered by the ventral muscles. In the strong phenotype, the third leg disc is completely covered by the ventral muscles due to the lack of a stalk.

(B–F) Left, stage-specific overexpression of the dome gene in epidermal cells driven by tsh-Gal4 and tub-Gal80^ts. Dashed lines indicate the period during which embryos or larvae are raised at 18 °C; bold lines indicate the period during which they are raised at 29°C. dome was overexpressed from the early first-instar to late third-instar larval stage (B), from the mid-second-instar to late third-instar larval stage (C), from the early to late third-instar larval stage (D), or from the mid to late-second instar larval stage (E), or was not overexpressed during any stage (F). Right, corresponding percentages of each stalk phenotypes shown as bar graphs. Shading in the bar graphs corresponds to the categories of stalk phenotypes shown in A. Average score ± standard deviation and number of observations are shown on the right.

(G) Percentages of stalk phenotypes in dC1GalT1^EY/EY and dC1GalT1^EY/EY; Stat92E^06346/+ mutants. Stat92E⁰⁶³⁴⁶ is a loss of function allele. Shading in the graphs represents the categories of stalk phenotypes shown on the right. See also Figure S4.

To further investigate whether the hyperactivation of the JAK/STAT signaling in dC1GalT1 mutants disrupts stalk formation, we inhibited this pathway specifically in the epidermis of dC1GalT1 mutants. Contrary to our expectations, knockdown of Stat92E or dome did not rescue the short stalk phenotype (Figure S4A), probably because downregulation of the epidermal JAK/STAT signaling may also cause stalk defects. To test this possibility, we inhibited the epidermal JAK/STAT signaling in WT and observed that knockdown of Stat92E or dome induced the short stalk phenotypes (Figure S4B). This finding suggests that the impairment of the epidermal JAK/STAT signaling also disrupts stalk formation. Furthermore, the short stalk phenotypes were partially rescued in dC1GalT1 mutants carrying a heterozygous loss of Stat92E (dC1GalT1^EY/EY; Stat92E^06346/+; Figure 4G). Collectively, therefore, these findings support the idea that the moderate activation of the JAK/STAT signaling in the epidermis is required for proper stalk formation.

Loss of hemocyte-derived Ppn leads to JAK/STAT hyperactivation and stalk defects

To identify core proteins carrying T antigen that regulate stalk formation, we performed an RNA interference (RNAi) screen specifically targeting proteins with a mucin-like domain in hemocytes. Knockdown of Ppn, encoding a highly glycosylated ECM protein, resulted in a prominent short stalk phenotype (Figure 5A). Analysis of four P element insertion alleles (Ppn^G4259, Ppn^MI03189, Ppn^MB07666, and Ppn^MB01784) and one chromosome deficiency line (Df(3R)BSC322) confirmed that Ppn is required for stalk formation (Figure S5).

Figure 5.

Ppn knockdown in hemocytes induces a short stalk phenotype and epidermal JAK/STAT hyperactivation

(A) Quantification of stalk phenotypes in larvae with hemocyte-specific Ppn knockdown using four RNAi lines (108005KK, 16523GD, 18436R-1, and 18436R-4). Knockdown of Ppn was driven by Hml-Gal4 and He-Gal4. Stalk phenotypes are divided into four categories, as shown in Figure 4A. Average score ± standard deviation and number of observations are shown on the right.

(B–Cʹ) JAK/STAT activity in control (B and Bʹ) and Ppn knockdown (C and Cʹ) larvae at 48 h AEH (stage 8). The activity of JAK/STAT signaling was visualized with 10×STAT-dGFP. Optical sections of epidermis are shown in B and C. Optical sections of disc cells in the body cavity are shown in Bʹ and Cʹ. Double-headed arrow in Bʹ and Cʹ indicates the width of the disc cell cluster. KO, Keilin’s organ. Scale bars, 50 μm. See also Figures S5–S7.

Knockdown of Ppn in hemocytes also led to elevated JAK/STAT activity in the epidermis and broader disc cell clusters underneath the KO (Figures 5B–5Cʹ), closely mimicking the dC1GalT1 mutant phenotype. Notably, overexpressing dC1GalT1 in the hemocytes of Ppn knockdown flies failed to restore stalk morphology (Figure S6), suggesting that Ppn operates downstream of, or cooperatively with, dC1GalT1 in this pathway. This also raises the possibility that Ppn is a candidate core-protein carrying T antigen.

As Ppn has been reported to carry sulfated glycosaminoglycans,⁴⁹ we investigated whether this glycan modification is required for Ppn function in stalk formation by inhibiting the synthesis of the glycosaminoglycan linkage region specifically in hemocytes. However, knockdown of dXylT, β4GalT7, or β3GalTII did not induce the short stalk phenotypes (Figure S7), indicating that the glycosaminoglycan modification on Ppn may be independent of its role in the stalk formation.

Ppn secretion depends on its mucin-type O-glycosylation

Previous studies have shown that Ppn is a heavily glycosylated ECM protein and contains a Ser/Thr-rich polypeptide domain (Figure 6A).^49,50 To examine whether Ppn carries T antigen, we expressed a recombinant 6×His-tagged Ppn (Ppn-His) protein in Drosophila Kc167 cells, an embryonic hemocyte-like cell line.⁵¹ The construct covered the consensus region of multiple splicing isoforms (Ppn-PC, Ppn-PE, Ppn-PF, and Ppn-PG; FlyBase: https://flybase.org/reports/FBgn0003137). Secreted Ppn-His was purified from the culture medium using Ni-NTA agarose. Western blotting with anti-6×His tag antibody confirmed successful purification (Figure 6B, left, arrowhead). Lectin blotting using peanut agglutinin (PNA), which binds T antigen, revealed a corresponding band (Figure 6B, right, arrowhead), indicating that Ppn is O-glycosylated with T antigen in hemocyte-derived cells.

Figure 6.

Mucin-type O-glycosylation of Ppn is required for its secretion

(A) Domain structure of Drosophila Ppn-PC (full length) and the recombinant 6×His-tagged Ppn (Ppn-His) protein. SS, signal sequence. TSR, thrombospondin type-1 repeat domain.

(B) Ni-NTA-purified fractions from the culture medium of Kc167 cells were subjected to 5% SDS-PAGE followed by western blot analysis using anti-His tag antibody (left) and lectin blot analysis using PNA (right). Arrowhead indicates the band corresponding to Ppn-His. WB, Western blot. LB, Lectin blot. Non TF, the sample derived from non-transfected cells (negative control). Ppn TF, the sample derived from transfected cells. Results are representative of three experiments.

(C) Confocal images of circulating hemocytes at the embryonic stage in WT and dC1GalT1 mutants. Ppn, T antigen, and Tn antigen are stained with anti-Ppn antibody (magenta), PNA (green), and HPA (green), respectively. Arrowheads indicate partial colocalization between Ppn and T or Tn antigen. Scale bars, 5 μm.

(D and E) Relative levels of Ppn-His (D) and dC1GalT1 (E) mRNA in Ppn-His-overexpressing and dC1GalT1 knockdown Kc167 cells. mRNA levels were normalized to Gapdh1 mRNA, and those of Ppn-His and dC1GalT1 in the control cells were set as 1.0. Data represent the mean ± standard error (n = 3).

(F) Cell lysates and culture media of control and dC1GalT1 KD Kc167 cells were subjected to 5% SDS-PAGE followed by Western blot analysis using anti-His tag antibody (left) and lectin blot analysis using PNA (right). The α-tubulin internal control is shown in lower left. Open arrowhead indicates the band corresponding to Ppn-His in the cell lysate. Closed arrowhead indicates the band corresponding to Ppn-His in the culture medium. Results are representative of three experiments.

(G and H) Relative band intensity of Ppn-His (G) and PNA (H) in the cell lysate, normalized to the band intensity of α-tubulin. Data represent the mean ± standard error (n = 3).

(I and J) Relative band intensity of Ppn-His (I) and PNA (J) in the culture medium. Data represent the mean ± standard error (n = 3). In (D, E, and G–J), statistical significance was assessed by the Dunnett test: ∗p < 0.05; ∗∗p < 0.01, n.s., not significant. See also Figures S8–S10.

Ppn is known to be expressed in embryonic hemocytes.^50,52,53 To clarify the localization of Ppn and mucin-type O-glycans within hemocytes, we generated a custom antibody against its N-terminal domain and performed immunostaining of embryonic and larval hemocytes. Ppn was strongly expressed in circulating embryonic hemocytes of both WT and dC1GalT1 mutants (Figure 6C), but its expression in larval hemocytes was considerably weaker (Figure S8). Previous studies have shown that dC1GalT1 mutants exhibit reduced T antigen levels and increased Tn antigen levels in embryonic hemocytes.^6,34 Consistent with those findings, lectin blotting revealed reduced T antigen^6,34 and increased Tn antigen levels (Figure S9) in dC1GalT1 mutant embryos. To visualize mucin-type O-glycans in embryonic hemocytes, we labeled T antigen in WT and Tn antigen in dC1GalT1 mutants using PNA and Helix pomatia agglutinin (HPA), respectively. Partial colocalization of Ppn with T antigen and Tn antigen was observed in circulating embryonic hemocytes (Figure 6C, arrowheads), suggesting that Ppn is a bona fide carrier of mucin-type O-glycans.

To investigate whether Ppn secretion is affected by the loss of dC1GalT1, we performed dC1GalT1 knockdown in Kc167 cells overexpressing Ppn-His and determined the amount of secreted Ppn-His protein. We confirmed that Ppn-His mRNA levels were comparable between control and dC1GalT1 knockdown cells (Figure 6D), while dC1GalT1 mRNA levels were significantly lower in dC1GalT1 knockdown cells than in control cells (Figure 6E). Western blotting with anti-6×His tag antibody revealed that intracellular Ppn-His levels were similar between control and dC1GalT1 knockdown cells, but secreted Ppn-His levels were significantly lower in dC1GalT1 knockdown cells (Figures 6F, 6G, and 6I). The molecular weight of secreted Ppn-His was higher than that of intracellular Ppn-His (Figure 6F), probably due to extensive glycosylation.

Furthermore, lectin blotting using PNA confirmed that levels of both intracellular and secreted T antigen were reduced (Figures 6F, 6H, and 6J). Notably, the PNA-positive band corresponding to secreted proteins, including Ppn-His, was markedly reduced in dC1GalT1 knockdown cells relative to that in control cells (Figure 6F, closed arrowheads). In contrast, PNA blotting did not detect a band corresponding to intracellular Ppn-His (Figure 6F, open arrowhead), even upon overexposure, indicating that the intracellular Ppn-His may not be modified by T antigen. Collectively, these findings suggest that defective synthesis of mucin-type O-glycan on Ppn impairs its secretion from hemocyte-derived cells (Figure 7B). Therefore, a reduction in Ppn secretion resulting from hypoglycosylation of core 1 glycans may cause disruption of stalk formation in dC1GalT1 mutants. Taken together, these results support a model in which hemocyte-secreted, T antigen-modified Ppn modulates epidermal JAK/STAT signaling to facilitate stalk formation and proper leg development (Figure 7A).

Figure 7.

Model of the mechanism underlying stalk formation regulated by Ppn carrying mucin-type O-glycans

(A) In WT flies, hemocyte-secreted Ppn, which carries T antigen, promotes the formation of the stalk of third leg discs by suppressing the JAK/STAT signaling in adjacent epidermal cells. Ppn may facilitate the downregulation of this pathway by modulating the composition of the epidermal basement membrane (BM).

(B) In dC1GalT1 mutant flies, the loss of dC1GalT1 in embryonic hemocytes leads to reduced Ppn secretion. This results in aberrant upregulation of JAK/STAT signaling in the epidermis, ultimately causing the failure of stalk formation. Reduced Ppn secretion may disrupt the integrity of the epidermal BM, thereby triggering JAK/STAT hyperactivation. KO, Keilin’s organ.

Next, to examine whether glycan engineering can rescue stalk formation, we overexpressed human ST6 N-acetylgalactosaminide α2,6-sialyltransferase 1 (hST6GALNAC1) in hemocytes of dC1GalT1 mutants. Because Tn antigen expression was elevated in dC1GalT1 mutants (Figure S9),⁶ we expected that hST6GALNAC1 would synthesize sialylated Tn antigen (Siaα2,6αGalNAcα1-Ser/Thr) instead of T antigen in dC1GalT1 mutants (Figure S10A). We confirmed hST6GALNAC1 expression in larvae using the Act5C-Gal4 driver (Figure S10B). Intriguingly, hemocyte-specific overexpression of hST6GALNAC1 partially rescued the short stalk phenotype (Figure S10C), raising the possibility that sialylated Tn antigen may partially substitute for the function of core 1 glycans in stalk formation.

Discussion

Tubulogenesis is a fundamental developmental process critical for shaping epithelial architecture across organs. Although stalk-like epithelial tubes connecting imaginal discs to the epidermis have long been observed in Drosophila, their developmental origin, molecular regulation, and physiological importance have remained unexplored. In this study, we have defined the peripodial stalk as a regulated epithelial tube whose formation is essential for proper leg development. Our findings uncover a previously unrecognized morphogenetic mechanism in which mucin-type O-glycans synthesized by dC1GalT1 in embryonic hemocytes modulate the local signaling environment of epidermis to control stalk formation.

We have demonstrated that the failure of stalk formation in dC1GalT1 mutants leads to the mislocalization of the third leg discs and subsequent leg malformation. Importantly, tissue-specific rescue experiments revealed that the expression of dC1GalT1 in embryonic hemocytes—but not in the leg tissue itself—is sufficient to restore both stalk morphogenesis and leg development. These findings establish that hemocyte-derived glycoconjugates have an unexpected non-cell-autonomous role in guiding epithelial tube formation in adjacent tissues. By visualizing cytoskeletal dynamics and junctional markers during early disc development, we have delineated the stepwise process of disc formation. We also clarified that convergent extension contributes to the stalk elongation, allowing us to propose that the constriction of the disc cell cluster is required for initial stalk formation. The disruption of dC1GalT1 impaired constriction of the disc cell cluster beneath the KO, which is likely to indicate failure of this epithelial remodeling process. Mechanistically, we discovered that epidermal JAK/STAT signaling is aberrantly upregulated in dC1GalT1 mutants and can be restored to normal levels by hemocyte-specific dC1GalT1 expression. Furthermore, a slight reduction in JAK/STAT activity partially rescued stalk formation in dC1GalT1 mutants. We confirmed that both forced activation and inactivation of epidermal JAK/STAT activity in WT were sufficient to disrupt stalk formation. Therefore, these data identify moderate JAK/STAT activity as a critical target of glycan-mediated regulation during stalk morphogenesis. Lastly, through a focused RNAi screen, we identified that Ppn, an ECM protein with a mucin-like domain, is an essential mediator of this pathway. Hemocyte-specific knockdown of Ppn phenocopied dC1GalT1 mutants, displaying short stalks, JAK/STAT hyperactivation, and failure of disc cluster constriction. Overexpression of dC1GalT1 did not rescue the short stalk phenotype caused by the loss of Ppn, positioning Ppn downstream of dC1GalT1 in the pathway. Moreover, Ppn was robustly expressed in embryonic hemocytes and partially colocalized with T antigen. Biochemical assays confirmed that secreted Ppn carries mucin-type O-glycans, including T antigen. Furthermore, loss of dC1GalT1 impaired the secretion of Ppn from hemocyte-derived cells. These findings suggest that dC1GalT1-mediated glycosylation of Ppn facilitates its secretion from hemocytes and enables it to function as an ECM regulator that modulates epidermal signaling.

Our work identifies a glyco-regulatory axis in which embryonic blood cells secrete a glycosylated ECM protein (Ppn) that tunes local signaling pathways in neighboring epithelial tissues. This represents a previously unappreciated mechanism by which inter-tissue glycan signaling shapes organ morphogenesis. We demonstrate that mucin-type O-glycans function as non-cell-autonomous modulators of JAK/STAT signaling in the context of tubulogenesis.

During metamorphosis, mature third leg discs form sac-like structures comprising a folded columnar epithelium and an overlying squamous peripodial epithelium.³⁶ The columnar epithelium ultimately gives rise to most of the adult leg structure. At the early pupal stage, the leg disc formed by the columnar epithelium everts through the opened stalk, and then elongates into the space of the pupal case. Our ultrastructural analyses revealed that the third leg discs in dC1GalT1 mutants were ectopically located between the ventral muscles and the cuticle, and appeared compressed in contrast to WT larvae, where the discs are freely exposed to the body cavity (Figures 2C and 2Dʹ). This misplacement may mechanically prevent disc eversion and appendage elongation, leading to structural failure of the leg. Indeed, some dC1GalT1 mutants displayed partially buried or entirely missing third legs (Figures 1H and 1I), strongly supporting this model. We also found that dC1GalT1 mutant larvae exhibited short stalks anchoring other imaginal discs, as well as those of the third leg discs. However, the positioning of discs other than the third leg discs remained normal, probably because the stalks anchoring them are inherently shorter than those of the third leg discs. In contrast, the positioning of third leg discs is strongly influenced by their stalk length.

Stalk elongation seems to involve convergent extension—a conserved morphogenetic mechanism in which the cells intercalate along a narrowing axis, driving tissue elongation.^39,54 The mechanism of convergent extension has been well-characterized in germband extension, where it is driven by myosin-II-dependent apical contraction and basolateral protrusions.^45,46,55,56 We found that the disc cell cluster failed to constrict beneath the KO in dC1GalT1 mutants (Figures 3Dʹ and 3Eʹ), and knockdown of the gene encoding myosin light chain (sqh) recapitulated this phenotype (Figures S2C–S2E), indicating defective convergent extension. Notably, JAK/STAT signaling is known to regulate convergent extension in other contexts, including gut elongation and germband extension.^42,57 Thus, our findings position JAK/STAT signaling as a key modulator of epithelial remodeling during stalk tubulogenesis and highlight mucin-type O-glycans as upstream regulators of this pathway.

Ppn is well known as a heavily glycosylated ECM protein with a Ser/Thr-rich domain that carries glycosaminoglycans.^49,50 We found that impaired synthesis of glycosaminoglycans in hemocytes did not disrupt stalk formation (Figure S7), suggesting that glycosaminoglycans on Ppn may be dispensable for its secretion from hemocytes and its bioactivity for stalk formation. Our biochemical analysis demonstrated that Ppn also carries mucin-type O-glycans, including T antigen, in Drosophila. This finding is consistent with previous reports indicating that the human ortholog Papilin is modified by mucin-type O-glycans.^{58,59,60,61,62} Expression analysis confirmed that Ppn is highly expressed in embryonic hemocytes (Figure 6C), coinciding with the developmental window in which dC1GalT1 function is required for stalk formation. Our observed expression of Ppn is consistent with previous studies showing that hemocytes mainly produce Ppn during embryonic stages, whereas the lymph gland mainly produces Ppn during larval stages.^50,53,63,64 We also found that dC1GalT1 expression during embryonic stages, but not larval stages, is required for stalk formation (Figure S1). Together, these data support a model in which glycosylated Ppn secreted from embryonic hemocytes modulates local signaling and tissue architecture.

Our further analysis revealed that defective synthesis of mucin-type core 1 glycans impaired Ppn secretion from hemocyte-derived cells (Figures 6F–6J), suggesting that core 1 glycans on Ppn are required for its secretion. These data are consistent with previous studies showing that loss of GALNT genes results in the impaired secretion of ECM components, including Ppn.^65,66 Our data also showed that loss of dC1GalT1 does not increase intracellular Ppn levels, despite impaired Ppn secretion. This raises the possibility that hypoglycosylation of core 1 glycans may destabilize Ppn, thereby promoting its intracellular degradation. Therefore, we propose that impaired Ppn secretion due to hypoglycosylation is the main cause of the stalk defects in dC1GalT1 mutants (Figure 7B). Drosophila is known to possess 10 GALNT paralogs. According to DGET (Drosophila Gene Expression Tool: https://www.flyrnai.org/tools/dget/web/; accessed on 14 November 2025), five of these paralogs—Pgant1, Pgant3, Pgant5, Pgant6, and Pgant7—are moderately expressed in Kc167 cells. These paralogs may therefore contribute to the mucin-type O-glycosylation of Ppn in embryonic hemocytes.

Interestingly, dGlcAT-P—the enzyme responsible for glucuronylation of T antigen—has been previously shown to regulate neuromuscular junctions and VNC morphology,^10,67 and dGlcAT-P mutant phenotypes are reminiscent of those observed in dC1GalT1 mutants.^8,34,53 The VNC phenotype in dGlcAT-P mutants is rescued by hemocyte-specific expression of dGlcAT-P,⁶⁷ indicating that glucuronylated T antigen from hemocytes has a key role in VNC formation. Furthermore, Ppn mutants also exhibit VNC elongation,⁵³ suggesting that Ppn may act as a functional carrier of this glycan. It is therefore plausible that the short stalk phenotype of dC1GalT1 mutants may result from loss of glucuronylated T antigen, highlighting the importance of specific glycan structures for Ppn secretion. Future analysis will reveal whether the glucuronylated T antigen is involved in Ppn secretion and stalk formation.

Our analysis revealed that the forced expression of hST6GALNAC1 in hemocytes partially rescued stalk formation in dC1GalT1 mutants (Figure S10), indicating that the synthesis of sialylated Tn antigen (Siaα2,6GalNAcα1-Ser/Thr), instead of core 1 glycans, may facilitate Ppn secretion, leading to stalk elongation. This raises the possibility that sialylated Tn antigen possesses bioactivity similar to that of core 1 glycans, including glucuronylated T antigen. Given that sialylated Tn antigen carries a negative charge, similar to glucuronylated T antigen, the negative charge from the terminal monosaccharide might contribute, at least in part, to Ppn stability and secretion.

Ppn localizes in the basement membrane (BM) of multiple Drosophila tissues, including the VNC, gut, malpighian tube, and proventriculus.^49,50,52 Hemocytes contribute to the BM formation of the embryonic VNC by secreting Ppn that is incorporated into the VNC BM.⁵⁰ Although the domain structure of Ppn is well conserved from nematodes to humans, only Drosophila Ppn contains a Ser/Thr-rich domain.^50,68 Drosophila Ppn also includes a papilin cassette with one thrombospondin type-1 repeat (TSR) domain, a cysteine-rich spacer domain, and several partial TSR domains at the amino-end, homologous to ADAMTS-like proteins but lacking proteolytic function (Figure 6A).^50,68,69 Notably, Drosophila Ppn can bind and inhibit the procollagen-processing activity of vertebrate ADAMTS2 in vitro,⁵⁰ and C. elegans Ppn regulates collagen remodeling by modulating the access of ADAMTS proteases to the BM.⁷⁰ These observations suggest that Ppn functions as a regulator of BM composition by controlling the localization and activity of collagen-modifying enzymes during development. Ppn also colocalizes with major ECM components—namely, collagen IV, laminin, nidogen, and Trol (Drosophila perlecan)—and is required for their proper organization in Drosophila larval lymph glands.⁶⁴ In addition, it negatively regulates EGFR signaling in the lymph gland via interaction with perlecan, indicating that Ppn may contribute to the sequestration of EGFR ligands. In the context of leg disc development, we propose that hemocyte-secreted glycosylated Ppn contributes to BM organization of epidermal cells around the KO, thereby limiting JAK/STAT signaling by sequestering its ligands and promoting stalk tubulogenesis. In contrast, defective mucin-type O-glycosylation of Ppn impairs its secretion from hemocytes, leading to excessive ligand exposure caused by epidermal BM deformation, which in turn inhibits stalk tubulogenesis. Consistent with this hypothesis, our previous study demonstrated that dC1GalT1 mutants exhibit BM deformation on larval muscles, as revealed by ultrastructural analysis.⁸

Taken together, our findings establish a glycan-mediated inter-tissue signaling axis in which mucin-type O-glycans, synthesized by dC1GalT1 in embryonic hemocytes and presented on the ECM protein Ppn, regulate epithelial morphogenesis by modulating the BM environment. By limiting epidermal JAK/STAT signaling around the KO, glycosylated Ppn promotes the formation of the peripodial stalk—a previously overlooked epithelial structure essential for proper leg disc positioning and adult leg morphogenesis. Our work expands the functional scope of mucin-type O-glycans beyond cell-autonomous glycoprotein modification, positioning them as long-range modulators of signaling and tissue architecture. Such principles may extend to vertebrate systems, where glycan-modified ECM components are increasingly recognized as key regulators of morphogenesis, regeneration, and disease.

Limitations of the study

Although our PNA lectin blotting indicated that Ppn was modified by T antigen, this was not directly confirmed by mass spectrometry. Biochemical analysis also revealed that loss of dC1GalT1 impaired extracellular Ppn levels in the culture medium of Kc167 cells, suggesting reduced secretion of Ppn. However, we cannot exclude the possibility that Ppn undergoes extracellular degradation after secretion due to its destabilization. Furthermore, it is not currently clear whether Ppn secretion from hemocytes is impaired in vivo.

Lastly, we found that the hemocyte-specific overexpression of hST6GALNAC1 partially rescued stalk formation in dC1GalT1 mutants. However, synthesis of sialylated Tn antigen in the flies was not demonstrated, although the expression of hST6GALNAC1 was confirmed.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Shoko Nishihara (shoko@soka.ac.jp).

Materials availability

Transgenic flies, antibody, and plasmid generated in this study are available from the lead contact without restriction.

Data and code availability

All data reported in this article will be shared by the lead contact upon request. This article does not report original code. Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We thank Dr. Daisuke Yamamoto, Dr. Krzysztof Jagla, Dr. Erika A. Bach, and Dr. Shigeo Hayashi for fly stocks. We also thank Bloomington Drosophila Stock Center, Vienna Drosophila Resource Center, Fly Stocks of the National Institute of Genetics, Kyoto Drosophila Stock Center, Developmental Studies Hybridoma Bank, and Drosophila Genomics Resource Center. This work was partially supported by funds from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), the Matching Fund for Private Universities, S0901095, 2009–2014 to SN and JSPS KAKENHI Grant Number JP23K14147 to KI. Shoko Nishihara and Kazuyoshi Itoh are supported by the “Human Glycome Atlas Project” as one of the “Large-Scale Academic Frontier Promotion Projects” of the MEXT. GaLSIC is also supported by J-GlycoNet, one of the official Joint Usage/Research Centers of Japan by the MEXT.

Author contributions

Conceptualization, T.J.F., K.I., and S.N.; methodology, T.J.F., K.I., and S.N.; validation, T.J.F., K.I., T.I., Y.A., and S.N.; investigation, T.J.F., K.I., T.I., and Y.A.; writing – original draft, K.I. and S.N.; writing – review and editing, T.J.F., K.I., T.I., Y.A., and S.N.; visualization, T.J.F., K.I., T.I., Y.A., and S.N.; supervision, S.N.; project administration, S.N.; funding acquisition, K.I. and S.N.

Declaration of interests

The authors declare no conflicts of interest.

Declaration of generative AI and AI-assisted technologies in the writing process

The authors declare no AI was used to generate or evaluate data or to write the article.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-Fas III	Developmental Studies Hybridoma Bank	Cat# 7G10; RRID: AB_528238
Rabbit polyclonal anti-Ppn	Sigma-Aldrich (custom service)	N/A
Mouse monoclonal anti-6×His tag	Proteintech	Cat# 66005-1-Ig; RRID: AB_11232599
Rabbit polyclonal anti-α-tubulin	GeneTex	Cat# GTX112141; RRID: AB_10722892
Cy5-conjugated anti-mouse IgG	Jackson ImmunoResearch Laboratories	Cat# 715-175-151; RRID: AB_2340820
Cy3-conjugated anti-rabbit IgG	Jackson ImmunoResearch Laboratories	Cat# 111-165-003; RRID: AB_2338000
Cy5-conjugated anti-rabbit IgG	Jackson ImmunoResearch Laboratories	Cat# 111-175-144; RRID: AB_2338013
HRP-conjugated anti-mouse IgG	Cell Signaling Technology	Cat# 7076S; RRID: AB_330924
HRP-conjugated anti-rabbit IgG	Cell Signaling Technology	Cat# 7074S; RRID: AB_2099233
Chemicals, peptides, and recombinant proteins
TOTO-1 iodide	Thermo Fisher Scientific	Cat# T3600
Biotinylated PNA	Seikagaku	N/A (discontinued)
FITC-conjugated PNA	Seikagaku	N/A (discontinued)
HRP-conjugated PNA	Seikagaku	300301 (discontinued)
Rhodamine-conjugated HPA	EY Laboratories	Cat# R-3601-1
HRP-conjugated HPA	EY Laboratories	Cat# H-3601-1
Alexa Fluor-647-conjugated streptavidin	Thermo Fisher Scientific	Cat# S21374
EcoRI-HF	New England Biolabs	Cat# R3101S
KpnI-HF	New England Biolabs	Cat# R3142S
XhoI	New England Biolabs	Cat# R0146S
Cellfectin™ II Reagent	Thermo Fisher Scientific	Cat# 10362100
Sf-900TM II SFM medium	Thermo Fisher Scientific	Cat# 10902096
Ni-NTA agarose	QIAGEN	Cat# 30210
Platinum™ SuperFi II DNA Polymerase	Thermo Fisher Scientific	Cat# 12361010
Protease inhibitor cocktail	Nacalai Tesque	Cat# 04080-24
ECL Prime western blotting detection reagent	Cytiva	Cat# RPN2232
TRI Reagent®	Molecular Research Center	Cat# TR118
Critical commercial assays
In-Fusion® HD Cloning Kit	Takara Bio	Cat# 639649 (discontinued)
MEGAscript™ T7 Transcription Kit	Thermo Fisher Scientific	Cat# AMB13345
SuperScript™ VILO™ MasterMix	Thermo Fisher Scientific	Cat# 11755250
FastStart Universal SYBR Green Master	Roche	Cat# 4913914001
Experimental models: Cell lines
Kc167	Drosophila Genomic Resource Center	Cat# 1; RRID: CVCL_Z834
Experimental models: Organisms/strains
Drosophila melanogaster: Canton S	Dr. Daisuke Yamamoto	N/A
Drosophila melanogaster: 1151-Gal4	Dr. Krzysztof Jagla	N/A
Drosophila melanogaster: UAS-dome5.1	Dr. Erika A. Bach	N/A
Drosophila melanogaster: 10×STAT-dGFP	Dr. Erika A. Bach	N/A
Drosophila melanogaster: UAS-GFPπRas;Dll-Gal4	Dr. Shigeo Hayashi	N/A
Drosophila melanogaster: UAS-dC1GalT1	Fuwa et al.7	N/A
Drosophila melanogaster: dC1GalT1EY13370	Bloomington Drosophila Stock Center	Cat# 20876; RRID: BDSC_20876
Drosophila melanogaster: dC1GalT1c01812	Bloomington Drosophila Stock Center	N/A (discontinued)
Drosophila melanogaster: dC1GalT1KG02976	Bloomington Drosophila Stock Center	Cat# 13500; RRID: BDSC_13500 (discontinued)
Drosophila melanogaster: Df(2L)Exel7040	Bloomington Drosophila Stock Center	Cat# 7811; RRID: BDSC_7811
Drosophila melanogaster: PpnG4259	Bloomington Drosophila Stock Center	Cat# 31782; RRID: BDSC_31782
Drosophila melanogaster: PpnMI03189	Bloomington Drosophila Stock Center	Cat# 36216; RRID: BDSC_36216
Drosophila melanogaster: PpnMB07666	Bloomington Drosophila Stock Center	Cat# 25277; RRID: BDSC_25277
Drosophila melanogaster: PpnMB01784	Bloomington Drosophila Stock Center	Cat# 23361; RRID: BDSC_23361
Drosophila melanogaster: Stat92E06346	Bloomington Drosophila Stock Center	Cat# 11681; RRID: BDSC_11681
Drosophila melanogaster: Df(3R)BSC322	Bloomington Drosophila Stock Center	Cat# 24347; RRID: BDSC_24347
Drosophila melanogaster: Act5C-Gal4	Bloomington Drosophila Stock Center	Cat# 3953; RRID: BDSC_3953
Drosophila melanogaster: arm-Gal4	Bloomington Drosophila Stock Center	Cat# 1560; RRID: BDSC_1560
Drosophila melanogaster: Hml-Gal4	Bloomington Drosophila Stock Center	Cat# 6395; RRID: BDSC_6395 Cat# 30140; RRID: BDSC_30140
Drosophila melanogaster: He-Gal4	Bloomington Drosophila Stock Center	Cat# 8699; RRID: BDSC_8699
Drosophila melanogaster: C855a-Gal4	Bloomington Drosophila Stock Center	Cat# 6990; RRID: BDSC_6990
Drosophila melanogaster: Ubx-Gal4	Bloomington Drosophila Stock Center	Cat# 45198; RRID: BDSC_45198
Drosophila melanogaster: tsh-Gal4	Bloomington Drosophila Stock Center	Cat# 3040; RRID: BDSC_3040
Drosophila melanogaster: GugAGiR-Gal4	Bloomington Drosophila Stock Center	Cat#: 6773; RRID: BDSC_6773
Drosophila melanogaster: elav-Gal4	Bloomington Drosophila Stock Center	Cat# 458; RRID: BDSC_458
Drosophila melanogaster: sd-Gal4	Bloomington Drosophila Stock Center	Cat# 8609; RRID: BDSC_8609
Drosophila melanogaster: esg-Gal4	Bloomington Drosophila Stock Center	Cat# 93857; RRID: BDSC_93857
Drosophila melanogaster: tub-Gal80ts	Bloomington Drosophila Stock Center	Cat# 7017; RRID: BDSC_7017
Drosophila melanogaster: UAS-Stat92E RNAi (HMS00035)	Bloomington Drosophila Stock Center	Cat# 33637; RRID: BDSC_33637
Drosophila melanogaster: UAS-tdTomato	Bloomington Drosophila Stock Center	Cat# 36327; RRID: BDSC_36327
Drosophila melanogaster: UAS-GFP	Bloomington Drosophila Stock Center	Cat# 1522; RRID: BDSC_1522
Drosophila melanogaster: sqh-GFP	Bloomington Drosophila Stock Center	Cat# 57145; RRID: BDSC_57145
Drosophila melanogaster: UAS-Ppn RNAi (108005KK)	Vienna Drosophila Resource Center	Cat# 108005
Drosophila melanogaster: UAS-Ppn RNAi (16523GD)	Vienna Drosophila Resource Center	Cat# 16523
Drosophila melanogaster: UAS-sqh RNAi (109493KK)	Vienna Drosophila Resource Center	Cat# 109493
Drosophila melanogaster: UAS-β3GalTII RNAi (7949GD)	Vienna Drosophila Resource Center	Cat# 7949
Drosophila melanogaster: UAS-Ppn RNAi (18436R-1)	Fly Stocks of the National Institute of Genetics	Cat# 18436R-1 (discontinued)
Drosophila melanogaster: UAS-Ppn RNAi (18436R-4)	Fly Stocks of the National Institute of Genetics	Cat# 18436R-4
Drosophila melanogaster: UAS-Stat92E RNAi (4257R-2)	Fly Stocks of the National Institute of Genetics	Cat# 4257R-2
Drosophila melanogaster: UAS-dome RNAi (HMS01293)	Fly Stocks of the National Institute of Genetics	Cat# HMS01293
Drosophila melanogaster: UAS-dXylT RNAi (17772R-3)	Fly Stocks of the National Institute of Genetics	Cat# 17772R-3
Drosophila melanogaster: UAS-β4GalT7 RNAi (11780R-1)	Fly Stocks of the National Institute of Genetics	Cat# 11780R-1
Drosophila melanogaster: cno-GFP	Kyoto Drosophila Stock Center	Cat# 115111; RRID: DGGR_115111
Drosophila melanogaster: UAS-hST6GALNAC1 (1)	This study	N/A
Drosophila melanogaster: UAS-hST6GALNAC1 (2)	This study	N/A
Recombinant DNA
pUAST	Drosophila Genomic Resource Center	Cat# 1000; RRID: DGRC_1000
pAc5.1/V5-His	Thermo Fisher Scientific	Cat# V411020
Software and algorithms
ImageJ	National Institute of Health, USA	RRID: SCR_003070; https://imagej.net/ij/
Microsoft Excel	Microsoft	RRID: SCR_016137; https://www.microsoft.com/en-gb/microsoft-365/excel
Other
Nunc™ Lab-Tek™ II Chamber Slide™	Thermo Fisher Scientific	Cat# 154534PK
Amicon® Ultra - 0.5 mL centrifugal filters (30K)	Millipore	Cat# UFC503024

Experimental model and study participant details

Drosophila husbandry

Fly stocks were maintained on a standard cornmeal-yeast-glucose diet at 25 °C, except as noted below. Flies carrying the temperature-sensitive tub-Gal80^ts transgene were maintained at 18 °C to suppress Gal4 activity and shifted to 29 °C to induce transgene expression (Figures 4A–4F and S1). For the experiments in Figures 5A, S5, and S6, flies were reared at 29 °C. Both male and female larvae were used in the experiments, and no sex-specific differences were observed in the phenotypes or signaling outcomes analyzed. Accordingly, sex was not treated as a biological variable in the analyses.

Culturing of Drosophila Kc167 cells

Kc167 cells were grown at 28 °C in Sf-900 II SFM medium (Thermo Fisher Scientific, Waltham, MA, USA).

Method details

Fly stocks

Canton-S (kindly provided by Dr. Daisuke Yamamoto) was used as WT Drosophila melanogaster. We used the following strains: 1151-Gal4 (kindly provided by Dr. Krzysztof Jagla), UAS-dome^5.1 and 10×STAT-dGFP (both kindly provided by Dr. Erika A. Bach), UAS-GFPπRas;Dll-Gal4 (kindly provided by Dr. Shigeo Hayashi), and UAS-dC1GalT1⁷; dC1GalT1^EY13370, dC1GalT1^c01812, dC1GalT1^KG02976, Df(2L)Exel7040, Ppn^G4259, Ppn^MI03189, Ppn^MB07666, Ppn^MB01784, Stat92E⁰⁶³⁴⁶, Df(3R)BSC322, Act5C-Gal4, arm-Gal4, Hml-Gal4, He-Gal4, C855a-Gal4, Ubx-Gal4, tsh-Gal4, Gug^AGiR-Gal4, elav-Gal4, sd-Gal4, esg-Gal4, tub-Gal80^ts, UAS-Stat92E RNAi (HMS00035), UAS-tdTomato, UAS-GFP, and sqh-GFP (all from Bloomington Drosophila Stock Center); UAS-Ppn RNAi (108005KK), UAS-Ppn RNAi (16523GD), UAS-sqh RNAi (109493KK), and UAS-β3GalTII RNAi (7949GD) (all from Vienna Drosophila Resource Center); UAS-Ppn RNAi (18436R-1), UAS-Ppn RNAi (18436R-4), UAS-Stat92E RNAi (4257R-2), UAS-dome RNAi (HMS01293), UAS-dXylT RNAi (17772R-3), and UAS-β4GalT7 RNAi (11780R-1) (all from Fly Stocks of the National Institute of Genetics); and cno-GFP (from Kyoto Drosophila Stock Center).

Generation of UAS-hST6GALNAC1 transgenic flies

The construct, pUAST-hST6GALNAC1, was generated by In-Fusion® cloning (Takara Bio, Shiga, Japan). The full-length coding sequence of hST6GALNAC1 (GenBank: NM_018414) was amplified from cDNA prepared from LNCaP cells (human prostate cancer cell line) using the forward primer 5′-AGGGAATTGGGAATTCAAAATGAGGTCCTGCCTGTGG-3′ and the reverse primer 5′-AAAGATCCTCTAGAGTCAGTTCTTGGCTTTGGCA-3′. The amplified DNA fragment was inserted into pUAST vector, which was cut by EcoRI-HF (New England Biolabs, Ipswich, MA, USA) and KpnI-HF (New England Biolabs) using an In-Fusion® HD Cloning Kit (Takara Bio) to yield pUAST-hST6GALNAC1. Microinjection of the vector into Drosophila embryos and generation of UAS-hST6GALNAC1 transgenic flies were performed by BestGene (Chino Hills, CA, USA). Two lines of transgenic flies, UAS-hST6GALNAC1 (1) (inserted in 3rd chromosome) and UAS-hST6GALNAC1 (2) (inserted in 3rd chromosome), were used for the experiments.

Transmission electron microscopy imaging

Transmission electron microscopy was carried out as described previously.^8,71 The body walls of third instar larvae from each genotype were fixed with 2.5% glutaraldehyde in PBS at 4 °C overnight. The tissues were then postfixed with 1% OsO₄ in 100 mM phosphate buffer (pH 7.3) at 4 °C for 1 h and dehydrated in a graded series of ethanol. After passage through propylene oxide, the specimens were embedded in Epon 812. Ultrathin sections were cut, stained with uranyl acetate and lead citrate, and observed with a JEM-1010C transmission electron microscope (JEM-1010C; JEOL, Tokyo, Japan).

Immunostaining and lectin staining

Imaginal discs and stalks of third-instar larvae were fixed and stained by a previously described method⁷² using mouse monoclonal anti-Fas III antibody (1:200; 7G10; Developmental Studies Hybridoma Bank, Iowa City, IA, USA) and TOTO-1 iodide (1:1,000; Thermo Fisher Scientific). Cy5-conjugated anti-mouse IgG antibody (1:300; Jackson ImmunoResearch Laboratories, West Grove, PA, USA) was used as a secondary antibody.

For staining of circulating hemocytes in embryos, WT embryos were fixed and stained at 0–20 h by a previously described method⁶ using the following antibody and lectins: rabbit polyclonal anti-Ppn antibody, generated against the Ppn N-terminal peptide C-LRQKRQYGANMYLPE lacking the signal sequence (1:2,000; Sigma-Aldrich, St. Louis, MO, USA); biotinylated PNA (1:400; Seikagaku, Tokyo, Japan); and rhodamine-conjugated HPA (1:200; EY Laboratories, San Mateo, CA, USA). Cy3-conjugated anti-rabbit IgG antibody (1:300; Jackson ImmunoResearch Laboratories), Cy5-conjugated anti-rabbit IgG antibody (1:300; Jackson ImmunoResearch Laboratories), and Alexa Fluor-647-conjugated streptavidin (1:400; Thermo Fisher Scientific) were used for secondary staining.

For staining of circulating hemocytes in third-instar larvae, hemolymph was harvested by bleeding the larvae in Ca²⁺-free Drosophila Ringer’s solution (pH 7.5). The hemolymph was transferred to a Nunc™ Lab-Tek™ II Chamber Slide™ (Thermo Fisher Scientific), and the hemocytes were fixed in PLP (2% paraformaldehyde, 0.01 M NaIO₄, 0.075 M lysine, 0.037 M sodium phosphate, pH 7.2). After washing with PBS-DT (0.3% sodium deoxycholate, 0.3% Triton X-100 in PBS), the hemocytes were incubated with rabbit anti-Ppn antibody (1:2000; Sigma-Aldrich) and FITC-conjugated PNA (1:400; Seikagaku) at 4 °C overnight. Cy3-conjugated anti-rabbit IgG antibody (1:300; Jackson ImmunoResearch Laboratories) was used for secondary staining. All images were collected on an LSM510 confocal laser microscope (Carl Zeiss, Jena, Germany).

Imaging of larval epidermal/disc cells and pupal legs during development

A living late-second instar larva (36–48 h AEH) or pupa (0 h APF) was gently sandwiched between two cover glasses that had been moistened with PBS. GFP-labeled larval epidermal/disc cells or pupal legs were observed by an LSM510 confocal laser microscope (Carl Zeiss).

Recombinant plasmid construction

The consensus sequence (1–5340 bp) of four splicing variants of Ppn was amplified from cDNA prepared from Drosophila third instar larvae using the forward primer 5ʹ-AGACCCCGGATCGGGGTACCGATATGGATTTATCGAGGCG-3ʹ and the reverse primer 5ʹ-GCCCTCTAGACTCGAGGCAGTTGTAGTTGCAGGC-3ʹ. The amplified fragments were inserted into vector pAc5.1/V5-His (Thermo Fisher Scientific), which was cut by KpnI-HF (New England Biolabs) and XhoI (New England Biolabs), using an In-Fusion® HD Cloning Kit (Takara Bio).

Plasmid transfection

Kc167 cells were seeded in a 6-cm dish at 7 × 10⁶ cells and incubated for 24 h. Both the 2 μg of expression vector and 20 μL of Cellfectin™ II Reagent (Thermo Fisher Scientific) were diluted separately in 500 μL of Sf-900^TM II SFM medium and then mixed for 15 min at room temperature. The culture medium was removed and 1 mL of transfection mixture was added. After 5 h of incubation using a seesaw shaker at room temperature, 4 mL of Sf-900^TM II SFM medium was added. The cells were incubated for 3 days at 28 °C and the culture medium was harvested.

Purification of recombinant protein

After addition of 250 μL of 8 × equilibration buffer (400 mM NaH₂PO₄, 2.4 M NaCl; pH 8.0) to 1750 μL of harvested culture medium, the supernatant was collected by centrifugation at 13,200 × g for 20 min at 4 °C. The supernatant was then incubated with 200 μL of Ni-NTA agarose (50% slurry; QIAGEN, Venlo, Netherlands) equilibrated by equilibration buffer (50 mM NaH₂PO₄, 300 mM NaCl, pH 8.0) using a rotator overnight at 4°C. After centrifugation at 200 × g for 5 min at 4 °C, Ni-NTA agarose was washed three times with 500 μL of wash buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10 mM imidazole; pH 8.0). After removal of the wash buffer by centrifugation at 200 × g for 5 min at 4 °C, remaining Ni-NTA agarose was incubated with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer (62.5 mM Tris-HCl, 10% glycerol, 2% SDS, 1% 2-mercaptoethanol, 0.01% bromophenol blue; pH 6.8) for 5 min at 99 °C. The supernatant containing denatured protein was harvested by centrifugation at 200 × g for 5 min at room temperature.

Generation of dsRNA for dC1GalT1 knockdown in Kc167 cells

The target DNA sequence was amplified by standard PCR with Platinum™ SuperFi II DNA Polymerase (Thermo Fisher Scientific) using Drosophila genomic DNA as a template. The following primers containing T7 promoter sequence at the 5ʹ end were used for dC1GalT1 KD1 (forward, 5′-TAATACGACTCACTATAGGGAGACGGAGCTGTTCGTCTACTCC-3′; reverse, 5′-TAATACGACTCACTATAGGGAGACCGGACATGTAGCCTTGTTT-3′) and dC1GalT1 KD2 (forward, 5′-TAATACGACTCACTATAGGGAGAGGAACAGGACGTGGGTGGACA-3′; reverse, 5′-TAATACGACTCACTATAGGGAGATGGGCAGGGCTTCGACCACAA-3′). dsRNAs were generated from the amplified DNA fragment using MEGAscript™ T7 Transcription Kit (Thermo Fisher Scientific) in accordance with the manufacturer’s instructions.

Treatment of Kc167 cells with dsRNA

Kc167 cells were seeded in a 6-well plate at 2 × 10⁶ cells in Sf-900^TM II SFM medium and incubated for 24 h. Next, 0.8 μg of expression vector and 8 μL of Cellfectin™ II Reagent (Thermo Fisher Scientific) were diluted separately in 300 μL of Sf-900^TM II SFM medium and then mixed together for 15 min at room temperature. The culture medium was removed and 600 μL of the transfection mixture was added. After 5 h of incubation using a see-saw shaker at room temperature, 400 μL of medium containing 10 μg of dsRNA was added with incubation for 1 h at room temperature. Next, 1 mL of medium was added, followed by 2 days of incubation at 28 °C. The culture medium was then replaced with 1 mL of medium containing 10 μg of dsRNA with incubation for 1 h at room temperature. Subsequently, 1 mL of medium was added, followed by 2 days of incubation at 28 °C, after which the culture medium and cells were harvested and stored at −20 °C and −80 °C, respectively, until analysis. Thawed culture medium was concentrated by Amicon® Ultra - 0.5 mL centrifugal filters (30 K; Millipore, Billerica, MA, USA) before western blot and lectin blot analyses.

Protein extraction from Kc167 cells or Drosophila embryos

The Kc167 cell pellet was resuspended in 200 μL of lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100 [Sigma-Aldrich], and protease inhibitor cocktail [Nacalai Tesque, Kyoto, Japan]; pH 7.4) and incubated on ice for 10 min. Separately, approximately 60 Drosophila embryos (17–22 h after egg laying; stage 17) were frozen using liquid nitrogen, homogenized in 50 μL of lysis buffer, and incubated on ice for 30 min. After centrifugation at 20,000 × g for 5 min at 4 °C, the supernatant was harvested and stored at −20 °C until analysis.

Western blot and lectin blot analyses

Samples were solubilized in 62.5 mM Tris-HCl (pH 6.8) containing 10% glycerol, 2% SDS, 1% 2-mercaptoethanol, and 0.01% bromophenol blue, and then denatured at 99 °C. The samples were subjected to 5% or 6% SDS-PAGE, and the separated proteins were transferred to Immobilon-P membranes (Millipore). For western blot analysis, the membrane was incubated with 1% BSA for 1 h at room temperature, and then with mouse monoclonal anti-6×His tag antibody (1:2,000; Proteintech, Rosemont, IL, USA) or rabbit polyclonal anti-α-tubulin antibody (1:2,000; GeneTex, Irvine, CA, USA) overnight at 4 °C. After three washes with TBS-T (0.1% TrintonX-100 in TBS), the membrane was incubated with horseradish peroxidase (HRP)-conjugated anti-mouse IgG antibody (1:20,000; Cell Signaling Technology, Danvers, MA, USA) or HRP-conjugated anti-rabbit IgG antibody (1:20,000; Cell Signaling Technology) for 1 h at room temperature. For lectin blot analysis, the membrane was incubated with HRP-conjugated PNA (1:20,000; Seikagaku, Tokyo, Japan) or HRP-conjugated HPA (1:1,000; EY Laboratories, San Mateo, CA, USA) overnight at 4 °C. After three washes with TBS-T, the membranes were visualized with ECL Prime western blotting detection reagent (Cytiva, Tokyo, Japan). Relative band intensities were quantified by using ImageJ software.

Real-time PCR analysis

Total RNA was extracted from Kc167 cells or wandering third-instar larvae using TRI Reagent® (Molecular Research Center, Cincinnati, OH, USA). cDNA was synthesized using SuperScript™ VILO™ MasterMix (Thermo Fisher Scientific). Real-time PCR was carried out using FastStart Universal SYBR Green Master (Roche, Basel, Switzerland) and a QuantStudio™ 12K Flex Real-Time PCR System (Thermo Fisher Scientific). Gene-specific primer sets are listed in Table S3.

Quantification and statistical analysis

The percentage of stalk phenotypes was calculated by dividing the number in each category by the total number of stalks. Stalk width was measured using ImageJ software. Real-time PCR was performed using the ΔΔCt method; Gapdh1 was the endogenous control for normalization. Relative expression levels were calculated by comparing the Ct value of the target gene to that of the reference gene; fold changes were determined using the formula 2ˆ(-ΔΔCt). Band intensities from western blot or lectin blot analyses were quantified using ImageJ software. Signal intensities of target proteins were normalized to the endogenous control detected with anti-α-tubulin antibody.

Statistical evaluation of the difference between groups was performed by Student’s t-test or Dunnett test implemented in Microsoft Excel. Differences were considered to be significant at a p-value of less than 0.05.

Published: February 18, 2026