Developmental patterning of adipose tissue by abd-A and Abd-B homeotic genes in Drosophila melanogaster

Abstract

The Bithorax Complex (BX-C) homeobox proteins specify segmental identities along the anterior-posterior axis during Drosophila embryogenesis. Differential expression of the BX-C genes abd-A and Abd-B distinguishes abdominal from thoracic adipocytes, yet the mechanism regulating this heterogeneity remains poorly understood. Here, we identify cis-regulatory elements (CREs) and transcription factors that direct abdominal-specific expression of abd-A and Abd-B in the larval fat body. Fine-mapping analyses identified a 627-bp CRE within the Abd-B locus and a ~6-kb CRE within the abd-A locus sufficient to drive heterogeneous expression. Yeast one-hybrid screening combined with functional analyses identified Lola, Lolal, and Combgap as key repressors of Abd-B, whereas Piragua (Prg) and Seven up (Svp) function as transcriptional activators, indicating that adipocyte heterogeneity in postembryonic adipose tissue is actively regulated. In turn, lola and prg are repressed by Abd-B, whereas lolal and svp are activated, forming a feedback circuit further modulated by Wnt signaling, which promotes lola and lolal expression while repressing svp. CUT&RUN analyses suggest that these interactions are direct, with dTCF/Pan and Abd-B occupancy detected at target loci. Together, our findings define a transcriptional circuit that regulates Abd-B gene transcription to pattern adipose tissue and may establish the developmental basis of fat depot specialization.

One-sentence summary:

Transcriptional regulation of abd-A & Abd-B in adipocytes

Introduction

Pioneering studies using Drosophila melanogaster have elucidated fundamental mechanisms governing body plan formation in metazoans. These processes rely on precisely coordinated spatial and temporal regulation of maternal and zygotic gene expression following fertilization, culminating in the formation of a multicellular blastoderm (Saunders 2021; Martin 2020; Martin et al. 2009). Maternal effect genes, including bicoid and nanos, establish the anterior-posterior axis of the early embryo (Kimelman and Martin 2012; Saunders 2021), whereas segmentation genes, such as gap genes and pair-rule genes, further subdivide the embryo into discrete segments (Jaeger 2011; Diaz-Cuadros et al. 2021). These early patterning cues converge on the regulation of downstream target genes, most notably the homeobox (Hox) genes. Precise control of Hox gene expression is essential for proper axial patterning and the formation of segment-specific structures (Soshnikova and Duboule 2009; Vasanthi and Mishra 2008). During postembryonic development, Polycomb group (PcG) and Trithorax group (TrxG) proteins preserve Hox gene expression states via epigenetic modification of histone tails (Schuettengruber et al. 2017; Kassis et al. 2017). However, establishing a permissive chromatin landscape alone is insufficient to generate the precise spatial and temporal transcriptional patterns in vivo. Accurate expression requires the concerted action of sequence-specific transcription factors, co-activators, and co-repressors, which engage cis-regulatory elements (CREs), recruit the transcriptional machinery, and integrate contextual signaling cues.

In Drosophila, Hox genes are organized into two clusters on the third chromosome: the Antennapedia Complex (ANT-C) and the Bithorax Complex (BX-C). The ANT-C comprises five Hox genes: labial (lab), proboscipedia (pb), Deformed (Dfd), Sex combs reduced (Scr), and Antennapedia (Antp), which specify the identities of the anterior segments (Struhl 1981). The BX-C contains three Hox genes: Ultrabithorax (Ubx), abdominal-A (abd-A), and Abdominal-B (Abd-B), which define posterior segment identities (Lewis 1978). Ubx specifies the third thoracic (T3) and first abdominal (A1) segments, whereas abd-A and Abd-B pattern the remaining abdominal segments. During early embryogenesis, the BX-C is subdivided into parasegment-specific chromosomal domains containing modular CREs that drive precise gene expression in defined posterior segments (Simon et al. 1990; Shimell et al. 2000; Qian et al. 1991; Barges et al. 2000; Zhou et al. 1999; Maeda and Karch 2006). Consistent with this organization, abd-A is predominantly expressed in abdominal segments A2 to A7 (Casares and Sánchez-Herrero 1995; Akam 1998), whereas Abd-B is expressed in all abdominal segments (A1-A7) and posterior thoracic segments during larval and adult stages (Sánchez-Herrero et al. 1985; Karch et al. 1985). In contrast, Ubx expression is restricted to T3 and A1 (Lewis 1978; Weatherbee et al. 1998). BX-C gene expression is further refined by tissue- and cell-type-specific enhancers embedded within these segment-specific regulatory domains (Simon et al. 1990; Busturia and Bienz 1993; Pirrotta et al. 1995). However, whether distinct CREs control BX-C gene expression in specific larval tissues, such as the fat body, remains unclear. Notably, BX-C proteins Abd-A and Abd-B are differentially expressed along the anterior-posterior axis of the larval fat body (Marchetti et al. 2003; Hemba-Waduge et al. 2025; Hemba-Waduge et al. 2026), suggesting regional heterogeneity within this tissue. Our previous work revealed that this differential heterogeneous expression defines distinct adipocyte populations with regional identities (Hemba-Waduge et al. 2025). However, the regulatory mechanisms underlying this heterogeneity, and the extent to which it is actively controlled at the transcriptional level, remain unresolved.

Adipose depots in vertebrates are similarly distributed across multiple anatomical regions and are commonly classified as subcutaneous or visceral. The species-specific placement of these depots suggests regulation by intrinsic developmental programs (Gesta et al. 2007; Parra-Peralbo et al. 2021). In mammals, adipose tissue comprises three major classes: white adipose tissue, which stores energy; brown adipose tissue, which dissipates energy through non-shivering thermogenesis; and beige (brite) adipocytes, inducible thermogenic cells that emerge within white depots in response to environmental or hormonal cues (Gesta et al. 2007; Parra-Peralbo et al. 2021; Berry et al. 2013; Sebo and Rodeheffer 2019). Molecular profiling over the past two decades has characterized gene programs underlying functional differences among these depots and revealed depot-specific Hox gene expression patterns (Hemba-Waduge et al. 2025).

In vertebrates, HOX genes are organized in four clusters, HoxA, HoxB, HoxC, and HoxD. HOX paralog groups 6–8 (HOX6–8) are homologous to the arthropod central Hox genes Antp, Ubx, and abd-A, whereas posterior HOX9–13 genes correspond to Abd-B (Hubert and Wellik 2023; Mark et al. 1997; Hueber et al. 2010). These genes are differentially expressed among white and brown adipose depots, although reported patterns vary across studies, potentially reflecting differences in age, sex, or physiological state (Cantile et al. 2003; Gesta et al. 2006; Cleal et al. 2017). Notably, Clemons et al. recently demonstrated that each human adipose depot exhibits a distinct HOX expression signature that correlates more strongly with anatomically related tissues than with adipocyte type, supporting the idea that adipocytes may retain positional identity (Clemons 2024). Despite these associations, defining causal roles for HOX genes in mammalian adipose depot specification remains challenging. Functional analyses are complicated by extensive paralogy within the HOX gene family and by the essential roles of these genes in embryonic patterning, making it difficult to disentangle their direct effects on adipocyte metabolism from broader functions in adipogenesis and development.

Here, we identify CREs that control abd-A and Abd-B transcription specifically in abdominal adipocytes. We define a 627-bp fragment within the Abd-B locus and an ~6-kb region within the abd-A locus that drives expression in larval adipocytes. Using the 627-bp Abd-B element as bait in a yeast one-hybrid (Y1H) screen, we identify three transcription factors, including Lola (Longitudinals lacking), Lolal (Lola like), and Cg (Combgap), that directly activate this CRE. Functional analyses show that Lola and Lolal regulate Abd-B transcription, and to a lesser extent abd-A, in the larval fat body. We further show that Prg (Piragua) and Svp (Seven up) are required for Abd-B activation. Expression of lola, lolal, svp, and prg is modulated by Wnt signaling and by Abd-B itself. Together, these results indicate that heterogeneous Abd-B expression in larval adipocytes is actively fine-tuned by a genetic circuit involving gene-specific CREs, multiple transcriptional factors, and Wnt signaling. In contrast, the regulatory mechanisms underlying heterogeneous abd-A expression are more complex and remain unresolved. These findings provide a mechanistic understanding of how the spatial patterning of abd-A and Abd-B expression influences the cellular identities across different subregions of the Drosophila larval fat body.

Materials and Methods

Drosophila stocks and maintenance

Flies were maintained at 25°C on a standard diet of cornmeal, molasses, and yeast. Stocks used in this study are listed in Supplementary Table S3. The Abd.A-Gal4 line was obtained from Dr. Samir Merabet, and the UAS-TransTimer line from Dr. Norbert Perrimon (Hudry et al. 2011; He et al. 2019). All other stocks were obtained from the Bloomington Drosophila Stock Center.

Generation of AbdB-Gal4 and abdA-Gal4 lines

AbdB-Gal4 and abdA-Gal4 driver strains were generated using a slightly modified version of the method described by Pfeiffer et al. (Pfeiffer et al. 2008). The enhancer region used to generate the AbdB^GMR34G07-Gal4 driver (1,427 bp) was subdivided into three fragments: two 400-bp fragments (RJ1 and RJ2) and one 600-bp fragment (RJ3). The RJ3 fragment was further subdivided into two overlapping 400-bp fragments (Fig. 4A). All fragments were PCR amplified using PrimeSTAR Max DNA polymerase (Takara, R045A) with primers containing 5’-CACC overhangs to enable directional cloning into the Gateway pENTR/D-TOPO vector (Thermo Fisher Scientific, K240020). LR recombination reactions (Thermo Fisher Scientific, 11791100) were used to transfer enhancer fragments into the Gateway destination vector pBPGUw (Addgene plasmid #17575). Following sequence validation, constructs were microinjected into Drosophila embryos (Rainbow Transgenic Flies and GenetiVision), and genomic integration was targeted to the attP40 site on chromosome 2. Balanced fly stocks carrying each Gal4 driver were generated using standard genetic crosses. Genomic DNA from each line was isolated and validated by PCR and sequencing.

Fig. 4. AbdA^JJ08-Gal4 and AbdB^GMR34G07-Gal4-driven depletion of Axn increases abd-A and Abd-B transcription within the Wnt-induced small adipocytes.

(A) Schematic illustrating how abd-A and Abd-B potentiate Wnt signaling through a feedforward loop that reinforces the interaction between Wnt signaling and abd-A/Abd-B. (B-E”) G-TRACE analysis following Axn depletion using AbdA^JJ08-Gal4 and AbdB^GMR34G07-Gal4 drivers. Genotypes: (B-B”) AbdA^JJ08-Gal4/+; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+; (C-C”) AbdA^JJ08-Gal4/UAS-Axn^RNAi; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+; (D-D”) +; AbdB-Gal4^GMR34G07/UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP; and (E-E”) UAS-Axn^RNAi/+; AbdB-Gal4^GMR34G07/UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP. Cells outlined by dotted lines in (B-B”) and (D-D”) represent adipocytes lacking AbdA^JJ08-Gal4 and AbdB^GMR34G07-Gal4 expression, respectively. Dotted outlines in (C-C”) and (E-E”) indicate large adipocytes with low Wnt activities. Green indicates lineage-traced expression, red indicates transient expression, and blue indicates DAPI nuclear staining. Scale bar in (B), 10μm (applies to panels B-E). Scale bar in (B’), 20μm (applies to panels B’/B”- E’/E”).

The AbdB-Gal4^{gypsy-RJ3-gypsy} driver strain was generated using the pBPGUw vector containing the RJ3 enhancer fragment. Two flanking gypsy insulator sequences were inserted using preamplified gypsy fragments with overlapping homology to the pBPGUw backbone at the SgrDI (Thermo Scientific, ER2031) and NsiI (NEB, R3127S) restriction sites. Assembly was performed using the NEBuilder HiFi DNA Assembly Cloning Kit (NEB, E5520S).

For abdA-Gal4 drivers, a 20 kb enhancer region was subdivided into 1.5–2.0 kb segments (Fig. 2A) and cloned following the same procedure used for AbdB-Gal4 generation. The bacterial artificial chromosome BACR24L18 (BACPAC Resources Center) was used as the template for generating both abd-A and AbdB-Gal4 constructs. Primer sequences are listed in Table S4.

Fig. 2. Identification of the CREs responsible for Abd-A expression in the larval fat body.

(A) Schematic representation showing the relative positions of twelve CREs identified in abd-A-Gal4 driver lines generated using the pBPGUw Gal4 vector. The CREs highlighted in green (JJ06-JJ08) exhibited distinct fat body expression. Two representative abd-A isoforms (RB and RD) are shown below. (B-I”) Analysis of the expression patterns of Abd-A (JJ06-JJ08) Gal4 drivers and Abd-A^JJ11-Gal4 driver using G-TRACE system. Genotypes: (B-C”) abdA^JJ06-Gal4/+; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+; (D-E”) abdA^JJ07-Gal4/+; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+; (F-G”) abdA^JJ08-Gal4/+; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+; and (H-I”) abdA^JJ11-Gal4/+; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+. In panels B-I”, green indicates lineage-traced expression, red indicates transient or current expression, and blue indicates DAPI nuclear staining. The thoracic and abdominal regions are labeled on the left side of the images. Scale bar (I”): 20μm (applies to all panels).

Yeast one-hybrid (Y1H) screening

Y1H screening was performed by Hybrigenics Services (http://www.hybrigenics-services.com). The sequence for Abd-B (RJ3/627bp in size) fragment was PCR-amplified and cloned into the vector pB303 which was constructed by inserting the HIS3 gene into the pFL39 backbone (Bonneaud et al. 1991). Screening was performed against a random-primed Drosophila whole embryo cDNA library constructed into pP6 that derives from the original pGADGH plasmid (Bartel et al. 1993). 100 million clones (10-fold the complexity of the library) were screened using a mating approach with the YHGX13 (Y187 ade2–101::loxP-kanMX-loxP, matα) and Y1H300 (mata) strains as previously described (Fromont-Racine et al. 1997). 258 His+ colonies were selected on a medium lacking leucine, tryptophan and histidine and supplemented with 5 mM 3-Aminotriazol to handle bait autoactivation. The prey fragments of the positive clones were amplified by PCR and sequenced at their 5’ and 3’ junctions. The resulting sequences were used to identify the corresponding interacting proteins in the GenBank database (NCBI) using a fully automated procedure. A confidence score (PBS, for Predicted Biological Score) was attributed to each interaction as previously described (Formstecher et al. 2005; Rain et al. 2001; Wojcik et al. 2002) Twenty-one candidate proteins were identified as interacting with the 627 bp fragment. Fourteen candidates (66.7%) with moderate-confidence interaction (class D, based on global PBS scores) were excluded from downstream analysis due to potential false-positives.

Cell biological analyses

Staining of larval fat bodies with BODIPY, DAPI, and phalloidin was performed as previously described (Zhang et al. 2017b; Li et al. 2022). Quantification for Fig. 1E, Fig. 5H, and Fig. 6H–6I were performed using ImageJ. Relative fluorescent intensity (RFI) was measured in thoracic and abdominal fat body regions using separate red and green channels. For Fig. 3M and Fig. 7G–7H, individual cells were manually delineated using the polygon selection tool in ImageJ. Cells located at tissue boundaries were excluded from analysis. Average cell size was calculated from 5–20 cells per image panel. All experiments were performed with three independent biological replicates. Statistical significance was assessed using one-tailed unpaired t-tests, based on a priori directional hypotheses specifying the direction of the expected outcomes.

Fig. 1. Identification of genomic regions regulating Abd-B expression in the larval fat body.

(A) Schematic showing the relative positions of six CREs identified in Abd-B-Gal4 driver lines. The CRE highlighted in green (GMR34G07/BL-49801) shows fat body-specific expression. Two representative Abd-B isoforms (RE and RB) are shown below. (B) Schematic overview of the G-TRACE system. The corresponding CRE drives Gal4 expression, activating RFP and FLP recombinase. FLP excises a STOP cassette flanked by FRT sites, removing the block between the Ubip63E promoter and nEGFP, thereby enabling EGFP expression. The real-time or current expression is marked in red, and the past or lineage-traced expression is in green. (C-G”) Lineage-tracing using the G-TRACE system with dCg-Gal4 and AbdB^GMR34G07-Gal4. Genotypes: (C-D”) dCg-Gal4/+; UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP/+ (F-G”) +; AbdB^GMR34G07-Gal4/UAS-RedStinger, UAS-FLP, Ubi-p63E-STOP-GFP. The thoracic and abdominal regions are labeled on the left side of the images. Blue indicates DAPI nuclear staining. (E) Quantification of mean fluorescence intensity of RFP and GFP signals in thoracic and abdominal regions, respectively. Signal intensity was measured across images from three independent biological replicates. Data are shown as violin plots; and genotypes are indicated below the graph. P-values were calculated using one-tailed unpaired t-tests based on directional experimental predictions. Statistical significance: P < 0.01 (**) and P < 0.001 (***). Scale bar in (G”): 20μm (applies to all images).

Fig. 5. Adipocyte-specific expression of Abd-B is regulated by a 627bp CRE.

(A) Schematic representation showing the relative positions of the RJ1, RJ2 and RJ3 fragments. The RJ3 fragment was further subdivided into two subfragments: RJ3a and RJ3b. (B-G’) Expression patterns of newly generated Abd-B-Gal4 drivers (RJ1-RJ3, RJ3a, and RJ3b) were assessed using the UAS-RedStinger reporter and compared with Abd-B-Gal4⁴⁹⁸⁰¹. Genotypes: (B-B’) UAS-RedStinger/X; +; AbdB-Gal4^GMR34G07/+ ; (C-C’) UAS-RedStinger/X; Abd.B-Gal4^RJ1 ;+; (D-D’) UAS-RedStinger/X; Abd.B-Gal4^RJ2 ;+; (E-E’) UAS-RedStinger/X; Abd.B-Gal4^RJ3 ;+; (F-F’) UAS-RedStinger/X; Abd.B-Gal4^RJ3a ;+; and (G/-G’) UAS-RedStinger/X; Abd.B-Gal4^RJ3b ;+. In all panels, blue indicates DAPI nuclear staining and red indicates RedStinger expression. The thoracic and abdominal regions are labeled on the left side of the images. The scale bar in (G’) applies to all images: 20μm. (H) Quantification of mean fluorescence intensity of the red channel within thoracic and abdominal regions. Signal intensity was measured from images obtained from three independent biological replicates. Data are presented as bar graphs, with genotypes indicated below the chart. Red represents UAS-RedStinger expression, and blue corresponds to DAPI nuclear staining. Statistical significance was assessed using one-way ANOVA, which identified three significantly different groups labeled a-c. Groups assigned different letters differ significantly, whereas groups sharing the same letter are not significantly different. Error bars represent standard deviation.

Fig. 6. Identification of transcription factors regulating differential Abd-B expression in the fat body.

(A) Candidate transcription factors identified from the Y1H screen were further genetically validated for their effects on Abd-B reporter expression. Candidate proteins were classified from ‘I’ to ‘IV’ based on the global PBS score (I: very high-confidence interaction; II: high-confidence interaction; III: good-confidence interaction; and IV: moderate-confidence interaction. (B-G’) Representative confocal images showing validation of these transcription factors using the UAS-RFP/+; AbdB-Gal4^GMR34G07/+ reporter. Genotypes: (B/B’) UAS-RFP/+; AbdB-Gal4^GMR34G07/+; (C/C’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lola^RNAi; (D/D’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lolal^RNAi; (E/E’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-cg^RNAi; (F/F’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-prg^RNAi; and (G/G’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-svp^RNAi. In all panels, nuclei are labeled with DAPI (blue), and RFP expression is shown in red. The thoracic and abdominal regions are labeled on the left side of the images. The scale bar in G’: 50μm (applies to all images). (H, I) Quantification of the average fluorescence intensity of the red channel in the thoracic and abdominal regions. Red channel fluorescence intensity was measured from image obtained across three independent biological replicates. Data are presented as bar graphs, with the corresponding genotypes indicated below each graph. P-values were determined using one-tailed unpaired t-tests. Asterisks (*) indicate comparisons between the thoracic region (orange) of dCg-Gal4/+ and the thoracic regions (blue) of the indicated genotypes, whereas daggers (†) denote comparisons between the abdominal region of dCg-Gal4/+ with the abdominal regions of the indicated genotypes. Error bars represent standard deviations. Statistical significance is indicated as follows: P < 0.01 (**/††) and P < 0.001 (***/†††).

Fig. 3. AbdB^GMR34G07-Gal4 and AbdA^JJ08-Gal4-driven depletion of Axn causes severe adipocyte defects.

(A)

Schematic illustrating how Wnt activation in adipocytes leads to reduced triglyceride accumulation and elevated free fatty acid levels, resulting in smaller adipocytes. (B-D’) Representative confocal images of larval adipocytes stained with DAPI (blue; nuclei) and BODIPY (green; lipid droplets). fz3-RFP expression (red) is increased in small adipocytes in (C) and (D) but is low or absent in control adipocytes (B) and in large adipocytes. The red channel from the genotype shown in (C) is shown separately in (D’). Genotypes: (B) dCg-Gal4/+; fz3-RFP/+; (C) dCg-Gal4/UAS-Axn^RNAi; fz3-RFP/+; and (D/D’) UAS-Axn^RNAi/+; AbdB^GMR34G07/fz3-RFP. Scale bar in (D’): 10 μm. (E-L) Representative confocal images of larval adipocytes from the indicated genotypes stained with DAPI (blue; nuclei), Phalloidin (Phall, red; cortical actin/microfilament bundles beneath the plasma membrane), and BODIPY (green; lipid droplets). Genotypes: (E) +; AbdB^GMR34G07-Gal4 /+; (F) UAS-Axn^RNAi/+; AbdB^GMR34G07-Gal4/+; (G) AbdA^JJ06-Gal4/+; +; (H) UAS-Axn^RNAi/+; AbdA^JJ06-Gal4/+; (I) AbdA^JJ07-Gal4/+; +; (J) UAS-Axn^RNAi/+; AbdA^JJ07-Gal4/+; (K) AbdA^JJ08-Gal4/+; +; and (L) UAS-Axn^RNAi/+; AbdA^JJ08-Gal4/+. Scale bar in (L): 10 μm. (M) Quantification of adipocyte size in panel E-L. Data represent three independent biological replicates, and 5–20 cells measured per replicate. Error bars indicate standard deviation. P-values were calculated using one-tailed unpaired t-tests based on directional experimental predictions. Significance levels: P < 0.01 (**) and P < 0.001 (***).

Fig. 7. Depletion of Lola or Lolal under active Wnt signaling extends adipocyte defects into the thoracic region.

(A-F’) Confocal images of larval adipocytes from the indicated genotypes stained with DAPI (blue; nuclei), Phalloidin (Phall, red; cortical microfilament bundles beneath the plasma membrane), and BODIPY (green; lipid droplets). Genotypes: (A/A’) dCg-Gal4/+; +; (B/B’) dCg-Gal4/UAS-Axn^RNAi; +; (C/C’) dCg-Gal4/+; UAS-lola^RNAi/+; (D/D’) dCg-Gal4/ UAS-Axn^RNAi; UAS-lola^RNAi/+; (E/E’) dCg-Gal4/+; UAS-lolal^RNAi/+; and (F/F’) dCg-Gal4/UAS-Axn^RNAi; UAS-lolal^RNAi/+. The thoracic and abdominal regions are labeled on the left side of the images. Scale bar in panel (F’): 10μm (applies to all images). (G-H) Quantification of adipocyte size in the thoracic regions (G) corresponding to panels (A-F) and the abdominal regions (H) corresponding to panels (A’-F’). Data are presented as violin plots, with the corresponding genotypes indicated below each graph. P-values were determined using one-tailed unpaired t-tests based on predefined directional hypothesis for the expected outcomes. Asterisks (*) indicate comparisons between the thoracic or abdominal regions of dCg-Gal4/+ and the corresponding regions of the other genotypes. Daggers (†) indicate comparisons between the thoracic or abdominal regions of dCg-Gal4/UAS-Axn^RNAi and the corresponding regions of the other genotypes. Statistical significance is indicated as follows: P < 0.05 (*/†), P < 0.01 (**/††), and P < 0.001 (***/†††).

The HCR RNA-FISH assay

The multiplexed in situ HCR RNA-FISH assay was conducted as described previously (Li et al. 2022). The following specific probe sets were obtained from Molecular Instruments: B1-Alexa Fluor 488 amplifiers with probe sets for abd-A (lot number PRO337); B2-Alexa Fluor 594 amplifiers with the probe sets for Abd-B (lot numbers PRQ416 and RTG947). Confocal images were captured using a Zeiss LSM900 confocal microscope system, and representative images for each experiment were presented.

For quantification of HCR data in larval adipocytes (Fig. 8), relative fluorescence intensity (RFI) was measured across entire cells, with background signal subtracted based on measurements from regions outside the cells. RFI quantifications for HCR experiments in adipocytes is shown in Fig. 8G–J. For HCR analyses in the VNC (Fig. S7), five independent measurements were taken per VNC to quantify the distance between Ubx and Abd-B expression domains. Three independent biological replicates were analyzed. Statistical significance was assessed using one-tailed unpaired t-tests, based on predefined directional hypotheses regarding the expected outcomes.

Fig. 8. Depletion of lola or lolal under active Wnt signaling induces ectopic transcription of abd-A and Abd-B in the thoracic adipocytes.

(A-F’) HCR RNA-FISH assays detecting abd-A (green) and Abd-B (orange) mRNA transcripts in larval adipocytes. White arrowheads indicate thoracic adipocytes exhibiting ectopic abd-A or Abd-B expression. Genotypes: (A/A’) dCg-Gal4/+; +; (B/B’) dCg-Gal4/UAS-Axn^RNAi; +; (C/C’) dCg-Gal4/+; UAS-lola^RNAi/+; (D/D’) dCg-Gal4/ UAS-Axn^RNAi; UAS-lola^RNAi/+; (E/E’) dCg-Gal4/+; UAS-lolal^RNAi/+; and (F/F’) dCg-Gal4/UAS-Axn^RNAi; UAS-lolal^RNAi/+. The thoracic and abdominal regions are labeled on the left side of the images. The scale bar in panel (E) applies to panels from (A-E’) and represents 10μm. The scale bar in panel (F) applies to panels (F-F’) and represents 10μm. (G-J) Quantification of the relative fluorescence intensity (RFI) of abd-A and Abd-B mRNA transcripts in the thoracic regions corresponding to panels (A-F) and the abdominal regions corresponding to panels (A’-F’). The corresponding genotypes are indicated below each graph. P-values were determined using one-tailed unpaired t-tests. Asterisks (*) indicate comparisons between the thoracic or abdominal regions of dCg-Gal4/+ and corresponding regions of the other genotypes. Daggers (†) indicate comparisons between the thoracic or abdominal regions of dCg-Gal4/Axn^RNAi and corresponding regions of the other genotypes. Error bars represent standard deviations. Statistical significance is indicated as follows: P < 0.05 (*/†), P < 0.01 (**/††), and P < 0.001 (***/†††).

RNA-seq analysis

RNA-seq data from previously published work (Hemba-Waduge et al. 2026) are available in the Gene Expression Omnibus (GEO) under accession number GSE280511. Newly generated RNA-seq data in this study was deposited in GEO under accession number GSE317450. RNA libraries were sequenced by BGI Americas Corporation. Sequencing reads were aligned to the Drosophila melanogaster dm6 reference genome using the RNASTAR aligner (Widmann et al. 2012). Gene-level count matrices were generated using featureCounts from the Subread package (Liao et al. 2014). Differential expression analysis was performed using DESeq2 (Love et al. 2014), and genes with adjusted P value < 0.05 were considered significantly differentially expressed. Heatmaps were generated for using the ComplexHeatmap package (Gu 2022).

CUT&RUN assay

CUT&RUN experiments using the dTCF^EGFP line together with anti-GFP antibody (Abcam, ab290) or negative control rabbit monoclonal IgG were previously described (Liu et al. 2024), and the corresponding data are available in the GEO under accession number GSE264356. CUT&RUN experiments using the Abd.B^EGFP line in larval CNS were performed using the same protocol, and the resulting data was deposited in GEO under accession number GSE317446. Alignment and deduplication were performed by nf-core/cutandrun v2.0, an integrated analysis pipeline for CUT&RUN data (https://nf-co.re/cutandrun/). Downstream peak calling was performed using MACS3 with a significance threshold of q < 0.0001, followed by normalization using MACS3 (https://macs3-project.github.io/MACS/). For anti-GFP samples, alignments were provided using the “-t” option, whereas IgG controls were provided using the “-c” option. For visualization in IGV_2.14.1 (Integrative Genomics Viewer) using the Drosophila melanogaster dm6 genome assembly, normalized pileup bedGraph files generated by MACS3 (“-B” option) were converted to BigWig format using bedGraphicToBigWig tool from the UCSC toolkit (Liu et al. 2024).

Results

The identification of the genomic region responsible for the heterogeneous expression of Abd-B in larval fat body

To identify the cis-regulatory elements (CREs) responsible for the heterogeneous expression of abd-A and Abd-B in larval adipocytes, we screened transgenic Gal4 lines containing DNA fragments from the BX-C locus for their ability to drive UAS-RedStinger reporter expression in the larval fat body. These Gal4 lines, generated by the Rubin laboratory, include seven within the Abd-B locus (Table S1), but none within the abd-A locus (Jenett et al. 2012; Pfeiffer et al. 2008). Through this screen, we identified one Gal4 line within the Abd-B locus, AbdB^GMR34G07-Gal4 (Fig. 1A, Table S1) that exhibited specific activity in the larval fat body (see below in Fig. 1F–1G).

To further characterize the CRE activity, we used the G-TRACE (Gal4 Technique for Real-time and Clonal Expression) lineage tracing system, which provides spatial and temporal resolution of Gal4 activity, with nuclear EGFP labeling lineage expression and RedStinger marking current expression (Fig. 1B) (Evans et al. 2009). In this system, lineage (past) CRE-Gal4 activity induces FLP recombinase expression, resulting in permanently GFP activation through FLP-out recombination, whereas RFP, expressed directly under UAS control, labels cells with ongoing Gal4 activity (Fig. 1B). Compared to the fat body-specific dCg-Gal4 driver, which showed uniform expression throughout the tissue (Fig. 1C–C” cf. Fig. 1D–D”; quantified in Fig. 1E), the AbdB^GMR34G07-Gal4 driver was specifically active in the abdominal fat body and exhibited pronounced heterogeneity among abdominal adipocytes, with substantially lower activity in the thoracic region (Fig. 1F–F” cf. Fig. 1G–G”; quantified in Fig. 1E). This expression pattern was recapitulated using the dual-color fluorescent transcriptional timer reporter UAS-TransTimer, which combines a fast-folding destabilized GFP (edGFP) and a slow-folding, long-lived RFP to sensitively monitor dynamic transcriptional activity at single-cell resolution (Fig. S1A) (He et al. 2019). Compared to the dCg-Gal4 driver, which displayed a uniform expression pattern throughout the larval fat body (Fig. S1B–B” cf. Fig. S1C–C”), AbdB^GMR34G07-Gal4 exhibited significantly higher activity within the abdominal adipocytes (Fig. S1E–E” cf. Fig. S1D–D”). We observed heterogeneity within the abdominal region adipocyte population, together with elevated lineage expression rather than real-time expression (Fig. S1E–E”). Together, these observations suggest that the genomic region driving AbdB^GMR34G07-Gal4 expression contains a discrete CRE responsible for heterogeneous Abd-B expression in the abdominal fat body.

BX-C proteins have been reported to be expressed in the larval fat body by immunostaining (Marchetti et al. 2003; Duffraisse et al. 2020; Banreti et al. 2014). Using EGFP-tagged endogenous Ubx, abd-A, and Abd-B lines, we previously showed that Abd-A and Abd-B proteins, but not Ubx, are expressed in the larval fat body (Hemba-Waduge et al. 2026). Accordingly, our subsequent analyses focused on the regulatory mechanisms controlling the heterogeneous transcription of abd-A and Abd-B in the larval fat body.

The identification of the genomic region responsible for the heterogeneous expression of abd-A in larval fat body

The expression pattern of an abdA-Gal4 driver during embryogenesis was characterized in a previous study (Hudry et al. 2011). This abdA-Gal4 line was inactive in the larval fat body (Fig. S2), and no other enhancer-driven Gal4 lines derived from the abd-A locus have been reported. Given the crucial role of Abd-A in regulating adipocyte heterogeneity (Hemba-Waduge et al. 2025) and motivated by the identification of the AbdB^GMR34G07-Gal4 line (Fig. 1), we employed a “promoter bashing” approach to pinpoint the regulatory region of abd-A responsible for its fat body-specific expression. We divided an approximately 20-kb genomic region, encompassing the abd-A locus, into 12 fragments, each between 1.5kb and 2.0kb. These regions were designated JJ01 through JJ12 (Fig. 2A). Each genomic fragment was cloned into the pBPGUw Gal4 vector (Pfeiffer et al. 2008) and used for transgene-mediated germline transformation to generate individual fly lines reporting potential enhancer activity within JJ01-JJ12. Using the UAS-RedStinger reporter, we analyzed the Gal4 activities driven by each fragment and identified three out of 12 fragments (abdA^JJ06, abdA^JJ07, and abdA^JJ08) that showed specific expression in larval fat body (Table S2).

The abdA^JJ06-Gal4 line exhibited region-specific expression, with strong lineage labeling (GFP-positive cells) in the abdominal fat body but no detectable current expression (RFP-positive cells) (Fig. 2C/C’). No expression was detected in the thoracic region (Fig. 2B/B’). In contrast, abdA^JJ07-Gal4 displayed current, but not lineage, expression in both abdominal and thoracic adipocytes (Fig. 2D/D’ and Fig. 2E/E’). The abdA^JJ08-Gal4 line exhibited both lineage and current activity throughout the larval fat body, with comparable expression levels in abdominal and thoracic regions (Fig. 2F/F’ and Fig. 2G/G’). Notably, although abdA^JJ06-Gal4 showed no detectable activity in the thoracic region (Fig. 2B/B’), both abdA^JJ07-Gal4 and abdA^JJ08-Gal4 exhibited heterogeneous expression within abdominal and thoracic adipocytes (Fig. 2D–2G). In contrast, the remaining nine lines (abdA^JJ01-Gal4 through abdA^JJ05-Gal4 and abdA^JJ09-Gal4 to abdA^JJ12-Gal4) showed no detectable expression in the larval fat body (Fig. 2H and Fig. 2I, and Fig. S3A–E, S3I–L). Together, these observations define an approximately 6-kb CRE within the abd-A locus (from 3R: 16,824,029 to 16,817,958; Fig. 2A) that is sufficient to drive expression in abdominal adipocytes. These results are consistent with the endogenous expression pattern of abd-A, which is enriched in the abdominal larval fat body and exhibits heterogeneous expression among adipocytes (Hemba-Waduge et al. 2026).

Depleting Axin/Axn using RNA interference controlled by AbdB^GMR34G07-Gal4 or abdA^JJ08-Gal4 causes severe adipocyte defects

Having identified distinct Abd-B and abd-A CREs capable of driving tissue-heterogeneous gene expression in the larval fat body, we next examined their functional contributions to adipocyte heterogeneity. As reported previously, mutation or RNAi-mediated depletion of Axin (Axn) activates Wnt/Wingless (Wg) signaling, leading to lipid mobilization and the formation of small adipocytes (Liu et al. 2024; Zhang et al. 2017a) (Fig. 3A). Axn encodes a core scaffold protein of the β-catenin destruction complex required for degradation of β-catenin/Armadillo (Arm) in Drosophila (Bejsovec 2018). We therefore used AbdB^GMR34G07-Gal4 and newly generated abdA-Gal4 drivers with larval fat body to deplete Axn and analyzed adipocyte heterogeneity as a functional readout.

Consistent with previous findings (Hemba-Waduge et al. 2026; Liu et al. 2024), depleting Axn in the larval fat body using dCg-Gal4 specifically induced the Wnt/Wg reporter Fz3-RFP in small adipocytes (Fig. 3C compared with control in Fig. 3B). Similarly, Axn depletion driven by AbdB^GMR34G07-Gal4 significantly increased Fz3-RFP expression in small adipocytes (Fig. 3D/D’), indicating that AbdB^GMR34G07-positive adipocytes are associated with elevated Wnt activity. Compared to controls (Fig. 3E), Axn depletion using AbdB^GMR34G07-Gal4 produced pronounced heterogeneity among abdominal adipocytes (Fig. 3D and 3F; quantified in Fig. 3M). Likewise, Axn depletion driven by abdA^JJ08-Gal4 (Fig. 3L) induced adipocyte heterogeneity (Fig. 3L), whereas depletion using abdA^JJ06-Gal4 (Fig. 3H) or abdA^JJ07-Gal4 (Fig. 3J) did not produce detectable effects (quantified in Fig. 3M). Corresponding control genotypes (Fig. 3G/I/K) showed no detectable alterations in lipid metabolism or adipocyte morphology. The absence of phenotypes in abdA^JJ06-Gal4 and abdA^JJ07-Gal4 likely reflects their weaker lineage or real-time expression in abdominal adipocytes (Fig. 2C’ and Fig. 2E’). These observations suggest that sustained Gal4 activity, as observed with AbdB^GMR34G07-Gal4 and abdA^JJ08-Gal4, is required to elicit detectable effects of Axn depletion. Together, these results indicate that the CREs asscoated with AbdB^GMR34G07-Gal4 and abdA^JJ08-Gal4 confer stronger functional activity in the larval fat body, thereby enhancing the effects of Axn depletion and promoting adipocyte heterogeneity relative to other abd-A and Abd-B CREs.

Although these observations reveal a correlation between AbdB^GMR34G07-Gal4 and abdA^JJ08-Gal4 activity and Axn depletion-induced adipocyte heterogeneity, they do not directly establish how lineage or real-time driver activity relates to Wnt-dependent regulation of lipid homeostasis at single-cell resolution. We previously showed that Wnt signaling acts together with Abd-A and Abd-B to regulate adipocyte heterogeneity in the larval fat body (Hemba-Waduge et al. 2026). Specifically, Wnt signaling promotes transcription of abd-A and Abd-B, whereas Abd-A and Abd-B function permissively to enable Wnt-mediated repression of target genes involved in lipogenesis and lipolysis (Fig. 4A) (Liu et al. 2024; Hemba-Waduge et al. 2026).

To determine whether abdA^JJ08-Gal4 and AbdB^GMR34G07-Gal4 activity correlates with Wnt-dependent adipocyte heterogeneity, we combined G-TRACE lineage analysis with Wnt activation induced by Axn depletion. In control animals, adipocytes positive or negative for GFP or RFP exhibited comparable cell sizes (Fig. 4B–B”). In contrast, depletion of Axn in the G-TRACE background revealed a strong correlation between adipocyte size and lineage activity of abdA^JJ08-Gal4 (Fig. 4C–C”). Large adipocytes lacked detectable GFP or RFP signals, whereas small adipocytes exhibited strong RFP expression, either with or without GFP (Fig. 4C’/C”). These observations indicate that abdA^JJ08-Gal4-active adipocytes are responsive to Wnt signaling-induced lipid mobilization, leading to reduced cell size. Similar results were obtained using AbdB^GMR34G07-Gal4 (Fig. 4E–E” compared with control in Fig. 4D–D”).

Together, these results indicate that adipocytes with active abd-A and Abd-B expression respond to increased Wnt signaling following Axn depletion by reducing lipid accumulation and cell size. In contrast, adipocytes lacking abdA^JJ08-Gal4 and AbdB^GMR34G07-Gal4 activity appear refractory to Wnt-induced lipid mobilization and consequently remain enlarged, with no apparent change in lipid droplet morphology. Because Abd-A and Abd-B are required for adipocytes to respond to Wnt signaling (Hemba-Waduge et al. 2026), these observations further suggest that abdA^JJ08-Gal4 and AbdB^GMR34G07-Gal4 activity closely reflects endogenous abd-A and Abd-B expression and protein function.

Identification of a 627bp CRE of Abd-B controlling adipocyte-specific expression

The robust and heterogeneous activity of the 1427 bp Abd-B CRE, contained within as revealed by the AbdB^GMR34G07-Gal4 line, prompted us to further delineate the specific elements responsible for Abd-B expression in the larval fat body. We divided the 1427 bp region into three fragments (Fig. 5A), each cloned into the pBPGUw vector to generate distinct Gal4 lines: AbdB^RJ1-Gal4, AbdB^RJ2-Gal4, and AbdB^RJ3-Gal4. The activity of these potential enhancer elements was examined in larval fat body using the UAS-RedStinger reporter. As shown in Fig. 5C/C’ and Fig. 5D/D’, AbdB^RJ1-Gal4 and AbdB^RJ2-Gal4 exhibited no detectable Gal4 activity in the fat body, whereas AbdB^RJ3-Gal4 (Fig. 5E/E’) faithfully recapitulated the expression pattern of the original 1427 bp fragment (Fig. 5B/B’): AbdB^RJ3-Gal4 drove strong and heterogeneous expression in the abdominal region of the fat body, with little to no activity in the thoracic region, consistent with the pattern observed in the AbdB^GMR34G07-Gal4 line (quantified in Fig. 5H). These results indicate that the regulatory sequences mediating AbdB^GMR34G07-Gal4 activity in the larval fat body are localized within the 627 bp AbdB^RJ3 fragment.

To further refine this CRE, we subdivided AbdB^RJ3 into two overlapping 400 bp fragments, AbdB^RJ3a and AbdB^RJ3b (Fig. 5A), and generated corresponding Gal4 driver lines (AbdB^RJ3a-Gal4 and AbdB^RJ3b-Gal4). Using the UAS-RedStinger reporter, we observed that neither fragment fully reproduced the expression pattern of AbdB^RJ3-Gal4 or AbdB^GMR34G07-Gal4: both AbdB^RJ3a-Gal4 and AbdB^RJ3b-Gal4 drove markedly weaker expression, and in fewer adipocytes (Fig. 5F/F’ and Fig. 5G/G’). These observations indicate that further subdivision of the 627-bp AbdB^RJ3 region disrupts CRE integrity and reduces activity, consistent with the presence of a single CRE within this sequence. Moreover, we generated another Gal4 transgenic line in which the AbdB^RJ3 enhancer is flanked by two Ty3 (formerly gypsy) insulator sequences (Abd.B-Gal4 ^{gypsy_RJ3_ gypsy}) to minimize the effects of the surrounding chromatin environment. We observed strong and heterogeneous expression in the abdominal region of the fat body (Fig. S4A/A’), and little activity in the thoracic region (Fig. S4B/B’). Therefore, the CRE responsible for heterogeneous Abd-B expression in larval adipose tissue resides within the AbdB^RJ3 (627 bp) fragment, which drives stronger expression in abdominal adipocytes and weaker or no expression in thoracic adipocytes.

Identification of transcription factors controlling differential heterogeneous Abd-B expression in the fat body

To identify the specific transcription factors that regulate Abd-B expression in the fat body, we performed a yeast one-hybrid (Y1H) screen (Reece-Hoyes and Marian Walhout 2012; Reece-Hoyes et al. 2011), using the 627 bp AbdB^RJ3 fragment as bait. We analyzed approximately 99.6 million interactions, from which 258 clones were identified as potential interactions. Among these, seven transcription factors were identified with high to moderate confidence (groups A-C in Fig. 6A), including Combgap (Cg), Charlatan (Chn), Lola (Logitudinals lacking), Lolal (Lola-like), Piragua (Prg), TBP-associated factor 1 (Taf1), and Seven up (Svp). In addition, 14 more clones exhibited moderate to low confidence interactions with the AbdB^RJ3 fragment.

To functionally assess the roles of these candidate factors in regulating CRE activity, we utilized the recombined AbdB^GMR34G07-Gal4; UAS-RFP reporter (Fig. 6B/B’). If a given transcription factor affects the activity of this CRE in vivo, then its depletion should alter the tissue expression pattern or cell heterogeneity of RFP expression pattern driven by AbdB^GMR34G07-Gal4 (Fig. 6B/B’). Our analyses revealed that knockdown of lola (Fig. 6C/C’; multiple independent transgenic RNAi lines for lola in Fig. S5A–S5D) and lolal (Fig. 6D/D’) markedly increased RFP expression in both abdominal and thoracic adipocytes (quantified in Fig. 6H), although regional heterogeneity in the abdominal region persisted. Lola is known to regulate multiple developmental processes in Drosophila, including axonal guidance (Crowner et al. 2002), stem cell maintenance, and germ cell differentiation (Bass et al. 2007). lolal encodes a BTB/POZ domain-containing protein that interacts with Trithorax-like (Trl/GAGA associated factor), a component of Polycomb- Repressive Complexes (PRC1 and PRC2) (Faucheux et al. 2003; Mishra et al. 2003). Similarly, the depletion of combgap (cg) resulted in a pronounced increase in Abd-B reporter expression in abdominal adipocytes, accompanied by a modest elevation in the thoracic region (Fig. 6E/E’ cf. Fig. 6B/B’; quantified in Fig. 6H). Cg encodes a sequence-specific DNA-binding protein implicated in the recruitment of PcG proteins (Ray et al. 2016; Kassis et al. 2017), although its function in adipose tissue has not been previously characterized. These observations suggest that Lola, Lolal, and Cg transcription factors act as negative regulators of Abd-B promoter activity, in both thoracic and abdominal adipocytes.

In contrast, knockdown of piragua (prg) (Fig. 6F/F’) and seven up (svp) (Fig. 6G/G’) significantly reduced Abd-B reporter activity (quantified in Fig. 6H), indicating that both factors are required for the activation of the AbdB^GMR34G07 CRE. prg encodes a protein containing Zinc-Finger-Associated-Domain (ZAD) and C2H2 zinc-finger (ZF) domains, which are implicated in diverse developmental morphogenic processes (Nazario-Yepiz and Riesgo-Escovar 2017). Svp functions in neuronal development (Benito-Sipos et al. 2011), embryonic segmental patterning (Ponzielli et al. 2002), and insulin signaling regulation (Musselman et al. 2018; Begemann et al. 1995). To our knowledge, neither Prg nor Svp has been previously linked to gene regulation in larval adipose tissue. By contrast, depletion of TBP-associated factor 1 (Taf1), charlatan (chn), and hairy (hry), three additional transcription factors identified in the Y1H screen (Fig. 6A), did not produce any detectable change in AbdB^GMR34G07-Gal4 activity (Fig. S5A/A’–S5C/C’), thus these transcription factors were not pursued in further analysis.

Collectively, these results identify Lola, Lolal, and Cg as negative regulators, and Prg and Svp as positive regulators, of Abd-B enhancer activity in the larval fat body. Notably, Lola and Lolal appear to play key roles in mediating the region-specific heterogeneity (thoracic vs. abdominal regions) of AbdB^GMR34G07-Gal4 activity.

Lola and Lolal limit region-specific Wnt responses in adipocytes through Abd-B repression

As reported previously, active Wnt/Wg signaling stimulates lipid mobilization primarily in the abdominal adipocytes but not in the thoracic adipocytes (Hemba-Waduge et al. 2026). We showed that abd-A and Abd-B, which are expressed in abdominal adipocytes, act permissively to modulate Wnt target gene expression in these cells (Fig. 4A). Ectopic expression of abd-A or Abd-B in thoracic adipocytes under Wnt-activated background is sufficient to induce adipocyte defects in thoracic adipocytes (Hemba-Waduge et al. 2025; Hemba-Waduge et al. 2026). We therefore hypothesized that if depletion of lola or lolal increases Abd-B expression in thoracic adipocytes, then Wnt activation in this context should extend the adipocyte defects into thoracic region.

Depletion of Axn using the dCg-Gal4 driver activated Wnt signaling and caused severe adipocyte defects, including a prominent heterogeneity in cell population within abdominal but not thoracic adipocytes (Fig. 7B/B’ cf. control in Fig. 7A/A’; quantified in Fig. 7G and 7H). Consistent with our predictions, depletion of lola in a Wnt-activated background not only exacerbated abdominal adipocyte defects but also produced pronounced effects in thoracic adipocytes (Fig. 7D/D’; quantified in Fig. 7G and 7H). Depletion of lola alone did not substantially alter adipocyte morphology, aside from subtle membrane-associated changes (Fig. 7C/C’). Depletion of lolal under a Wnt-activated background caused qualitatively similar but weaker effects, with thoracic adipocytes showing mild heterogeneity (Fig. 7F/F’ cf. Fig. 7E/E’; quantified in Fig. 7G and 7H).

To directly assess the effects of lola and lolal depletion on abd-A and Abd-B transcription, we performed multiplexed HCR RNA-FISH (hybridization chain reaction RNA fluorescence in situ hybridization) imaging (Choi et al. 2018). Consistent with previous observations (Hemba-Waduge et al. 2026), Wnt activation caused by Axn depletion increased abd-A transcription in abdominal adipocytes, with weaker effects on Abd-B expression (Fig. 8B’ cf. Fig. 8A’; see Fig. S6A–D for separate channels; quantified in Fig. 8H). Thoracic adipocytes under Wnt-activated conditions showed a modest increase in abd-A expression compared to controls (Fig. S6A’ cf. Fig. S6B’). Depletion of lola significantly increased abd-B expression in abdominal adipocytes (Fig. 8C’ cf. Fig. 8A’) and induced a marked up-regulation of abd-B in thoracic adipocytes compared to controls (Fig. 8C cf. Fig. 8A; Fig. S6E and S6F; quantified in Fig. 8G). Notably, combing lola depletion with Wnt activation resulted in pronounced adipocyte defects in both abdominal and thoracic regions, accompanied by strong upregulation of both abd-A and Abd-B transcripts (Fig. 8D/D’ cf. Fig. 8A/A’; Fig. S6G and S6H; quantified in Fig. 8G and 8H). Under these conditions, Abd-B expression in abdominal adipocytes was further elevated relative to Wnt activation alone (Fig. 8D’ vs. Fig. 8A’).

Depletion of lolal produced similar, but weaker effects. Transcription of abd-B was markedly increased in thoracic adipocytes compared to controls (Fig. 8E cf. Fig. 8A; Fig. S6I and S6J), and Wnt activation in a lolal-depleted background extended adipocyte defects into the thoracic region (Fig. 8F/F’ vs. Fig. 8A/A’; Fig. S6K and S6L; quantified in Fig. 8G and 8H). Taken together, these results identify Lola and Lolal as physiological repressors of abd-B transcription in vivo and reveal that they contribute to the region specificity of Wnt-dependent adipocyte responses.

Role of Wnt signaling in potentiating Abd-B transcription in the larval fat body

We previously reported that active Wnt signaling potentiates, rather than activates, the expression of abd-A and Abd-B (Hemba-Waduge et al. 2026). Specifically, although abd-A and Abd-B are expressed in most abdominal adipocytes and their transcription is further enhanced by active Wnt signaling, these genes are undetectable in thoracic adipocytes. This indicates that Wnt signaling alone is insufficient to induce their de novo transcription (Hemba-Waduge et al. 2026), but that the severity of fat accumulation defects in the abdominal regions correlates with the degree of Wnt signaling activation. Axn-depletion driven by SREBP-Gal4 produces milder phenotypes and a smaller transparent abdominal region of the larval fat body, whereas the axn¹²⁷ mutant causes stronger effects and a broader transparent abdominal region (Hemba-Waduge et al. 2026).

Having identified Lolal, Lola, Svp, Prg, and Cg as transcription factors responsible for directly regulating the transcription of Abd-B, we next asked whether their expression levels are affected by Wnt signaling. To address this, we analyzed mRNA levels of these transcription factors in control thoracic fat body samples (designated as ‘WT-Th’ for wild-type thoracic) and abdominal fat body samples as ‘WT-Ab’ (Hemba-Waduge et al. 2026). We observed no obvious differences between ‘WT-Th’ and WT-Ab’ samples (Fig. 9A). We then examined the effects of Wnt signaling activation by expressing an RNAi to Axn and analyzing thoracic (‘Axn-Th’) and abdominal (‘Axn-Ab’) fat bodies expressing Axn^RNAi. As shown in Fig. 9A, Wnt activation specifically increased the mRNA levels of lola and lolal, while significantly decreasing svp expression in ‘Axn-Ab’ samples, but not in ‘Axn-Th’ samples. In contrast, prg and cg expression levels were not markedly affected by Wnt signaling activation (Fig. 9A).

Fig. 9. A genetic circuit regulating Abd-B transcription in the larval fat body.

(A) Heatmap showing the effects of Wnt signaling activation on mRNA levels of five transcription factors identified from the Y1H screen. Expression levels of abd-A, Abd-B, and Axn are included as positive controls. WT-Th, wild-type thoracic fat body; WT-Ab, wild-type abdominal fat body; Axn-Th, thoracic fat bodies expressing Axn^RNAi; Axn-Ab, abdominal fat body expressing Axn^RNAi. (B-D) Genomic tracks showing dTCF binding peaks at the svp (B), lola (C), and lolal (D) loci in wing discs. Genes activated by Wnt signaling are shown in red, and genes inhibited by Wnt signaling are shown in blue. (E) Heatmap showing the effect of Abd-B depletion in fat body on mRNA levels of five transcription factors identified from the Y1H screen. Abd-B expression is included as a positive control. (F-F”) Confocal images showing Abd-B mRNA transcripts (magenta) in the larval ventral nerve cord detected by HCR RNA-FISH together with Abd-B^EGFP expression (green). (F’) Single-channel image (green) showing Abd-B^EGFP expression, and (F”) single-channel image showing Abd-B mRNA transcripts (magenta) in the larval ventral nerve cord. The white dotted line marks the posterior tip of the larval ventral nerve cord. (G-J) Genomic tracks showing Abd-B binding peaks at the tsh (G), dac (H), lolal (I), and lola (J) loci in the larval CNS. (K) Model illustrating that Abd-B transcription in the larval fat body is regulated by multiple transcription factors, including Lola, Lolal, Svp, Prg, and Cg. Wnt signaling may further potentiate Abd-B expression by modulating the expression of lola, lolal, and Svp.

To determine whether Wnt signaling directly regulate transcription of lola, lolal, and svp, we analyzed previously published dTCF/Pan CUT&RUN datasets from Drosophila wing discs (Liu et al. 2024). dTCF/Pan is the sole transcription factor downstream of canonical Wnt/Wg signaling (Franz et al. 2017). Multiple dTCF binding peaks were detected within intronic regions of the svp locus (Fig. 9B), the promoter and first introns of lola (Fig. 9C), and the first intron of lolal (Fig. 9D). To validate these observations, we performed CUT&RUN targeting Pan/dTCF in larval fat body nuclei and identified multiple prominent peaks near the transcription start sites or within gene bodies of svp, lola, and lolal that overlapped with previously identified Pan/dTCF peaks in wing disc cells (Fig. S7). We also detected Pan/dTCF binding at both Abd-B and abd-A loci in wing discs and fat body nuclei (Fig. S8) (Hemba-Waduge et al. 2026). Although CUT&RUN from adipocyte nuclei exhibited suboptimal signal-to-noise ratio (Liu et al. 2024), the major Pan-binding sites were largely consistent across datasets.

Next, we asked whether transcription factors identified in the Y1H screen are themselves regulated by Abd-B. RNA-seq analysis of fat bodies with Abd-B depletion showed increased expression of lola and prg, accompanied by reduced expression of lolal and svp (Fig. 9E). Because CUT&RUN in larval fat body tissue is technically challenging (Hemba-Waduge et al. 2026; Liu et al. 2024), we performed CUT&RUN in the larval central nervous system (CNS), where Abd-B is expressed in the posterior ventral nerve cord (Fig. 9F–F”) (Allen et al. 2020; Duckhorn 2022).

Although CNS-derived CUT&RUN may not fully recapitulate larval adipocyte chromatin states, it provides complementary and supportive evidence. To further assess regulatory relationships of lola and lolal on Abd-B expression in an independent context, we performed additional analyses in the larval CNS. Specifically, lola and lolal were depleted using Elav-Gal4, followed by HCR analysis of abd-A and Abd-B expression. Because Elav-Gal4-driven depletion of lola using resulted in lethality with the transgenic RNAi strain FBst0035721, two alternative VDRC RNAi lines were used. The larval VNC (ventral nerve cord) is well suited for these analyses, as expression domains of abd-A, Abd-B, and Ubx are well characterized in this tissue (Allen et al. 2020; Duckhorn 2022). As shown in Fig. S9, depleting lola and lolal in the larval CNS altered Hox domain organization in a manner consistent with effects observed in the fat body. Specifically, Elav-Gal4-driven knockdown reduced the distance between the Ubx and Abd-B expression domains in the VNC (Fig. S9B””–S9D”” compared to control in Fig. S9A””; quantified in Fig. S9E), indicating an anterior expansion of the Abd-B domain. These findings parallel our fat body data, where depleting lola and lolal in a Wnt-activated background expanded Wnt-induced adipocyte phenotypes into more anterior (thoracic) regions. Together, these results support a role for lola and lolal in restricting Abd-B expression boundaries across multiple larval tissues.

Abd-B has previously been reported to negatively regulate teashirt (tsh) and dachshund (dac) in posterior tissues across different developmental contexts, although whether these interactions are direct remains unclear (Wang et al. 2013; Estrada and Sanchez-Herrero 2001). Using CUT&RUN profiling using the Abd-B^EGFP line, we identified at least one prominent binding peak within each of the tsh and dac loci (Fig. 9G,H), serving as positive controls and supporting the possibility that Abd-B directly regulates these targets. CUT&RUN analysis of larval CNS further revealed significant Abd-B occupancy at the lolal (Fig. 9I), and lola loci (Fig. 9J), suggesting potential direct regulation these transcription factors. In contrast, genome browser tracks at the svp, prg, and cg loci did not meet peak-calling thresholds in our analyses (Fig. S10).

Collectively, RNA-seq and CUT&RUN data support a regulatory model in which Abd-B expression is positively regulated by Lola, Lolal, and Cg, and negatively regulated by Svp and Prg (Fig. 6). Conversely, lola and prg are negatively regulated by Abd-B, whereas lolal and svp are positively regulated by Abd-B, together forming a feedback regulatory network. This circuit is further modulated by Wnt signaling, which promotes lola and lolal transcription while repressing svp expression (Fig. 9A). Finally, CUT&RUN analyses of dTCF/Pan and Abd-B indicate that regulation of lola, lolal, and svp is likely direct, as evidenced by the presence of binding peaks for both factors at these loci (Fig. 9K).

Discussion

Hox proteins encoded in the Bithorax Complex, including Abd-A and Abd-B, are central determinants of posterior segment identity in Drosophila. We previously showed that differential expression of endogenous abd-A and Abd-B in larval fat bodies defines adipocyte heterogeneity (Hemba-Waduge et al. 2026; Hemba-Waduge et al. 2025). However, how this heterogeneity is transcriptionally established has remained unresolved. Here, we identify cis-regulatory elements (CREs) and transcription factors that, together with Wnt signaling, form a genetic circuit controlling heterogeneous Abd-B expression in larval adipocytes (Fig. 9K). These findings provide mechanistic insight into how spatially patterned Hox gene expression contributes to the cellular identity of subregions of the Drosophila larval fat body.

cis-regulatory control of Abd-B transcription in larval adipocytes.

By mapping CREs within the Abd-B locus, we identified a 627 bp region within Abd-B that drives region-specific and heterogeneous expression in the larval fat body. This enhancer activity is consistent with our previous transcriptomic analyses showing enriched Abd-B expression in abdominal relative to thoracic adipocytes. Y1H screen and functional genetic analyses further revealed that multiple transcription factors, including Lola, Lolal, and Svp, modulate both reporter activity and endogenous Abd-B, indicating that Abd-B transcription is controlled by a combinatorial regulatory network in the larval fat body (Fig. 9K).

Lola and Lolal are paralogous BTB (BR-C, ttk and bab) or POZ (Pox virus and Zinc finger) domain-containing proteins that share high sequence similarity but have diverged functionally (Zollman et al. 1994; Bardwell and Treisman 1994; Costoya 2007; Hu et al. 2021; Hu et al. 2011). Lola has been implicated in axon guidance and ovarian cell death (Crowner et al. 2002; Bass et al. 2007), whereas Lolal functions as an effector of the Trl-GAF (Trithorax-like/GAFA factor)-mediated Polycomb Group (PcG) regulation (Faucheux et al. 2003; Mishra et al. 2003). Our data suggest a distinct role for these paralogs in larval adipocytes: Lolal is required for proper regulation of endogenous Abd-B expression, whereas Lola contributes primarily to maintaining transcriptional difference between thoracic and abdominal fat body regions.

Lolal (also known as Batman) has been shown to act in both activation and repression of homeotic genes, including repression of Sex combs reduced (Scr) (Faucheux et al. 2003). Through its BTB/POZ domain, Lolal directly interacts with Trl/GAF protein, a factor implicated in PcG recruitment and chromatin organization (Mishra et al. 2003). Lolal also genetically interacts with Polyhomeotic (Ph), and both are required for PRE (Polycomb Response Element)-mediated silencing (Faucheux et al. 2003; Mishra et al. 2003), suggesting that Lolal may facilitate PcG-dependent repression at the Abd-B locus. In addition, recent work has shown that Trl/GAF can promote long-range chromatin interactions via its POZ/BTB domain (Li et al. 2023), raising the possibility that Lolal contributes to high-order chromatin looping between Abd-B promoters and distal enhancers.

In addition to Lola and Lolal, Y1H screening identified Cg, Svp and Prg as direct binders of the 627 bp Abd-B enhancer fragment. Functional analysis using the AbdB^GMR34G07-Gal4 reporter revealed distinct regulatory effects: depletion of cg increased reporter activity, whereas depletion of svp and prg reduced it. Like Lolal, Cg is also involved in recruiting PcG proteins (Kassis et al. 2017). Notably, depletion of Cg resulted in more uniform AbdB^GMR34G07-Gal4 reporter expression within the abdominal fat body together with comparatively milder effects within the thoracic region, suggesting a role for Cg in generating interregional heterogeneity. In contrast, depleting either Lola or Lolal expanded reporter expression into thoracic regions while preserving heterogeneity, indicating that these factors primarily regulate the regional identity of larval adipocytes. Further evidence suggested that depleting lola and lolal under an activated Wnt signaling regime extended the adipocyte defects into the thoracic region (Fig. 7D and Fig. 7F cf. Fig. 7B), potentially driven by increased Abd-B transcription within thoracic adipocytes (Fig. 8D and Fig. 8F cf. Fig. 8B). This aligns with our previous observations, in which we have demonstrated that ectopic expression of Abd-B in a Wnt-activated regime is sufficient to cause adipocyte defects in the thoracic region (Hemba-Waduge et al. 2026).

Together, our findings support a model in which heterogeneous Abd-B transcription in larval adipocytes emerges from the integration of enhancer-specific transcription factor inputs and higher-order chromatin organization. Long-range promoter-enhancer interactions, potentially mediated by BTB/POZ domain-containing proteins, may enable precise spatial control of Abd-B transcription. The incomplete overlap between reporter activity and endogenous expression likely reflects the modular and hierarchical regulatory architecture of the Bithorax Complex. Indeed, Abd-B expression is controlled by multiple parasegment-specific regulatory domains separated by boundary elements that both restrict enhancer crosstalk and permit long-distance regulatory interactions (Postika et al. 2018). Tissue-specific outputs further depend on clustered transcription factor binding sites within cis-regulatory modules (Starr et al. 2011). Thus, while AbdB^GMR34G07-Gal4 serves as a useful qualitative proxy for Abd-B transcriptional activity, it captures only a subset of the regulatory logic governing endogenous Abd-B expression.

Establishing transcription factor-binding site relationships.

Many eukaryotic transcription factors recognize short DNA motifs that occur frequently throughout the genome (Ekker et al. 1991; Choi and Sinha 2006). However, only a subset of these sites is functionally engaged in transcriptional regulation. Specific gene regulation is achieved, in part, through the combinatorial context of neighboring transcription factor-binding sites and through protein-protein interactions among bound factors. Consequently, identifying and characterizing clusters of transcription factor-binding sites is essential for defining the rules that govern binding specificity and for determining the regulatory outcomes of transcription factor-DNA complex formation, including transcriptional activation or repression. Characterization of CREs, such as those identified here, is therefore critical for decoding the mechanisms underlying complex transcriptional control of key target genes.

Transcriptional regulation of abd-A in larval fat bodies.

In contrast to the better-defined mechanisms regulating heterogeneous Abd-B transcription, the transcriptional control of abd-A in larval adipocytes remains poorly understood. Using a promoter-bashing approach, we identified an ~6-kb CRE within the abd-A locus (abdA^JJ06, abdA^JJ07, and abdA^JJ08) that drives region-specific and heterogeneous expression in the larval fat body. Within this region, the ~2-kb fragment (abdA^JJ08) most closely recapitulates endogenous abd-A expression, as its activity is sufficient to drive adipocyte heterogeneity upon Axn depletion (Fig. 3). Notably, G-TRACE analyses suggest that abdA^JJ06 is active in early adipocyte lineages but show low activity in third-instar larval fat bodies, whereas abdA^JJ07 exhibited opposite pattern. In contrast, abdA^JJ08 is active in both lineage tracing and third-instar larval fat body (Fig. 2), potentially providing a sufficient temporal window for effective RNAi-mediated Axn depletion by this CRE (Fig. 3).

Despite these advances, we were unable to further delineate the minimal regulatory elements controlling abd-A expression in larval adipocytes, highlighting the complexity of its regulatory architecture. Notably, depletion of lola and lolal produced comparable effects on abd-A and Abd-B expression (Fig. 8), suggesting that these Hox genes share partially overlapping regulatory inputs while also being subject to gene-specific control mechanisms. Further comparative analyses of the three CREs (abdA^JJ06, abdA^JJ07, and abdA^JJ08) may therefore provide an entry point to dissect the regulatory logic underlying heterogeneous abd-A transcription in larval fat bodies. Given the essential role of abd-A, and its major contribution to adipocyte heterogeneity (Hemba-Waduge et al. 2026; Hemba-Waduge et al. 2025), elucidating the regulatory mechanisms governing its transcription remains an important objective for future studies.

A genetic circuit regulating Abd-B transcription in larval fat bodies.

Our previous work showed that abd-A and Abd-B expression is potentiated by active Wnt/Wg signaling, while Abd-A and Abd-B themselves act permissively to enable Wnt target gene expression (Fig. 4A) (Hemba-Waduge et al. 2026). Although the mechanistic basis of this reciprocal relationship remains unresolved, our current findings suggest that Wnt signaling promotes Abd-B transcription through a transcriptional regulatory circuit summarized in Fig. 9K. This model integrates genetic, transcriptomic, and chromatin-binding data to explain how Wnt/Wg signaling further potentiates spatially heterogeneous Abd-B expression. While key components of this circuit are supported by Abd-B CUT&RUN data generated from larval central nervous tissue, Abd-B binding may be tissue-specific, and future chromatin profiling in larval adipocytes will be required to fully validate this model.

Consistent with this framework, CUT&RUN analysis in the larval central nervous tissue revealed Abd-B occupancy at multiple nodes of the proposed regulatory network, including the svp, lola, and lolal loci, indicating that Abd-B may directly participate in shaping its own transcriptional landscape. Integration of these binding data with RNA-seq and functional perturbation analyses supports a model in which Abd-B expression is governed by a balance of activating and repressive inputs. Lola, Lolal, and Cg function as positive regulators of Abd-B transcription, whereas Svp and Prg provide inhibitory inputs. In turn, Abd-B feeds back onto this network by repressing the expression of lola and prg while activating lolal and svp, thereby establishing interconnected feedback loops that constrain Abd-B expression in a spatially heterogeneous manner. Wnt/Wg signaling further biases this circuit toward Abd-B activation by promoting lola and lolal expression while repressing svp. Supporting direct regulation, CUT&RUN profiles for both dTCF/Pan and Abd-B show binding within the lola, lolal, and svp loci (Fig. 9K), indicating that Wnt signaling and Abd-B converge on shared transcriptional targets to stabilize patterned Abd-B expression.

Taken together, this study advances our understanding of the regulatory mechanisms controlling BX-C gene expression in larval adipocytes. By delineating the relevant CREs and transcription factors, it highlights the complexity of postembryonic BX-C gene regulation. The observed heterogeneity in abd-A and Abd-B expression, as well as regional differences in larval fat body, suggest a potential role for these Hox genes in regulating adipocyte development and metabolic homeostasis. Future studies will be essential to define how Lolal and Lola modulate heterogeneous abd-A and Abd-B transcription and to determine the functional consequences for adipocyte heterogeneity and lipid metabolism.

In higher organisms, adipose tissues are distributed across multiple anatomical depots defined by intrinsic developmental programs. Extensive studies have characterized depot-specific gene expression under diverse physiological and pathological conditions, revealing molecular bases for functional heterogeneity. However, depot-specific expression of Hox gene expression and its functional consequences remain poorly understood. Although several studies have reported depot-restricted Hox expression in mice and humans, results vary across depots and studies (see recent review (Hemba-Waduge et al. 2025)), highlighting a complex, context-dependent regulatory landscape. Given the high evolutionary conservation of the transcription factors identified here from Drosophila to humans, these regulatory mechanisms may be broadly conserved.

Limitations of the study.

In addition to the limitations discussed above, our study has several constraints. First, although our data implicate CREs and transcription factor networks in controlling heterogeneous abd-A and Abd-B expression, we lack direct evidence for enhancer-promoter interactions or higher-order chromatin looping. Analysis of chromatin architecture in larval adipocytes remains technically challenging due to polyploid, limited cell numbers, and abundant lipid droplets, which currently preclude high-resolution chromosome conformation assays such as Micro-C. Second, Polycomb group (PcG) and Trithorax group (TrxG) complexes are classically considered key regulators for maintaining postembryonic Hox gene expression (Kassis et al. 2017; Schuettengruber et al. 2017; Beck et al. 2010; Kuroda et al. 2020). However, our findings suggest that BX-C gene regulation may be active and compensatory rather than passive maintenance, with tissue-specific mechanisms involving novel regulators such as Lola and Lolal predominating in this context. Third, our proposed regulatory model (Fig. 9K) is derived primarily from genetic perturbations and bulk RNA-seq analyses, which capture steady-state transcriptional outputs rather than dynamic regulatory behavior. Given that the circuit includes multiple feedforward and feedback interactions, time-resolved and quantitative approaches, including mathematical modeling, will be required to fully understand how heterogeneous abd-A and Abd-B expression is established and maintained in the larval fat body.

Supplementary Material

Supplement 1

Fig. S1 Validation of CREs responsible for Abd-B expression in the larval fat body. (A) Schematic diagram illustrating the major steps underlying the UAS-TransTimer system. Upon CRE-driven Gal4 expression, the “TransTimer” cassette, containing destabilized GFP (dGFP) and stable RFP, is activated. In this system, current/real-time expression is marked by green fluorescence, whereas lineage or past expression is indicated by red fluorescence. To further analyze the activity of the dCg-Gal4 and AbdB^GMR34G07-Gal4 drivers, we employed the UAS-TransTimer reporter system. Genotypes: (B-C”) dCg-Gal4/+; pUAST-edGFP::2A::RFP/+. (D-E”) +; pUAST-edGFP::2A::RFP/AbdB^GMR34G07-Gal4. (B-B” and D-D”) thoracic region; (C-C” and E-E”) abdominal region. Green indicates transient or current expression, red indicates lineage or past expression, and blue corresponds to DAPI staining marking nuclei. Scale bar in panel C: 20μm (applies to all images).

Fig. S2 Expression pattern of abdA-Gal4 in the larval fat body. Lineage tracing analysis of the abdA-Gal4 driver using the G-TRACE system. Genotype: +; abdA-Gal4/UAS-RedStinger, UAS-FLP,Ubi-p63E(^FRTSTOP^FRT)GFP. Red fluorescence represents RedStinger expression, indicating current Gal4 activity, whereas green fluorescence indicates GFP expression, marking lineage-tracing cells. (A-A”) thoracic region; (B-B”) abdominal region. Scale bar in panel A: 20μm (applies to all images).

Fig. S3 Expression patterns of AbdA-Gal4 drivers generated using the pBPGUw-Gal4 vector. The UAS-RedStinger reporter system was used to examine the expression patterns of the abdA-Gal4 drivers. Genotypes: (A/A’) UAS-RedStinger/X; abdA^JJ01-Gal4/+; (B/B’) UAS-RedStinger/X; abdA^JJ02-Gal4/+; (C/C’) UAS-RedStinger/X; abdA^JJ03-Gal4/+; (D/D’) UAS-RedStinger/X; abdA^JJ04-Gal4//+; (E/E’) UAS-RedStinger/X; abdA^JJ05-Gal4/+; (F/F’) UAS-RedStinger/X; abdA^JJ06-Gal4/+; (G/G’) UAS-RedStinger/X; abdA^JJ07-Gal4/+; (H/H’) UAS-RedStinger/X; abdA^JJ08-Gal4/+; (I/I’) UAS-RedStinger/X; abdA^JJ09-Gal4/+; (J/J’) UAS-RedStinger/X; abdA^JJ10-Gal4/+; (K/K’) UAS-RedStinger/X; abdA^JJ11-Gal4/+; and (L/L’) UAS-RedStinger/X; abdA^JJ12-Gal4/+. The thoracic and abdominal regions are labeled on the left side of the images. SG: Salivary glands. Scale bar in panel A: 50μm (applies to all images).

Fig. S4 The Abd-B cis-regulatory element RJ3 flanked by gypsy insulators exhibits region-specific expression patterns similar to those of RJ3 without gypsy insulators. The UAS-RedStinger reporter system was used to examine the expression pattern of the abdB-Gal4 driver containing the RJ3 fragment flanked by gypsy insulators. Genotype: UAS-RedStinger/X; AbdB-Gal4^{gypsy-RJ3-gypsy}/+; +. Red fluorescence indicates RedStinger expression driven by the corresponding AbdB-Gal4 construct. Scale bar in panel B: 20μm (applies to all images).

Fig. S5 Validation of additional Y1H assay candidates using the UAS-RFP combined with the AbdB-Gal4^GMR34G07 driver. (A-D’) The effects of depleting lola were further validated using two additional RNAi lines from the Vienna Drosophila Resource Center (VDRC: #101925 and #12573). The UAS-RFP/+; AbdB-Gal4^GMR34G07/+ reporter system was used. In all panels, blue indicates DAPI staining marking nuclei, and red indicates RFP expression. Scale bar in panel D’: 50μm (applies to all images in A-D’). Genotypes: (A/A’) UAS-RFP/+; AbdB-Gal4^GMR34G07/+; (B/B’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lola^RNAi; (C/C’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lola^RNAi [VDRC# 101925] ; and (D/D’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lola^RNAi [VDRC# 12573]. (E-G’) Analysis of Taf1, chn, and hry depletion using the UAS-RFP recombined with the AbdB-Gal4^GMR34G07 driver. Genotypes: (E/E’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-Taf1^RNAi; (F/F’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-Chn^RNAi; and (G/G’) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-Hry^RNAi. The thoracic and abdominal regions are labeled on the left side of the images. Scale bar in panel G’: 50 μm (applies to images E-G’).

Fig. S6: Analysis of abd-A and Abd-B transcript levels following depletion of lola or lolal under conditions of activated Wnt signaling. HCR RNA-FISH was used to detect abd-A (green) and Abd-B (orange) mRNA transcripts in larval adipocytes. White arrowheads indicate thoracic adipocytes expressing abd-A or Abd-B transcripts. Genotypes: (A-A”) and (B-B”) dCg-Gal4/+; +; (C-C”) and (D-D”) dCg-Gal4/UAS-Axn^RNAi; +; (E-E”) and (F-F”) dCg-Gal4/+; UAS-lola^RNAi/+; (G-G”) and (H-H”) dCg-Gal4/UAS-Axn^RNAi; UAS-lola^RNAi/+; (I-I’) and (J-J”) dCg-Gal4/+; UAS-lolal^RNAi/+; and (K-K”) and (L-L’') dCg-Gal4/ UAS-Axn^RNAi; UAS-lolal^RNAi/+. The thoracic and abdominal regions are labeled on the left side of the images. Scale bar in panel A: 10μm (applies to all images).

Fig. S7: Genome browser tracks displaying dTCF/Pan binding peaks at the svp (A), lola (B), and lolal (C) loci. Peaks were identified by CUT&RUN analysis performed on wing imaginal discs (upper panels, dark blue indicating downregulated expression and red indicating upregulated expression upon Wnt/Wg activation) and purified adipocyte nuclei from larval fat bodies (lower panels, orange). Tracks are visualized in the IGV_2.14.1 browser. The y-axis is autoscaled, and four gene isoforms per locus are displayed in magenta. Shared prominent peaks between datasets are indicated by arrows. (*) indicates transcription start sites.

Fig. S8: Genome browser tracks showing dTCF/Pan binding peaks at the Abd-B (A) and abd-A (B) loci. Peaks were identified by CUT&RUN analysis performed on wing imaginal discs (upper panels, red indicating upregulated expression upon Wnt/Wg activation) and purified adipocyte nuclei from larval fat bodies (lower panels, orange). Tracks are visualized in the IGV_2.14.1 browser. The y-axis is autoscaled, and four gene isoforms per locus are displayed in magenta. Shared prominent peaks between datasets are indicated by arrows. (*) indicates transcription start sites. Scale bars: 2.5 kb.

Fig. S9: Elav-Gal4-driven depletion of lola or lolal markedly expands the Abd-B expression domain in the larval ventral nerve cord (VNC). HCR RNA-FISH was used to detect abd-A (green), Abd-B (orange), and Ubx (magenta) mRNA transcripts in the larval VNC. Scale bar in panel D: 10μm (applies to all images). Genotypes: (A-A””) UAS-RFP/+; AbdB-Gal4^GMR34G07/+; (B-B””) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lolal^RNAi; (C-C””) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lola^RNAi [VDRC #101925]; and (D-D””) UAS-RFP/+; AbdB-Gal4^GMR34G07/UAS-lola^RNAi [VDRC #12573]. (E) Quantification of the normalized gap between Abd-B and Ubx expression domains in the VNC (n = 3 independent biological replicates, with 5 independent measurements obtained from each VNC). P values were calculated using one-tailed unpaired t-tests, based on predefined directional predictions in the experimental design. Error bars represent standard deviations. ***: P<0.001.

Fig. S10: Genome browser tracks showing Abd-B binding peaks at the svp (A), prg (B), and cg (C) loci. Peaks were identified by CUT&RUN analysis performed on larval central nervous system (CNS), with enrichment primarily in the ventral nerve cord (VNC). Tracks are visualized in the IGV_2.14.1 browser. The y-axis is autoscaled. (*) indicates transcription start sites.

Supplemental material available at GENETICS online.

Data availability

Drosphila strains generated in this study are available upon request.