Autocrine and paracrine Wingless signalling in the Drosophila midgut by both continuous gradient and asynchronous bursts of wingless expression

Wg establishes a gradient in the adult Drosophila posterior midgut

To observe the pattern of wg expression in the adult Drosophila posterior midgut (regions R4 and R5, (Buchon et al., 2013)), we used the Wg antibody and several wg transcriptional reporters. An enhancer trap insertion in the wg locus (wg-Gal4 > UAS-GFP) (Pfeiffer et al., 2000) coincided with the anti-Wg antibody, and revealed that wg is highly expressed at the pylorus (Figure 1A–B), in agreement with previous reports (Singh et al., 2011; Takashima et al., 2008). We did not observe Wg expression in the midgut cells adjacent to the pylorus, as recently reported (Tian et al., 2016). Our observation was further confirmed with a wg-lacZ enhancer trap insertion (wg⁰²⁶⁵⁷), which showed a strong pyloric expression (Figure S1). To evaluate whether this Wg protein could be emanating from the pylorus, or instead came from areas of wg expression in the midgut that were not recapitulated by the enhancer trap insertions, we used wg-Gal4 to drive the expression of a GFP-tagged Wg (wg-Gal4 > UAS-GFP:wg) (Packard et al., 2002). We observed GFP:Wg accumulation at the pylorus, and a decreasing gradient from the pylorus towards the anterior end of the midgut (Figure 1C–D and Figure S2). Notably, GFP:Wg diffused from the pylorus much further into the posterior midgut than into the hindgut epithelium. Interestingly, Wg travels in the midgut tissue to a distance longer than the width of the third larval instar imaginal wing primordium, where a Wg gradient is also established (Neumann & Cohen, 1997).

Figure 1. wg expression shows graded pattern along the posterior midgut.

(A–A’) Anti-Wg (red, A; grey, A’) (B–B’) wg-Gal4, UAS-GFP (green, B; grey, B’) show high levels at the pylorus (arrowheads). (C–C’) wg-Gal4, UAS-GFP:wg (green, C; grey, C’) shows signalling gradient with high levels at the pylorus (arrowheads). (D) Intensity values of wg-Gal4, UAS-GFP:wg along thirteen parallel lines (as in the arrow in C’), averaged and smoothened with a Gaussian filter (5 µm wide) (black line). The corresponding raw data is represented in Figure S2. The limits of the grey area mark one standard deviation from the average value. (E–E’) Flies expressing only NRT-Wg showed high Wg signals (anti-Wg) (red, E; grey, E’) at the pylorus (arrowheads). EEs are marked by nuclear anti-Prospero staining (red, E; grey, E’). Scale bars: 25 μm.

We also looked at flies expressing a membrane-tethered form of Wg (NRT-Wg) from the endogenous wg locus (Alexandre et al., 2015). Anti-Wg again showed high Wg signals at the pylorus, with no obvious signalling gradient (Figure 1E,E'). These results suggest that the pylorus is the main source of Wg signal for the adult Drosophila posterior midgut, and that the Wg ligand can diffuse from this region, forming a gradient.

Wg is expressed in asynchronous bursts within the adult Drosophila midgut epithelium

Our results so far indicate that most of the Wg protein in the posterior midgut comes from the pylorus. However, we still detected faint Wg antibody staining in midgut areas anterior to the lowest point of the Wg gradient, and we also observed low activity of wg transcriptional reporters (Figure 1). To ascertain whether there was wg expression in the homeostatic midgut epithelium, as reported during regeneration (Cordero et al., 2012), we made use of the wg{KO; Gal4} line. In this line, the wg locus was edited to express Gal4 instead of Wg (Alexandre et al., 2015), and therefore is expected to accurately reproduce the expression of wild-type wg.

To amplify the signal of expression, wg{KO; Gal4} was crossed to UAS-Flp, act<stop<lacZ; tub-Gal80^ts (wg{KO}^ts>Flp, act<<lacZ). After switching to the restrictive temperature, Gal4 activates and over time all Wg-producing cells and their offspring will be strongly labelled with LacZ, even at minimal levels of wg expression. When the wg{KO}^ts>Flp, act<<lacZ flies were cultured at 18°C until 13–20 days of adulthood, some background activation was detected, as scattered, individual LacZ+ cells (Figure 2A,A’) plus a field of cells in the posterior R5 region, abutting the pylorus (Figure S3A). However, after 2 days of incubation at 29°C (6–10 days old flies at dissection), the pylorus was marked with LacZ expression (Figure S3B–C), as expected, and the midgut epithelium showed more, sparse LacZ+ cells, either isolated or in pairs (Figure 2B,B’). When cultured for 8 days at 29°C (11–18 days old flies at dissection), more cells expressed lacZ in a salt and pepper pattern, with bigger groups that included polyploid ECs as well as diploid cells (Figure 2C,C’). These observations are likely the result of additional clonal induction accompanied by clonal expansion of previously labelled ISC, rather than of wg being expressed coordinately by patches of cells of multiple differentiated cell types. For the flies cultured for 8 days at 29°C, we also inspected the muscle layer of the entire posterior midgut, which showed no wg expression from the muscle cells (Figure 3). Taken together, our results indicate that under homeostatic conditions, wg is expressed intermittently and asynchronously in the Drosophila midgut epithelium, possibly in diploid cells including the ISCs.

Figure 2. Wg expression is demonstrated by lineage tracing in the midgut epithelium.

(A,A’) wg[KO]-Gal4, UAS-Flp, tub-Gal80^ts, Act<stop<lacZ guts at 18°C (13–20 day old flies). (B–C’) wg[KO]-Gal4, UAS-Flp, tub-Gal80^ts, Act<stop<lacZ guts after 2 days (6–10 day-old flies) or 8 days (11–18 day-old flies) of incubation at 29°C for Gal4 activation. (B,B’)Small LacZ⁺ (anti-βGal) (grey, B; green, B’) clusters of 1~2 cells, mostly diploid, after 2 days at 29°C. (C,C’) Larger LacZ⁺ (anti-βGal) (grey, C; green, C’) clusters of both diploid and polyploid cells after 8 days at 29°C. DAPI nuclear staining is shown in magenta. Inset boxes show selected regions at higher magnification. Scale bars: 50 μm. Fields of view correspond to the anterior R5 region.

Figure 3. Wg is not expressed in the muscle layer of the posterior midgut.

The muscle layer of the wg[KO]-Gal4, UAS-Flp, tub-Gal80^ts, Act<stop<lacZ gut after 8 days of incubation at 29°C for signal induction. No LacZ signals (anti-βGal, green) are detected in the muscle cells. DAPI nuclear staining is shown in magenta. Arrowhead points to the pylorus. Scale bar: 50 μm.

Expression of other components of the Wg signalling pathway

In order to further characterise the spatial organisation of the Wg signalling pathway in the Drosophila midgut, we observed the expression patterns of the receptor frizzled2 and the component of the destruction complex shaggy (sgg). Using a frizzled-2 enhancer trap insertion (fz2-GFP), we found that fz2-GFP was highly expressed at the pylorus (Figure 4A,A’). fz2-GFP could also be observed in diploid cells (presumably ISCs and EBs) and EEs away from the pylorus (Figure 4A’). There was no fz2-GFP expression in ECs.

Figure 4. Characterisation of fz2 and sgg expression patterns in the adult Drosophila posterior midgut.

(A,A’) fz2-GFP (green) is expressed in the small cells and the pylorus (arrowheads). (B,B’) sgg-GFP (green) is expressed in all cell types, with strong signals at the pylorus (arrowheads). Prospero (red) marks the EEs. DAPI nuclear staining is shown in grey. Scale bars: 25 μm.

To examine the expression pattern of sgg, we used a Sgg:GFP protein trap. Sgg:GFP could be detected throughout the posterior midgut, in all cell types, with strong expression at the pylorus (Figure 4B,B’).

The adult Drosophila midgut displays asynchronous stabilisation of Armadillo

Next we wanted to examine the activation of Wg signalling in the adult Drosophila midgut. We used a frizzled-3 (fz3) reporter to monitor Wg signalling activity, fz3-RFP, since fz3 is a direct target of the Wg pathway (Olson et al., 2011). We inspected flies expressing fz3-RFP and wg-Gal4 > UAS-GFP, which both showed high expression at the pylorus (Figure 5A,C–C’), with fz3-RFP exhibiting a signalling gradient, culminating at the pylorus (Figure 5C). fz3-RFP also displayed strong epithelial expression in the region abutting the gastric zone (Figure 5A,C’). To identify the cell types expressing fz3-RFP, we used esg-lacZ to label ISCs/EBs, and GBE-Su(H)-GFP to distinguish EBs. We observed that in the regions of the gut away from the pylorus, fz3-RFP is only expressed in ISCs and EBs (Figure 6). However, the stability of RFP in this tissue is unknown, presumably long, and therefore fz3-RFP expression could be indicative of long-past Wg pathway activation.

Figure 5. fz3-RFP and wg-Gal4>UAS-GFP together show the patterns of Wg activity and expression in the adult Drosophila posterior midgut.

fz3-RFP (red, A; grey, C) and wg-Gal4>UAS-GFP (green, A; grey, C’) are both highly expressed at the pylorus (arrowheads). fz3-RFP also displays epithelial expression in the region abutting the gastric zone (red, A; grey, B). (B,B’) and (C,C’) each show the region of the midgut within the dashed box above. Scale bars: 100 μm.

Figure 6. Fz3-RFP shows that ISCs/EBs away from the pylorus are responsive to Wg signalling.

(A) Su(H)GBE-GFP (green) labels EBs. (A’) Fz3-RFP (red) marks the cells responding to Wg signalling. (A’’) Esg-lacZ (anti-βGal, blue) labels ISCs and EBs. (A’’’) Merged. Fz3-RFP (red) is detected only in the ISCs and EBs, and all the ISCs and EBs showed fz3-RFP expression. DAPI nuclear staining is shown in grey. Inset boxes show selected regions at higher magnification. Scale bar: 50 μm.

A more instantaneous reporter for Wnt signalling is the cytosolic levels of the overexpressed nuclear effector of the pathway, Armadillo/β-catenin (Alexandre et al., 2015; Hayward et al., 2005; Lin et al., 2008). When full-length armadillo (arm^FL) is overexpressed in the epithelium of the imaginal wing disc, Armadillo protein is only accumulated close to the domains of wg expression (Hayward et al., 2005). Arm is so rapidly degraded that its overexpression is only detected in cells where Wg signalling stabilises the protein. Therefore, the accumulation of UAS-Arm^FL correlates with presently active Wg signalling. We tested whether ISCs/EBs showed stabilisation of UAS-arm^FL using esg-Gal4, UAS-GFP. Surprisingly, only a proportion of the GFP-expressing ISCs and EBs showed elevated levels of Arm (Arm⁺GFP⁺/GFP⁺ per field of view: average% = 29%, highest% = 52%, lowest% = 19%) (Figure 7 and Table S1). This suggests that Wg signalling is activated in selected ISCs and EBs only in short periods of time, without any obvious spatial pattern. This is in good agreement with our observations of wg expression, especially as reported with the “tracer amplifier”, wg{KO}^ts>Flp, act<<lacZ. The results indicate that Wg production from the ISCs and EBs occurs randomly, eliciting a paracrine/autocrine response that is similarly unpatterned.

Figure 7. Only a fraction of ISCs/EBs shows stabilisation of Arm^FL.

(A–A’’) esg-Gal4[TS], UAS-arm, UAS-GFP intestines at 18°C show no GFP nor Arm^FL induction. (B–B’’) GFP (anti-GFP, green) is detected in esg⁺ cells without Arm^FL accumulation (anti-Arm) (grey, B; red, B’) (arrows). DAPI nuclear staining is shown in grey. Inset boxes show selected regions at higher magnification. Scale bars: 50 μm.

We also overexpressed arm^FL in all cells with the tub<stop<GAL4 driver, which was induced at adulthood by hs-Flp. Only a fraction of the cells, which included both differentiated and undifferentiated cells (ISCs/EBs, marked by anti-HRP), showed high levels of both cytoplasmic and nuclear Arm (Figure 8B–B’’). Arm antibody detected irregular Arm distribution at the cell membranes (Figure 8B–B’’). By contrast, uninduced tissue showed regular wild-type Arm staining at the cell membranes, with higher levels in ISCs and EBs (Figure 8A–A’’). Global arm overexpression suggests that cells in the midgut tissue respond to Wg signalling asynchronously. Also, this forced global Arm induction appeared to perturb regular Arm distribution in the midgut, and possibly disrupting cell packaging, leading to tissue dysplasia. Taken together, the adult Drosophila midgut tissue appears to respond to Wg signalling and stabilise Arm in an unpatterned way.

Figure 8. Midgut tissue responds to Wg signalling in a non-uniform way.

(A–A’’) heat-shock FLP, tub<STOP<Gal4, UAS-arm without heat-shock induction of UAS-Arm^FL overexpression. (B–B’’) With the tub<stop<GAL4 driver, UAS-Arm^FL is induced in all cells, but only a proportion of the cells shows elevated Arm (anti-Arm, red) signals in the cytoplasm and nucleus. HRP (anti-HRP, green) labels ISCs and EBs. DAPI nuclear staining is shown in grey. Inset boxes show selected regions at higher magnification. Scale bars: 50 μm.

Fly strains maintenance

Flies were raised and maintained at 18, 25 or 29°C with 65–75% humidity and a 12 hour light/12 hour dark cycle on standard cornmeal/yeast medium (consisting of 1.25% agar, 10.5% dextrose, 10.5% maize, 2.1% killed yeast, and 3.5% nipagin. Supplier: Brian Drewitt, Cambridge, UK) seeded with live yeast. Stocks were obtained from the Bloomington Drosophila Resource Center unless otherwise stated. The NRT-Wg and UAS-GFP:Wg lines were provided by J. Vincent. The fz3-RFP line is from R. DasGupta. Experiments were conducted in well-fed, mated females, 3–20 days old of the following genotypes:

Figure 1:

Oregon R.

+; wg-Gal4/ CyO; UAS-GFP/ TM3, Ser

+/ w; wg-Gal4/ UAS-GFP:Wg; +/ MKRS or TM6B

+; NRT-wg; +

Figure 2,3:

w; wg{KO; Gal4}/ UAS-Flp; tub-Gal80^ts/ Act<stop<lacZ

Figure 4:

w; +; fz2^CB02997

w, sgg^CPTI000023; +; +

Figure 5:

w; fz3-RFP/ wg-Gal4; UAS-GFP/ TM6B, Sb, Tb

Figure 6:

y, w; esg-lacZ/ fz3-RFP; Su(H)GBE-GFP/ TM6B

Figure 7:

y, w; esg^NP7397/ +; UAS-arm/ tub-Gal80^ts, UAS-GFP

Figure 8:

y, w, hs-FLP_1.22; tub<GFP, stop<Gal4 / +; UAS-arm / +

Figure S1:

+; wg⁰²⁶⁵⁷ cn¹/ CyO; ry⁵⁰⁶

Histology and tissue analysis

The following fixation methods were used: (1) Adult intestines were dissected and collected in BBS for up to 30 minutes, then fixed for 2 hours at room temperature in 4% PFA diluted in BBS. This method was used for most of the experiments. (2) Adult intestines were dissected in ice-cold “wash solution” (ddH₂O + 0.7% NaCl + 0.05% Triton) for up to 15 minutes and collected within a mesh basket. The basket was submerged in a “double beaker” with wash solution at 90°C for 5 seconds, and then immediately placed in ice-cold wash solution for 2 minutes. The double beaker was prepared placing a 250 ml beaker inside a 600 ml beaker, both containing wash solution and set on a hotplate until the temperature in the 250 ml beaker reached 90°C (about 30–45 minutes). This method was used for stainings of Armadillo.

After fixation, the tissue was rinsed three times in PBT (PBS containing 0.1% Triton X-100), then washed three times with blocking buffer (PBT containing 2% BSA and 2% FCS), each time 15 minutes on the rotator at room temperature. Primary antibody incubations were overnight at 4°C. After washing with PBT (3 × 15 minutes), secondary antibodies were incubated for 2–4 hours rotating in the dark at room temperature. DAPI (1 μg/ml) was added after the final wash.

Primary antibodies were mouse monoclonal anti-Wg (1:100, gift from J. Vincent) (Alexandre et al., 2015), goat polyclonal anti-HRP (1:500, Jackson, code number 123-001-021) (Hönigsmann et al., 1975), rabbit polyclonal anti-βGal (1:10,000, Cappel) (de Navascués et al., 2012), mouse monoclonal anti-Pros (1:200, Developmental Studies Hybridoma Bank) (de Navascués et al., 2012), mouse monoclonal anti-N27 Arm (1:20, made in the Martinez-Arias lab), chicken polyclonal anti-GFP (1:200, Abcam, ab13970) (de Navascués et al., 2012). Alexa fluorophor-conjugated secondary antibodies (1:500) were from Invitrogen: anti-mouse 568 (Catalog #A-11004, A10037), anti-rabbit 488 (Catalog #A-11034, R37118), anti-rabbit 633 (Catalog #A-21071), anti-goat A488 (Catalog #A-11055), anti-goat A633 (Catalog #A-21082), anti-chicken A488 (Catalog #A-11039). DNA dye was DAPI (Invitrogen).

Tissues were mounted in Vectashield and imaged on Zeiss LSM 700 confocal system using 40× objective and numerical aperture of 1.2.

Transgene activation by temperature-sensitive TARGET system and heat-shock flip out system

To induce gene expression, the temporal and regional gene expression targeting (TARGET) method was used with the GAL4, UAS and GAL80^ts elements (McGuire et al., 2004). The flies were crossed at the restrictive temperature (18°C), and then the progeny of the desired genotype was allowed to age at 18°C for 3 to 20 days post-eclosion to reach homeostatic condition. The flies were then incubated at 29°C to allow GAL4 activity, inducing the transcription of the UAS transgenes.

The heat-shock flip out system was also used to induce transgene expression (Gordon & Scott, 2009). Adult flies were raised at 25°C until they were 3–20 days old, then they were treated with heat-shock for 30 minutes in a 37°C water-bath. The hs-FLP recombinase was activated upon heat treatment, eliminating the GFP gene and the stop cassette. This activates tub-GAL4, leading to the stimulation of UAS-regulated genes.

Data analysis

Images and figures were assembled using ImageJ (1.47v). Images are maximum intensity projections or selected, representative layers of confocal stacks. The intensity values of wg-Gal4, UAS-GFP:wg were obtained by ImageJ (1.47v) (Figure 1C’), then plotted using RStudio (0.99.491) and Adobe Illustrator CS6 (Figure 1D and Figure S2). Cell counts were conducted using the PointPicker plugin in ImageJ (1.47v) (Table S1). The cellular percentages were calculated using Microsoft Excel 2011 (Table S1).