Phosphatase of Regenerating Liver-1 Regulates Wing Vein Formation through TGF-β Pathway in Drosophila melanogaster

¹ Division of Human Reproduction and Developmental Genetics, The Women's Hospital, Zhejiang University School of Medicine, 310006 Hangzhou, Zhejiang, China

² Institute of Genetics, Zhejiang University, and Department of Genetics, Zhejiang University School of Medicine, 310058 Hangzhou, Zhejiang, China

³ Joint Institute of Genetics and Genomic Medicine between Zhejiang University and University of Toronto, Zhejiang University, 310058 Hangzhou, Zhejiang, China

^*Correspondence: xyyongm@zju.edu.cn (Yongmei Xi); xhyang@zju.edu.cn (Xiaohang Yang)
Academic Editor: Ranieri Bizzarr

Front. Biosci. (Landmark Ed) 2022, 27(6), 176; https://doi.org/10.31083/j.fbl2706176

Submitted: 25 March 2022 | Revised: 15 May 2022 | Accepted: 27 May 2022 | Published: 1 June 2022

This is an open access article under the CC BY 4.0 license.

Download PDF

Brower Figures

Cite

Abstract

Background: Drosophila Phosphatase of Regenerating Liver-1 (PRL-1) is the only homolog of the mammalian PRLs with which it shares high sequence and structural similarities. Whilst PRLs are most notable for their high expression in malignant cancers and related promotion of cancer progression, the specific biological functions of the PRLs remain largely elusive. Methods: Here, using a gain-of-function approach, we found that PRL-1 functions during wing vein development in Drosophila melanogaster (Drosophila). Overexpression of Drosophila PRL-1 caused dose-dependent wing vein proliferation. Results: Genetic screening of the main TGF- $\beta{}$ signaling factors, Mad and Smox, showed that the RNAi-mediated knockdown of Mad could alleviate the extra vein phenotype caused by overexpressed PRL-1 and lead to loss of the posterior section of longitudinal veins. However, knockdown of Smox resulted in an identical phenotype with or without the overexpression of Drosophila PRL-1. Clonal analyses revealed that overexpression of PRL-1 led to decreased expressions of activated phospho-Mad protein, as measured by immunostaining. Real-time PCR showed that the transcriptional levels of Smox were significantly increased upon overexpression of the Drosophila PRL-1 in wing discs, with a dose dependent effect. Conclusions: We propose that the main function of Drosophila PRL-1 in wing development is to affect the phospho-Mad levels and Smox transcriptional levels, therefore influencing the competitive balance for Medea between Mad and Smox. Our study demonstrates the novel role for Drosophila PRL-1 in regulating TGF- $\beta{}$ signaling to influence wing vein formation which may also provide insight into the understanding of the relationship between PRLs and TGF- $\beta{}$ signaling in mammals.

Keywords

PRL-1

Smox

TGF-β

wing vein

Drosophila

1. Introduction

Drosophila PRL-1 is the only homolog for the three human PRL proteins, PRL1, PRL2, and PRL3. The PRLs belong to a family of dual specificity phosphatases (DSP) classified as a subgroup of protein tyrosine phosphatases (PTP). They were originally discovered as immediate-early gene products in the regenerating liver [1, 2, 3]. Mammalian PRLs commonly possess a tyrosine phosphatase domain and a CAAX prenylation motif (where C is the cysteine, A is an aliphatic amino acid and X is any amino acid) which is located next to a polybasic region in the C-terminal for membrane anchoring [3]. Many studies on clinical samples have demonstrated the connection between PRL overexpression and the malignancy of various types of cancers such as colorectal, breast, and gastric carcinomas [4], and the implantation of cultured cancer cells, PRL1 or PRL3-overexpressed cells into nude mice, resulted in metastatic tumors [5].

After the recognition of its intrinsic involvement in metastatic cancers and in actively driving malignant progression, various animal models such as Drosophila [6], Mus musculus [7, 8], and Rattus norvegicus [9] were then used to gain insight into the biological function and the underling molecular mechanisms of the PRLs. Drosophila PRL-1 was identified to be a growth inhibitor that could counteract the oncogenic activity of Src. Drosophila PRL-1 localizes to the lateral membrane of the wing discs and has also been reported as ubiquitously expressed and localized to both the cytoplasm and the plasma membrane in various tissues including embryo epidermis, larval midgut, gastric caecum, developing eye and in the wing discs themselves. The ability of Drosophila PRL-1 to inhibit growth was confirmed to be dependent on the CAAX motif that is required to localize Drosophila PRL-1 to the apical edge of the lateral membrane in the wing discs [6]. Several other recent studies have showed that Drosophila PRL-1 functions in the central nervous system. The absence of Drosophila PRL-1, for example, specifically reduces synapse organization in the terminal arbor of one target area of mechanosensory neurons. It has also been reported that untranslated Drosophila PRL-1 mRNA (the long UTRs sequences of Drosophila PRL-1 mRNA) mediate the local translation of Drosophila PRL-1 in local synaptic arbor formation and that Drosophila PRL-1 provides a specificity factor to restrict InR-Akt signaling and synapse formation in a subcellular compartment of neurons [10]. Notably, knockdown of Drosophila PRL-1 in PDF clock neurons dramatically lengthens the circadian period. Drosophila PRL-1 could also mediate the dephosphorylation of TIM to set period length and behavioral phase in darkness, enabling behavioral adjustment to changes in day-night cycles [11]. Our previous study found that Drosophila PRL-1 is involved in protecting the nervous system against olfactory CO ${}_{2}$ stimulation. Mutation of Drosophila PRL-1 flies exhibit a permanently held-up wing phenotype upon CO ${}_{2}$ exposure due to a neural dysfunction. We then showed that Drosophila PRL-1 could interact with its downstream gene Unextended wing (Uex) and control Uex expression to maintain neuroprotective functions [12].

Despite there being a rapidly growing amount of literature on PRLs, our knowledge of the specific biological function of the PRL family of proteins remains limited. Our study used the Drosophila wing to examine the overexpression of PRL-1 in genetically controlled models. This system reveals that Drosophila PRL-1 functions as a newly discovered regulator of the transforming growth factor- $\beta{}$ (TGF- $\beta{}$ ) signaling pathway to control wing vein pattern.

The adult wing of Drosophila has a stereotypical pattern of five longitudinal veins (LVs, L1–L5) and two crossveins (ACV, PCV) to act as the rigid supports necessary for flight [13]. Wing veins consist of rows of cells that differentiate into heavily pigmented cuticles and smaller apical sized cells. These are formed through specification of cell fate, not cell migration or cell shape change [14, 15]. During metamorphosis several signaling pathways, including those mediated by EGF, BMP/Dpp, Hedgehog, Notch, and Wnt, define the expression of many transcription factors involved in the partition of the wing disc epithelium into proveins and interveins [14]. For example, in the third instar larvae imaginal wing disc, EGFR signaling is activated within the proveins to direct the subdivision of each provein in a central region that will then differentiate as a vein [16, 17]; Notch signaling is involved in establishing the correct width of various wing veins [18]; and BMP signaling is required for the specification of both longitudinal proveins and intervein regions in the imaginal disc at the larval stage, also later required for crossvein formation during pupariation [14, 17].

BMP (Bone Morphogenetic Protein) is another member of the TGF- $\beta{}$ superfamily. BMP signaling in Drosophila includes the ligands Decapentaplegic (Dpp) and Glass bottom boat (Gbb), the type I receptor Thickveins (Tkv), the type II receptor Punt, and the transcription factor mothers against Dpp (Mad) [19, 20]. Dpp is secreted either as a homodimer or a heterodimer with Gbb. These ligands bind to Tkv and Punt receptors to phosphorylate the C-terminal of Mad. Phosphorylated Mad (pMad) is then, together with co-Smad Medea, translocated into the nucleus for the transcriptional regulation of various downstream target genes such as spalt-major (salm), spalt-related (salr), optomotor-blind (omb), spalt (sal), vestigial (vg), brinker (brk), and Ultrabithorax (Ubx) [19, 21, 22, 23]. Other signals mediated by TGF- $\beta{}$ ligands include those of the TGF- $\beta{}$ /Activin pathways. In Drosophila, Activin signaling is mediated by the secreted ligands Activin (dAct) and Dawdle (Daw), the type I receptor Baboon (Babo), the type II receptor Punt, and Smox (Smad on X, Drosophila Smad2). Activin signals are transduced in a similar fashion as in BMP signaling via their binding to type I and II receptors, the C-terminal phosphorylation of Smox, followed by heterotrimeric complex formation with Medea and then transcriptional activation of target genes in the nucleus [24]. It should be noted that the dAct/Smox signal pathway shares several components with the Dpp/Mad branch including Punt and Medea [25]. Both dAct/Smox and Dpp/Mad pathways play important roles in fly wing vein development [22, 23, 26, 27]. It has been reported that Medea functions as the limiting factor between Mad and Smox signaling pathways. The phosphorylated transcription factors Mad and Smox compete for association with rate-limiting quantities of Medea to control vein formation [26].

In this study, we used a gain-of-function approach to investigate Drosophila PRL-1 functions during wing development. Overexpression of Drosophila PRL-1 caused the formation of extra vein tissue at the posterior end of the PCV along the longitudinal vein 5 (L5), with a dose dependent effect. Genetic screening found that RNAi-mediated knockdown of Mad could alleviate the extra vein phenotype caused by overexpressed Drosophila PRL-1 and lead to loss of the posterior of longitudinal vein 4. Knockdown of Smox on the background of overexpressing Drosophila PRL-1 in the wing disc showed a phenotype similar to the knockdown of Smox alone. Clonal analyses revealed that overexpression of Drosophila PRL-1 resulted in lower expression of the activated phospho-Mad protein as measured by immunostaining. Real-time PCR showed that the transcriptional levels of Smox were significantly increased upon the overexpression of Drosophila PRL-1 in the wing disc, with a dose dependent effect. We propose that the main function of PRL-1 in Drosophila wing development is to affect the phospho-Mad levels and Smox transcriptional levels, therefore potentially influencing the competitive balance for Medea between Mad and Smox. Our study demonstrates a novel role for PRL-1 in regulating TGF- $\beta{}$ signaling during wing vein formation. This may provide additional valuable insight into the understanding of the relationship between PRLs and TGF- $\beta{}$ signaling in mammals.

2. Materials and Methods

2.1 Comparison of PRL Sequences

The Drosophila PRL-1 amino acid sequence was downloaded from Flybase (http://flybase.org/) and human PRL1-3 amino acid sequences were downloaded from the National Center for Biotechnology Information (NCBI, https://www.ncbi.nlm.nih.gov/). The Clustal X 2.0 software (European Bioinformatics Institute, Hinxton, Cambridgeshire, CB10 1SD, UK) was used for the multiple amino acid sequence alignments of Drosophila PRL-1 and human PRL1-3. Results were presented using the DNAman software (version 9.0, Lynnon Corporation, 116 RUE DU MILICIEN, VAUDREUIL-DORION QC J7V 9M5, Canada).

2.2 Fly Strains and Genetics

Flies were reared on a standard cornmeal medium at 25 °C. The following strains were used: w ${}^{1118}$ , Engrailed-GAL4 (Chao Tong, Zhejiang University, Hangzhou, China), UAS-PRL-1 (generated by our lab, Guo et al. [12]), hsFlp; Act5C $>$ y ${}^{+}$ $>$ Gal4 UAS-green fluorescent protein (GFP)/CyO (from J.C. Pastor-Pareja, Tsinghua University, Beijing, China); PRL-1 ${}^{16}$ (as in Guo et al. [12]); and UAS-Mad RNAi (THU4727, THU5761, THU5760, THU4801), UAS-Smox RNAi (THU2384), UAS-Med RNAi (THU2306, THU4652) (all from the Tsinghua Fly Center, Beijing, China).

2.3 Clonal Analysis

We used the flip-out system to generate overexpression clones in the brain. To induce ectopic expression of Drosophila PRL-1, we crossed hsFlp;Act5C $>$ y ${}^{+}$ $>$ Gal4 UAS- GFP/CyO flies with UAS-PRL-1 flies or wild type (as the control) and the larvae were heat shocked for 1 hr at 24–48 hr after egg laying (AEL). The larvae were then incubated at 25 °C until 120 hr AEL at which time they were dissected. Overexpression clones with the genotype hsFLP/+;Act5C $>$ y ${}^{+}$ $>$ Gal4 UAS-GFP/UAS-PRL-1 were marked by the presence of GFP.

2.4 Adult Wing Analysis

Adult female or male wings (N = 20 for each genotype, repeated three times) were removed and mounted in 80% glycerol mounting medium. Images of wings were obtained on an Eclipse 80i microscope (Nikon, Tokyo, Japan). Wing posterior and total areas were measured using ImageJ (version 1.51, LOCI, University of Wisconsin, Madison, Wisconsin, USA). Statistical analysis was performed using a student’s t-test.

2.5 Staining Procedures and Microscopy

Drosophila brains from third instar larvae were dissected in ice cold 1 $\times{}$ PBS (10 mM NaH ${}_{2}$ PO ${}_{4}$ /Na ${}_{2}$ HPO ${}_{4}$ , 175 mM NaCl, pH 7.4) and fixed for 20 min in PBS with 4% paraformaldehyde. Fixed brains were washed three times with 0.1% Triton X-100 in PBS (PBT) and blocked in PBT containing 3% BSA for 0.5 h at room temperature. Brains were then incubated with the primary antibodies overnight at 4 °C, followed by three washes with PBT before incubating with secondary antibodies for 2 h. DAPI was added for the last 20 min. After three further washes with PBT, brains were mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA). The primary antibody used was rabbit anti-phospho-Mad [1:800, Phospho-Smad1/5 (Ser463/465) (41D10) Rabbit mAb, GST, #9516]. The secondary antibody was Alexa Fluor 555 conjugated (1:250 or 1:500; Thermo Fisher Scientific, Waltham, MA, USA). DAPI (1 g/mL; Sigma-Aldrich, St. Louis, MO, USA) was used to stain for nuclei. GFP was directly observed under a confocal microscope. Images were taken using an Olympus FV1000 confocal microscope (Olympus, Tokyo, Japan) and processed using Adobe Photoshop CS6 (Adobe, San Jose, CA, USA). ImageJ was used to generate a sum-intensity z stack projection and a measurement of the total fluorescence intensity.

2.6 Quantitative Real-Time PCR

Total RNA was isolated from dissected disc cells using Trizol® Reagent (Invitrogen, Carlsbad, CA, USA, #15596018). A 1 $\mu{}$ g RNA and First-Strand cDNA synthesis kit (Invitrogen) was used to produce single-stranded cDNA. Real-time PCR was performed using a Power SYBR Green PCR Master Mix (ABI) and an ABI 7900HT Fast real-time PCR system. All experiments were performed in triplicate.

Rp49 was used as an endogenous control. The following PCR primers (5 ${{}^{\prime}}$ -3 ${{}^{\prime}}$ ) were used; rp49 Forward, GCTAAGCTGTCGCACAAA, rp49 Reverse, TCCGGTGGGCAGCATGTG; Snoo Forward, ACCAGATCATCCTCGATCAG, Snoo Reverse, TTGTGGCTGCTTCATTGGC; Smox Forward, ACCGTTACAGATCAGAAGTG, Smox Reverse, GGTGAATGGCAACATCTC; myo Forward, ACGATGAGGAGTACGAGAG, myo Reverse, TGATGAGCACTTATCCAATCC; EcR Forward,ACAGGACGTAGATGCACAG, and EcR Reverse, CATATGCGTCGATATAGTGCG.

2.7 Statistical Analysis

Data shown in the Figs. 2B,2B’,2C,4E,5A,5B are reported as mean $\pm{}$ SD. Statistical analysis was performed using the paired, two-tailed Student’s t-test contained in the Microsoft-Excel program. Significance levels are indicated in the figure legends.

3. Results

3.1 Structural Characteristics of Drosophila PRL-1

The Drosophila PRL-1 gene (CG4993), encoding the single Drosophila PRL protein (Drosophila PRL-1), is located on the left arm of the second chromosome. It consists of 176 amino acids in approximately 20-kDa protein (http://flybase.org/). To evaluate the conservative properties of PRL phosphatases between humans and Drosophila, we aligned their amino acid sequences using public databases. Results showed that Drosophila PRL-1 shares an overall 75.7% similarity in comparison to all three human PRLs (Fig. 1A). More specifically, the amino acid similarity of Drosophila PRL-1 shows a high positive comparison with each of the human PRLs individually, with a 73% similarity between Drosophila PRL-1 and human PRL-1, 75% for human PRL-2, and 74% for human PRL-3. We then performed bioinformatics analysis on the structural characteristics of Drosophila PRL-1. Drosophila PRL-1 contains several domains also characteristic of mammals including the putative ‘CXnE’ motif [3], the conserved catalytic PTP domain that is critical to phosphatase activity [28], the polybasic region, and the CAAX prenylation motif for targeting to the plasma membrane [29] (Fig. 1A,B, red rectangles), showing that that Drosophila PRL-1 shares high sequence and structural similarities with human PRLs.

Fig. 1.

Structural characteristics of PRLs. (A) Multiple sequence alignments of PRLs from Homo sapiens and PRL-1 from Drosophila melanogaster. Identical residues are marked with a black background, conserved residues with a blue background, and similar residues with a salmon background. The putative ‘CXnE’ motif, WPD-loop, P-Loop, a polybasic region and a CAAX prenylation motif are indicated with red squares. (B) Schematic diagrams of Drosophila PRL-1. The putative ‘CXnE’ motif begins at C54.

3.2 Overexpressing Drosophila PRL-1 Increases Wing Vein Formation

We had previously generated a Drosophila PRL-1 deletion mutant which showed a wing hold-up phenotype when stimulated under a high concentration of CO ${}_{2}$ treatment. After inhibition of the CO ${}_{2}$ neuron receptor Gr63a or Gr21a, the wind hold-up phenotype could be rescued [12]. We did not notice any morphological changes in the wings of these Drosophila PRL-1 mutants. Using transgenic flies containing the full-length of Drosophila PRL-1 [12], we then performed experiments by crossing UAS-PRL-1 flies to several lines that expressed the transcriptional activator GAL4 in a tissue-specific manner. Among these, the overexpression of PRL-1 in developing larval wing discs, as driven by engrailed-Gal4 (En-Gal4) expressed in the posterior compartment of the wing, led to reduction of adult wings size (Fig. 2A). This had also been noted in the Pagarigan et al. [6] study. Here, the ratio of posterior compartment to the entire wing showed a 2.7% reduction (p = 5.79475 $\times{}$ 10 ${}^{-7}$ ) in the single copy overexpressed PRL-1 male, a 6.7% reduction (p = 4.88867 $\times{}$ 10 ${}^{-9}$ ) in the double copy overexpressed PRL-1 male, a 1.9% reduction (p = 6.75841 $\times{}$ 10 ${}^{-5}$ ) in single copy overexpressed PRL-1 female, and a 4.6% reduction (p = 8.18251 $\times{}$ 10 ${}^{-7}$ ) in double copy overexpressed PRL-1 female (Fig. 2A,B’). This indicates that Drosophila PRL-1 has dose dependent effect on growth inhibition.

Fig. 2.

Overexpression of Drosophila PRL-1 caused dose-dependent wing vein proliferation. (A) Adult wings from flies expressing the following transgenes under the control of En-Gal4; Left panes are wings from males; Right panes are wings from females. The lines divide wings into anterior (A) and posterior (P) parts. (B,B’) The ratio of posterior compartment (P) to the entire wing (A+P) (n = 20 in each genotype, and with experiments repeated three times). (B) Male, (B’) female. (C) The percentage of adult wings with ectopic vein branches; n = 20, repeated three times. (D) Wing vein pattern in the PRL-1 mutants. Data are presented as means $\pm{}$ SD. ** p $<$ 0.01, *** p $<$ 0.001, **** p $<$ 0.0001.

Interestingly, we observed that the posterior wing compartment of both male and female showed an extra branch at the posterior end of the PCV along the longitudinal vein 5 (L5) in En-Gal4 $>$ UAS-PRL-1 fly wings, as compared to an adult wild-type wings (Fig. 2A, red arrows). It was notable that the ectopic vein branch of the wings of flies with double copy PRL-1 overexpression was longer than that in those with a single copy (Fig. 2A). In addition, the percentage of flies with extra wing veins was higher upon double PRL-1 overexpression than that upon single PRL-1 overexpression (Fig. 2C), with 70% and 95.4% upon double overexpression, and 15% and 17.3% upon single overexpression, in males and females, respectively. We also analyzed the wing vein pattern in the PRL-1 mutants and failed to notice any abnormal morphological or wing vein phenotypes (Fig. 2D). These results suggest that Drosophila PRL-1 is a dosage-dependent regulator of wing vein formation.

3.3 Drosophila PRL-1 is a Novel Regulator of TGF-

\beta{}

Pathway in Wing Vein Development

To further investigate the functions and molecular mechanism of Drosophila PRL-1 in wing vein development, we obtained several RNA interference (RNAi) transgenic fly lines that were known to be involved in the pathways that control vein formation. We performed screening experiments using the flies with a background of PRL-1 overexpression driven by En-Gal4. When the antisense RNA of the target gene is expressed by the En-Gal4 driver, the RNA transcribed from the target DNA forms double-stranded RNA with the antisense RNA that is disrupted by the RNAi machinery and depleted from the target gene product at the posterior of wing. Using RNAi, we could therefore compare the phenotype resulting from a specific gene knockdown in the posterior of wing tissue and attempt to clarify the relationship between the specific gene and PRL-1.

We found that RNAi-mediated knockdown of Mad could alleviate the extra vein phenotype caused by overexpressed PRL-1 (Fig. 3A–D, yellow arrows), but this also led to a loss of the posterior portion of longitudinal vein 4 (Fig. 3D, red arrows), which was an identical phenotype seen in some Dpp partial loss-of-function mutants [26, 30]. Notably, downregulating Mad only in the posterior portion of wing tissue did not result in any obvious phenotype (Fig. 3C), while knock down of Mad together with simultaneous overexpression of PRL-1 could induce vein loss. This demonstrates that the efficiency of the knockdown of Mad alone is not sufficient to lead vein loss, and that this only occurs in combination with PRL-1. In this way the data clearly indicates a Drosophila PRL-1 genetic interaction with Mad. We then wondered whether overexpressing PRL-1 has any effect on Mad levels. We carried out a flip-out method using heat shock induced FLP (hsFLP); Act5C $>$ y ${}^{+}$ $>$ Gal4 UAS-GFP flies and then generated overexpression PRL-1 clones. We failed to find any ideal wing imaginal disc clones, but found mitotic clones in the brain. Therefore, we continued to analyze the correlation between PRL-1 and phospho-Mad in the brain. We used an anti-phospho-Mad antibody for the GFP detection of the phosphorylated levels of Mad in the PRL-1 overexpression cells and found that the activated phospho-Mad levels were down-regulated by almost 50% in overexpression PRL-1 cells as measured by immunostaining (Fig. 4). This may explain the observation of the combination of the overexpression of PRL-1 together with Mad RNAi down-regulating Dpp signaling to a certain degree, and that this is sufficient to induce the loss of wing veins.

Fig. 3.

Adult wings from flies expressing the following transgenes under the control of En-Gal4. (A) En-Gal4 $>$ WT; n = 20. (B) En-Gal4 $>$ UAS-PRL-1; n = 20. (C) En-Gal4 $>$ UAS-Mad RNAi (THU4727); n = 20. (D) En-Gal4 $>$ UAS-PRL-1; UAS-Mad RNAi; n = 14. (E) En-Gal4 $>$ UAS-Smox RNAi; n = 20. (F) En-Gal4 $>$ UAS-PRL-1; UAS-Smox RNAi; n = 16.

Fig. 4.

Overexpressing PRL-1 led to decreased expressions of activated phospho-Mad protein. (A–D) Overexpression of PRL-1 marked by GFP in the brain. Lower expression of phospho-Mad is observed in overexpressing PRL-1 clone cells as compared with the wild-type cells around the clone. (A) Phospho-Mad (red). (B) GFP (green). (C) DAPI (blue). (D) Merge. (E) Mean fluorescence of phospho-Mad; n = 6. Data are presented as means $\pm{}$ SD. *** p $<$ 0.001.

As both Dpp/Mad and dAct/Smox pathways both play important roles in fly wing vein development [26, 27] and share several components [25], we therefore performed RNAi-mediated knockdown of Smox to confirm the phenotype. We found that down-regulating the Smox gene led to ectopic vein tissue formation in the vicinity of L5 and an extra branch at the posterior end of the PCV (Fig. 3E, green arrows), which showed a similar phenotype as noted in Sander et al. [26]. When we knocked down Smox on the background of overexpressing PRL-1, the ectopic vein phenotype was almost same as that in the down-regulation of Smox alone (Fig. 3F, green arrows). This may suggest that PRL-1 functions upstream of Smox. To further confirm this hypothesis, we carried out Real-time PCR experiments to check Smox levels. Results showed that, except for Smox, the levels of genes such as Sno oncogene (Snoo), Myoglianin (Myo), and EcR, that are also involved in Activin signaling pathway, remained unchanged upon PRL-1 overexpression (Fig. 5A). Additionally, the transcriptional levels of Smox were shown to be increased in a dose-dependent manner in the PRL-1-overexpressed wing discs (Fig. 5B). The overexpression of PRL-1 triggering the up-regulation of Smox probably acts to compensate for decreased Activin signaling. These results suggest that Drosophila PRL-1 is a regulator of TGF- $\beta{}$ pathway and genetically interacts with Mad and Smox, affecting the phosphorylation of Mad, as well as the transcriptional levels of Smox, and therefore influencing the competitive balance for Medea between Mad and Smox during wing vein development.

Fig. 5.

Overexpressing PRL-1 increases the transcriptional levels of Smox. (A) Real-time PCR analysis of the relative mRNA levels of genes involved in the Activin signaling pathway showing that only Smox is obviously increased in overexpressing PRL-1 wing discs; n = 3. (B) The transcriptional levels of Smox in the discs from different copy numbers of En-gal4 and UAS-PRL-1 animals indicate that the transcriptional levels of Smox is a dose dependent effect in the overexpression of PRL-1 wing disc s; n = 3. Data are presented as means $\pm{}$ SD. * p $<$ 0.05, *** p $<$ 0.001, **** p $<$ 0.0001.

4. Discussion

Drosophila PRL-1 belongs to the Phosphatase of Regenerating Liver (PRL) family, also known as protein tyrosine phosphatase 4A (PTP4A). In mammals it consists of three members, PRL-1, 2 and 3. Numerous studies using cultured cells have demonstrated that the PRLs can induce cell proliferation, migration, and invasiveness in cancers [29, 31]. Using the Drosophila wing as a model system, here we report that the homolog of mammalian PRLs, Drosophila PRL-1, functions during wing vein development through its regulation of the TGF- $\beta{}$ signaling pathway. By gain of function analysis, overexpression of Drosophila PRL-1 in the wing was noted to result in extra wing veins. By contrast, loss of function experiments involving PRL-1 mutant lines failed to yield any wing vein deficiencies (Fig. 2D). This indicates that whilst the loss of PRL-1 may fail to negatively affect cell proliferation, that its overexpression positively promotes it in certain circumstance. This is reminiscent of the role of its mammalian counterparts PRL-1, PRL-2 and PRL-3 which reportedly have high expressions in a several human cancers [4]. Here, in particular, we reveal a new mechanism of PRL-1 and its involvement in regulating the TGF- $\beta{}$ pathway.

Using bioinformatics methods to analyze the conservative properties of PRL phosphatases between human and Drosophila we found that Drosophila PRL-1 shares a 75.7% similarity between all three human PRLs and with many specific structural similarities noted (Fig. 1). This is consistent with Zeng et al. [32] and Lin et al. [33] who found that Drosophila PRL-1 shares high similarities (74%–76%) to all three mammalian PRLs. In our previous study, we found that a holding-up of wings phenotype in the PRL-1 deletion mutant flies could be rescued by overexpressing human PRL-1 or human PRL-2, which implies that human PRL phosphatases may function in a similar neuroprotective capacity [12]. However, in the present study, overexpressing human PRL-1 or human PRL-2 by En-gal4 in the wing failed to result in any similar ectopic wing vein anomalies as seen upon Drosophila PRL-1 overexpression (data not shown). This could be because the expression level is not sufficient, or that some of signal transduction members for human PRLs are absent in Drosophila. Pagarigan et al. [6] reported that overexpressing Drosophila PRL-1 could inhibit growth in a manner dependent on the CAAX motif that is integral to the localization of Drosophila PRL-1 to the apical edge of the lateral membrane.

The Drosophila adult wing has a stereotypical pattern of longitudinal veins and crossveins which are constructed mainly of two cell types, vein cells and intervein cells. The specification of vein territories and the differentiation of wing veins and intervein regions are regulated by a high number of complex signaling pathways including those of EGFR, transforming growth factor- $\beta{}$ , Hedgehog, Notch, Wnt, and JAK-STAT [18, 26, 34, 35, 36, 37, 38, 39]. The various roles and interrelationships among these are far from completely understood. In Drosophila, the TGF- $\beta{}$ signaling pathway can be divided into two broad signaling families, those of the Dpp and those of the Activin pathways, to control various cellular processes including those of cell proliferation, differentiation, apoptosis, cell plasticity, and migration [20, 26].

In this scenario, ectopic PRL-1 lowers the expression of the activated phospho-Mad protein, in which a decrease in Dpp signaling and elevation of the transcriptional levels of Smox may act to downregulate the Activin pathway. It has been previously reported that RNAi-mediated Babo knockdown also resulted in decreased Activin signaling and led to a significant increase in the transcription of Smox [40]. These increased transcriptional levels of Smox could lead to a compensation role upon decreased Babo expression. The Activin/Smox signal pathway shares its component Medea with the Dpp/Mad pathway, and Medea has been reported to function as a limiting factor between the Mad and Smox signaling pathways [26]. Overexpressing PRL-1 could potentially cause downregulation of both the Dpp pathway and Activin pathway, therefore, influencing the competitive balance for Medea between Mad and Smox during wing vein pattern formation, as shown in the schematic of Fig. 6.

Fig. 6.

Schematic representation of the putative mechanism by which PRL-1 could regulate the TGF- $\beta{}$ pathway. Overexpression of PRL-1 affects the phosphorylation of Mad as well as the transcriptional levels of Smox. This may therefore influence the competitive balance for Medea between Mad and Smox during wing vein development and lead to extra wing veins.

Our genetic experiments show that the knockdown of Mad on the background of ectopic PRL-1 leads to the loss of the extra vein, thus reversing the phenotype. This suggests that PRL-1 and Mad may function in parallel pathways to regulate vein formation. However, the detailed molecular mechanism of how PRL-1 affects the phosphorylation of Mad awaits further investigation. PRL-1 and Smox may act in a linear pathway in the regulation of vein formation. The ectopic vein phenotype caused by knocking down Smox was nearly identical to that of the combination of Smox knockdown and overexpression of PRL-1 in the posterior of the wing (Fig. 3E,F). Together, these experiments suggest that dPRL-1 acts, at least in part, upstream of Smox.

5. Conclusions

Our study identified a novel function of PRL-1 in wing vein formation. We propose that the main function of PRL-1 in Drosophila wing development is to affect the phospho-Mad levels and Smox transcriptional levels, therefore influencing the competitive balance for Medea between Mad and Smox. This study demonstrates a novel role for PRL-1 in regulating TGF- $\beta{}$ signaling to influence wing vein formation. It could also lead to valuable clarification of the relationship between PRLs and TGF- $\beta{}$ signaling in mammals.

Author Contributions

YX, HZ and ZL designed the experiments and interpreted the data. HZ, ZL, XY (Xin Yuan) and HW contributed to experiment completion and data analysis. HZ, ZL and YX wrote the initial manuscript. YX and XY (Xiaohang Yang) revised the manuscript. All authors contributed helpful suggestions for this manuscript.

Ethics Approval and Consent to Participate

Not applicable.

Acknowledgment

We thank Chris Wood of the College of Life Sciences, Zhejiang University for checking the English of this manuscript.

Funding

This work was supported by National Key R&D Program of China (2018YFC1004900) and the National Basic Research Program of China (2013CB945600, X.Yang.), National Natural Science Foundation of China (31801230).

Conflict of Interest

The authors declare no conflict of interest.

References

[1]

Kitamura K, Mizuno Y, Sasaki A, Yasui A, Tsuiki S, Kikuchi K. Tsuiki and K. Kikuchi: Molecular-Cloning and Sequence-Analysis of Cdna for the Catalytic Subunit 1-Alpha of Rat-Kidney Type-1 Protein Phosphatase, and Detection of the Gene-Expression at High-Levels in Hepatoma-Cells and Regenerating Livers as Compared to Rat Livers. Journal of Biochemistry. 1991; 109: 307–310.

| Google Scholar PubMed | Crossref

[2]

Kitamura T, Nakamura K, Mizuno Y, Kikuchi K. Gene Expressions of Protein-Tyrosine Phosphatases in Regenerating Rat-Liver and Rat Ascites Hepatoma-Cells. Japanese Journal of Cancer Research. 1995; 86: 811–818.