Zebrafish nephrogenesis is regulated by interactions between retinoic acid, mecom, and Notch signaling

Abstract

The zebrafish pronephros provides a conserved model to study kidney development, in particular to delineate the poorly understood processes of how nephron segment pattern and cell type choice are established. Zebrafish nephrons are divided into distinct epithelial regions that include a series of proximal and distal tubule segments, which are comprised of intercalated transporting epithelial cells and multiciliated cells (MCC). Previous studies have shown that retinoic acid (RA) regionalizes the renal progenitor field into proximal and distal domains and that Notch signaling later represses MCC differentiation, but further understanding of these pathways has remained unknown. The transcription factor mecom (mds1/evi1 complex) is broadly expressed in renal progenitors, and then subsequently marks the distal tubule. Here, we show that mecom is necessary to form the distal tubule and to restrict both proximal tubule formation and MCC fate choice. We found that mecom and RA have opposing roles in patterning discrete proximal and distal segments. Further, we discovered that RA is required for MCC formation, and that one mechanism by which RA promotes MCC fate choice is to inhibit mecom. Next, we determined the epistatic relationship between mecom and Notch signaling, which limits MCC fate choice by lateral inhibition. Abrogation of Notch signaling with the y-secretase inhibitor DAPT revealed that Notch and mecom did not have additive effects in blocking MCC formation, suggesting that they function in the same pathway. Ectopic expression of the Notch signaling effector, Notch intracellular domain (NICD), rescued the expansion of MCCs in mecom morphants, indicating that mecom acts upstream to induce Notch signaling. These findings suggest a model in which mecom and RA arbitrate proximodistal segment domains, while MCC fate is modulated by a complex interplay in which RA inhibition of mecom, and mecom promotion of Notch, titrates MCC number. Taken together, our studies have revealed several essential and novel mechanisms that control pronephros development in the zebrafish.

INTRODUCTION

Vertebrate kidney organogenesis proceeds through the formation and regression of several successive structures, each comprised of excretory units known as nephrons (Dressler, 2006). The first structure is the pronephros, composed of rudimentary nephrons formed next to the nephric cord, a bilateral epithelial tubule derived from the intermediate mesoderm (IM). Whereas the pronephros is a vestigial organ in mammals, it serves as the embryonic excretory organ in lower vertebrates such as fishes and frogs (Dressler, 2006). During mammalian development, a mesonephros forms posteriorly to the pronephros and functions transiently in fetal life, then subsequently a third structure, the metanephros, is formed that functions as the definitive adult kidney (Dressler, 2006). The metanephros arises when the ureteric bud grows out of the caudal end of the nephric duct, invades the surrounding metanephric mesenchyme, and induces a mesenchyme-to-epithelial transition (MET) in cell aggregates adjacent to the ureteric bud tips (Little and McMahon, 2012). Mesenchymal cells undergoing MET form a polarized renal vesicle, which develops sequentially into a comma-shaped body, S-shaped body, and eventually into a segmented nephron tubule (Little and McMahon, 2012). Highly elaborated branching of the ureteric bud along the radial axis of the metanephric mesenchyme generates a complicated network within the collecting duct system, with thousands of nephrons situated in an intricate, arborized three-dimensional arrangement.

There is currently a limited understanding of how nephron tubules are patterned into segments, due in part to the complexity of mammalian nephrogenesis and kidney anatomy (Constantini and Kopan, 2010). However, in recent years the zebrafish has emerged as a useful vertebrate model to study nephron segmentation (Gerlach and Wingert, 2012). Zebrafish embryos form a pronephros with two nephrons that originate from bilateral stripes of IM, from which renal progenitors are generated (Drummond, et al., 1998). The rostral renal progenitors give rise to a single glomerulus that the nephrons share, while the remaining renal progenitors undergo a MET to generate tubules that fuse at the cloaca (Drummond, et al., 1998; Serluca and Fishman, 2001). Gene expression profiling, largely based on genes encoding solute transporter proteins that account for the exquisite functions of each segment, revealed that the zebrafish pronephros segment composition is analogous to other vertebrates (Wingert, et al., 2007; Wingert and Davidson, 2011). By 24 hours post fertilization (hpf), the zebrafish pronephros exhibits eight segments: the glomerulus (G), neck (N), proximal convoluted tubule (PCT), proximal straight tubule (PST), distal early (DE), corpuscle of Stannius (CS), distal late (DL), and pronephric duct (PD) (Fig. 1A) (Wingert, et al., 2007).

Fig. 1. mecom transcripts mark an early caudal domain of the renal progenitors, and the mecom expression domain is dynamic during nephrogenesis.

(A) Schematic depictions of zebrafish pronephros at 24 hpf, shown in lateral view. Enlargement represents segmental organization of the nephron at 24 hpf. Abbreviations: G (glomerulus), N (neck), PCT (proximal convoluted tubule), PST (proximal straight tubule), DE (distal early), CS (corpuscle of Stannius), DL (distal late), PD (pronephric duct), C (cloaca). (B) At the 2 and 3 somite stages, the renal progenitor field was labeled by pax2a (purple) and forming somites were marked by dlc (red). Onset of mecom expression in renal progenitors could be detected at the 3 somite stage, in a caudal domain exclusive to the dlc-expressing rostral domain. Inset shows non-overlapping expression territory of dlc (red) and mecom (purple) at 3 somites in the renal progenitor field. (C) Expression of mecom (purple) and myod1 (red) at various time points between 6 to 14 somite stages in wild types. At 14 somites, the expression domains of solute transporters slc4a4a and slc12a3 indicate premature patterning of the pronephros proximal versus distal segment regions. (D) Upper panel: genomic structure of zebrafish mecom202 (dark purple, bottom) and mecom201 (evi1) (light purple, top). The mecom202-5’ and mecom202-3’ riboprobes (orange filled lines) were designed to distinguish mecom202 and mecom201 transcripts by targeting the 5’ and 3’ region exclusively present in mecom202 transcripts. Lower panels: in 24 hpf wild type embryos, WISH using a full-length mecom202 probe and mecom202-specific probes showed that mecom expression was restricted to the DL and PD regions of the pronephros.

Studies of zebrafish nephrogenesis have identified several transcription factors and signaling pathways that are crucial for renal progenitor patterning (Gerlach and Wingert, 2012). Among them, the diffusible morphogen retinoic acid (RA) is essential for proximal-distal regionalization of the renal progenitor field (Wingert, et al., 2007; Wingert and Davidson, 2011). In target tissues, RA regulates gene expression by entering the nucleus and binding to its nuclear receptors, which upon RA interaction directly bind to retinoic acid response elements (RARE) to modulate transcription (Duester, 2008). Zebrafish genetic mutants lacking key RA synthesizing enzymes or wild types treated with diethylaminobenzaldehyde (DEAB), a chemical that blocks RA biosynthesis, develop a pronephros with reduced proximal segments and expanded distal segments (Wingert, et al., 2007; Wingert and Davidson, 2011). These studies established that RA induces proximal segment identities during the early somite stages, and may directly repress distal segments. Downstream of RA, the terminal boundaries of each segment are defined by the expression of domain-specific genes and appear to be controlled by the activity of multiple downstream transcription factors, presently known to include irx3b and hnf1fJ (Wingert and Davidson, 2011; Naylor, et al., 2013). Further, Notch signaling restricts MCC number by modulating the fate choice between transporting epithelium and MCC during mid-to-late somitogenesis (Ma and Jiang, 2007; Liu, et al., 2006). Despite these findings, many questions remain concerning how each nephron segment is precisely established during nephrogenesis—including the identity of other key factors involved in segmentation.

In searching for additional factors that may control nephron segmentation, we identified the zinc finger transcription factor Mecom as an intriguing candidate gene. Mecom is a splicing variant of the Evi1 (ecotropic virus integration site 1) and Mds1 (myelodysplastic syndrome 1) genes, which results in a N-terminal extension of the Evi1 protein (Wieser, 2007). Targeted disruptions that result in the loss of both transcripts causes embryonic lethality in mice associated with defects in neural, heart, and blood development—which suggests this locus has multiple essential roles during ontogeny (Wieser, 2007; Hoyt, et al., 1997). More recent work has demonstrated that Mecom is required for long-term hematopoietic stem cell maintenance (Zhang, et al., 2011).

With respect to vertebrate kidney development, transcripts encoding Mecom have been detected in the pronephros distal tubule and duct in Xenopus (Mead, et al., 2005) and zebrafish mecom has also been reported in the pronephros (Wingert, et al., 2007; Wingert and Davidson, 2011). In the zebrafish, mecom is initially expressed in the renal progenitor field, but its domain changes dynamically during nephrogenesis (Wingert, et al., 2007; Wingert and Davidson, 2011). mecom marks a broad caudal domain in early stages, then later is restricted to the DL and PD at 24 hpf. A genome-scale in situ analysis of mammalian transcriptional regulatory factors reported expression of murine Mecom in nascent nephron S-shaped bodies in the developing metanephric kidney (Yu, et al., 2012), thus further suggesting it could be involved in nephron patterning across vertebrates. However, the mechanism of how mecom modulates vertebrate nephron segmentation and the signaling pathways that may interact with mecom in renal progenitors remain unclear.

Through the present study, we found that interactions between RA, mecom, and Notch signaling are essential for zebrafish pronephros development. We show that mecom expression is extremely dynamic in zebrafish renal progenitors and is negatively regulated by RA. Using both loss and gain-of-function approaches, we found that mecom is necessary for proper DL segment formation, and that the absence of mecom activity expands the PST segment and MCC numbers. Moreover, mecom and RA have opposing roles in PST and DL formation, as mecom morphants treated with exogenous RA had a more expanded PST and an abrogated DL, while DEAB treatment rescued segmentation in mecom morphants. Consequently, we discovered a previously unappreciated role for RA in MCC development, since DEAB treatment prevented MCC formation while mecom knockdown in DEAB-treated embryos rescued MCCs. These data indicate that RA regulates MCC fate choice by inhibiting mecom. Furthermore, we established the epistatic relationship between Notch signaling and mecom during MCC differentiation, where mecom acts upstream to promote Notch activity. Taken together, our data suggest a model in which mecom and RA function during early nephrogenesis stages to arbitrate proximodistal segment pattern, and reveal that MCC fate choice is modulated by a complex interplay between RA, mecom, and Notch signaling to precisely define the MCC domain and density in the nephron.

MATERIALS AND METHODS

Zebrafish husbandry and ethics statement

Zebrafish were maintained in the Center for Zebrafish Research at the University of Notre Dame Freimann Life Science Center. Wild type embryos were raised and staged as described (Kimmel, et al., 1995). The Institutional Animal Care and Use Committee at the University of Notre Dame supervised experimental procedures under protocols 13-021 and 16-025.

Morpholino knockdown, cRNA synthesis, and heat shock experiments

All morpholinos were purchased from Gene Tools, LLC (Philomath, OR), and were solubulized as recommended and stored at a 4 mM concentration. The mecom morpholino e1SD (5’-CTGAGTGACTTACATATGAAGGGCT-3’) was designed to target the splicing donor site of zebrafish mecom (XM_001920912) exon 3, and e1SA (5’- TTGTGGCAGAC-CTCACGACGGTGTT-3’) targets the splicing acceptor of exon 4. The mecom mismatch morpholinos (5’-CTGATTGACGTACAAATGATGGGCA-3’, 5’-TTGTAGCAGGCCTCGC-GACTGTGTA-3’) were used as controls. A combination of mecom e1SD and e1SA morpholinos produced fully penetrant effects, and was used for all mecom knockdown experiments. One-cell stage wild type embryos were injected with 1-5 nl 0.2 mM mecom morpholinos and raised to the desired stages at 28°C. For gene expression analysis, embryos were fixed in 4% paraformaldehyde/1×PBST and stored in methanol at −20°C. Synthetic mecom cRNA was synthesized from a mecom.pCS2 plasmid using the mMessage mMachine sp6 kit (Ambion). For rescue experiments, one-cell stage embryos were co-injected with 1-5 nl 5 ng/μl mecom cRNA and 0.2 mM mecom morpholinos. For NICD heat shock experiments, embryos were incubated at 37°C for 1 hour beginning at 90% epiboly, allowed to develop to the desired stage, fixed and processed for WISH, and then ultimately genotyped as described (Scheer and Campos-Ortega, 1999).

RT-PCR

RNA was extracted from mecom morphant and wild type embryos at 24 hpf with Trizol reagent (Invitrogen) according to the manufacturer’s instructions. RT-PCR was performed using the SuperScript First-Strand Synthesis System (Invitrogen). To confirm mis-splicing of mecom in morphant embryos, PCR primers were designed to specifically amplify the mecom mRNA region corresponding to 180 bp of sequence spanning exon 3 and 4 (exon 3 forward primer 5’-AACTTGCAGAGGCATATCCG-3’, and exon 4 reverse primer 5’-GGCAGAGATTGG-AGAACTGC-3’). Primers located in intron 3-4 were used to amplify a 190 bp intronic fragment (intron 3-4 forward primer 5’-CCTATGCTAGGGACACCTGG-3’, and intron 3-4 reverse primer 5’-GCCTCGAAAACGTAGAATGC-3’). To assess the exon 3-intron 3-4 fusion in mecom morphants, the exon 3 forward primer and intron 3-4 reverse primer were used in combination to amplify a ~1.3 kb cDNA fragment.

Dextran injection

To assay kidney function, 40 kDa fluorescent dextran-FITC (Invitrogen) was injected into an axial somite of 1-phenyl-2-thiourea (PTU) treated wild type embryos or mecom morphants that were anesthetized with 0.02% tricaine at 48 hpf. Embryos were revived and incubated in PTU in the dark until later time points for observation. Dextran clearance was observed with a fluorescent microscope at 72 hpf and 98 hpf.

Benzidene staining

o-dianisidine stock was made by dissolving 0.07 g of o-dianisidine (Sigma D9134) in 50 mls of 100% ethanol. Live embryos were incubated in the dark for 15 minutes in a working solution comprised of 2 ml o-dianisidine stock, 500 ul of 0.1M sodium acetate pH 4.5, 2 ml of distilled water, and 100 ul of hydrogen peroxide. Embryos were rinsed three times in E3 and then fixed in 4% PFA for scoring and imaging.

Chemical treatments

RA and DEAB (Sigma-Aldrich) were dissolved in 100% DMSO to make 1M stocks and the aliquots were stored at −80°C (Wingert, et al., 2007). For RA treatments, tailbud stage embryos were incubated in 1×10⁻⁷M RA/DMSO made with E3 embryo media in the dark until 24 hpf, washed 3 times with E3, then fixed. For DEAB treatments, the embryos were incubated in 1.6×10⁻⁵M DEAB/DMSO diluted in E3 media from 75% epiboly to 24 hpf. Control embryos were allowed to develop in 1×10⁻⁷M or 1.6×10⁻⁵M DMSO over corresponding developmental intervals. DAPT was dissolved in DMSO to make a 10mM stock and stored at −80°C. Bud stage embryos were incubated in 100 μM DAPT/DMSO or DMSO alone in E3 media to 24 hpf at 28°C in the dark. These chemical treatments were fully penetrant and produced consistent results.

WISH

For our reported WISH expression studies and images, representative results are provided based on analysis of at least 20 embryos, and gene expression domains as reported by somite boundaries were based on counts of at least 5 separate embryos for accuracy. Zebrafish WISH was performed as previously described (Wingert, et al., 2007). To generate antisense probes for pax2a, dlc, egr2a, mecom, myod1, smyhc1, slc4a4a, slc20a1a, trpm7, clcnk, slc12a1, slc12a3, wt1b, and nr5a1a, we used IMAGE clone template plasmids for in vitro transcription, as previously reported (Wingert, et al., 2007; Wingert and Davidson, 2011). Probes for odf3b were transcribed from PCR templates amplified with primers (5’-ATGTCA-CCTGTGGATGTATG-3’, 5’-AATTAACCCTCACTAAAGGGTTAATCTTCACC-3’). To distinguish mecom from other transcript variants, we generated antisense probes mecom829 and mecom555 targeting the first 829 bp and the last 555 bp fragments specifically present in mecom using PCR templates obtained with mecom829 primers (forward: 5’-TGGATTTTGAGGGACAGGAG-3’, reverse: 5’-AATTAACCCTCACTAAAGGGCATTGGGGCATTATGGGTAG-3’) and mecom555 primers (forward: 5’-CTTTGAGTCTGGTTCGGAGC-3’, reverse: 5’-AATTAACCCTCACTAAAGG GTTGGATTTCCAGTTGAAGGC-3’) respectively.

Cell counting and statistics

Quantification of MCC density in the pronephros was conducted by counting the number of MCCs in a length ranging from 50 μm to 150 μm per nephron. MCC density was determined by calculating the number of MCCs per 10 μm. At least 20 embryos were examined for each experiment. Student’s t-test was applied for quantitative results.

RESULTS

mecom delineates a dynamic caudal subdomain in the renal progenitor field

Based on the gene expression patterns of specific solute transporters, the molecular anatomy of the zebrafish pronephros consists of proximal and distal segments that are analogous to mammalian nephrons (Wingert, et al., 2007). Eight similar regions have been identified, including two proximal tubule domains (PCT, PST) and two distal tubule domains (DE, DL) (Fig. 1A). Prior studies have documented the appearance of mecom transcripts as occurring between the 6-8 somite stages during pronephros development (Wingert, et al., 2007; Wingert and Davidson, 2011). To further examine the onset of mecom expression in the renal progenitor field, we assessed the domain of mecom transcripts by performing double whole-mount in situ hybridization (WISH) with several combinations of riboprobes consisting of both segmentation and somitogenesis gene probes that would enable labeling of renal progenitor domains compared to the somite boundaries. To confirm the embryonic stage and define somite boundaries, emerging somites were labeled in embryos between 1-6 somites using deltaC (dlc), while embryos between 6-18 somites were labeled with myogenic differentiation 1 (myod1).

The domain of renal progenitors was visualized based on the expression of pax2a (Fig. 1B) (Pfeffer, et al., 1998). At the 2 somite stage, mecom expression was absent in the renal progenitor field, whereas transcripts were present in the developing brain (Fig. 1B, Fig. S1A). The onset of mecom expression in the renal progenitor domain occurred at the 3 somite stage, adjacent to where somite 4 was forming (Fig. 1B, Fig. S1A). In addition to marking the somites, the Notch ligand dlc is known to mark the rostral territory of the renal progenitors, which will eventually develop into proximal segments in the pronephros (Wingert, et al., 2007; Wingert and Davidson, 2011). During the onset of mecom expression, there was a clear delineation between the dlc-expressing rostral domain and the mecom-expressing caudal domain at the 3 somite stage (Fig. 1B, inset). Notably, the detection of non-overlapping domains of dlc and mecom at the 3 somite stage is the earliest proximodistal distinction that has been observed within the renal progenitor field during zebrafish pronephros development.

Between the 6 and 15 somite stages, dynamic mecom expression was observed in the developing pronephros. The domain of mecom transcripts gradually shifted toward the caudal end of the renal progenitor field (Fig. 1C, Fig. S1B). By 14 somites, pre-mature pronephric segments were established: the proximal domain was labeled by the solute transporter solute carrier family 4, member 4a (slc4a4a), while the distal domain was marked by solute carrier family 12 (sodium/chloride transporters), member 3 (slc12a3) (Fig. 1C, lower two panels). At this time point, mecom transcripts were detected throughout the distal domain and partially overlapped with the proximal domain (Fig. 1C). The progressive shifting of the mecom expression domain continued from 18-28 somite stages, until its expression became restricted to the DL and PD (Fig. 1D, Fig. S2). The presence of mecom transcripts in early renal progenitors, and its ongoing dynamic expression domain, suggested to us that mecom could be functioning at multiple times in developing renal progenitors.

In zebrafish, two mecom splice variants have been identified. The 3129 basepair (bp) mecom202 (mecom, also known as prdm3) (Sun, et al., 2008) possesses additional extensions in both the 5’ and 3’ ends of the 1807 bp mecom201 (evi1). To distinguish expression of mecom202 from mecom201, we utilized the 5’ and 3’ end fragments present exclusively in mecom202 transcripts and designed riboprobes mecom202-5’ and mecom202-3’ (Fig. 1D). WISH using a full-length mecom riboprobe indicated mecom was restricted in the DL and PD regions of the pronephros at 24 hpf, while mecom202-5’ and mecom202-3’ probes further demonstrated the mecom202 transcripts were expressed in the pronephros and restricted to the DL and PD at 24 hpf as well (Fig. 1D; Fig. S1, Fig. S2).

mecom morphants exhibited renal edema and dysfunction

To assess the role of mecom in zebrafish pronephros segmentation, we performed morpholino knockdown studies. One-cell stage wild type embryos were injected with a mecom morpholino targeting the splice donor site of exon 3, the splice acceptor of exon 4, or a combination of these two morpholinos (Fig. 2A, Fig. S3A, Fig. S3B). We found that the morpholino combination induced stronger phenotypes than either morpholino alone (Fig. 2A, Fig. S3A, Fig. S3B). Based on these results, we co-injected the mecom splice morpholinos for the remainder of our studies on mecom loss of function. Next, we performed a series of RT-PCR experiments to assess changes in mecom mRNA splicing in co-injected morphant embryos compared to uninjected wild type controls. Primers were designed to amplify the mecom sequence spanning exon 3 and exon 4 to identify the properly spliced mecom mRNA fragment (Fig. 2A). We found that cDNA isolated from wild-type embryos contained mecom transcripts with the predicted exon 3-4 band size of 180 base pairs (bp), which indicated appropriate splicing between these exons (Fig. 2A). However, mecom morphant embryos injected with the combination of splice morpholinos showed low, if any, correctly processed mecom mRNA across exons 3-4 (Fig. 2A). This suggested that there was a significant reduction in normal mecom mRNA as a consequence of the morpholino injection. In mecom morphants, these primers failed to amplify a mecom cDNA product containing the entire intron 3-4 (data not shown), but this could be due to the size of that intron, which is predicted to be 4.4 kilobases (kb) in length.

Fig. 2. mecom morphants exhibit pericardial edema and symptoms of renal failure.

(A) Schematic indicates targeting sites of mecom morpholinos (blue lines) blocking splice sites of mecom mRNA between exon 3 and 4. Primers (red arrowheads) were designed to amplify the 180 bp linkage region between properly spliced exon 3 and 4. Right panels: cDNA isolated from wild type embryos showed the 180 bp band indicating proper splicing to remove intron 3. In contrast, amplification of this product was abrogated in mecom morphants. eef1a1l1 was used as internal control. (B) mecom morphants showed gross developmental defects when compared with mismatch controls. Pericardial edema could be visualized at 30 hpf in mecom morphants indicating fluid accumulation (left panels). At 50 hpf, mecom morphants displayed severe pericardial edema and body curvature (right panels). (C) Fluorescent 40 kDa dextran was injected into the somite of wild types or mecom morphants at 48 hpf. A failure of renal clearance, indicated by dextran accumulation, was observed in the yolk and edema of mecom morphants at 72 and 96 hpf.

To further address how the mecom transcript was processed in the morphants, we designed primers that would amplify smaller portions of the intronic interval. In mecom morphants, but not in wild type uninjected embryos, we detected the presence of transcripts corresponding to mecom intron 3-4 sequences (example in Fig. S3C, data not shown). Further, in mecom morphants we were able to amplify a 1.3 kb mecom cDNA sequence spanning the 3’ end of exon 3 and the 5’ end of intron 3-4; importantly, this intronic fragment contains a series of in-frame stop codons (data not shown). Thus, we hypothesize that the mecom morpholinos generate a mis-spliced mRNA that in turn results in a truncated Mecom protein. While we were unable to determine if mecom morphants retain the entire intron 3-4 after mecom transcript processing, our data nevertheless confirm that these splice junction mecom morpholinos lead to mis-spliced mecom mRNA.

Next, to broadly evaluate the role of mecom during zebrafish development, we followed the development of mecom morphants compared to control embryos injected with the mismatch morpholinos. At 30 hpf, mecom morphants showed pericardial edema, mild tail-axis curvature, and brain and eye defects (Fig. 2B). In contrast, wild type embryos injected with the mismatch morpholinos had no gross developmental abnormalities (Fig. 2B). At 50 hpf, mecom morphants displayed severe pericardial edema, suggesting possible pronephros dysfunction and fluid imbalance compared to wild types (Fig. 2B). Fluid accumulation could also result from cardiac defects and impaired circulation. However, the heart rate was normal in morphants compared with wild type embryos, circulation ensued normally at 24 hpf, and circulation continued to appear normal through 48 hpf (data not shown). To further scrutinize the circulation, we used benzidene staining, which labels hemoglobin in differentiated erythrocytes starting just after 32 hpf, and can more precisely assay if blood pooling transpired in mecom morphants (Fig. S4) (Wingert, et al., 2004). We found that 8.6% (4/46) of mecom morphants had cranial blood pooling at 36 hpf, and that the incidence of this phenotype increased to 54.5% (12/22) of mecom morphants by 48 hpf, while no wild type embryos displayed a pooling phenotype (out of 51 animals assayed at these time points) (Fig. S4). Taken together, these data suggest the mecom knockdown is associated with circulation and/or vascular integrity defects that worsen over developmental time. We hypothesize that this contributed to the pericardial effusion phenotype in mecom morphants. To assess renal function in mecom knockdowns, we explored renal absorption and clearance properties with dextran tracing experiments. We injected 40 kDa dextran-FITC into the somites of wild type embryos or mecom morphants at 48 hpf. From 72 hpf to 98 hpf, wild types exhibited clearance of the fluorescent molecule, whereas abundant dextran-FITC accumulation was visualized in mecom morphants (Fig. 2C). These results indicate that mecom is necessary for proper development and function of the pronephros.

mecom knockdown induced PST specific expansion in the zebrafish pronephros

To further characterize nephron development in mecom morphants, we used double WISH with domain-specific markers and the somite marker smyhc1 to precisely define each segment boundary relative to each somite. Interestingly, at 24 hpf, mecom morphants had expanded proximal tubules (Fig. 3). The pan-proximal marker slc4a4a, encoding an electrogenic Na⁺ bicarbonate cotransporter, was expressed in the PCT and PST segments from somite 4 to 11 in wild-type embryos, whereas in mecom morphants, the slc4a4a domain was expanded from somite 4 to 14, and was rescued by mecom overexpression (62.5%, 20/32) (Fig. 3, Fig. S5). Scrutiny of each segment domain revealed that expansion of the proximal tubule was attributed to an enlarged PST. The PCT expresses the sodium-dependent phosphate transporter slc20a1a, a domain that was situated adjacent to somites 4 to 8 in both wild type embryos and mecom morphants (Fig. 3, Fig. S5). However, the PST labeled by the transient receptor potential cation channel gene trpm7 was expanded from somite 9 to 14 in mecom morphants, while its expression in wild types was located next to somites 9 to 11 (Fig. 3, Fig. S5). The PST expansion was consistent in more than 90% (49/52) of mecom morphants. In addition, more than 60% (23/36) of mecom morphants could be successfully rescued by co-injecting the mecom morpholinos along with full-length zebrafish mecom cRNA (Fig. 3, Fig. S5). This result supports the conclusion that the segmentation defects observed after mecom knockdown are specific to the abrogation of mecom activity during nephrogenesis. The development of other proximal nephron cell types like the podocytes, or the interrenal lineage that emerges next to the podocytes, was not affected in mecom morphants (Fig. S6).

Fig. 3. mecom knockdown leads to an expanded PST segment.

(A) At 24 hpf, WISH indicates proximal segments marked by slc4a4a elaborated in mecom morphants. Within proximal domains, PCT labeled by slc20a1a was not affected comparing with wild type embryos. Knockdown of mecom induced a 3-somite expansion in trpm7-expressing PST in morphant pronephros at 24 hpf, which could be rescued by co-injecting mecom cRNA along with mecom morpholinos. (B) Schematic summary of proximal segment organization in wild type, mecom morphant, and mecom rescued embryos, with PST alterations highlighted in yellow. Abbreviations: P (podocytes), PCT (proximal convoluted tubule), PST (proximal straight tubule), DE (distal early), DL (distal late), PD (pronephric duct).

mecom morphants display PST expansion at the expense of the DL segment

To determine how the abnormal PST in mecom morphants affected distal segment patterning, we used double WISH with distal markers and smyhc1. Indeed, the increased PST in the morphant pronephros was accompanied by a robust reduction in the distal pronephros domain (Fig. 4, Fig. S5). In wild type embryos at 24 hpf, the expression pattern of the chloride conductance channel gene clcnk could be seen in the DE, DL and PD regions from somite 12 to 18 (Fig. 4, Fig. S5). In contrast, the clcnk expression domain was reduced in the mecom morphants, being located from somite 15 to 18 starting adjacent to somite 15; however, this aberrant phenotype was rescued by mecom transcript overexpression (66.7%, 22/33) (Fig. 4, Fig. S5). The DE domain marked by slc12a1, which encodes a Na⁺/K⁺/Cl⁻ cotransporter, had a similar length in wild types and mecom morphants (Fig. 4, Fig. S5). Surprisingly, the distal pronephros region was reduced specifically at the cost of the DL segment. The DL domain indicated by slc12a3, a Na⁺/Cl⁻ transporter gene, showed a significant reduction in mecom morphants, and was restricted to somites 17 to 18, whereas wild type embryos showed slc12a3 expression from somite 14 to 17 (Fig. 4, Fig. S5). The abrogation of the DL domain was not associated with heightened cell death in this domain at 24 hpf based on acridine orange staining (data not shown). Finally, we found that DL segment formation was rescued in nearly 70% (26/37) of embryos that were co-injected with mecom morpholinos and mecom cRNA (Fig. 4A, Fig. S5), further supporting the specificity of mecom morpholinos. We conclude from these findings that mecom is essential for normal nephron segmentation, such that mecom activity promotes the DL and restricts the PST.

Fig. 4. mecom knockdown leads to the formation of a reduced DL segment.

(A) WISH using distal segment markers showed a 3-somite reduction in DL region by 24 hpf. The distal segments and pronephros ducts were labeled by clcnk. Expression of slc12a1 and slc12a3 marked DE and DL respectively. Pronephros segment boundaries were evaluated relative to the somites, which were shown by smyhc1 expression (red). The reduced DL could be ameliorated by co-injection of mecom morpholino and mecom cRNA. (B) Schematic depiction of distal segment alterations, with DL domains highlighted in orchid. Abbreviations: P (podocytes), PCT (proximal convoluted tubule), PST (proximal straight tubule), DE (distal early), DL (distal late), PD (pronephric duct).

RA negatively regulates mecom during pronephros segmentation

Next, we sought to identify possible developmental pathways relative to mecom in establishing the pronephric segmentation pattern. RA signaling plays important roles during segmentation: RA is required to induce proximal segments and to prevent distal segment expansion at early somitogenesis stages (Wingert, et al., 2007; Wingert and Davidson, 2011). To further explore the relationship between mecom and RA signaling, we treated wild type embryos and mecom morphants with exogenous all-trans RA at the concentration of 1×10⁻⁷M from bud stage to 24 hpf, and examined segmentation changes using WISH (Fig. 5, Fig. S7). At this RA dosage and time window, wild type embryos developed a minor ‘proximalized’ pronephros phenotype with an expanded PST, indicated by trpm7 expression, and a reduced DL segment marked by slc12a3 (Fig. 5, Fig. S7). However, at the same treatment condition, mecom morphants displayed a severe ‘proximalized’ pronephros indicated by the expansion of the trpm7 expressing PST domain throughout the entire tubule and an almost complete loss of the slc12a3 expressing DL segment (Fig. 5, Fig. S7). Furthermore, as shown previously (Wingert, et al., 2007), treating wild types with DEAB, an inhibitor of RA aldehyde dehydrogenase (aldh) synthesizing enzymes, resulted in a ‘distalized’ pronephros with an expanded DL and reduced PST (Fig. 5, Fig. S7). When compared to mecom morphants or mecom morphants treated with RA, mecom morphants incubated with DEAB showed a partially reduced PST, as shown by trpm7 expression from somite 9 to 13, and partially expanded slc12a3 expression from somite 15 to 17 (Fig. 5, Fig. S7). Altogether, these results show that mecom and RA have antagonistic activities in PST and DL formation: mecom inhibits the PST while RA promotes the PST, and mecom promotes the DL while RA inhibits the DL.

Fig. 5. mecom and RA have opposing roles in PST and DL formation during proximodistal segmentation of the pronephros.

(A) Wild type embryos or mecom morphants were incubated with 1×10⁻⁷M RA, 1.6×10⁻⁵M DEAB or DMSO. WISH using the PST marker trpm7 and DL marker slc12a3 showed that exogenous RA resulted in a more severe segmental phenotype in the mecom morphant pronephros, with further expanded PST and reduced DL. DEAB treatment partially rescued the segmentation phenotype in mecom morphants. (B) Schematic summary of segmentation changes in wild type embryos, mecom morphants, and wild type embryos or morphants treated with RA or DEAB. Yellow and blue boxes highlight the PST and DL segments, respectively.

Based on these findings, we hypothesized that RA could negatively regulate mecom, which would provide a mechanism to account for the patterning change in the PST and DL segments. Interestingly, RA treatment beginning at 60% epiboly has been shown to reduce the mecom expression domain at the 6-8 somite stage, while DEAB exposure beginning at this time leads to an expansion of the mecom domain at the 6-8 somite stage—changes that correlate with reduced and expanded distal segments, respectively (Wingert, et al., 2007). To further assess the relationship between RA signaling and mecom, wild type embryos were treated with RA or DEAB and mecom expression was evaluated using WISH (Fig. 6A, Fig. S7). Wild type embryos treated with RA showed a dramatically reduced mecom expression domain at 24 hpf, while embryos treated with DEAB had a dramatically expanded mecom domain, consistent with the notion that RA negatively regulates mecom expression during nephrogenesis (Fig. 6A, Fig. S7).

Fig. 6. RA negatively regulates mecom, which enables MCC formation.

(A) Wild type embryos treated with 1×10⁻⁷ M RA, 1.6×10⁻⁵ M DEAB or DMSO. Analysis of mecom transcripts using WISH shows that RA treatment diminishes the mecom domain while abrogation of RA signaling expands the mecom domain. (B) MCCs were labeled by odf3b in wild types and mecom morphants treated with RA or DEAB. mecom knockdown resulted in a caudal shift of MCC domain in the DL and PD regions, and this effect was partially rescued by treating morphant embryos with DEAB. Notably, wild type embryos treated with DEAB had abolished MCC formation, while RA induced ectopic MCC formation in the distal region of the pronephros.

The activities of RA and mecom also regulate epithelial cell fate choice in tubule segments

In the zebrafish pronephros, intercalated along the PCT, PST, DE, and the anterior-most DL transporting epithelia are distinct multiciliated cells (MCCs), which can be evaluated based on the expression of various ciliogenesis genes, such as odf3b, via WISH (Ma and Jiang, 2007; Liu, et al., 2006). How the MCC domain is defined within the renal progenitor field has yet to be understood, but MCC cell fate choice is limited through Notch signaling (Ma and Jiang, 2007; Liu, et al., 2006). Interestingly, we found that mecom morphants exhibited an expanded MCC domain and a noticeable overall increase of MCC density at 24 hpf (Fig. 6B, Fig. S7, Fig. S8). These observations indicate that mecom negatively regulates MCC development.

While RA is vital for patterning pronephros segmentation (Wingert, et al., 2007; Wingert and Davidson, 2011), prior studies have not addressed how RA might affect MCC formation. Given the relationship between RA and mecom, we assessed whether RA is also involved in MCC patterning. Wild type embryos were treated with 1×10⁻⁷M all-trans RA from tailbud stage to 24 hpf. The RA-treated embryos had a caudal expansion of the MCC domain (Fig. 6B, Fig. S7). When mecom morphants were treated with the same dosage of RA during this time window, they showed a slightly further expanded MCC domain (Fig. 6B, Fig. S7). These data suggested that RA has dual roles in patterning of the MCC domain and stimulating MCC formation, possibly via the inhibition of mecom.

To determine if RA is required for MCC progenitor formation during early nephrogenesis, we exposed embryos to DEAB to inhibit RA synthesis. Wild type embryos treated with DEAB from 75% epiboly to 24 hpf, at the concentration documented to induce proximodistal segmentation changes in the pronephros (Wingert, et al., 2007), displayed a complete loss of MCCs at 24 hpf (Fig. 6B, Fig. S7). Consistent with the notion that RA is required for MCC formation, lightbulb (lib) embryos that have a mutation in the RA biosynthesis gene aldehyde dehydrogenase 1a2 (Wingert and Davidson, 2011) showed a reduction in the MCC domain (Fig. S9). Next, we further tested the relationship between mecom and RA in MCC formation. Since our findings thus far indicated that mecom represses MCCs, and RA represses mecom, we hypothesized that DEAB treatment had abrogated MCC formation due to the elevated mecom expression (Fig. 6A). In keeping with this, DEAB-treated mecom morphants developed MCCs (Fig. 6B, Fig. S7). These results indicate that alleviating the inhibitory effect of mecom on MCC differentiation is sufficient to reverse the MCC deficiency that results from an absence of RA signaling. Taken together, these data suggest that there is an exquisite interplay involving RA and mecom that titrates MCC formation: essentially, that RA acts to inhibit mecom, while mecom inhibits MCC fate choice.

mecom and Notch signaling coordinate MCC density in zebrafish pronephros

In the pronephros tubule, MCCs are dispersed among the single-ciliated transporting epithelia in a ‘salt-and-pepper’ pattern (Ma and Jiang, 2007; Liu, et al., 2006). Notch signaling creates this cell distribution by controlling the MCC fate decision through a mechanism of Notch-mediated lateral inhibition (Ma and Jiang, 2007; Liu, et al., 2006). During pronephros differentiation, MCC progenitors expressing the Notch ligand jag2a exclusively repress MCC fate in neighboring cells, leading them to acquire a transporting epithelial fate (Ma and Jiang, 2007; Liu, et al., 2006). The y-secretase inhibitor DAPT has been shown to effectively block Notch signaling and causes increased MCC numbers to form during nephrogenesis (Ma and Jiang, 2007; Liu, et al., 2006). Based on the discovery that RA levels affect MCC formation, we first addressed the relationship between RA and Notch signaling in modulating MCC fate. Changes in RA levels during mid-gastrulation (between 60-70% epiboly) alter proximodistal segmentation (Wingert, et al., 2007; Wingert and Davidson, 2011), while changes in Notch affect MCC differentiation at slightly later times (90% epiboly to bud stages) (Ma and Jiang, 2007; Liu, et al., 2006). Thus, it is very likely that RA acts upstream of Notch in MCC formation. MCC formation was evaluated when embryos were treated with RA or a combination of RA and expression of the Notch1a intracellular domain (NICD) under heat-shock control using a transgenic line Tg(hsp70:gal4; uas:notch1a-intra) (Fig. S10). RA treatment shifted the MCC domain posteriorly in non-transgenic NICD siblings (Fig. S10). This was consistent with the change in proximodistal segment domains after RA treatment in wild types (Fig. 5), and confirmed our prior observation (Fig. 6B). Dual RA treatment and NICD heat-shock caused a similar MCC domain posterior shift, but the number of MCCs was reduced (Fig. S10). This confirms that Notch acts as a repressor of MCC fate, and is consistent with the notion that RA treatment causes a diminution of Notch signaling that increases MCC formation in the pronephros.

Next, we hypothesized that mecom might interact with Notch to mediate the MCC epithelial fate choices because mecom morphants showed an increase in MCC density similar to the effect of blocking Notch signaling (Ma and Jiang, 2007; Liu, et al., 2006). To compare the effects of mecom and Notch on MCC differentiation, MCC number was evaluated in mecom morphants and DAPT treated wild types using the ciliogenesis marker odf3b. As noted previously, the MCC domain shift in mecom morphants correlated with the alteration in segment sizes, and there was a significant increase of MCC density compared to wild types at 24 hpf (Fig. 7). To assess MCC quantity, the number of MCCs was quantified per every 10 μm in single nephrons. MCC density in mecom morphants was increased by 50% compared to wild types (Fig. 7). As reported (Ma and Jiang, 2007; Liu, et al., 2006), wild type embryos treated with DAPT exhibited an increased MCC density (Fig. 7). In addition, whereas MCCs are regularly dispersed in a salt-and-pepper array within the wild type nephron, large MCC aggregates were observed both in DAPT-treated wild types and mecom morphant nephrons (Fig. 7). mecom morphants incubated with DAPT did not show any further increase in MCC density compared to mecom morphants or embryos treated with DAPT treatment alone (Fig. 7). Further, disrupting Notch signaling in mecom morphants with DAPT did not induce additional condensed MCC numbers (Fig. 7). Taken together, these results show that there is no additive effect of mecom and Notch signaling in modulating MCC differentiation. As such, we hypothesized that mecom and Notch signaling may collaborate in the same pathway to limit MCC formation, with additional roles of mecom to precisely define MCC territory in the pronephros.

Fig. 7. mecom acts upstream of Notch signaling to modulating MCC differentiation and regulates the MCC domain.

(A) Wild type embryos treated with 100 μM DAPT showed a significant increase of MCC number without ectopic MCC formation. mecom knockdown led to a caudal expansion of MCC in the DL and PD, and exhibited increased MCC density compared to wild types. A similar condensed MCC arrangement could also be seen in mecom morphants treated with DAPT. The overexpression of Notch1a resulted in decreased MCCs in heatshock induced Tg(hsp70:gal4; uas:notch1a-intra) embryos. Ectopic MCC formation associated with mecom knockdown was abolished by Notch signaling activation, though the domain of MCCs was still expanded. (B) Differentiated MCCs at 24 hpf under 10× magnification in a single nephron from wild types, wild types treated with DAPT, mecom morphants and morphants treated with DAPT, and finally wild types and mecom morphants with NICD overexpression. Note MCCs displayed a condensed organization/cluster pattern in DAPT-treated wild type and mecom morphant pronephros, while DAPT treatment in mecom morphant failed to induce further MCC density. Arrows indicate large MCC aggregates observed in DAPT-treated wild types, mecom morphants, and DAPT-treated mecom morphants, which were absent from the wild type pronephros. For each experiment, at least 20 embryos were examined. (C) Quantification of MCC density in wild types, wild type embryos treated with DAPT, mecom morphants, and mecom morphants treated with DAPT. Student t test revealed significant increase of MCC density in DAPT treated wild types and mecom morphants relative to untreated wild types (***p=0.0005). Alterations of MCC density between mecom morphants and morphants treated with DAPT did not reach statistical significance. For each experiment, at least 20 embryos were examined.

To address whether mecom acts upstream or downstream of Notch signaling in modulating MCC differentiation, we utilized Notch transgenic lines to investigate whether MCC expansion could be ameliorated in mecom morphants with ectopic Notch signaling. We again used the double transgenic line Tg(hsp70:gal4; uas:notch1a-intra) to overexpress NICD with heat-shock induction. NICD activation by heat-shocking transgenic embryos at the bud stage resulted in fewer MCCs compared with control siblings (Fig. 7), consistent with published results (Liu, et al., 2006). Strikingly, transgenic NICD activation in mecom morphants reduced MCC numbers, thus successfully rescuing the aggregate MCC phenotype (Fig. 7). These results provide evidence that Notch signaling acts downstream of mecom to restrict MCC formation during the choice between MCC and transporting epithelia fates. Overall, these studies show that mecom is a vital component of the gene regulatory network that controls MCC development.

DISCUSSION

Herein we uncovered a fundamental role of mecom in modulating proximodistal segmentation and MCC fate choice during zebrafish nephrogenesis (Fig. 8A). Given the presence of mecom transcripts in a shifting caudal domain of the renal progenitors during the time period when nephron segmentation is established (through to 24 hpf), we speculate that mecom could act in multiple places and times to restrict the PST and promote the DL, respectively. The onset of mecom expression was detected at the 3-somite stage in the caudal domain of the renal progenitor field, in a non-overlapping pattern with the dlc-expressing rostral domain. This provides molecular evidence for renal progenitor specification several hours earlier in somitogenesis than previously documented (Gerlach and Wingert, 2012). However, this finding is consistent with the observation that modulations in RA levels during mid to late gastrulation have consequences for nephron proximodistal pattern (Gerlach and Wingert, 2012). The mecom expression domain later undergoes an incredibly progressive and dynamic shift toward the caudal end of the renal progenitor field before its restriction to the DL-PD at 24 hpf. Based on the presence of mecom transcripts in central and caudal renal progenitors, we hypothesize that mecom functions in proximodistal segmentation to promote DL induction and restrict the PST field during nephrogenesis (Fig. 8A).

Fig. 8. Model of mecom function during nephrogenesis.

(A) Role(s) of mecom during pronephros proximodistal segmentation. At early somitogenesis, forming somites generate a gradient of RA, which diffuses along the IM and modulates proximodistal patterning of the renal progenitors by promoting proximal segmentation and restricting distal fates. Initially expressed in the caudal domain of the renal progenitor field, mecom is negatively regulated by RA signaling and executes a contrary role to that of RA by favoring distal tubule formation and/or limiting proximal segmentation. Interplays between RA and mecom, as well as other transcription factors and signaling pathways precisely define the patterning of the renal progenitors during nephrogenesis, which develops into a pronephros with at least eight distinct segments by 24 hpf. mecom is specifically essential to promote the DL and possibly to restrict the PST. (B) RA, mecom, and Notch activities coordinate MCC formation. RA signaling is required for MCCs to develop, and one modality is that RA acts to negatively regulate mecom expression. RA likely has other targets that promote MCC formation, which remain unidentified. To further refine the balance between MCC and transporting epithelia fate choice, mecom represses ectopic MCC formation by promoting Notch signaling. Notch is the penultimate signal that inhibits MCC induction via lateral inhibition. By 24 hpf, MCCs are dispersed along the pronephric tubules in PST, DE, and DL regions adopting a proper ‘salt-and-pepper’ fashion. Abbreviations: RA (retinoic acid), PCT (proximal convoluted tubule), PST (proximal straight tubule), MCC (multiciliated cell), DE (distal early), DL (distal late), PD (pronephric duct).

In addition, we have determined several new roles and relationships between mecom, RA and Notch signaling during MCC fate choice along the pronephros (Fig. 8B). We found that mecom morphants exhibit increased MCC numbers, demonstrating that mecom is needed to repress MCC fate choice. Blocking RA signaling by treating embryos with DEAB completely prevented MCC formation, and this inhibitory effect was alleviated by knocking down mecom. These data suggest that RA signaling promotes/enables MCC formation through mecom inhibition. Further, our studies place mecom upstream of Notch, since MCC expansion in mecom morphants was rescued by inducing Notch signaling. Based on these findings, we conclude that RA is crucial for MCC progenitor formation by limiting mecom activity, and that mecom in turn provides an inhibitory effect on MCC formation through promoting Notch signaling to regulate the balance of MCC-transporting epithelia fate choice (Fig. 8B). Also, since RA treatment can expand the MCC domain in mecom morphants, this suggests that there could be other factor(s) downstream of RA, possibly in parallel with mecom/Notch, that contribute to the positive role of RA in promoting MCC differentiation (Fig. 8B). Additional studies are needed to address how mecom coordinates both proximodistal segmentation and MCC fate choice—such as whether these are actually independent or overlapping pathways. Nevertheless, our data suggest that the proper dosage of RA, mecom, and Notch is absolutely essential for normal nephrogenesis.

Mechanisms of mecom function during zebrafish nephrogenesis

The molecular mechanism of mecom in regulating pronephron segmentation and epithelial fate decision remains intriguing. In mammals, the Mecom gene is characterized as one of the splice variants of the gene Evi1, a conserved zinc finger transcription factor crucial for both development and oncogenesis (Wieser, 2007). The Mecom transcript consists of extra sequence from the upstream Mds1 gene, resulting in a 188 amino acid extension encoding a PR domain homologous to the SET protein methyltransferase domain (Wieser, 2007). The SET domain-containing proteins are a family of epigenetic regulators responsible for most histone lysine methylation (Dillon, et al., 2005). A genome-wide survey of zebrafish SET domain-containing gene revealed mecom (referred to as prdm3) as a member of the SET domain-containing proteins (Sun, et al., 2008). It is not known, however, whether the PR domain encoded by the zebrafish mecom sequence possesses histone methyltransferase activity.

The presence of zinc finger domains and their potentials to bind specific DNA sequences suggests mecom may function as a transcription factor to directly modify gene activities. However, target genes of mecom responsible for kidney development have yet to be identified. The presence of several consensus binding sites of co-activators and co-repressors in the Mecom protein argue additional biochemical role of mecom as a transcription cofactor (Wieser, 2007). Indeed, mecom has been shown to participate in chromatin modification by interacting with CtBP corepressor or histone deacetylases (HDACs) (Wieser, 2007). In Drosophila, the mecom homologous gene hamlet controls olfactory receptor neuron diversification via locus-specific histone methylation at the Notch target promoter (Endo, et al., 2012). In vertebrates, the highly related mecom variant evi1 has been shown to repress BMP/Smad-mediated activation of endogenous genes required for cell fate specification. In this case, Evi1 and CtBP are recruited to target gene promoter upon TGF? stimulation, leading to decreased histone acetylation and transcription (Alliston, et al., 2005). Notably, chromatin modification seems to be crucial for proper nephron development, as inhibition of HDAC in zebrafish embryos results in the expansion of the renal progenitor cell population (de Groh, et al., 2010).

Several studies suggest that mammalian Mecom affects cellular proliferation and differentiation in cell-type specific manners (Wieser, 2007). Therefore, in the developing pronephros, one possibility that we have not yet explored is that mecom fulfills its function by interfering with renal progenitor proliferation and/or differentiation. Further, the persistence of mecom transcripts in the DL and PD at 24 hpf could serve later roles in nephron morphogenesis—which might partly explain renal failure in mecom morphants. Recent studies have reported elevated epithelial proliferation in the distal tubule and duct at 3-4 days post fertilization, which serves as a compensatory mechanism allowing for complex morphological change in the proximal segments driven by collective cell migration toward the glomerulus (Vasilyev, et al., 2009; Vasilyev, et al., 2012). Thus, persistent mecom expression in the DL segment and PD at 24 hpf may account for active cell proliferation in these regions and in turn facilitate nephron morphogenesis post segmentation events. Interestingly, the overexpression of Mecom in Xenopus repressed proximal fates and glomerulus formation, while inactivation of Mecom by fusing the wild type protein with the VP16 activation domain disrupted pronephros duct development (van Campenhout, et al., 2006). In light of our findings, Mecom activity may be a conserved requirement in distal nephron development, though additional studies are needed to explore this possibility.

The elucidation of new roles that RA plays during zebrafish nephrogenesis

RA is a key modulator of pronephros development in the zebrafish (Gerlach and Wingert, 2012). RA signaling is required between gastrulation and early somitogenesis to induce podocytes and proximal segmental fates and to inhibit expansion of distal segments (Wingert, et al., 2007). RA generated in the anterior paraxial mesoderm (PM) is hypothesized to diffuse to the adjacent renal progenitors that give rise to the pronephros (Wingert, et al., 2007; Wingert and Davidson, 2011). Since the rostral-most PM produces RA first, the current model is that a gradient of RA is generated along the IM, which provides higher levels of RA in the rostral region that induce proximal fates, while low RA levels in the caudal region allow distal segment formation (Wingert, et al., 2007; Wingert and Davidson, 2011). Furthermore, growing evidence suggests transcription factors essential for nephron development act downstream of RA during pronephric segmentation patterning. For example, the homeodomain transcription factor irx3b has been shown to modulate DE formation and PCT/PST boundary establishment downstream of RA (Wingert and Davidson, 2011).

Several studies support the existence of a regulatory relationship between RA and Mecom. In mammalian cell culture studies, all-trans RA was shown to activate the Mecom locus via a consensus RARE located in exon 1a of the human Evi1 gene (Bingemann, et al., 2009). Meanwhile, RA also induced mecom expression indirectly through unknown mechanisms (Bingemann, et al., 2009). In the Xenopus pronephros, exogenous RA is associated with elevated Mecom transcript levels (van Campenhout, et al., 2006). In contrast, our analysis reveals that RA signaling inhibits mecom expression in the zebrafish pronephros. Since mecom is expressed in zebrafish pronephros renal progenitors starting from the onset of nephrogenesis, mecom could be a direct or indirect target of RA signaling. A better understanding of the relationship between RA and mecom will help to elucidate how these factors and/or their targets interact to precisely define the pronephros segment pattern.

The link between RA signaling and MCC development has not been previously reported. In the zebrafish pronephros, the majority of pronephric epithelial cells bear single apical cilium, while MCCs represent a subset of epithelial cells that display up to 16 apical motile cilia. It has been demonstrated that Notch signaling regulates differentiation of transporting epithelia and MCCs in the zebrafish pronephros via lateral inhibition, where ‘multiciliated progenitor cells’ suppress transporting epithelia to acquire MCC fates possibly by expressing high levels of jag2, which interacts locally with Notch3 receptors present on transporting epithelial surface (Ma and Jiang, 2007; Liu, et al., 2006). Our data reveals that MCC fate choice is titrated through a concert of RA, mecom, and finally Notch signaling. Further experiments are needed to ascertain how mecom promotes Notch activity in renal progenitors. However, our work here adds several levels of complexity to the knowledge about how the MCC-transporting epithelia balance is regulated.

In conclusion, our studies have revealed new insights into the processes of nephron segmentation and cell fate choice during nephrogenesis. Given the similarities that exist between nephron segmental structure and ontogeny between zebrafish and other vertebrates, future analysis of Mecom function during mammalian renal development may provide novel information about nephron patterning in humans that could be relevant to understanding renal stem cell biology, kidney birth defects and other kidney diseases.

Supplementary Material

Supplemental Fig. 1

Expression of mecom transcripts in the renal progenitor field at early somitogenesis in wild type emrbyos. (A) WISH using mecom202-3’ probe (left column), mecom202-5’ probe (middle column), or the combination of mecom202-specific probes (right column). Forming somites were labeled by dlc (red). (B) At the 15 somite stage, mecom202 (purple) was expressed in a caudal domain of the renal progenitor field.

Supplemental Fig. 10

The expansion of MCCs in RA-treated embryos can be rescued by NICD overexpression. MCC development comparison between wild type, heat shocked non- transgenic NICD siblings treated with RA, and heat shocked transgenic NICD embryos treated with RA. Notch activation in transgenic NICD was sufficient to reduce MCCs in the context of RA treatment. MCCs were labeled by odf3b (purple).

Supplemental Fig. 2

Expression of mecom transcripts during nephron segmentation establishment. WISH with mecom202-specific probes (purple) indicates that the mecom- expressing domain displayed a progressive shift toward the caudal end of the developing pronephros. Forming somites were marked by smyhc1 (red).

Supplemental Fig. 3

Comparison of mecom knockdown phenotypes when performed with individual splice morpholinos and detection of mispliced mecom intronic sequence. Wild type embryos were injected with mecom splicing donor morpholino (e1SD, middle panels) or splicing acceptor morpholino (e1SA, bottom panels) at the 1-cell stage and WISH was performed at 24 hpf to examine pronephros segmentation changes. Renal genes (purple) and forming somites were labeled by smyhc1 (red) (A) Expression domains of proximal markers slc4a4a (PCT, PST), slc20a1a (PCT), and trpm7 (PST) revealed mecom morpholino e1SD or e1SA alone induced a PST expansion, but not a change in the PCT. (B) mecom morpholino e1SD or e1SA also led to a reduced DL as shown by distal markers clcnk (DE, DL, PD), slc12a1 (DE) and slc12a3 (DL). (C) Co-injection of the mecom morpholinos results in mis-spliced mecom mRNA. In this example, primers designed to amplify a region of intron 3-4 located between exon 3 and 4 (left) amplify this sequence only in mecom morphant cDNA pools, but not in wild type embryos (right) or reverse transcriptase negative controls.

Supplemental Fig. 4

Analysis of blood circulation in mecom morphants. Benzidene staining was used to evaluate erythrocyte circulation and compare vascular integrity between wild type and mecom morphant embryos. (A) Wild type embryos showed circulation confined to the vasculature and no blood pooling as the same time points (top row; 51 wild types surveyed), whereas some mecom morphants were similarly normal (middle row) but others displayed blood pooling in the head (bottom row). (B) Quantification of blood pooling phenotype at 36 and 48 hpf in mecom morphants.

Supplemental Fig. 5

Bar charts quantifying the percentage of segmentation deficiencies in mecom morphants or embryos co-injected with mecom morpholino and cRNA in Figure 3 and 4. Proximal segmentation defect (A-C) was characterized by the 3-somite extension in slc4a4a or trpm7 expressing domains. Intermediate segmental defects of slc4a4a and trpm7 expressions represented expansions of segmentation markers less than 3 somites. Distal segmentation defect (D-F) was indicated by the 3-somite reduction in clcnk or slc12a3 expressing segments. Intermediate segmental defects of clcnk and slc12a3 expressions were identified by reduced expressions of segmentation markers less than 3 somites. For each quantification, at least 10 embryos were examined.

Supplemental Fig. 6

mecom knockdown does not alter podocyte formation or interrenal cell fate. WISH with wt1b (purple) and nr5a1a (red) at 24 hpf shows that podocyte formation and interrenal organ development were not affected by mecom knockdown.

Supplemental Fig. 7

Quantification of mecom knockdown phenotypes in epistasis studies with RA, DEAB, and Notch. The number of embryos examined for each study is provided. In the cases where an embryo did not show the phenotype reported in Figure 5, 6, or 7, this judgment was based on over or under-stained WISH signal which prevented us from properly evaluating that individual embryo.

Supplemental Fig. 8

Individual mecom splice intereference morpholinos lead to expanded MCC numbers. Splicing morpholinos of mecom, e1SD (middle panel) and e1SA (bottom panel), induced an expanded MCC domain. MCCs were labeled by odf3b (purple).

Supplemental Fig. 9

RA-deficient aldh1a2 (lib) mutants exhibit a reduction in MCC formation. MCCs were labeled by odf3b (purple) in aldh1a2−/− and wild type siblings.

Highlights.

• -zebrafish nephrons are divided into segments by the interplay of RA and mecom
• -mecom is essential to form the distal segment and inhibit the proximal segment
• -the balance of RA, mecom and Notch controls nephron multiciliated cell (MCC) fate
• -RA negatively regulates mecom in renal progenitors to promote MCC formation
• -mecom attenuates MCC number by positively regulating Notch activity downstream of RA

Abbreviations

CS

corpuscle of Stannius

DE

distal early

DEAB

4-diethylamino-benzaldehyde

DL

distal late

dpf

days post fertilization

G

glomerulus

hpf

hours post fertilization

mecom

mds1/evi1 complex

IM

intermediate mesoderm

MCC

multiciliated cell

MET

mesenchymal to epithelial transition

N

neck

NICD

Notch intracellular domain

PCT

proximal convoluted tubule

PD

pronephric duct

PM

paraxial mesoderm

PST

proximal straight tubule

RA

retinoic acid

RARE

retinoic acid response element