Nephron proximal tubule patterning and corpuscles of Stannius formation are regulated by the sim1a transcription factor and retinoic acid in zebrafish

Abstract

The mechanisms that establish nephron segments are poorly understood. The zebrafish embryonic kidney, or pronephros, is a simplified yet conserved genetic model to study this renal development process because its nephrons contain segments akin to other vertebrates, including the proximal convoluted and straight tubules (PCT, PST). The zebrafish pronephros is also associated with the corpuscles of Stannius (CS), endocrine glands that regulate calcium and phosphate homeostasis, but whose ontogeny from renal progenitors is largely mysterious. Initial patterning of zebrafish renal progenitors in the intermediate mesoderm (IM) involves the formation of rostral and caudal domains, the former being reliant on retinoic acid (RA) signaling, and the latter being repressed by elevated RA levels. Here, using expression profiling to gain new insights into nephrogenesis, we discovered that the gene single minded family bHLH transcription factor 1a (sim1a) is dynamically expressed in the renal progenitors—first marking the caudal domain, then becoming restricted to the proximal segments, and finally exhibiting specific CS expression. In loss of function studies, sim1a knockdown expanded the PCT and abrogated both the PST and CS populations. Conversely, overexpression of sim1a modestly expanded the PST and CS, while it reduced the PCT. These results show that sim1a activity is necessary and partially sufficient to induce PST and CS fates, and suggest that sim1a may inhibit PCT fate and/or negotiate the PCT/PST boundary. Interestingly, the sim1a expression domain in renal progenitors is responsive to altered levels of RA, suggesting that RA regulates sim1a, directly or indirectly, during nephrogenesis. sim1a deficient embryos treated with exogenous RA formed nephrons that were predominantly composed of PCT segments, but lacked the enlarged PST observed in RA treated wild-types, indicating that RA is not sufficient to rescue the PST in the absence of sim1a expression. Alternately, when sim1a knockdowns were exposed to the RA inhibitor diethylaminobenzaldehyde (DEAB), the CS was abrogated rather than expanded as seen in DEAB treated wild-types, revealing that CS formation in the absence of sim1a cannot be rescued by RA biosynthesis abrogation. Taken together, these data reveal previously unappreciated roles for sim1a in zebrafish pronephric proximal tubule and CS patterning, and are consistent with the model that sim1a acts downstream of RA to mitigate the formation of these lineages. These findings provide new insights into the genetic pathways that direct nephron development, and may have implications for understanding renal birth defects and kidney reprogramming.

INTRODUCTION

Organogenesis of the vertebrate kidney involves the formation of up to three distinct structures that develop in succession from the renal progenitors that emerge from the intermediate mesoderm (IM) (Saxen, 1987; McCampbell and Wingert, 2012; Romagnani, et al., 2013). Across species, each kidney functions to various degrees in the regulation of waste excretion, fluid balance, and osmolarity. The pronephros is the first embryonic kidney to arise. While lower vertebrates (fish and amphibians) use the pronephros to perform vital excretory tasks, it is a transient, vestigial organ in higher vertebrates (reptiles, birds, and mammals). The mesonephros is the second vertebrate kidney structure that forms concomitant with the degeneration and/or remodeling of pronephric tissues, and performs excretory roles during embryonic life. In higher vertebrates, the mesonephros degenerates upon the formation of the third and final kidney, known as the metanephros, which functions throughout adulthood. Interestingly, fish and other lower vertebrates utilize their mesonephros during adulthood and never develop a metanephric kidney (Gerlach, et al., 2011).

Despite these developmental differences, vertebrate kidney structures are all comprised of tubular functional units known as nephrons (Cheng and Wingert, 2014). Nephrons connect to the vascular system at their proximal end by surrounding a small bundle of capillaries, while the opposing distal end links to collecting ducts that drain waste into the urinary system. Nephrons are regionalized along their proximo-distal length with segments of discrete segment populations of epithelial cells that are specialized to perform precise modifications of fluid as it transits through the tubule (Reilly, et al., 2007). Recent research has demonstrated that the segmental nature of the nephrons is fundamentally conserved across pro-, meso- and metanephric nephrons in humans and popular animal research models including the zebrafish, frog and mouse (Wingert and Davidson, 2008; Wessely and Tran, 2011; Kroeger and Wingert, 2014). At present, however, there is only a rudimentary understanding of the mechanisms that regulate renal progenitor patterning into the specific segment identities found within the nephron (Kopan, et al., 2007; Schedl, 2007).

The zebrafish pronephros, in particular, is an excellent model for studying segmental patterning during nephrogenesis (Gerlach and Wingert, 2013). The pronephros is anatomically simple, as it consists of two nephrons (Drummond, et al., 1998). These pronephric nephrons emerge from bilateral fields of renal progenitors that derive from the IM during somitogenesis, and then undergo a mesenchymal to epithelial transition (MET) to form tubular structures (Gerlach and Wingert, 2014). Each pronephric nephron contains the following segments: a blood filter comprised of podocytes (P), neck (N), proximal convoluted and straight tubules (PCT, PST), distal early and late tubules (DE, DL), and a pronephric duct (PD) (Figure 1A) (Wingert, et al., 2007). In addition, a small subset of cells situated in each renal progenitor field within the domain occupied by DL precursors will become the corpuscles of Stannius (CS), endocrine organs that regulate calcium and phosphorus in teleosts (Camp, et al., 2003; Elizondo, et al., 2005; Wingert, et al., 2007). After the CS precursors emerge from the renal progenitor field, they undergo morphogenesis events that situate them into bilateral lobes that are located dorsal to the distal tubules (Camp, et al., 2003; Elizondo, et al., 2005; Wingert, et al., 2007). The marked conservation of nephron segmentation between zebrafish and other vertebrates, in combination with the structural simplicity inherent to a two-nephron kidney, makes the zebrafish both a relevant and feasible experimental system for conducting genetic interrogations to discover nephrogenesis mechanisms.

Figure 1. sim1a expression is dynamic during nephrogenesis.

(A) Schematics depict the IM and PM domains in a young zebrafish embryo. At 24 hpf, the zebrafish pronephros consists of two segmented nephrons as indicated by the colored regions. Corresponding numbers indicate the boundaries of each segment respective to the somites. (B) Illustrations denote the developmental time course of sim1a expression, which first appears within the IM (pax2a+ domain) at the 2 ss. Dark purple indicates strong expression, while light purple signifies weak expression. (C) Flat mounted embryos staged at young developmental time points were assayed by WISH for the kidney markers pax2a, sim1a, or mecom (purple). The somites and proximal renal progenitors were labeled by dlc or myod1 (red). Solid lines (purple and red) demarcate renal progenitors regions where high transcript expression was observed. The dotted line (purple) indicates a weak expression domain, with line color corresponding to the gene identity indicated in figure. (D) Embryos at older stages were analyzed by WISH for sim1a expression (purple). Embryo anterior is located to the left in all panels. Abbreviations: CS – corpuscles of Stannius; DE – distal early; DL – distal late; hpf – hours post fertilization; IM – intermediate mesoderm; N – neck; P – podocytes; PCT – proximal convoluted tubule; PD – pronephric duct; PM – paraxial mesoderm; PST – proximal straight tubule; ss – somite stage; WISH – whole mount in situ hybridization.

To date, several major events that direct renal progenitor patterning during zebrafish pronephros formation have been identified. Recently, we found that renal progenitors undergo a regionalization that divides them into rostral and caudal domains, and that the materialization of this rostral domain requires a local source of retinoic acid (RA) secreted by the adjacent paraxial mesoderm (PM), which forms the embryonic somites (Wingert and Davidson 2011; Li, et al., 2014). Interestingly, RA positively regulates the expression domains of rostral domain markers, such as the transcription factor wt1a (Wingert, et al., 2007; Bollig, et al., 2009), which is required for podocyte development (Perner, et al., 2007; O’Brien, et al., 2011). In addition, RA negatively regulates the transcription factor mecom, a caudal domain gene that is required for PST and DL development (Li, et al., 2014). Further subdivisions, demarcated by the nested expression of other transcription factors and other genes, precede the emergence of discrete segments by 24 hours post fertilization (hpf), which corresponds to approximately the 28 somite stage (ss) (Wingert and Davidson, 2011; McKee, et al., 2014). However, while numerous transcription factors have been mapped to the emerging renal progenitor domains (Wingert and Davidson, 2011), the functional roles of most remain an enigma.

Among these, previous studies have documented the expression of the transcription factor sim1a in the zebrafish pronephric renal progenitors (Serluca and Fishman, 2001). sim1a encodes a basic helix-loop-helix and Period-Arnt-Sim (bHLH-PAS) transcription factor that is homologous to the Drosophila single-minded (Sim) gene, a master regulator of midline cell development in the central nervous system (Linne, et al., 2012). In zebrafish, sim1a is likewise expressed in the developing central nervous system (Wen, et al., 2002) where it is requisite for the formation of dopaminergic neurons from neural progenitors (Borodovsky, et al., 2009; Mahler, et al., 2010; Wolf and Ryu, et al., 2013), the creation of the neuroendocrine system (Eaton and Glasgow, 2006; Eaton, et al., 2008; Löhr, et al., 2009), as well as axon guidance (Schweitzer, et al., 2013). Despite the assignation of these various sim1a functions in the nervous system across invertebrate and vertebrate species, the role of sim1a has not yet been explored during kidney establishment.

Here, we used a combination of expression and functional studies of sim1a to gain new insights into nephrogenesis, and elucidated essential roles of sim1a both in pronephros segmentation and CS development. Using whole mount in situ hybridization (WISH) to profile renal progenitor gene expression, we discovered that sim1a expression is highly dynamic during nephron construction. sim1a is one initial marker of the renal progenitor caudal domain, and that its expression later is maintained in both proximal tubule segments before becoming restricted to the CS. Since these findings suggested that sim1a might contribute to segment patterning and CS formation, we performed loss and gain of function studies to explore the role(s) of sim1a in renal ontogeny. sim1a deficiency caused an expansion of the PCT, which was minimally functional as indicated by a dextran-FITC uptake assay, and an abrogation of the PST and CS populations. However, the domains of both the DE and DL segments remained unchanged. These results suggest that sim1a activity is necessary to pattern the PST and CS, and that sim1a may negotiate the PCT/PST boundary. Consistent with these findings, sim1a overexpression was sufficient to some extent in promoting the formation of the PST and CS populations at the expense of the PCT. Further, we evaluated the relationship between sim1a and RA, and discovered that sim1a expression is reliant on RA levels. We found that elevations in RA were not sufficient to rescue PST formation in sim1a morphants, and that the abrogation of RA synthesis using 4-diethylaminobenzadehyde (DEAB) was not sufficient to rescue CS formation in sim1a morphants—indicating that RA cannot substitute for normal sim1a function in PST and CS development. Taken together, these data are consistent with the hypothesis that sim1a functions downstream of RA during renal progenitor patterning. In sum, these studies show for the first time that sim1a is essential for several aspects of nephron segmentation, and establish that sim1a is an essential component of CS formation.

MATERIALS AND METHODS

Zebrafish husbandry and ethics statement

Zebrafish were cared for and maintained in the Center for Zebrafish Research at the University of Notre Dame, with experimental procedures approved under protocols 13-021 and 16-025. Wild-type embryos of the Tübingen strain were raised and staged as described (Kimmel, et al., 1995). Embryos were incubated at 28 °C, anesthetized with 0.02% tricaine, and then fixed for analysis using 4% paraformaldehyde (PFA)/1× PBS before being stored in methanol at −20 °C.

WISH, flat mount preparation, and imaging

Whole mount in situ hybridization (WISH) was performed on zebrafish embryos as described and then imaged (Galloway, et al., 2008; Lengerke, et al., 2011; Cheng, et al., 2014). Antisense RNA probes were either digoxigenin-labeled (clcnk, mecom, pax2a, sim1a, slc12a1, slc12a3, slc20a1a, slc4a4, stc1, trpm7, and wt1b) or fluorescein-labeled (dlc, myod1, and smyhc1) and were synthesized utilizing IMAGE clone template plasmids for in vitro transcription as formerly described (Wingert, et al., 2007; Wingert and Davidson, 2011). Embryos younger than the 18 ss were flat mounted for better visualization of gene expression patterns as described (Cheng, et al., 2014). The domains of gene expression were reported with respect to somite boundaries. Representative results for studies assayed by WISH were based on the analysis and counts of typically >20 embryos, and percentages of phenotypes observed were quantified for documentation and comparison of phenotypes between control and experimental groups. Photographs were taken using a Nikon eclipse Ni with a DS-Fi2 camera. Images were processed using Adobe Photoshop CS5.

Morpholino knockdown

Antisense morpholino oligonucleotides (MOs) were synthesized (Gene Tools, LLC) for gene knockdown procedures. MOs were suspended in DNase/RNase free water to make stock concentrations of 4 mM, and each stock was stored at −20 °C. The sim1a ATG MO (5’ – TCGACTTCTCCTTCATGCTCTACGG – 3’) (Eaton and Glasgow, 2006) provided fairly penetrant effects and was therefore utilized for the majority of loss of function experiments. Of note, embryos injected with the sim1a ATG MO are herein referred to as sim1a morphants unless otherwise stated. Additionally, as a specificity control for the sim1a ATG MO, a five-base ATG mismatch morpholino (mmMO) was designed where lowercase nucleotides indicate altered bases (5’-TCcACaTCTCCaTCATGCTgTAgGG-3’). Additionally, a previously described sim1a splice MO targeting the exon 2 and intron 2 boundary (5’-TGTGATTGTGTACCTGAAGCAGATG-3’) was used as an independent MO control to further verify the morphant phenotypes obtained during sim1a ATG MO studies (Löhr, et al., 2009). Approximately 1 nl of sim1a ATG MO (1:30), mmMO (1:30), and splice MO (1:10) was injected into wild-type embryos at the one-cell stage of development and then embryos were incubated to the desired time point(s) for observation and phenotypic analyses. For p53 knockdowns, we utilized the morpholino (5’-AGAATTGATTTTGCCGACCTCCTCT-3’) (Rentzsch, et al., 2003), which has been implemented in numerous studies. Approximately 7 ng of the p53 MO was injected into wild-type embryos at the one-cell stage. For the concomitant knockdown of both p53 and sim1a, 7 ng of the p53 MO was co-injected with the sim1a ATG MO (Löhr, et al., 2009). For knockdown studies, at least 20 embryos were evaluated per renal marker, though cohorts typically consisted of 25-40 embryo samples.

Acridine orange staining

Acridine orange (AO), a fluorescent cell-permeable derivative, was used to label apoptotic cells in zebrafish wild-type or morphant embryos (Hammerschmidt, et al., 1996; van Ham, et al., 2010; Westerfield, 2000). In brief, a 1:50 AO solution was made from a 50× AO stock (250 μg/ml) diluted in 0.003% 1-phenyl-2-thiourea (PTU)/E3 embryo media. At the desired time points, embryos were incubated in the AO solution for 1 hour in the dark, rinsed three times with 0.003% PTU, and then imaged under the GFP channel in 2% methylcellulose/0.02% tricaine.

Dextran injection

To assess renal function, clearance assays using the fluorescent 40 kDa dextran-fluorescein (FITC) (Invitrogen) were performed on wild-type controls and sim1a morphants. Each experimental group was treated with 0.003% PTU between 24-30 hpf prior to being anesthetized with 0.02% tricaine for the dextran-FITC injection at 38 hpf. Embryos were then revived and allowed to develop in 0.003% PTU at 28 °C in the dark. Live fluorescent imaging of embryos in a solution comprised of 2% methylcellulose/0.02% tricaine was utilized to visualize dextran clearance at later developmental stages.

Cardiovascular function and statistical analysis

o-dianisidine staining was performed as described (Wingert, et al., 2004, 2005; Dooley, et al., 2008; Fraenkel, et al., 2009). At 38, 48, and 72 hpf, the number of heartbeats per minute were measured in wild-type embryos and sim1a morphants. Between 13-15 embryos were assessed for each time point. Two-tailed Student’s t-tests were used to calculate significance. Embryos at 38 and 48 hpf were treated with 0.003% PTU and then anesthetized by 0.02% tricaine prior to heartbeat measurements.

sim1a subcloning, cRNA synthesis, and microinjections

A pUC57 plasmid containing the sim1a transcript open reading frame was obtained from GenScript USA Inc. (Piscataway Township, NJ). Primers for sim1a were designed where the forward primer for the sim1a ATG site contained a 4 bp anchor (italicized) followed by the EcoRI sequence (underlined) and Kozak consensus sequence (lowercase). Alternately, the reverse primer contained a 10 bp anchor (italicized) and the sim1a stop sequence. The 60mer sim1a forward primer was: 5’ – GACTGAATTCgccgccaccATGAAGGAGAAGTCGAAAAACGCGGGGCGCACGCGGCGGGA – 3’. The 60mer sim1a reverse PCR primer was as follows: 5’ – TGACCTCGAGTCAGCTGCCATTGGTGATGATGACGGAGGTGCCTTTATGGCCCGGCGCTT – 3’. The Expand PCR kit (Roche) was utilized in conjunction with these 60mer primers to obtain the sim1a insert. The pGEM-T Easy kit (Promega) was used to perform ligation reactions to insert sim1a transcript into the PCS2 expression vector into the EcoRI and XhoI sites. One Shot TOP10 competent cells (Invitrogen) were transformed with the sim1a/PCS2+ plasmid and sequenced. Wild-type full-length sim1a capped RNA (cRNA) was synthesized with the mMESSAGE mMACHINE SP6 Transcription kit (Ambion). For sim1a overexpression experiments, approximately 1 nl of sim1a cRNA (200 pg) was injected into wild-type embryos at the one-cell stage. Alternately, for rescue studies in sim1a deficient embryos, approximately 1 nl of a dilution containing the sim1a ATG or splice MO and cRNA dosages between 36 and 600 pg of sim1a cRNA were injected. The sim1a ATG MO and splice MO were not predicted to bind sim1a cRNA, due to only partial sequence overlaps of 17/15 bases and 12/25 bases, respectively.

sim1a overexpression quantification

Photographs were taken at a 10× magnification using a Nikon eclipse Ni with a DS-Fi2 camera. The 2 points line and polygon measurement tools in the Nikon imaging software were utilized to quantify (1) length - slc20a1a, trpm7 and (2) overall area and individual cell size - stc1, respectively. Significance was calculated by two-tailed Student’s t-tests. Images were processed in Adobe Photoshop CS5.

Chemical treatments

All-trans RA and DEAB (Sigma) were dissolved in 100% DMSO to make 1 M stocks as previously indicated (Wingert, et al., 2007; Li et al., 2014). Either a low (1×10⁻⁸ M, denoted as + RA) or high (1×10⁻⁷ M, designated as ++ RA) dose of RA/DMSO diluted in E3 was used during these chemical experiments. For sim1a expression studies, wild-type embryos were treated with + RA from 50-60% epiboly (epi) to the 5-7 somite stage, at which point embryos were washed three times with E3 and then fixed at desired developmental stages. In additional studies, wild-type embryos and sim1a morphants were either incubated with + RA from 90% epi to the 15 ss or with ++ RA from 90% epi to the 5-7 ss. For DEAB treatments, 1.65×10⁻⁵ M DEAB/DMSO diluted in E3 was applied to wild-type embryos and sim1a morphants from 75% epi to the 5-7 ss. Control embryos were incubated with corresponding concentrations of DMSO, and all embryos were kept in the dark during chemical treatments.

RESULTS

sim1a demarcates the caudal subdomain in the renal progenitor field, and is dynamic during nephrogenesis and the formation of the associated CS

At the onset of zebrafish pronephric development, pax2a delineates the entire renal progenitor field that arises from the IM adjacent to the PM in the embryonic trunk (Fig. 1A, B) (Pfeffer, et al., 1998; Gerlach and Wingert, 2013). Furthermore, dlc is a Notch ligand that has been documented to mark both the somites that arise from the PM and the rostral renal progenitors, which will ultimately form the proximal segments of the pronephros (Fig. 1A, C) (Wingert, et al, 2007; Wingert and Davidson, 2011; Li, et al., 2014). The renal progenitors undergo a MET by the 20-22 somite stage (ss) that gives rise to two epithelial tubules (Gerlach and Wingert, 2014) which become joined at their proximal ends by connecting to a single glomerulus at the midline and link distally to the cloaca, forming a common exit portal for urine (Fig. 1A) (Drummond, et al., 1998). By 24 hpf, these epithelial tubules, termed nephrons, begin to terminally differentiate into structures comprised of highly specialized proximal and distal segments that can be distinguished from neighboring populations based on the unique expression signature of various solute transporter or transcription factor genes as follows: P – wt1b; N – pax2a; PCT – slc20a1a; PST – trpm7; DE – slc12a1; DL – slc12a3; PD –gata3 (Fig. 1A) (Wingert, et al., 2007). Further, as the CS emerges from the DL progenitor domain and comes to occupy a position dorsal to this segment, the CS cells specifically express the glycoprotein hormone encoded by stanniocalcin 1 (stc1) (Camp, et al., 2003; Elizondo, et al., 2005; Wingert, et al., 2007; Elizondo, et al, 2010).

While the mechanisms directing the intricate process of nephrogenesis remain largely unknown, recent research has suggested that the onset of nephron segmentation patterning occurs at the 3 ss where the renal progenitors can be divided into distinct rostral and caudal domains as indicated by dlc and mecom transcripts respectively (Fig. 1C) (Li, et al., 2014). Since sim1a has been documented in the developing zebrafish kidney (Serluca and Fishman, 2001), we first sought to further define the spatiotemporal expression of sim1a with respect to these rostral/caudal renal progenitor domains during early somitogenesis and throughout the time of nephron establishment.

Whole mount in situ hybridization (WISH) followed by flat mounting was used to label and then visualize sim1a-expressing cells in the renal progenitors of wild-type embryos between the tailbud stage and 28 ss (Fig. 1). sim1a was exclusively expressed in the caudal domain at the 2 ss, being situated adjacent to the newly forming somite and extending down the trunk, but did not encompass the entire pax2a+ field that surrounds the tailbud (Fig. 1C). As somitogenesis progressed, the sim1a domain was dynamically altered in renal progenitors. Typically, the evolving caudal boundary of the sim1a expression domain was located adjacent to the newly forming somite, while the evolving rostral boundary of sim1a remained mutually exclusive of the rostral dlc-expressing region within the renal progenitor field (Fig. 1C). Specifically, during the 3 and 5 ss, sim1a expression began to extend more caudally and was down-regulated in rostral renal progenitors, eventually encompassing a broad domain that spanned an area located adjacent to somites 5-13 by the 14 ss (Fig. 1B, C). While sim1a was highly expressed in this region at the 14 ss, faint levels of sim1a were detectable in caudal renal progenitors (Fig. 1B, C). At the 20-22 ss, sim1a was localized to the proximal regions of the pronephros next to somites 5-11, which corresponds to the domain where the PCT and PST segments emerge (Fig. 1A, B, D, Fig. S1). However, at the 24 ss, sim1a expression was also found in the CS precursors that are located next to somite 15 (Fig. S1). By the 28 ss, sim1a transcripts were restricted to the CS cells and were no longer detectable in proximal tubule segments (Fig. 1D, Fig. S1). Overall, these data show that a discrete subdivision of the renal progenitor domain is present as early as the 2 ss, and reveal that the dynamic sim1a expression domain shows a striking correlation with the emergent location of the PCT and PST, followed by the CS during nephrogenesis.

sim1a morphants have defective renal clearance associated with reduced heart rate

To interrogate the functional role(s) of sim1a in pronephros ontogeny, we performed morpholino (MO) knockdown studies using two previously published MOs to independently target sim1a protein translation or transcript splicing (Eaton and Glasgow, 2006; Löhr, et al., 2009). A developmental time course of live sim1a morphants was examined to observe the morphological changes associated with the loss of sim1a mediated by MO knockdown, which revealed developmental defects consistent with published observations. Darkening within the head region of sim1a morphants was typically observed during early embryonic development through 24 hpf (Fig. 2A). At 48 hpf, sim1a morphants were characterized by possessing small head and eye phenotypes, which persisted throughout development as documented (Eaton, et al., 2008) (Fig. 2A). Furthermore, assessment of cell death at these time points using acridine orange staining showed that sim1a morphants exhibited higher levels of apoptosis compared to control embryos, including throughout the central nervous system, which was expected based on the roles of sim1a in neurogenesis (Borodovsky, et al., 2009; Löhr, et al., 2009) (Fig. S2A). Additionally, acridine orange positive cells were also detected in sim1a morphants throughout other tissues in the trunk (Fig. S2A). This most likely suggests that non-specific cell death had occurred, which is associated with off-target effects of MOs in zebrafish (Nasevicius and Ekker, 2000; Robu, et al., 2007). Even so, it is also possible that the loss of sim1a expression led indirectly to disruptions of development in other tissues. The formation of severe pericardial edema became apparent in sim1a morphants at approximately 72 hpf (data not shown) and continued through 120 hpf (Fig. 2A). These distinguishing characteristics were not present in wild-type controls or embryos injected with the sim1a ATG mismatch morpholino (Fig. S2A, B).

Figure 2. sim1a morphants display drastic developmental defects and abnormal kidney function.

(A) A live time course shows the typical morphological phenotypes exhibited by sim1a morphants at different time points – open arrowhead demarcates darkened regions within the head; solid arrowhead indicates small head and eye phenotypes; solid arrow marks the presence of edema. (B) Kidney function in sim1a morphants was analyzed by a 40 kDa dextran-FITC uptake assay where injections were performed at approximately 38 hpf. sim1a morphants displayed partial PCT reabsorption (dorsal inset view) and reduced renal clearance of dextran-FITC over time, with persistent fluorescence throughout the head, trunk and pericardium when compared to wild-types. White lines indicate either the PCT domain in relation to the somites or the PCT morphology at given time points. Embryo anterior is to the left. Abbreviations: hpf – hours post fertilization; hpi – hours post injection; PCT – proximal convoluted tubule.

Since the appearance of edema is suggestive of cardiac and/or renal system failure (Li, et al., 2014), kidney function was assessed in sim1a morphants by performing a dextran-FITC excretion assay. Under normal circumstances, the kidney tubules collect dextran from the vascular system during blood filtration, where it can undergo endocytosis specifically in the PCT and can also be cleared through urinary production such that fluorescence within the embryo diminishes over time (Anzenberger, et al., 2006; Li, et al., 2014). Wild-type embryos injected with 40 kilodalton (kDa) dextran-FITC at approximately 38 hpf showed subsequent PCT uptake at 24 hours post injection (hpi) through 72 hpi (Fig. 2B). Interestingly, while sim1a morphants similarly injected with 40 kDa dextran-FITC were able to uptake the dextran by 24 hpi, they were unable to completely clear this substance from their systems as indicated by the excessive levels of fluorescence at 72 hpi (Fig. 2B). Of further note, the PCT segment undergoes coiling morphogenesis between the 48-120 hpf (Wingert, et al., 2007), which proceeded normally in dextran-FITC injected wild-types (Fig. 2B). However, at 72 hpi, sim1a morphants lacked the distinctive coiling of the PCT, indicating disrupted morphogenesis within this particular region of the nephron (Fig. 2B). As previously mentioned, pericardial edema was detected at 24 hpi (~60 hpf) and became more drastic by 72 hpi (~110 hpf) in sim1a morphants (Fig. 2B).

We also assessed the integrity of the circulatory system in sim1a morphants, since compromised vascular infrastructure can result in hematomas, which is when blood pools outside vessels in the body, contributing to the dysregulation of fluid flow that culminates with edema. To detect whether such blood pooling occurred in sim1a morphants, o-dianisidine staining was performed, as this technique provides a sensitive assessment to localize embryonic blood that is not apparent using light microscopy alone (Wingert, et al., 2004, 2005; Dooley, et al., 2008; Fraenkel, et al., 2009). Wild-type controls and sim1a morphants were examined between 38-96 hpf, and no evidence of blood pooling was found in either group, suggesting that the vascular integrity within sim1a morphants was normal (data not shown). Next, since edema can result from alterations in cardiac development, heart morphology and rate were compared between wild-types and sim1a morphants at 38, 48, and 72 hpf. While heart development appeared to be morphologically normal in sim1a morphants (data not shown), the heartbeat frequencies of sim1a morphants were slightly lower, and this reduction was statistically significant according to two-tailed Student’s t-tests (Fig. S2C). Taken together, these data suggest that sim1a loss of function is not associated with dramatic deficits in vascular or heart development, but that the hemodynamic forces in sim1a may be diminished as a consequence of the reduced heartbeat frequency.

sim1a loss of function leads to an abrogation of the PST segment and the CS, and a concomitant expansion of the PCT segment

To interrogate the possible function(s) of sim1a during nephron segmentation within the pronephros, we next performed a combination of gene knockdown and expression studies. Wild-type embryos were injected at the 1-cell stage with a sim1a MO that targeted the start codon of sim1a in order to block translation, or a sim1a splice site MO to abrogate splicing (Eaton and Glasgow, 2006; Löhr, et al., 2009) (Fig. 3, Fig. S3, Fig S4). The MO injected embryos were then evaluated by WISH to assess nephron composition. Overall, WISH analysis revealed that sim1a knockdown caused drastic alterations in the proximal regions of the nephron and the CS at 24 hpf (Fig. 3, Fig. S3, Fig. S4). Similar alterations in the PCT, PST, and CS populations were observed in wild-type embryos injected with either the sim1a start site or splice site MO, but the percentage of morphants with segment phenotypes was more consistent with the start site MO, suggesting only partial penetrance of gene knockdown through interference with transcript splicing (Fig. S3B, C). Specifically, the translation-targeting sim1a MO mediated loss of sim1a led to a 2-3 somite expansion of the PCT, which was marked by solute carrier family 20, member 1a (slc20a1a) (Fig. 3A, Fig. S3B). Of note, slc20a1a expression at the eleventh somite in sim1a morphants was typically absent in most cases (80%, Fig. S3B), while some morphants showed weak slc20a1a expression in this area (20%, Fig. S3B). This area is symbolized by the striped region of the PCT segment depicted in the schematic (Fig. 3B). Alternately, the expression domains of transient receptor potential cation channel, subfamily M, member 7 (trpm7) in the PST and stanniocalcin 1 (stc1) in the CS were typically abrogated in the majority of sim1a morphant embryos (78%, and 85%, respectively) (Fig. 3A, Fig. S3B). In a subset of sim1a morphants, the trpm7 domain was minimally present, with a small area of trpm7+ cells located adjacent to somite 11 (13%) (Fig. S3B). However, the distal tubule was normal, with both the DE and DL segments remaining unaffected in sim1a morphants as indicated by analysis of solute carrier family 12, members 1 (slc12a1) and 3 (slc12a3) expression, respectively (Fig. 3A, Fig. S3B). As an experimental control, WISH for these markers was performed on embryos injected with a sim1a ATG mismatch MO, and in these cases, a wild-type segmentation phenotype was consistently observed (Fig. 3A, Fig. S3A).

Figure 3. The loss of sim1a results in the abrogation of the PST segment and the CS.

(A) Segmentation patterning changes within the pronephros following sim1a knockdown were independently assessed by the utilization of MOs targeting the sim1a ATG and splice sites. Control 5 bp ATG mmMO injections were performed to further confirm the specificity of the sim1a translation-targeting MO. WISH of embryos at 24 hpf indicates the expression of specific kidney segment markers (purple) in relation to the somites, which were labeled with smyhc1 (red). Segment domains are further demarcated by black lines with corresponding somite numbers as indicated. Embryo anterior is to the left. (B) Diagram indicates the dominant segmentation phenotypes displayed by sim1a morphants in comparison to the wild-type controls at 24 hpf. Colored bars indicate the respective domains of each nephron segment in relation to the somite numbers located above. The striped regions within the colored bars signify areas with lower transcript expression. Abbreviations: CS – corpuscles of Stannius; DE – distal early; DL – distal late; hpf – hours post fertilization; mmMO – mismatch morpholino; MO – morpholino; PCT – proximal convoluted tubule; PST – proximal straight tubule; WISH – whole mount in situ hybridization.

Based on the abnormal phenotypes observed in the proximal regions of the nephron and CS in sim1a morphants, additional WISH analyses were performed to further confirm these segmental defects in the context of sim1a translation-targeting. We utilized a riboprobe cocktail of kidney markers to concomitantly label alternating segments (Kroeger, et al., 2014). The riboprobe mix labeled the P with wt1b, the PCT with slc20a1a, and the DE with slc12a1 (Fig. S4A). A consistent decrease in the distance between the PCT and DE among the majority of sim1a morphants was again observed, further suggesting that there was a prevalent reduction but not always a complete loss of the PST segment at the 24 hpf stage (Fig. S4A). This gap may reflect variability in gene knockdown or incomplete penetrance of the sim1a start site MO, or may be due to an issue of WISH sensitivity (e.g. difficulty in detecting very low transcript levels). Nevertheless, it is reasonable to conclude from these studies that sim1a is essential for proper development of the PST segment within the zebrafish pronephros.

Past research has reported some developmental delay associated with the loss of sim1a (Eaton, et al., 2008). This concern prompted the evaluation of the CS at later time points to validate its loss in sim1a morphants. Analysis by WISH of sim1a morphant embryos at 36 and 48 hpf confirmed that stc1 expression was entirely lost, indicating the importance of sim1a in the establishment of the CS (Fig. S4B). In sum, for the most part, sim1a loss of function leads to an abrogation of the PST segment and the CS, in conjunction with a concomitant expansion of the PCT segment (see color-coded schematic map of segment domains and somite location, Fig. 3B).

As sim1a morphants displayed some cell death throughout the embryo, suggestive of off-target MO effects, we explored whether concomitant knockdown of p53 and sim1a would provide a more informative context in which to evaluate sim1a loss of function. Wild-type embryos were injected with a MO to knockdown p53 (Rentzsch, et al., 2003) or with the double combination of p53 and sim1a ATG MOs. Combined knockdown of p53+sim1a, compared to sim1a knockdown alone, was associated with less severe defects in head, eye, and brain development based on morphological observations (Fig. S2B, Fig. S5A). Knockdown of p53 was not associated with overt differences in embryonic development, consistent with previous observations (Robu, et al., 2007) (Fig. S5A). Using acridine orange staining, we observed that p53 knockdown caused minimal cell death and that p53+sim1a morphants also had reduced cell death in neurological tissues and trunk mesoderm (Fig. S5B) compared to sim1a morphants (Fig. S2A).

Next, we assessed nephron segmentation in p53 deficient embryos using WISH. Interestingly, knockdown of p53 led to an expansion of the PCT and PST segments, and a concomitant reduction of the DL segment (Fig. S6, Fig. S7A). In parallel to these observations, we found that nephron segment development in the context of combined p53+sim1a knockdown led similarly to the formation of an expanded PCT and reduced DL, and additionally these embryos formed nephrons with no detectable presence of the PST or CS populations (Fig. S6, Fig. S7B). The finding that p53 deficiency alters nephrogenesis suggests that cell turnover plays a critical role in regulating segment size during zebrafish pronephros ontogeny. To date, how cell proliferation and death affect pronephric segment pattern has remained largely unexplored. Our previous studies have indicated that cell proliferation after the 15 ss is not required for normal pronephros segmentation (Wingert and Davidson, 2011), but cellular dynamics in renal progenitors prior to this stage have not been examined. Our new discovery that p53 deficiency alters nephrogenesis highlights the fact that further research on cell death/proliferation is needed to better understand segmentation. However, the role for p53 in nephrogenesis precludes the use of p53 knockdown studies to address the issue of sim1a MO toxicity and to examine higher sim1a MO dosages to assess nephron segment phenotype penetrance.

sim1a gain of function expands the PST at the expense of the PCT, and results in a dramatic increase in the CS lineage

To further explore the roles of sim1a in the development of the PST and the CS within the pronephros, overexpression studies were conducted (Fig. 4, Fig. S8). For this line of inquiry, approximately 200 pg of sim1a capped messenger RNA (cRNA) was injected into wild-type embryos at the 1-cell stage. WISH analysis at 24 hpf in sim1a cRNA injected embryos revealed that the PST, marked by trpm7, expanded by one somite into the region where the nephron would normally possess a PCT identity (Fig. 4A, Fig. S8B). Thus, the enlarged PST was discovered to span the area from somite 8 to 11 (Fig. 4A). In accordance with this finding, a corresponding reduction was observed in the slc20a1a expressing PCT segment, which was formed adjacent from somites 5 to 7 (Fig. 4A, Fig. S8A). Although these alterations were somewhat subtle, when the length of these respective segments was quantified, the segmental changes in both the PCT and PST were found to be significant (with p-values < 0.05) (Fig. 4B, C). Additional proximal tubule alterations were not observed with other cRNA amounts, suggestive that the dosage of sim1a expression is critical during nephrogenesis (data not shown). Overall, these results (Fig. 4F) suggest that sim1a is partially sufficient to promote PST identity at the expense of PCT fate, and may indicate that other factors are required as well. Further, sim1a activity may directly inhibit the PCT fate or perhaps be involved in the regulation of the boundary between the PCT and PST segments.

Figure 4. sim1a overexpression is partially sufficient to promote the PST and CS while inhibiting PCT fate.

(A) Wild-type embryos injected with 200 pg of sim1a cRNA were assayed by WISH at 24 hpf with specific kidney segment markers, (purple), and with smyhc1 to demarcate the somites (red). Black lines indicate segment domains relative to somite numbers. Embryo anterior is to the left. Quantitative analyses utilizing the two-tailed Student’s t-test were performed to assess changes in the expression domains of (B) slc20a1a, (C) trpm7, (D) the size of individual stc1+ CS cells and (E) the size of the stc1+ area. (F) Summary of the sim1a overexpression phenotype. Colored bars indicate segment domains in relation to the corresponding somite numbers indicated above. Abbreviations: cRNA – capped RNA; CS – corpuscles of Stannius; DE – distal early; DL – distal late; hpf – hours post fertilization; OE – overexpression; PCT – proximal convoluted tubule; PST – proximal straight tubule; WISH – whole mount in situ hybridization.

In addition to proximal tubule segment alterations, sim1a overexpression led to a substantial expansion of the area occupied by the CS population, demarcated by stc1 expression, which was significant as well (p < 0.001) (Fig. 4A, D, E, Fig. S8C). As with the aforementioned analysis of the tubule segments, further CS alterations were not observed with varying dosages of sim1a cRNA (data not shown). To assess whether the expanded CS was composed of more cell or cells with an increased size, we quantified the size of stc1+ CS cells in wild-types and sim1a morphants. The average size of stc1+ CS cells was similar in both wild-type embryos and sim1a morphants, and indeed no statistical difference was found (Fig. 4D). Even though the CS was significantly expanded in sim1a morphants, the overall distal domain remained unaffected by this increased CS phenotype as indicated by the pan-distal marker clcnk (Fig. 4A, F, Fig. S8D). Consequently, this data in combination with that from the previous knockdown studies (Fig. 3, Fig. S3), suggests that sim1a is necessary and in some measure sufficient for the formation of the CS during nephrogenesis.

Next, we explored whether sim1a cRNA overexpression was capable of rescuing proximal nephron patterning or CS formation in sim1a deficient embryos (Fig. S9, Fig. S10). For these studies, wild-type embryos were injected with either the sim1a MO that targeted the start codon or transcript splicing, along with varying amounts of sim1a cRNA (between 36 and 600 pg), then examined at 24 hpf with tubule segment and CS markers using WISH (Fig. S9, Fig. S10). In both sim1a knockdown models, the expanded PCT domain was reduced in a subset of the sim1a deficient embryos that concomitantly received either 36 or 200 pg of sim1a cRNA, indicating partial rescue of the PCT phenotype (Fig. S9, Fig. S10A, B, C, D). However, the PCT domain in sim1a deficient embryos that were co-injected with sim1a cRNA was still somewhat expanded, being located adjacent to somites 5-9, compared to wild-type embryos where the PCT domain normally occupies a domain adjacent to somites 5-8 (Fig. S9A, Fig. S10A, B, C, D). Other dosages of sim1a cRNA were not associated with further rescue of the PCT phenotype that results from sim1a deficiency (data not shown). In addition, formation of the PST was not rescued in sim1a deficient embryos that received sim1a cRNA at any dosage that was tested (Fig. S9A, Fig. S10A, B, C, D, data not shown). However, approximately 17% and 12% of embryos developed stc1+ CS cells sim1a adjacent to somite 15 when they were co-injected with a very low quantity of sim1a cRNA (36 pg) along with the sim1a ATG and splice MO, respectively (Fig. S9B, Fig. S10B, D). Interestingly, higher quantities of sim1a cRNA were not sufficient to enable even limited restoration of the CS lineage (Fig. S10A, C, data not shown). Taken together, our cumulative results strongly suggest that the precise dosage of sim1a is crucial for normal PCT, PST, and CS development. Overall these observations are consistent with the conclusion that sim1a is a key regulator of proximal tubule and CS cell lineages.

RA mediates the sim1a domain within the renal progenitor field of the pronephros

We next wanted to investigate sim1a with regard to developmental pathways that have known roles in the establishment of nephron segmentation within the pronephros. One such factor is RA, a secreted signaling molecule that complexes with retinoic acid receptors (RXR, RAR) located at the cell membrane of target cells (Duester, 2008). Once internalized, this RA complex traffics to the nucleus where it alters gene transcription by binding retinoic acid response elements (RAREs) located in gene regulatory regions (Duester, 2008). Previous research has established the importance of RA signaling during nephrogenesis, where RA secreted from the cervical region of the PM influences renal progenitors by promoting proximal fates and restricting the distal segments during early somitogenesis (Wingert, et al., 2007; Wingert and Davidson, 2011). Exogenous RA treatment leads to a dose-dependent expansion of the gene expression territories of transcription factors that demarcate the rostral domain of the renal progenitors, along with a reduction of the caudal domain. This ultimately results in the expansion of the proximal segments and reductions in the distal segments, respectively (Wingert, et al., 2007; Wingert and Davidson, 2011). Conversely, blocking RA synthesis through DEAB treatment leads to an expansion of the caudal domains in the renal progenitor field at the expense of rostral domains, causing expanded distal segments and an abrogation of proximal segment fates (Wingert, et al., 2007; Wingert and Davidson, 2011). Interestingly, we discovered the presence of a consensus RARE sequence (Bastien and Rochette-Egly, 2004) through sequence analysis of the putative promoter region of sim1a (Fig. S11A), suggesting the possibility that RA might directly regulate sim1a expression. Based on this observation, we consequently sought to establish the potential relationship between RA and sim1a during nephron and CS patterning within the renal progenitor field.

To determine whether RA gain or loss of function had an effect on sim1a expression, wild-type embryos were either treated with exogenous RA or DEAB (Wingert, et al., 2007; Wingert and Davidson, 2011). At the 18 ss, when sim1a transcripts are normally present in a region of renal progenitors located adjacent to somites 5-13, which includes the future proximal tubule segments (PCT, PST), RA exposure led to a 3 somite expansion of the sim1a domain (Fig. 5, Fig. S11, Fig. S12). Moreover, at the 28 ss, sim1a expression in the CS was located adjacent to somite 16, being shifted distally by 1 somite following RA treatment, compared to wild-types where the sim1a + CS cluster was located next to somite 15 (Fig. 5, Fig. S11, Fig. S12A). Alternately, when RA synthesis was inhibited in wild-type embryos by DEAB treatment, a reduction of the sim1a expression domain in the developing proximal segments was observed at the 18 ss and later there was a rostral shift of the sim1a CS domain at the 28 ss (Fig. 5, Fig. S11, Fig. S12B). The identities of these regions were verified by the WISH expression analysis of the pan-proximal marker, solute carrier family 4, member 4a (slc4a4a), and the CS marker, stc1, in wild-type embryos treated with DEAB (Fig. S11B). Overall, these findings show that the domain of sim1a is altered when RA levels are changed, suggesting that sim1a acts downstream of RA during renal progenitor development.

Figure 5. RA mediates the sim1a domain within the pronephric renal progenitor field.

(A) Wild-type embryos were treated with either a DMSO control at 75% epi, exogenous RA (1×10⁻⁸ M, denoted as + RA) at 50-60% epi, or 1.67×10⁻⁵ M DEAB at 75% epi. Groups were treated with their respective chemical until the 5-7 ss and then analyzed for sim1a expression (purple) by WISH at the 18 and 28 ss. Transcripts encoding smyhc1 (red) mark the somites. Black lines with corresponding numbers indicate segment domains and somites, respectively. Embryo anterior is located to the left. Abbreviations: DEAB – 4-diethylaminobenzaldehyde; epi – epiboly; RA – retinoic acid; ss – somite stage; WISH – whole mount in situ hybridization.

RA promotes the proximalization of the nephron at the expense of the distal segments, and elevated RA levels are not sufficient to rescue PST fate in sim1a morphants

As aforementioned, RA is essential to form the PCT and PST segments, as the reduction of RA biosynthesis leads to an abrogation of proximal nephron segments and a corresponding expansion of distal nephron segments, including the DE, CS, and DL regions (Wingert, et al., 2007). There is a temporal requirement for RA signaling to promote proximal nephron fates and restrict distal ones, and RA levels also have dose-dependent temporal effects on the formation of these lineages from the renal progenitor field (Wingert, et al., 2007; Wingert and Davidson, 2011). Therefore, we next explored whether reductions in RA levels at discrete developmental times was sufficient to rescue PST or CS formation in sim1a morphants, and restore the PCT segment to its normal domain within the pronephric nephrons.

To perform this analysis, wild-type embryos and sim1a morphants were treated with exogenous RA, and pronephros patterning was assessed by WISH using segment-specific riboprobes. The effect of RA on nephron segmentation was determined first in wild-type embryos that were treated with either a low (+ RA) or high (++ RA) dose of RA and then analyzed by WISH (Fig. 6, Fig. S13). In comparison to wild-type controls treated with DMSO, + RA treatment led to a 1 somite expansion in the PCT (slc20a1a) resulting in a slight distal shift of the PST (trpm7), which was also expanded by 1 somite (Fig. 6A, Fig. S13A, B). Upon evaluation of the distal regions in + RA treated embryos, a distal shift of the DE (slc12a1) as well as a shift and a 1 somite reduction in the DL (slc12a3) was seen (Fig. 6A, Fig. S13B). These effects were further compounded when after higher dose of RA (++ RA) was applied to wild-type embryos. Under these circumstances, the PCT (slc20a1) displayed a 2 somite expansion which caused a distal shift in the neighboring PST (trpm7) segment that had, once again, expanded by 1 somite in wild-type embryos treated with ++RA (Fig. 6A, Fig. S13C). This overall expansion of the proximal region caused a shift in the DE (slc12a1) and DL (slc12a3) where the latter segment was also reduced by 2 somites (Fig. 6A, Fig. S13C). CS formation was very weak or was entirely absent in embryos treated with either dose of RA (+RA or ++RA) during these respective time intervals (Fig. 6A, Fig. S13B, C). This is likely due to developmental delay, which has been noted in prior studies with exogenous RA (Wingert, et al., 2007). In sum, these results were documented in a color-coded schematic to represent the dominant phenotypes observed in RA treated wild-type embryos (Fig. 6C).

Figure 6. Increased RA levels promote the proximalization of the nephron but are not sufficient to rescue PST formation in sim1a morphants.

(A) Wild-type embryos or (B) sim1a morphants were treated with either a 1×10⁻⁷ M DMSO control, a low dose of RA (1×10⁻⁸ M; denoted as + RA), or with a higher RA concentration of 1×10⁻⁷ M (denoted as ++ RA). Both the DMSO control and ++ RA dose were given from the 90% epi to 5-7 ss while the + RA dose was applied from 90% epi to the 15 ss. Changes to nephron segmentation after the addition of RA was visualized by WISH in embryos at 24 hpf. The expression patterns of segment specific transcripts (purple) and smyhc1 transcripts, which marks the somites (red), are shown. Black lines with corresponding somite numbers were used to further illustrate segment domains. Embryo anterior is located to the left. (C) Schematic delineating the effects of low (+) and high (++) RA concentrations on nephron segmentation in wild-type embryos and sim1a morphants at 24 hpf. Nephron segments are represented by colored bars in relation to their corresponding somite numbers. Regions with lower transcript expression are indicated by stripes. Abbreviations: CS – corpuscles of Stannius; DE – distal early; DL – distal late; epi – epiboly; hpf – hours post fertilization; PCT – proximal convoluted tubule; PST – proximal straight tubule; RA – retinoic acid (+ RA – low; ++ RA – high dose); ss – somite stage; WISH – whole mount in situ hybridization.

To determine the effect of RA on segmentation in the setting of sim1a deficiency, + RA and ++ RA treatments were applied to sim1a morphants in the same manner (Fig. 6, Fig. 13). As expected based on wild-type controls, the PCT (slc20a1a) was expanded, either by 1 or 3 somites following the addition of + RA and ++ RA, respectively (Fig. 6B, Fig. S13E, F). However, RA treatment at either dose was unable to rescue the formation of the PST (trpm7) segment (Fig. 6B, Fig. S13D, E, F), which is typically absent in sim1a morphants (refer to Fig. 3). Regardless, consistent shifts of the DE (slc12a1) and reductions in the DL (slc12a3) still occurred in RA treated sim1a morphants (Fig. 6B, Fig. S13E, F). Notably, a 1 somite gap adjacent to somite 12, and thus corresponding to where the PST should be, was apparent between the boundaries of the PCT and DE in sim1a morphants treated with a low concentration of RA (+ RA) (Fig. 6B, C, Fig. S13E). However, when a higher dose of RA (++ RA) was utilized, an overall conversion of the expanded proximal region to slc20a1a (PCT) identity at the expense of the distal regions was revealed (Fig. 6B, C, Fig. S13F). Furthermore, consistent with prior findings (refer to Fig. 3), the CS (stc1) was still abrogated in RA-treated sim1a morphants (Fig. 6B, C, Fig. S13D, E, F). Taken together, these findings demonstrate that elevated levels of RA are not sufficient to rescue PST formation during nephrogenesis in sim1a-deficient embryos.

Chemical inhibition of RA biosynthesis by DEAB induces distal fates but is not sufficient to rescue CS formation in sim1a-deficient embryos

Previous studies have demonstrated that an abrogation of RA biosynthesis during renal progenitor development in zebrafish, due to Aldh enzyme activity inhibition with DEAB or genetic deficiency in aldh1a2, causes an expansion of distal nephron segment domains, including the DE, CS, and DL regions (Wingert, et al., 2007; Wingert and Davidson, 2011). One interpretation of these data is that RA acts to inhibit distal segment formation (Wingert, et al., 2007; Wingert and Davidson, 2011). Given our finding that sim1a knockdown eliminated CS formation, we next interrogated whether blocking RA production would be sufficient to rescue CS development in sim1a morphant embryos.

To investigate this, wild-type control embryos and sim1a morphants were treated with DEAB and assayed by WISH to evaluate the outcome in nephron segment patterning. In general, there was an overall distalization of the pronephros in DEAB treated wild-type embryos, though the degree to which this occurred varied by segment (Fig. 7A, Fig. S14A, B). DEAB treatment caused a 5 somite expansion in the DL (slc12a3) segment which ultimately caused a proximal shift of the slightly expanded DE segment indicated by slc12a1 expression (Fig. 7A, Fig. S14B). A corresponding proximal shift and domain expansion of the CS was observed as well (Fig. 7A, Fig. S14B). Similarly, sim1a morphants treated with DEAB displayed a greatly expanded DL leading to the consequent proximal shift of the DE that had an increased domain of approximately 2 somites (Fig. 7B, Fig. S14D). However, CS formation remained absent in sim1a morphants regardless of DEAB treatment, demonstrating that an abrogation of RA biosynthesis was not sufficient to rescue CS cell numbers in sim1a deficient embryos (Fig. 7B, Fig. S14C). These data ultimately suggest that sim1a is essential for the formation of the CS irrespective of RA levels within the embryo.

Figure 7. The inhibition of RA biosynthesis by DEAB induces distal fates but fails to rescue CS formation in sim1a morphants.

DMSO control or 1.67×10⁻⁵ M DEAB treatments of (A) wild-types or (B) sim1a morphant embryos from 75% epi to the 5-7 ss were performed. WISH analysis in embryos at 24 hpf for distal segment markers (slc12a1, stc1, and slc12a3) (purple) and the somites (smyhc1) (red). Black bars represent segment domains in correlation to somite numbers. Embryo anterior is located to the left. Insets show dorsal view of the CS. Abbreviations: CS – corpuscles of Stannius; DE – distal early; DEAB – 4-diethylaminobenzaldehyde; DL – distal late; epi – epiboly; hpf – hours post fertilization; ss – somite stage; WISH – whole mount in situ hybridization.

DISCUSSION

Here, we have discovered essential roles for the sim1a transcription factor in proximal tubule development in zebrafish and identified sim1a as the first gene known to be required for patterning of the CS (Fig. 8). Our studies are consistent with a model in which sim1a activity induces PST lineage formation within the field of renal progenitors that are competent to adopt proximal fates (Fig. 8). This may occur by sim1a inhibition of the PCT identity, and/or through the promotion of a PST-specific gene program. Although sim1a overexpression led to a modest expansion of the PST segment domain at the cost of the PCT, it was unable to induce the PST identity throughout the proximal tubule, suggesting that additional critical factors may collaborate with sim1a in order to promote PST fate. In addition to its expression in a renal progenitor domain that gives rise to proximal tubule fates, high levels of sim1a transcripts are subsequently expressed in a cluster of renal progenitors located adjacent to somite 15, where the CS emerges. Through our present study, we provide evidence that sim1a activity is indeed essential for the emergence of the CS lineage among distal renal progenitors, since in the absence of sim1a alone, the CS fails to develop. Furthermore, sim1a overexpression is sufficient to induce the CS. The temporal progression of sim1a transcript expression in the forming proximal tubule followed by the CS segment suggests the hypothesis that sim1a has distinct temporal roles in the development of these nephron regions, but additional studies are needed to further explore this notion.

Figure 8. The roles of sim1a during nephrogenesis in the zebrafish embryo.

Interactions between RA, sim1a, and additional transcription factors and mechanisms are vital for establishing normal nephron segmentation pattern by 24 hpf in the zebrafish pronephros. sim1a is essential for the formation of both the PST and the CS. It is possible that sim1a mediates the PCT/PST boundary by inhibiting the PCT domain as well. RA could also be modulating the expression of sim1a during the establishment of the PST while acting to restrict sim1a during CS formation. While sim1a appears to function downstream of RA, interactions between RA and sim1a may be indirect or direct during nephrogenesis. Abbreviations: CS – corpuscles of Stannius; DE – distal early; DL – distal late; hpf – hours post fertilization; N – neck; P – podocytes; PCT – proximal convoluted tubule; PD – pronephric duct; PST – proximal straight tubule; RA – retinoic acid.

Interestingly, we found that changes in RA levels alter the domain of sim1a expression in renal progenitors, positing RA as a potential regulator of sim1a expression during nephrogenesis (Fig. 8). Using chemical genetics, we learned that exogenous RA treatment, which promotes both PCT and PST fate at the expense of the distal segments in wild-type embryos, was unable to rescue the PST in sim1a morphants. Furthermore, abrogation of RA biosynthesis by exposure to DEAB, which promotes distal fates in wild-types, was unable to rescue CS formation in sim1a morphants. As changes in RA cannot compensate for sim1a deficiency, these data suggest that sim1a exerts its effects on renal progenitors subsequent to the activities controlled by RA signaling, thus supporting a model in which RA acts upstream of sim1a to pattern the proximal tubule and influence the CS lineage (Fig. 8).

Whether RA exerts direct or indirect effects on sim1a gene expression in renal progenitors to control proximal segmentation or CS formation was not elucidated in the present work. However, RA has been shown to directly regulate the rostral expression of wt1a (Bollig, et al., 2009; Miceli, et al., 2014), a transcription factor that is essential for proper podocyte specification from the rostral domain of renal progenitors (Perner, et al., 2007; O’Brien, et al., 2011). Given this precedence, the identification of an RARE in the upstream genomic sequence of sim1a (Fig. S11A) supports the hypothesis that RA may directly regulate sim1a transcription. Additional genetic and biochemical studies are needed to explore this potential mechanism of sim1a regulation, and to determine if it is operative during renal progenitor development. In sum, our studies are consistent with the conclusion that RA signaling is epistatic to the activities of sim1a in renal progenitors, and that RA exerts its patterning effects on the renal field partly by influencing the expression domain of sim1a.

Conservation of Sim1/2 factors and possible implications for kidney organogenesis across vertebrate species

There is at present only a rudimentary understanding of the molecular mechanisms that direct nephron segmentation during kidney organogenesis (Kopan, et al., 2007; Schedl, 2007; Cheng and Wingert, 2014). Recent studies have begun to ascertain key transcription factors that comprise the gene regulatory networks that control pro- and mesonephros development (Boualia, et al., 2013). Continuing to discover how renal lineages arise in development is an important topic in nephrology, with many implications for gaining new insights into kidney disease (Li and Wingert, 2013), and also has high relevance to understanding the mechanisms of renal regeneration in various contexts (McCampbell and Wingert, 2014).

The discovery that sim1a is an essential regulator of proximal tubule pattern is novel and provides a valuable launching point to evaluate the role(s) of this gene during renal development in other vertebrates. Indeed, Sim1/2 bHLH PAS transcription factors are found across vertebrate species, including fish, frog, chick, mouse, and human (Ema, et al., 1996; Fan, et al., 1996; Chrast, et al., 1997; Coumailleau, et al., 2000; Wen, et al., 2002; Coumailleau and Duprez, 2009). Functional domains within SIM proteins are highly conserved: prior analysis has discerned that zebrafish sim1a, and Sim1 in murines, Xenopus, and humans show over 95% sequence identity in the bHLH domain (Wen, et al., 2002). In addition, zebrafish sim1a shares 88-94% sequence identity in the PAS subdomains with mammalian and Xenopus SIM factors (Wen, et al., 2002). In mice, both Sim1 and Sim2 are imperative for fetal development, and animals with null mutations die shortly after birth (Michaud, et al., 1998; Goshu, et al., 2002, Schamblott, et al., 2002).

To date, while kidney phenotypes have not been reported in these or other various models of Sim1 haploinsufficiency in mice or humans (Holder, et al., 2000; Michaud, et al., 2001), there is evidence to suggest that the roles of Sim1/2 in the kidney of higher vertebrates warrants investigation. Notably, Sim1 expression has been previously reported in the mammalian kidney, with transcripts detected in the mouse E11.5 embryo by an RNA blot (Ema, et al., 1996). By section in situ hybridization, Sim1 has been spatially localized in the mesonephric ducts, and Sim2 has been reported in a similar pattern, but with weaker overall signal intensity (Fan, et al., 1996).

More recently, the genitourinary developmental molecular anatomy project (GUDMAP) (McMahon, et al., 2008; Harding, et al., 2011) has annotated Sim1 expression in the comma- and S-shaped body stages of mouse metanephric nephron formation. Intriguingly, within the S-shaped nephron body, Sim1 transcripts were noted within the medial region, suggestive of expression in the anlage of the loop of Henle (also termed the primitive loop of Henle), a region situated between the early proximal and early distal tubules. Additional spatiotemporal expression studies are needed to determine if the expression of Sim1 in the medial S-shaped body includes PST progenitors and to evaluate how Sim1 nephron localization in the mouse metanephros may be conserved or distinct from that of the zebrafish pronephros. As is typical of freshwater fish, zebrafish kidney nephrons do not contain a loop of Henle—they have no need to concentrate their urine (Wingert, et al., 2007). Consequently, it is possible that additional, unique functions of Sim1 may have evolved with respect to the mammalian kidney. Even so, the molecular evolution of nephron segments over the course of vertebrate genealogical history is a fascinating topic. At present, the loop of Henle is indeed believed to represent an evolutionary adaptation to the ancestral primitive nephron that occurred in clades such as birds and mammals—with the capacity for water conservation representing an emergent necessity accompanying the transition to the environmental rigors of terrestrial life (Romagnani, et al., 2013). Thus, while a fascinating topic for future study, the functional significance of these expression patterns ultimately remains unknown at present. Given our findings in the zebrafish pronephros, investigation of Sim1/2 functionality in mammals, as well as genetic fate mapping of Sim1-expressing cells in the murine nephron, are needed to assess the contributions of these factors during the formation of mammalian kidney structures.

These speculations do raise the broader question of the implications for genetic studies of nephrogenesis within the zebrafish pronephros in terms of elucidating renal patterning mechanisms in more complex excretory organs. The limitation in our zebrafish pronephros model is very likely to entail differences in local tissue environment. The zebrafish pronephric nephrons, for example, experience a rostro-caudal RA gradient, which is not presently thought to exist in the mammalian meso- or metanephros (Cheng and Wingert, 2014). However, the gene expression profiles of nephron segments in different vertebrates, ranging from Xenopus to humans, exhibit many common components (Wingert and Davidson, 2008). Thus, it is reasonable to speculate that while some environmental contexts of nephron ontogeny naturally differ across species, the gene regulatory networks that ultimately specify analogous renal lineages will have a number of conserved signatures. Thus, while the renal patterning in the mammalian metanephric kidney may have evolved to occur in the absence of RA’s influence, it may be the case that the subsequent activities of other transcriptional regulators and signaling pathways are nevertheless quite similar. The clear advantage of zebrafish for renal ontogeny research is the ability for rapid genetic analysis, which offers the ability to survey and identify components that warrant further studies in other kidney development models.

Proximal tubule development during nephrogenesis, and putative mechanisms of sim1a action in zebrafish pronephric renal progenitors

To date, the list of factors known to modulate proximal tubule development is rather limited—at present, the best understood component is Notch signaling. In mammals, proximal nephron segment formation during metanephros formation is reliant on Notch2, and conditional inactivation of Notch2 leads to the loss of proximal tubule segments (Cheng, et al., 2007). As our study now adds sim1a to the cast of molecules that can modulate proximal fates, establishing the genetic relationship between sim1a and Notch signaling in future studies will be important.

In addition, we have previously shown that the Iroquois (irx) homeodomain transcription factor irx3b is a vital segmentation factor during the development of the pronephric nephrons in zebrafish (Wingert and Davidson, 2011; Marra and Wingert, 2013), a role that is conserved in the pronephros of Xenopus (Reggiani, et al., 2007). In zebrafish, irx3b transcripts localize to the domain occupied by PST and DE progenitors, and irx3b deficiency leads to an abrogation of the DE segment and compensatory expansions in both the PCT and PST (Wingert and Davidson, 2011). While these data highlight a clear requirement for irx3b in DE segment differentiation, they also suggest that irx3b may normally modulate the domains of the PCT-PST, perhaps by regulating the boundary between these proximal segments (Wingert and Davidson, 2011). Now, given the role of sim1a in PST formation, it will be vital in future work to determine the genetic relationship between sim1a and irx3b in renal progenitors.

To identify the molecular mechanism of sim1a action in zebrafish renal progenitors, it will likely be necessary to identify its transcriptional activities. The activities of SIM1/2 proteins have been extensively studied in several developmental contexts, where they can act as transcriptional activators or repressors to direct cell fate and differentiation (Kewley, et al., 2004). Previous research has established that SIM proteins heterodimerize through their PAS domain with partners belonging to the Arnt (Arylhydrocarbon receptor nuclear translocator) and Arnt-2 protein families (Crews and Fan, 1999; Möglich, et al., 2009). Identification of Sim1a binding partners in the kidney will be one valuable avenue to gain insights into the ways by which this protein directs renal cell lineage decisions. Gaining further knowledge about renal lineage regulation is relevant to understanding both kidney development as well as regeneration (Diep, et al., 2011; Johnson, et al., 2011; Poureetezadi, et al., 2013; McCampbell, et al., 2014; Morales, et al., 2014).

Development of the enigmatic teleost-specific CS—sim1a at the helm as ‘master regulator’?

The present study has provided a seminal glimpse into the developmental patterning of the CS, a structure that has received relatively sparse experimental attention. First discovered by Hermann Friedrich Stannius (1839), the CS glands are known to be derived from pro- and mesonephric progenitors in teleostean fishes (Garrett, 1942, Krishnamurthy, 1976; Kaneko, et al., 1992). While the CS is not found in higher vertebrates, knowledge about the ontogeny and function of this organ is relevant to understanding fish renal physiology and cation homeostasis. The CS is vital for maintaining proper circulating levels of calcium and phosphate ions, and misregulation of ionic transporter expression in this gland leads to the development of kidney stones in zebrafish (Elizondo, et al., 2010). However, the genetic components that mediate the emergence of the CS lineage have remained unknown until now. Our data indicates that sim1a is required for the CS fate, thus ostensibly suggesting that sim1a may be the key regulatory component for the induction and/or survival of this lineage. Further work is needed to elucidate the precise mechanisms by which sim1a expression positively regulates CS development.

In summary, continuing to delineate the transcriptional networks that are essential for nephrogenesis in the vertebrate kidney is crucial for understanding the processes involved in renal cell fate and differentiation programs. The findings presented in this study provide useful new insights into the genetic pathways that direct nephron proximal segment patterning in the zebrafish, and have shed light on one essential genetic requirement for CS formation. Given the conservation between nephron composition and function across vertebrate kidney structures, these observations may have implications for understanding the causes of renal birth defects, other kidney diseases, the mechanisms of renal regeneration, and renal reprogramming.

Highlights.

1. -zebrafish nephron proximal tubule patterning is reliant on sim1a activity

2. - sim1a is essential to form the corpuscles of Stannius (CS), endocrine glands that emerge from the renal progenitor field

3. -retinoic acid (RA) regulates the expression domain of sim1a in renal progenitors

4. -interplay between RA and sim1a controls the balance of proximal tubule and CS fates

Abbreviations

AO

acridine orange

cRNA

capped RNA

CS

corpuscle of Stannius

DE

distal early

DEAB

4-diethylaminobenzaldehyde

DL

distal late

dpf

days post fertilization

epi

epiboly

hpf

hours post fertilization

hpi

hours post injection

IM

intermediate mesoderm

MET

mesenchymal to epithelial transition

mmMO

mismatch morpholino

MO

morpholino

N

neck

ORF

open reading frame

P

podocytes

PCT

proximal convoluted tubule

PD

pronephric duct

PFA

paraformaldehyde

PM

paraxial mesoderm

PST

proximal straight tubule

PTU

1-phenyl-2-thiourea

RA

retinoic acid

RARE

retinoic acid response element

RAR/RXR

retinoic acid receptors

sim1a

single-minded family bHLH transcription factor 1a

ss

somite stage

WISH

whole mount in situ hybridization

WT

wild-type