Heat Shock 70-kDa Protein 5 (Hspa5) Is Essential for Pronephros Formation by Mediating Retinoic Acid Signaling

Background: The function of Hspa5 in early embryonic development is not well understood.

Results: Hspa5 is involved in mediating retinoic acid signaling and is required for pronephros.

Conclusion: Hspa5 is essential for pronephros formation by mediating retinoic acid signaling.

Significance: This is the first report on the cross-talk between physiological ER stress and transduction of retinoic acid signaling.

Abstract

Heat shock 70-kDa protein 5 (Hspa5), also known as binding immunoglobulin protein (Bip) or glucose-regulated protein 78 (Grp78), belongs to the heat shock protein 70 kDa family. As a multifunctional protein, it participates in protein folding and calcium homeostasis and serves as an essential regulator of the endoplasmic reticulum (ER) stress response. It has also been implicated in signal transduction by acting as a receptor or co-receptor residing at the plasma membrane. Its function during embryonic development, however, remains largely elusive. In this study, we used morpholino antisense oligonucleotides (MOs) to knock down Hspa5 activity in Xenopus embryos. In Hspa5 morphants, pronephros formation was strongly inhibited with the reduction of pronephric marker genes Lim homeobox protein 1 (lhx1), pax2, and β1 subunit of Na/K-ATPase (atp1b1). Pronephros tissue was induced in vitro by treating animal caps with all-trans-retinoic acid and activin. Depletion of Hspa5 in animal caps, however, blocked the induction of pronephros as well as reduced the expression of retinoic acid (RA)-responsive genes, suggesting that knockdown of Hspa5 attenuated RA signaling. Knockdown of Hspa5 in animal caps resulted in decreased expression of lhx1, a transcription factor directly regulated by RA signaling and essential for pronephros specification. Co-injection of Hspa5MO with lhx1 mRNA partially rescued the phenotype induced by Hspa5MO. These results suggest that the RA-Lhx1 signaling cascade is involved in Hspa5MO-induced pronephros malformation. This study shows that Hspa5, a key regulator of the unfolded protein response, plays an essential role in pronephros formation, which is mediated in part through RA signaling during early embryonic development.

Introduction

The kidney is an essential organ that maintains homeostasis of the internal environment by filtering and excreting waste products and maintaining water and salt balance in the body. In vertebrates, the renal system develops from three successive developmental forms, the pronephros, mesonephros, and metanephros, originating from somatic and splanchnic intermediate mesoderm during embryonic development (1, 2). Although there are profound differences in anatomical structure between these three forms, the nephron is the basic structural and functional unit. The nephron contains three components: the glomerulus, tubule, and duct (3–5). From the pronephros to the metanephros, the successive forms become more complicated in the number and organization of nephrons. In higher vertebrates, the metanephros functions as the adult kidney, whereas in amphibians such as Xenopus, the mesonephros serves as the adult kidney, and the pronephros functions at embryonic stages. Although the developmental end point may differ in the vertebrate classes, the molecular mechanisms involved in kidney development are evolutionarily conserved in zebrafish, frog, mouse, rat, and humans (6).

Kidney development is a complex multistep process. It is initiated from the intermediate mesoderm at neurula stages. Many signaling pathways including bone morphogenetic protein (7), Wnt (8), FGF (9), Notch (10), and retinoic acid (RA)⁵ signaling have been implicated in pronephros development. These coordinated signaling cascades trigger the expression of a series of transcription factors that include Lim homeobox protein 1 (Lhx1), Wt1, and Pax8, which orchestrate the induction of the pronephros primordium.

RA signaling plays distinct roles during pronephros development. In early Xenopus gastrulae, treatment with all-trans-retinoic acid (atRA) and activin induces pluripotent ectodermal cells to differentiate into pronephros tissue (11–14). Inhibition of RA signaling impairs pronephros formation (15, 16). During early embryonic development, RA signaling is initiated mainly by atRA, a metabolite of vitamin A. atRA binds to a family of nuclear receptors, RARs (α/β/γ), which in turn form heterodimers with retinoid X receptors (RXRs) (α/β/γ) to activate expression of RA response element (RARE)-containing target genes (16, 17). lhx1 is one of the earliest genes that is expressed in the nephric mesenchyme and has been implicated in kidney development (18–22). The promoter region of lhx1 contains RAREs, and its expression is directly regulated by the RA signaling pathway (15). In Xenopus embryos, overexpression of lhx1 leads to the expansion of the pronephric field, whereas depletion of lhx1 severely reduces the pronephric field during organogenesis (15, 21). In Xenopus, it has been reported previously that RA can induce lhx1 expression in animal caps (23).

Heat shock 70-kDa protein 5 (Hspa5), also known as binding immunoglobulin protein (Bip) or glucose-regulated protein 78 (Grp78), belongs to the heat shock protein 70 kDa family of molecular chaperones (24). It functions in endoplasmic reticulum (ER) homeostasis and is a key regulator of the unfolded protein stress response. ER stress refers to any environmental condition such as thermal stressor heavy metal insults that interfere with the proper tertiary folding of proteins. In response to ER stress, cells activate the unfolded protein response, which is regulated in part by the IRE1α signaling cascade, to maintain homeostasis of the ER by attenuating protein translation and promoting proper protein folding, secretion, and protein degradation. Hspa5 expression is up-regulated during the unfolded protein response (see the diagram in Fig. 1U). Under normal conditions, Ire1α is constrained in an inactive form by binding to Hspa5. In response to ER stress, Hspa5 releases Ire1α, allowing free Ire1α to access full-length Xbp1 mRNA, generating the active form of Xbp1. Xbp1 in turn functions in regulation of biogenesis to recover homeostasis in the ER. It is notable that a number of key components including Xbp1 and Ire1β are highly expressed during embryonic development; however, their functional roles during embryogenesis are not well defined (25–27). In mammalian cells, Hspa5 is predominantly localized to the ER lumen and considered a chaperone, but recent reports indicate that it is also distributed at the cell membrane and participates in signal transduction (28–30).

FIGURE 1.

Expression pattern of hspa5 in developing Xenopus embryos. A–K, spatial expression pattern of hspa5 analyzed by whole mount in situ hybridization. Embryos are shown in different views (dorsal-lateral and vegetal). A and B, Hspa5 was expressed in the animal half prior to gastrulation stages. C–E, during gastrulation, hspa5 is strongly expressed in ventral ectoderm, dorsal blastopore lip, and the developing notochord but is absent from the prospective neural plate. In C and D, the red arrow indicates dorsal blastopore lip; in E, the white arrowhead indicates the notochord. F and G, at neurula stages, hspa5 expression is restricted to the notochord (white arrow) and the region anterior to the neural plate (red arrow) that corresponds to the future cement gland. H and I, at stage 22, strong expression is found in the cement and hatching glands. H, anterior view; I, lateral view. J and K, at the tail bud stage, hspa5 is expressed in the cement gland, pronephros (red arrow), ear vesicle, and epidermis. It is notable that the hspa5 signal can also be found in the liver primordium (K, black arrow). L and M, transverse sections of tail bud stages (J) showing hspa5 expression in pronephros (pn) and notochord (nc) tissue. N and O, transverse sections of a late tail bud stage embryo (K) show hspa5 expression in the pronephros (pn) and liver (li). P–S, Xbp1 expression in Xenopus embryos at the stages indicated. T, RT-PCR shows the temporal expression pattern of hspa5. Numbers above the lanes indicate the corresponding embryonic stages. A negative control (RT−) without reverse transcriptase was included. odc expression was used as the internal standard control. U, schematic diagram illustrating the interaction of Xbp1 and Hspa5 in the unfolded protein response. V, real time PCR indicates that hspa5 is up-regulated by overexpression of xbp1. xbp1 mRNA at the indicated doses was injected into Xenopus embryos at the two-cell stage. Animal caps were dissected at stage 9 and cultured for 2 h. The expression of hspa5 was analyzed by real time PCR, and its expression was normalized to odc. The animal cap assay was performed twice. St, stage.

Here, we report for the first time that Hspa5 plays an essential role in pronephros development in Xenopus embryos. We show that Xenopus hspa5 is strongly expressed in pronephros throughout tail bud stages and knockdown of Hspa5 function results in the inhibition of pronephros formation. We further show evidence that RA signaling and Lhx1 are involved in Hspa5MO-induced pronephros malformation. Supporting these results, overexpression of lhx1 partially rescued the morphant phenotype induced by Hspa5MO. Collectively, our studies demonstrate the importance of Hspa5 in kidney development.

EXPERIMENTAL PROCEDURES

Embryo Manipulations

Female frog ovulation was stimulated by injection with 500 IU of human chorion gonadotropin (Sigma-Aldrich) into the dorsal lymph sac. In vitro fertilization was performed by mixing male testis homogenates and eggs. After fertilization, embryos were dejellied with 2% cysteine solution, pH 8 and cultured in 0.2× MMR (20 mm NaCl, 0.2 mm KCl, 0.2 mm MgSO₄, 0.4 mm CaCl₂, 0.02 mm EDTA, 1 mm HEPES, pH 7.8). Embryos were staged according to Nieuwkoop and Faber (31).

Whole Mount in Situ Hybridization, β-Galactosidase (LacZ) Staining, and Vibratome Sectioning

Whole mount in situ hybridizations were performed according to the standard protocol (32, 33). LacZ staining was carried out as published previously (34). Briefly, 100 pg of LacZ mRNA with either the indicated morpholino (Hspa5MO1 or Hspa5MO2) or mRNA was injected into one blastomere of Xenopus embryos at the indicated stages for lineage tracing. After fixation with HEMFA (0.1 m HEPES, 2 mm EGTA, 1 mm MgSO₄, 3.7% formaldehyde), embryos were stained in X-Gal or Red-Gal staining solution (1 mg/ml X-Gal or Red-Gal, 5 mm K₃Fe(CN)₆, 5 mm K₄Fe(CN)₆, 2 mm MgCl₂ in 1× PBS) at room temperature. After fixation, embryos were dehydrated and stored in absolute ethanol at −20 °C until whole mount in situ hybridization. Vibratome sectioning was performed as described elsewhere (35, 36). After whole mount in situ hybridization, embryos were embedded in gelatin-albumin solution (5 g/liter gelatin, 380 g/liter chick egg albumin, 200 g/liter sucrose in 0.1 m phosphate buffer, pH 7.4) mixed with $\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$10$}\right.$ volume 25% (v/v) glutaraldehyde and sectioned with vibratome at a thickness of 20 μm.

Microinjection Experiments

Hspa5MOs and standard control MO were purchased from Gene Tools (Philomath, OR). For overexpression, mRNAs were synthesized in vitro with the mMessage mMachine SP6 kit (AM 1340, Ambion). Microinjection was performed as published previously (25).

In Vitro Induction of Pronephros

Animal caps were dissected from stage 9 embryos in 1× MMR and then cultured in animal cap culture solution (67% L-15 medium, 7.5 mm Tris-HCl, pH 7.5, 1 mg/ml BSA) until sibling embryos reached the desired stages. About 20 embryo animal caps were collected to extract RNA (33). For pronephros induction, the dissected animal caps were treated with activin at 10 mg/ml and atRA at 10⁻⁴ m for 3 h and then transferred into animal cap culture medium until the sibling embryos reached stage 15 or 32.

RNA Extraction, RT-PCR, and Real Time PCR

Total RNA was extracted from Xenopus embryos or animal caps using TRIzol (Invitrogen) and precipitated by isopropanol. The RNA extract was purified with RNeasy (Qiagen) after DNase I treatment. cDNA was synthesized by using Superscript III (Invitrogen) according to the manufacturer's instruction. Semiquantitative PCR was carried out as we described previously (36, 37). Real time PCR was performed with the ABI 7900HT Fast Real-Time PCR system (Applied Biosystem) using SYBR Green PCR Master Mix (TaKaRa) and specific primers. Xenopus odc and human β-actin were used as internal controls to normalize the target gene expression in Xenopus and HEK293T samples. All quantitative PCR experiments were performed in triplicate. The primers used in this study were either designed by Primer 3 or by following published studies. The primer sequences are listed in Table 1.

TABLE 1.

Primers to detect differential gene expression by RT-PCR

Fw, forward; Re, reverse.

Name	Sequence (5′–3′)	Refs.
cdx4 Fw	ATGTCAGGCTCCTTATACCA	63
cdx4 Re	ACAAACCCTTTCTCACTTGC
hoxd1 Fw	CAGCCCCGATTACGATTATTATGG
hoxd1 Re	CCGGGGAGGCAGGTTTTG
gbx2 Fw	CCCCCAAAACTCAAACCCTTCTAA	64
gbx2 Re	TGGCTCCGTATGGCAAACCTATT
pax2 Fw	CTGCTTTCCCTACTGACGTT	This study
pax2 Re	TAGCACGATAGTTTGTTCTT
smp30 Fw	TTAGACTGGTCTCTGGATCAC
smp30 Re	CGATAGGTAACTTTACAGTCTG
nephrin Fw	GCCACATTACAGTGGTCAAA
nephrin Re	ACAGGAGAACGACAGAGAGG
pax8 Fw	CCAACAGCAGCATCAGATC
pax8 Re	CAATGACACCTGGCCGGATA
hspa5 Fw	GAACCGTCTAACACCTGA
hspa5 Re	TAGCAGCCAGAGGCACAT
lhx1 Fw	AGGGGACCTAGGACCACTATC	23
lhx1 Re	CTCTTTGGATCGTCGGTTCTGGAACCA
odc Fw	CAGCTAGCTGTGGTGTGG	65
odc Re	CAACATGGAAACTCACACC
Human HSPA5 Fw	CGGGCAAAGATGTCAGGAAAG	66
Human HSPA5 Re	TTCTGGACGGGCTTCATAGTAGAC
Human β-actin Fw	AGGATGCAGAAGGAGATCACTG	67
Human β-actin Re	GGGTGTAACGCAACTAAGTCATAG
Human RARα Fw	CTGCCAGTACTGCCGACTGC	68
Human RARα Rw	ACGTTGTTCTGAGCTGTTGTTCGTA
Human RXRα Fw	CGACCCTGTCACCAACATTTGC	69
Human RXRα Re	GAGCAGCTCATTCCAGCCTGCC
Human RXRβ Fw	CGTGACATGAGGATGGACAAGA	70
Human RXRβ Re	TTGCAGTAGGTCTCCAGTGATG
Human CYP26A1 Fw	CATGTTCTCCAGAAAGTGCG	71
Human CYP26A1 Re	GGGATTCAGTCGAAGGGTCT
Human RALDH1 Fw	CTGCTGGCGACAATGGAGT	72
Human RALDH1 Re	GTCAGCCCAACCTGCACAG
Human RALDH2 Fw	AGACTTGGTGGAACGGGACA	73
Human RALDH2 Re	TATCAGCCCAGCCTGCGTAA
Human RALDH3 Fw	ACCTGGAGGTCAAGTTCACCAAGA	38
Human RALDH3 Re	ACGTCGGGCTTATCTCCTTCTTCC

Western Blot

Proteins were extracted from Xenopus embryos and HKE293T cells with lysis buffer (137 mm NaCl, 5 mm EDTA, 10 mm Tris-HCl, pH 7.5, 0.5% Triton X-100), and separated by SDS-PAGE. Antibodies for detecting protein were diluted at 1:1000 for anti-HSPA5 antibody (Cell Signaling Technology) and 1:5000 for anti-tubulin antibody (Cell Signaling Technology). The HRP-labeled secondary antibodies were diluted at 1:10,000 and detected with enhanced chemiluminescence (Super Signal West Dura Extended Duration Substrate, Thermo Fisher Scientific).

Luciferase Assay

DNA constructs of HSPA5 shRNA-pLKO.1 and a no-target control shRNA were purchased from Sigma. The HSPA5 shRNA sequence is CCGGCTTGTTGGTGGCTCGACTCGACTCGAGTCGAGTCGAGCCACCAACAAGTTTTT. shRNA-pLKO.1 was transfected into HEK293T cells along with pCMV-dR8.91 and vesicular stomatitis virus G. Forty-eight hours after transfection, lentiviral particles were collected and used to infect HEK293T cells. Twenty-four hours after infection, HEK293T cells were transfected with pGL3-RARE-luciferase and pRL-Renilla by Lipofectamine 2000 (Invitrogen). atRA (2 × 10⁻⁸ m) was added into culture medium 6 h after transfection. A luciferase assay was performed with the Dual-Luciferase reporter assay system (Promega) 48 h after transfection. Measurements were carried out in triplicate and are expressed as mean ± S.D. The following small interfering RNAs (siRNAs) were also used for knockdown of HSPA5 in HEK293T cells: siRNA1: sense, 5′-GGAGCGCAUUGAUACUAGATT-3′; antisense, 5′-UCUAGUAUCAAUGCGCUCCTT-3′; siRNA2: sense, 5′-GAGGCUUAUUUGGGAAAGATT-3′; antisense, 5′-UCUUUCCCAAAUAAGCCUCTT-3′. Two siRNAs were mixed in a 1:1 ratio at a final concentration of 60 nm. The siRNAs as well as pGL3-RARE-luciferase and pRL-Renilla were transfected into HEK293T cells using Lipofectamine 2000. Six hours after transfection, atRA was added to the cell culture medium. The luciferase assay was performed with the Dual-Luciferase reporter assay system (Promega) 48 h after transfection.

Immunostaining and Confocal Imaging

For immunostaining with Xenopus embryos, the embryos were bleached with 10% H₂O₂ in methanol for 3 h after fixation with HEMFA. After washing with 1× PBS, embryos were then blocked (155 mm NaCl, 10 mm Tris-HCl, pH 7.5, 10% FBS, 5% DMSO) for 1 h at room temperature. Embryos were incubated with the first antibody, 3G8 or 4A6 (European Xenopus Resource Centre), for 3 days at 4 °C. Subsequently, embryos were washed five times with 1× PBS and then incubated with the secondary anti-mouse antibody conjugated with Alexa Fluor 568 (Invitrogen). Embryos were photographed with a microscope (SZX16, Olympus).

For immunostaining with mammalian cells, SH-SY5Y cells were harvested 48 h after transfection with HSPA5 siRNA and fixed with 4% formaldehyde solution in 1× PBS. The endogenous HSPA5 and RARα were detected using rabbit anti-HSPA5 antibody (Cell Signaling Technology) and mouse anti-RARα (Abcam), respectively. Alexa Fluor 488-conjugated anti-rabbit and Alexa Fluor 568-conjugated anti-mouse antibodies were then used to visualize the localization and density of HSPA5 and RARα, respectively. Slides were mounted using Fluoromount (Southern Biotechnology) and viewed using an Olympus FluoView^TM FV1000 confocal microscope. Sequential scanning mode was utilized to prevent possible bleed-through of fluorescence emission.

Statistical Analysis

Data are shown as an average with a S.D. of measurements derived from at least three trials. One-way analysis of variance analysis was used to determine whether the difference was statistically significant.

RESULTS

Expression Pattern of hspa5 in Xenopus Embryos

Whole mount in situ hybridization was carried out to examine the spatial expression pattern of hspa5 in Xenopus embryos. Hspa5 was expressed in the animal half of the embryo from the two-cell stage to blastula stage (Fig. 1, A and B). During gastrulation, hspa5 expression was detected in the ventral ectoderm. In addition, a distinct expression domain was found in the dorsal blastopore lip (red arrow), whereas the prospective neural ectoderm lacked staining (Fig. 1, C–E). At neurula stage, staining was observed anterior to the neural plate (red arrow), corresponding to the future cement gland, as well as in ventral ectoderm. Consistent with early staining in the blastopore lip in the gastrula, distinct staining was also shown in the notochord (white arrow) (Fig. 1, F and G). At early tail bud stages (stage 22), hspa5 was expressed in the cement gland and hatching gland (Fig. 1, H and I). Embryos at late tail bud stages exhibited hspa5 expression in cement gland, pronephros (red arrow), liver (black arrow), ear vesicles, and epidermis (Fig. 1, J and K). Transverse sections confirmed hspa5 expression in the pronephros and notochord in early tail bud stages (Fig. 1, L and M) and in pronephros and the liver at late tail bud stages (Fig. 1, N and O). The spatial expression pattern of hspa5 was reminiscent of xbp1 (25), another key regulator of the unfolded protein response. A comparison of the expression pattern of xbp1 with that of hspa5 showed that xbp1 was co-expressed in the ventral ectoderm and dorsal blastopore lip (red arrow) during gastrulation (Fig. 1P), in the notochord and the region anterior to the neural plate (red arrow) at neurula stage (Fig. 1Q), and in the cement gland, pronephros (red arrow), and epidermis at tail bud stages (Fig. 1, R and S) (25). Our observations suggest that both hspa5 and xbp1 are as closely associated in embryonic development as they are in the ER stress response (39). Examination of the temporal expression pattern showed that hspa5 mRNA could be detected prior to gastrulation, and its expression was increased during gastrulation and maintained at higher levels onward (Fig. 1T). During the unfolded protein response, active Xbp1 can up-regulate Hspa5 expression (Fig. 1U). To demonstrate this interaction, we overexpressed Xenopus xbp1 in animal cap assays and found that hspa5 expression was indeed enhanced (Fig. 1V). Further confirmation of the up-regulation was observed in experiments repeated in Xenopus embryos (data not shown). Collectively, the complex expression of hspa5 in developing embryos suggests its importance during embryonic development.

Hspa5 Is Essential for Xenopus Embryonic Development

It has been reported that Hspa5^−/− mice are embryonic lethal (40), indicating an essential role for Hspa5 in early embryonic development. To further investigate the function of Hspa5 during early embryonic development in Xenopus, we used two Hspa5 morpholino antisense oligonucleotides (MO1 and MO2) (Fig. 2A) to reduce endogenous Hspa5 protein expression through translational inhibition. We found that injection of the two Hspa5MOs into Xenopus embryos at the two-cell stage resulted in a dose-dependent reduction of endogenous Hspa5 (Fig. 2, B–E), indicating that both Hspa5MOs could specifically attenuate Hspa5 protein translation in vivo. It is notable that injection of Hspa5MOs had no effect on the expression of β-tubulin. We examined the phenotypic effects of increasing doses of Hspa5MO1 injected into two blastomeres of two-cell stage embryos. Depletion of Hspa5 with 30 ng of Hspa5MO1 disrupted embryonic development. The injected embryos showed various degrees of developmental abnormalities including strong inhibition of head structures such as eyes as well as shortened anteroposterior axis. Lower injection doses (10 and 20 ng/embryo) induced a similar but less severely affected phenotype. We also examined the effects of Hspa5MO2 on embryonic development and observed the same phenotype. Based on the degree of severity, we categorized and quantified the phenotype of the injected embryos that developed to stage 32 into mild, moderate, and severe groups (Fig. 2F). The severity of the phenotype correlated with the doses of both Hspa5MO1 and Hspa5MO2 (Fig. 2G). Rescue experiments were performed using hspa5 mRNA containing six silent mutations (Fig. 2H) expected to be resistant to the binding of Hspa5MOs. The phenotype induced by 30 ng of Hspa5MO1 or Hspa5MO2/embryo was attenuated by co-injection with 200 pg of pCS2-hspa5 mRNA (Fig. 2I). In the Hspa5MO1 rescue experiments, the ratio of embryos exhibiting a severe phenotype was reduced from 58 to 13%, whereas the ratio of embryos in the mild and moderate categories increased from 41 to 87%. These observations indicate that the phenotype induced by Hspa5MO1 is specific. Consistent with Hspa5MO1, the phenotype induced by Hspa5MO2 was also partially rescued by pCS2-hspa5 mRNAs. In the Hspa5MO1 rescue experiments, the ratio of embryos exhibiting a severe or moderate phenotype was reduced from 69 to 47%, whereas the ratio of embryos in the mild category increased from 31 to 53%. These results indicate that both Hspa5MOs are effective in knocking down Hspa5 expression, which is essential for embryonic development.

FIGURE 2.

Hspa5MO specifically knocks down Hspa5 in Xenopus embryos. A, the two Hspa5MOs targeted to the ATG sequence are depicted in blue font and underlined. The translation start site of hspa5 is shown in black with the ATG in red. B–E, Hspa5MO1 and -MO2 reduced endogenous Hspa5 protein in Hspa5 morphants. Xenopus embryos at the two-cell stage were injected with the indicated doses of Hspa5MO1 (B) or Hspa5MO2 (D) and collected at stage 11. Endogenous Hspa5 was detected by Western blotting using α-Hspa5 antibody. Injection of Hspa5MO1 reduced endogenous Hspa5 in a dose-dependent manner in Xenopus embryos. The knockdown experiments were repeated twice for each MO, and a representative gel is shown in B and D. Western blots shown in B and D were quantified in C and E, respectively. The signal densities were acquired and analyzed using a GS-800 calibrated imaging densitometer. The relative signals of Hspa5 to tubulin from either Hspa5MO1- or Hspa5MO2-injected embryos were normalized to those from control embryos. The ratios indicate the decrease of Hspa5 expression. F, Hspa5MO1 and Hspa5MO2 induced the same phenotypes. The injected embryos show a range of phenotypes of short and curved body axis and disruption of eye and other head structures. The phenotypes of embryos injected with Hspa5MO1 or -MO2 were categorized into mild, moderate, and severe groups. G, quantification of phenotypes that were induced by increasing doses of Hspa5MO1 or Hspa5MO2. The numbers above the columns indicate the number of injected embryos from three experiments, and the phenotypes were categorized by investigators who were blinded to the treatment. H, silent mutations were introduced into the pCS2-hspa5 sequence to prevent binding of Hspa5MOs. I, the pCS2-hspa5 mRNA partially rescued the phenotype induced by Hspa5MOs. Quantification of phenotypes induced by the indicated Hspa5MOs or a combination of pCS2-hspa5 mRNA and Hspa5MO is shown in the column chart. Three independent experiments were performed, and hspa5 mRNA partially rescued the phenotype induced by Hspa5MO1. The number of injected embryos is listed on the top of the columns.

Hspa5 Is Required for Pronephros Formation in Xenopus Embryos

Our results show that Xenopus hspa5 is highly expressed in developing pronephric tissue (Fig. 1). In a previous mouse study, heterozygous animals carrying an Hspa5 mutation developed glomerulosclerosis and tubulointerstitial nephritis disease (41). It has also been reported that dysfunction of Hspa5 accelerated kidney injury (42), and expression of Hspa5 was up-regulated in diabetic kidneys (43). We therefore wanted to investigate the functional role of Hspa5 in Xenopus pronephros development, which may provide valuable insights into explaining the reported observations. To avoid possible interference with initial mesoderm formation, a mixture of Hspa5MO and LacZ mRNA was injected into one of the ventral-vegetal blastomeres of Xenopus embryos at the eight-cell stage that largely contribute to the formation of the pronephros. The LacZ mRNA was used as a lineage tracer to identify the injected side at later stages. When the ventral-vegetal blastomeres were targeted with either 15 or 30 ng of Hspa5MO2, the morphology of the injected embryos was predominantly normal (Fig. 3, A–C). Whole mount in situ hybridization and immunostaining experiments were performed to examine the formation of the pronephros in embryos targeted on one side with 30 ng of Hspa5MO2/embryo using marker genes that include pax2 (44), lhx1 (21, 22), and β1 subunit of Na/K-ATPase (atp1b1) (45). The marker genes pax2 (60%; 18 of 30), lhx1 (67%; 24 of 36), and atp1b1 (85%; 29 of 34) were strongly suppressed on the injected side but not on the uninjected side (Fig. 3, D–F and D′–F′), which was further confirmed by section analysis (Fig. 3, D″–F″). We also examined the expression of genes involved in pronephros tubule and glomus development including smp30 (46) and nephrin (47). Expression of smp30 (tubule; 64%; 21 of 33) and nephrin (glomus; 55%; 17 of 31) was decreased on the Hspa5MO2-injected side when compared with the uninjected side (Fig. 3, G, H, G′, and H′). Inhibition was confirmed by analysis on serial sections (Fig. 3, G″ and H″). The phenotypes were specific as the expression of pax2, lhx1, smp30, and nephrin remained normal when the same amount of standard control MO was injected into one vegetal-ventral blastomere of eight-cell embryos (Fig. 3, I–L′).

FIGURE 3.

Knockdown of Hspa5 inhibits pronephros formation. A–C, phenotype of Xenopus tadpoles injected with Hspa5MO2 into one ventral-vegetal blastomere at the eight-cell stage. Either 15 or 30 ng of Hspa5MO2 was co-injected with LacZ mRNA into one of the ventral-vegetal blastomeres of eight-cell stage embryos. A, anterior-dorsal view of mildly affected embryos at stage 16. B, lateral view of mildly affected embryos at stage 32. LacZ staining indicates the injected side; the uninjected side was used as a negative control. C, phenotype categorization of embryos injected with 15 or 30 ng Hspa5MO2. Over 90% of the embryos showed a normal or mild repression of anterior axis development after the injection. The severe phenotype refers to failure of gastrulation. D–F″, depletion of Hspa5 inhibits expression of pan-pronephros marker genes such as pax2, lhx1, and atp1b1. Eight-cell stage embryos were injected with LacZ mRNA (100 pg) into one of the vegetal-ventral blastomeres as a lineage tracer along with Hspa5MO2 (30 ng/embryo). The injected embryos were collected at stage 32, and the expression of pronephric marker genes as indicated was examined by whole mount in situ hybridization. D, E, and F, the uninjected side. D′, E′, and F′, expression of pax2, lhx1, and atp1b1 on the MO-injected side of the same embryos, respectively. D″, E″, and F″, transverse sections of embryos in D, E, and F accordingly. G–H″, expression of smp30 (tubule) and nephrin (glomus) at the Hspa5MO2-injected side was inhibited. G and H, the uninjected side. G′ and H′, Hspa5MO2-injected side. G″ and H″, transverse sections of embryos shown in G and H, respectively. I–L′, expression of pax2, lhx1, smp30, and nephrin in the tadpoles injected with 30 ng of standard control MO/embryo with 100 pg of LacZ mRNA into one vegetal-ventral blastomere at the eight-cell stage. I–L, uninjected side. I′-–L′, standard control MO-injected side (CMO-lacZ).

Similar to the whole mount in situ hybridization experiments, immunostaining with antibodies 3G8 and 4A6 that recognize pronephric tubules and ducts, respectively, also illustrated the inhibition of pronephros formation on the Hspa5MO-injected side (Fig. 4). The LacZ injection alone had little effect on pronephros formation (data not shown). Taken together, these data indicate that Hspa5 is required for pronephros formation.

FIGURE 4.

Hspa5 depletion inhibits of pronephric duct and tubule formation. Single vegetal-ventral blastomeres of eight-cell stage embryos were injected with LacZ mRNA (100 pg/embryo) and Hspa5MO2 (30 ng/embryo). Injected embryos were collected at approximately stage 39. Embryos were stained for LacZ and immunostained with 3G8 and 4A6 antibodies to detect pronephric ducts and tubules. A, A′, B, and B′, pronephric tubules in embryos detected by 3G8 antibody. C, C′, D, and D′, pronephros ducts in stage 39 embryos detected by 4A6 antibody. A′–D′, higher magnification of A–D. White arrow indicates pronephros. pn, pronephros. The white arrow indicates tubule in A–B′ and duct in C–D′. Uninjected, uninjected side; MO-lacZ, Hspa5MO2-injected side.

Depletion of Hspa5 Inhibits the Expression of Pronephros-specific Genes in Animal Cap Assays

Animal caps dissected from early gastrula embryos are pluripotent and can be induced to differentiate into pronephros tissue by treatment with atRA and activin (12–14). To further investigate the role that Hspa5 plays during pronephros development, we carried out an in vitro pronephros induction assay and examined two developmental time points, stages 15 and 32. In agreement with previous reports, treatment with activin and atRA induced animal caps to differentiate into pronephros tissue as revealed by up-regulation of lhx1, pax8, pax2, and wt1 when caps were collected at stage 15. Investigation of stage 32 treated caps showed an up-regulation of lhx1, pax8, pax2, smp30, and nephrin markers (Fig. 5, A–C). Conversely, knockdown of Hspa5 with Hspa5MO2 in animal caps resulted in decreased expression of the examined genes upon activin and atRA treatment at both stages (Fig. 5, B and C). These results were consistent with the inhibition of pronephros formation in Hspa5 morphants (Figs. 3 and 4). We also examined expression of lhx1 in Hspa5 morphants at stage 15. One dorsal blastomere of four-cell stage embryos was injected with either Hspa5MO1 (30 ng/embryo) or Hspa5MO2 (30 ng/embryo). The resulting lhx1 expression was reduced on the injected side (68% (13 of 19) in Hspa5MO1-injected embryos and 61% (14 of 23) in Hspa5MO2-injected embryos) when compared with the uninjected side (Fig. 5, D and E). These results support an important role for Hspa5 in pronephros formation both in vivo (embryos) and in vitro (animal cap assays).

FIGURE 5.

Pronephros induction in animal cap assays. A, schematic diagram of in vitro pronephros induction assay showing animal caps are induced to differentiate into pronephros tissues using atRA and activin. B and C, Hspa5MO2 (MO2) was injected into both blastomeres at the two-cell stage. Animal caps were dissected at stage 9 and treated with activin and atRA for 3 h. Subsequently, the animal caps were transferred into normal animal cap culture medium and cultured until stage 15 (St15) (B) or stage 32 (St32) (C). Expression analysis was performed by RT-PCR using total RNA extracted from the animal caps. AC, control animal cap; WE, sibling whole embryos. D and E, knockdown of Hspa5 suppresses lhx1 expression. Either 30 ng of Hspa5MO1 (MO1) (D) or Hspa5MO2 (MO2) (E) was co-injected with 100 pg of LacZ mRNA into one ventral animal blastomere of four-cell stage embryos. Injected embryos were collected at stage 15. The injected side was identified by red X-Gal staining, and the expression of lhx1 was examined by whole mount in situ hybridization.

Hspa5 Regulates Pronephros Formation by Mediating RA Signaling

Because knockdown of Hspa5 in animal caps impaired pronephros differentiation induced by activin and atRA, we sought to investigate how Hspa5 acts on transduction of RA or activin signaling. The animal caps dissected from stage 9 Hspa5MO2-injected embryos were treated with either atRA, activin, or a combination of both atRA and activin. In the control animal caps, atRA treatment led to an up-regulation of expression in RA-responsive genes including lhx1, cdx4, and gbx2 (Fig. 6A, lane 3) (37, 48). In contrast, RA-induced up-regulation was consistently reduced in the Hspa5MO2-injected caps (Fig. 6A, lane 4). Upon treatment with either 0.1 or 1 μm atRA, the RA-responsive genes including lhx1, gbx2, hoxd1, and cdx4 were up-regulated in a dose-dependent manner. Therefore, attenuation of Hspa5 activity causes reduced expression of these genes (Fig. 6B). Similarly, activin treatment also resulted in induction of lhx1, hoxd1, cdx4, gbx2, and chordin in the uninjected animal caps (Fig. 6A, lane 5). Interestingly, knockdown of hspa5 under the same condition did not affect the expression of hoxd1, cdx4, gbx2, and chordin but did lead to an increase in lhx1 expression (Fig. 6A, lane 6). These observations indicate that Hspa5 knockdown attenuates RA signaling but not activin signaling. When treated with both activin and atRA, the control animal caps showed increased expression of hoxd1, lhx1 and gbx2 but not chordin or cdx4 (Fig. 6A, lane 7), whereas the Hspa5MO2-injected animals caps exhibited unchanged expression of hoxd1, cdx4, and gbx2 except lhx1. Lhx1 is essential for pronephros formation, and its expression was reduced (Fig. 6A, lane 8). Collectively, these results provide evidence that RA signaling is attenuated when Hspa5 is depleted in the animal cap assay, suggesting that Hspa5 regulates pronephros formation at least in part through RA signaling.

FIGURE 6.

Hspa5 is required for transduction of RA signaling. A, Hspa5MO2 down-regulates lhx1, cdx4, and gbx2 gene expression in animal caps treated with atRA but not in animal caps treated with activin. Hspa5MO2 (MO2) (30 ng/embryo) was injected into both blastomeres of Xenopus embryos at the two-cell stage. Animal caps were dissected at stage 9, treated with either atRA or activin separately or a mixture of both for 3 h, and then cultured in normal animal cap culture medium for another 3 days. Expression of the indicated genes was examined by RT-PCR. Depletion of Hspa5 attenuates expression of lhx1, cdx4, and gbx2 induced by atRA. Activin treatment up-regulates lhx1, hoxd1, cdx4, gbx2, and chordin expression. Knockdown of Hspa5 has little effect on the indicated genes except that lhx1 expression is enhanced. Treatment of atRA and activin induces lhx1, hoxd1, and gbx2 expression in control animal caps, whereas knockdown of Hspa5 causes a reduction of lhx1 but not hoxd1 or gbx2. AC, control animal cap; WE, sibling whole embryos. B, Hspa5MO decreases expression of atRA-responsive genes. Animal caps were treated with 0.1 or 1 μm atRA for 4 h at room temperature. RA-responsive genes lhx1, gbx2, hoxd1, and cdx4 were induced by atRA treatment in a dose-dependent manner. Depletion of Hspa5 leads to decreased induction of these genes. C, real time PCR shows that HSPA5 shRNA effectively reduces HSPA5 mRNA levels in human HEK293T cells. Expression of HSPA5 in cells transfected with either control shRNA or HSPA5 shRNA was normalized to the control cells. D–G, endogenous HSPA5 expression is reduced in HSPA5 shRNA-infected (D and E) or HSPA5 siRNA-infected (F and G) HEK293T cells. The expression of HSPA5 was detected by Western blotting. β-Tubulin was used as a loading control (D and F). The quantified HSPA5 signal in either HSPA5 shRNA-infected cells or control shRNA-infected cells was normalized to the β-tubulin signal. The relative expression was normalized to that of control cells (E and G). The intensity of Western blot signals was quantified by a GS-800 calibrated imaging densitometer. H and I, depletion of HSPA5 by either Hspa5 shRNA (H) or Hspa5 siRNA (I) reduces the luciferase activity of RA luciferase reporter gene in HEK293T cells. atRA treatment enhances the luciferase activity, whereas HSPA5 shRNA reduces the luciferase activity in a dose-dependent manner. BMS453, an RA signaling inhibitor, also inhibits the luciferase activity of RA luciferase reporter gene. DMSO treatment was used as a control. The asterisks indicate the statistically significant difference. con, control; RLU, relative luciferase units. Error bars represent S.D.

We also examined the functional role of Hspa5 in RA signaling in cultured mammalian cells. Using shRNA and siRNA that specifically knock down HSPA5 in HEK293T cells, we carried out luciferase assays to quantitatively measure RA reporter activity in HSPA5-depleted cells. The knockdown efficiency was examined by real time PCR (Fig. 6C) and Western blotting (Fig. 6, D–G), which confirmed the reduction of HSPA5 expression at both the mRNA and protein levels. Luciferase assays were carried out in both control cells and the HSPA5-depleted cells to examine the activity of a RA reporter in the presence of atRA. As expected, atRA treatment significantly increased luciferase activity of the RA-responsive reporter gene in the control cells, whereas in HSPA5-depleted cells, the luciferase activity was lower compared with the control (Fig. 6, H and I). Moreover, the reduction was dose-dependent (Fig. 6H) and comparable with cells treated with BMS453, a known RA signaling inhibitor (49). These results indicate that knockdown of HSPA5 also attenuates RA signaling in HEK293T cells.

To further investigate the underlying mechanism as to how HSPA5 regulates RA signaling, we carried out real time PCR to examine the expression of genes involved in RA signaling transduction and atRA metabolism including RARα, RXRα, RXRβ, RALDH1, RALDH2, RALDH3, and CYP26A1 in HSPA5-depleted HEK293 cells. Knockdown of HSPA5 did not significantly alter the expression of the genes examined except that RALDH1 was moderately up-regulated (Fig. 7A). Thus it is likely that HSPA5 regulates expression of these genes at the protein level or affects protein localization. We then investigated the effects of HSPA5 depletion, if any, on RARα subcellular localization and its expression level at the protein level in SH-SY5Y cells. Although RARα is a nuclear receptor, it is mainly located in the cytoplasm, and only a small amount of RARα could be found in the nucleus (Fig. 7, C2 and E2), which is in line with what has been reported previously (50). We also found that the expression of RARα in HSPA5-depleted SH-SH5Y cell was clearly reduced (Fig. 7, D2 and F2). The HSPA5 depletion was validated by immunostaining detecting endogenous HSPA5 expression in the same cells (Fig. 7, C1–F1). The reduction of HSPA5 was further confirmed by Western blotting with the protein extracts from the cells under the same experimental conditions (Fig. 7B). It is worth noting that strong nucleus-localized HSPA5 signals were detected in SH-SH5Y cells, consistent with a previous study (51). Collectively, we have provided evidence to support that HSPA5 is required to maintain the RARα expression at the protein level that may contribute to the inhibition of RA signaling.

FIGURE 7.

Depletion of HSPA5 induces down-regulation of RARα at protein level. A, real time PCR indicates the expression of genes involved in RA signaling and atRA metabolism in control (blue column) or in HSPA5-depleted HEK293T cells (red column). B–F3, knockdown of HSPA5 reduced RARα protein in SH-SY5Y as revealed by confocal imaging. Either HSPA5 siRNA or control siRNA was transfected into SH-SY5Y cells. The knockdown efficiency is confirmed by Western blotting (B). C–F3, confocal imaging illustrates the subcellular expression of endogenous HSPA5 (green; C1–F1) and RARα (red; C2–F2) in SH-SY5Y cells. The nucleus was stained with DAPI (blue; C–F), and the merged images are shown in C3–F3. A clear reduction of RARα was identified in HSPA5-depleted cells (D2 and F2) compared with that in control (con) siRNA-transfected cells (C2 and E2). C–D3, lower magnification; E–F3, high magnification. Scale bars represent the length as indicated.

We have established that Hspa5 is required for Xenopus pronephros formation and that depletion of Hspa5 inhibits the expression of pronephros-specific genes in animal cap assays. To further assess the underlying mechanisms as to how Hspa5 contributes to pronephros formation, we focused on lhx1. lhx1 is one of the earliest genes expressed in the pronephric anlagen, and its expression was reduced in Hspa5MO-injected animal caps and embryos (Figs. 3, E–E″; 5, D and E; and 6A). Both hspa5 and lhx1 are expressed in the dorsal blastopore lip at gastrula stages (52). Our data show that RA signaling is inhibited when Hspa5 is depleted both in animal caps and in HEK293T cells and that lhx1 is responsive to atRA treatment (Fig. 6, A and B). It has been reported that several putative RAREs have been identified in the lhx1 promoter region (15, 53). This observation prompted us to investigate whether Hspa5 regulated pronephros formation through RA-Lhx1 signaling in whole embryos.

Rescue experiments were performed to investigate whether Lhx1 activity could rescue the phenotypes induced by Hspa5MO. Hspa5MO1 alone or in combination with lhx1 mRNA was injected into one ventral-vegetal blastomere of Xenopus embryos at the eight-cell stage with LacZ mRNA. Pronephros formation was evaluated by examining the expression of markers including lhx1, pax2, nephrin, and smp30. Injection of Hspa5MO1 caused severe inhibition of pronephros at the injected side, whereas co-injection of lhx1 mRNA and Hspa5MO1 partially rescued the phenotype induced by Hspa5MO1 (Fig. 8). These results suggest that the role of Hspa5 in pronephros development is mediated by RA-Lhx1 activity.

FIGURE 8.

Lhx1 expression partially rescues pronephros inhibition induced by Hspa5MO. A–H‴, expression of pronephric markers including lhx1, pax2, nephrin, and smp30 in embryos injected with either Hspa5MO1 (MO1) (30 ng) or a mixture of Hspa5MO1 (30 ng) and lhx1 (100 pg) were examined by whole mount in situ hybridization. Single ventral-vegetal blastomeres of eight-cell stage embryos were targeted, and LacZ mRNA was used as a lineage tracer. A, C, E, and G, Hspa5MO-injected side, A′, C′, E′, and G′, non-injected (Non-inj) side of the same embryo. B, D, F, and H, Hspa5MO1- and lhx1-injected side. B′, D′, F′, and H′, non-injected side. I, quantification of the rescue experiments of different pronephros markers. The phenotype assessments were performed blinded to the treatment. The numbers on the tops of the columns indicate the total number of injected embryos.

DISCUSSION

Hspa5 is a multifunctional protein that has been implicated in many biological processes. There is increasing evidence indicating that Hspa5 is involved in biogenesis and signal transduction and that it plays important roles in carcinogenesis (54, 55). In this study, we investigated Hspa5 function during early embryonic development in Xenopus and showed for the first time that Hspa5 plays an essential role in pronephros formation that is mediated by RA-Lhx1 signaling.

A number of genes expressed in the pronephros have been implicated in regulating its formation (46, 56, 57). We found that hspa5 was highly expressed in the pronephros in Xenopus embryos. Injection of Hspa5MO into one of the ventral-vegetal blastomeres at the 8-cell stage, which contribute to the pronephros, resulted in a decrease of pronephric markers such as lhx1, pax2, nephrin, atp1b1, and smp30 (Fig. 3). The inhibition of pronephros formation was further confirmed by whole mount immunostaining with tubule antibody 3G8 and duct antibody 4A6 (Fig. 4). Consistent with the observation in whole embryos, the expression of the pronephros marker genes in animal caps treated with activin and atRA was also significantly reduced (Fig. 5). Taken together, the data indicate that Hspa5 is required for pronephros formation.

RA signaling is crucial for pronephros formation. In Xenopus embryos, inhibition of RA signaling by either expression of a dominant-negative RA receptor, treatment with chemical inhibitors such as BMS453, or overexpression of the RA-catabolizing enzyme Cyp26 (15) results in the impairment of pronephros formation. Alternatively, the overactivation of RA signaling expands the size of the pronephros (15). In this study, we found that depletion of Hspa5 largely impaired pronephros formation. Moreover, the expression of retinoic acid-responsive genes such as hoxd1, gbx2, cdx4, and lhx1 were decreased in Hspa5-depleted animal caps treated with atRA (Fig. 6, A and B). Using confocal imaging, we also observed reduction of RARα in HSPA5-depleted SH-SY5Y cells that may contribute to the attenuation of RA signaling (Fig. 7). Interestingly, activin also induced the expression of hoxd1, gbx2, cdx4, and lhx1 in animal cap assays, but depletion of Hspa5 in activin-treated animal caps did not affect the expression of these genes (Fig. 6A). Several lines of evidence suggest that Hspa5 can interact with Cripto and promote Nodal signaling but attenuates activin A signaling (28, 58, 59). It is possible that knockdown of Hspa5 enhances activin signaling to a certain extent, resulting in an increase of lhx1 expression. Taken together, our observations suggest that Hspa5 regulates pronephros formation through RA signaling rather than activin signaling in in vitro pronephros induction assays.

It has been suggested that RA, but not activin, induces the generation of Ca²⁺ transients, which are involved in the process of pronephric tubulogenesis (60). The inhibition of RA signaling by Hspa5 was also confirmed in HEK293T cells. We found that the RA luciferase reporter activity was drastically reduced in either the HSPA5 shRNA- or HSPA5 siRNA-transfected cells (Fig. 6, H and I). Another line of evidence showing the Hspa5 regulates RA signaling is that lhx1 was down-regulated in Hspa5 morphants (Figs. 3, E′ and E″; 5, D and E; 6, A and B), and lhx1 expression could partially rescue the phenotypes induced by Hspa5MO1 (Fig. 8). lhx1 encodes one of the earliest transcription factors that is expressed in the pronephric anlagen and is directly regulated by RA signaling. During kidney organogenesis, Lhx1 is essential for tubule morphogenesis, duct extension, and glomerulus development (20, 61, 62). In Xenopus, a dominant-negative Lhx1 inhibited differentiation of pronephros derived from animal caps treated with activin and atRA (22). Similarly, depletion of Hspa5 also inhibited pronephros differentiation in this assay as revealed by decreased expression of lhx1 as well as other pronephros marker genes (Figs. 3–5). Taken together, our data suggest that Hspa5 is required for transduction of RA signaling during pronephros formation, and disruption of pronephros formation induced by Hspa5 depletion was at least in part due to the down-regulation of lhx1, a downstream effector of RA signaling.