A caveolin binding motif in Na/K-ATPase is required for stem cell differentiation and organogenesis in mammals and C. elegans

See allHide authors and affiliations

Science Advances  27 May 2020:
Vol. 6, no. 22, eaaw5851
DOI: 10.1126/sciadv.aaw5851


Several signaling events have been recognized as essential for regulating cell lineage specification and organogenesis in animals. We find that the gain of an amino-terminal caveolin binding motif (CBM) in the α subunit of the Na/K–adenosine triphosphatase (ATPase) (NKA) is required for the early stages of organogenesis in both mice and Caenorhabditis elegans. The evolutionary gain of the CBM occurred at the same time as the acquisition of the binding sites for Na+/K+. Loss of this CBM does not affect cell lineage specification or the initiation of organogenesis, but arrests further organ development. Mechanistically, this CBM is essential for the dynamic operation of Wnt and the timely up-regulation of transcriptional factors during organogenesis. These results indicate that the NKA was evolved as a dual functional protein that works in concert with Wnt as a hitherto unrecognized common mechanism to enable stem cell differentiation and organogenesis in multicellular organisms within the animal kingdom.


Embryonic development is characterized by the temporal and spatial regulation of cell proliferation, migration, differentiation, and tissue formation. Although these processes are genetically determined, several signaling mechanisms including Wnt have been recognized as essential in regulating cell lineage specification and organogenesis (13).

The Na/K–adenosine triphosphatase (ATPase) (NKA), discovered in crab nerve fibers by Skou (4), belongs to the P-type ATPase superfamily. It has an enzymatic function that couples adenosine 5′-triphosphate (ATP) hydrolysis to the transmembrane movement of Na+ and K+ in a cell lineage–dependent manner. For example, while the NKA is involved in the formation of action potentials in excitable cells, its polarized distribution is key to the functionality of the epithelium.

In addition to its canonical enzymatic function, we and others have shown that the NKA has an enzymatic activity–independent signaling function through its interactions with membrane cholesterol and proteins such as Src, epidermal growth factor (EGF) receptor, and caveolin-1 (58). We use the term signaling with liberty here, referring to the ability of NKA to work as a receptor, a scaffold, and a signal integrator by regulating the functions of its interacting proteins. This newly appreciated signaling function of the NKA has been implicated in several cellular processes (912). However, direct genetic evidence supporting a role for NKA signaling in animal physiology and disease progression is still lacking. This is due, in part, to the technical difficulties in studying its signaling separately from its ATPase-mediated pumping function because the latter is required for the survival of animal cells (13). Fundamentally, it is unknown whether the signaling function is an intrinsic property of the protein NKA, as its Na+- and K+-driven enzymatic activity has been recognized as. Therefore, we were prompted to address two important questions: (i) Were the signaling and Na+/K+ transport functions of the NKA coevolved? (ii) If so, does the signaling function of NKA represent a primordial yet common mechanism for the regulation of a fundamental process in animal biology?


The loss of α1 CBM does not affect NKA maturation or assembly as a full functional enzymatic complex, but perturbs its membrane subdomain distribution and abolishes its signaling function

Structurally, the NKA is composed of both α and β subunits. The α subunit contains the binding sites for Na+/K+ as well as ouabain, which are distinct from that of other P-type ATPases (14). It also has an N-terminal caveolin binding motif (CBM) proximal to the first transmembrane helix (fig. S1A). To assess the functionality of this motif, we made F97A and F100A mutations that map to the rat α1 NKA sequence. This strategy has been used by others to study the function of CBM in proteins other than the NKA (15). We used a knockdown and rescue protocol to generate a stable cell line (LW-mCBM) that essentially expresses just the CBM mutant α1, which was confirmed using [3H]ouabain binding assays (fig. S1B). Western blot and confocal imaging analyses showed that the expression of mutant α1 NKA in LW-mCBM was comparable to that in the control cell line, named AAC-19 cells (fig. S1, B and C). The expression of CBM mutant α1 was sufficient to restore the expression of the β1 subunit of the NKA, allowing normal plasma membrane targeting of the CBM mutant NKA in LW-mCBM cells (fig. S1, C and D). The successful generation of a stable CBM mutant α1 cell line suggests that the CBM is not essential for the enzymatic activity of the NKA because the ion-transporting function is necessary for animal cell survival (13). In further support, we conducted kinetic studies of the CBM mutant NKA. As shown in Fig. 1A, the overall enzymatic activity per unit of α1 NKA expression was identical between the control AAC-19 and LW-mCBM cells. The Km values of Na+, K+, and ouabain were comparable between the CBM mutant NKA and control (Fig. 1, B to D) (16). Together, these data indicate that the N-terminal CBM is not directly involved in the regulation of the enzymatic properties of the NKA.

Fig. 1 Pumping activity in LW-mCBM cells.

(A) Crude membrane preparations were made from AAC-19 and LW-mCBM cells and measured for ouabain-sensitive ATPase activity as described in Material and Methods. (B) Ouabain concentration curve. Crude membrane from LW-mCBM cells was prepared and measured for ATPase activity in the presence of different concentrations of ouabain. Data are shown as percentage of control, and each point represents three independent experiments. Curve fit analysis and IC50 (median inhibitory concentration) were calculated by GraphPad. (C and D) Measurements of Na+ and K+ Km. Assays were done as in (B). The combined data were collected from at least three repeats, and Km value (means ± SEM) was calculated using GraphPad.

On the basis of the above, we next turned our attention to determining the effects of the CBM mutation on signaling capabilities of the α1 NKA. Specifically, we first conducted immunoprecipitation experiments. As we reported previously in many types of cells (8), immunoprecipitation of caveolin-1 coprecipitated α1 in AAC-19 cells. In contrast, mutation of the CBM resulted in an over 80% decrease in coprecipitated α1 in LW-mCBM cells (Fig. 2A).

Fig. 2 Effects of mCBM on NKA-mediated signaling function.

(A) Cell lysates from AAC-19 and LW-mCBM were immunoprecipitated (IP) with polyclonal anti–caveolin-1 antibody. Immunoprecipitated complex was analyzed by Western blot for α1 and caveolin-1 (n = 4). **P < 0.01 compared to AAC-19. (B) Cell lysates from AAC-19 and LW-mCBM cells were subjected to sucrose gradient fractionation as described in Materials and Methods. A representative Western blot of three independent experiments was shown. **P < 0.01 in comparison to AAC-19. (C) AAC-19 and LW-mCBM cells were treated with different concentrations of ouabain for 10 min and analyzed by Western blot. A representative Western blot was shown (n = 4). *P < 0.05 versus 0 mM ouabain. (D) Cell growth curves of AAC-19 and LW-mCBM. *P < 0.05 versus AAC-19 cells. (E) BrdU assay of AAC-19 and LW-mCBM. The values are means ± SEM from at least three independent experiments. Photo credit: Xiaoliang Wang, Marshall Institute for Interdisciplinary Research at Marshall University.

To substantiate these observations, we next conducted a detergent-free and carbonate-based density gradient fractionation procedure and found that α1 NKA and its main signaling partners (Src and caveolin-1) were co-enriched in the low-density caveolar fractions, as previously reported in epithelial cells (8, 17). In sharp contrast, the expression of the CBM mutant α1 caused the redistribution of these proteins from low-density to high-density fractions (Fig. 2B). Quantitatively, when the ratios of fraction 4/5 of each protein versus total were calculated, we found that the low-density fraction 4/5 prepared from the control AAC-19 cells contained ~60, ~70, and 80% of caveolin-1, Src, and α1 NKA, respectively. However, in LW-mCBM cells, only ~20% of caveolin-1, Src, and α1 NKA were detected in fraction 4/5 (Fig. 2B).

To address the functional consequences of the dissociation of the α1 NKA from its signaling partners in LW-mCBM cells, we exposed these cells to ouabain, a specific agonist of the receptor NKA/Src complex. As shown in Fig. 2C, while ouabain stimulated phosphorylation of extracellular signal–regulated kinase (ERK), a downstream effector of the NKA/Src signaling pathway in AAC-19 cells (5, 8), it failed to do so in LW-mCBM cells.

We have previously shown that α1 NKA signaling is key to the dynamic regulation of cell growth (16, 18). As shown in Fig. 2D, LW-mCBM cells grew much slower than AAC-19 cells. 5-Bromo-2′-deoxyuridine (BrdU) incorporation assays further verified that the expression of CBM mutant α1 resulted in an inhibition of cellular proliferation (Fig. 2E). In short, the above in vitro experiments indicate that the gain of CBM enables the NKA to perform the enzymatic activity–independent signaling functions.

The loss of the α1 CBM results in embryonic lethality in mice

With the preceding in vitro data suggesting that the CBM is critically important to the signaling function of the NKA, we next set forth to test the physiological significance of this finding. Thus, we generated a knock-in mouse line expressing the aforementioned CBM mutant α1. The CBM mutant (mCBM) mouse was generated using the Cre/LoxP gene targeting strategy (19), as depicted in fig. S2A. The chimeric offspring were crossed to C57BL6 females to yield mCBM heterozygous mice, and the desired F97A and F100A substitutions were verified (fig. S2B). mCBM heterozygous mice were born fertile and survived to adulthood. Our attempts to generate mCBM homozygous mice yielded no viable homozygous pups (Fig. 3A) in nearly 400 young mice genotyped by polymerase chain reaction (PCR). These results document for the first time that the CBM in the α1 subunit of the NKA represents a fundamental signaling mechanism essential for mouse embryonic development and survival.

Fig. 3 Effect of mCBM on embryonic development.

(A) Early embryonic lethality of mCBM homozygous embryos. (B) Morphological comparison and body size of wild-type (WT) (top), heterozygous (middle), and homozygous (bottom) mCBM embryos at E9.5. Black bars, 0.3 mm. The arrows show the abnormal head morphology. Body size was measured from at least 12 embryos in different genotypes by ImageJ. Data are presented as means ± SEM. ***P < 0.01 versus the average of WT. (C) Sagittal sections of WT and homozygous (Homo) and heterozygous (Het) embryos at E9.5 with hematoxylin and eosin (H&E) staining. Homozygous embryos that had defective brain development indicated by open arrows. (D) Brain cross section of WT, homozygous, and heterozygous embryos at E9.5 with H&E staining. Homozygous embryos that had unclosed neural tube in forebrain, midbrain, and hindbrain were indicated by arrows; WT and heterozygous E9.5 embryos with closed neural tube were indicated by arrowhead. (E) Morphological comparison of WT and Na/K-ATPase α1 (+/−) embryos at E9.5. White bars, 0.3 mm (n = 5 to 7). Photo credit: Xiaoliang Wang, Marshall Institute for Interdisciplinary Research at Marshall University.

Inhibition of receptor NKA/Src complex does not affect embryonic development

There is evidence that endogenous ouabain is important in animal physiology because of its role in stimulating the signaling function of the NKA (10, 19, 20). Because the loss of the CBM abolishes ouabain-induced signal transduction in vitro, we tested whether administration of pNaKtide, a specific inhibitor of the receptor NKA/Src complex (21), would cause the same embryonic lethality as we observed in mCBM mice. As depicted in fig. S3, we observed no change in fetal survival after administration of pNaKtide to female mice before mating and continued until the end of pregnancy. It is important to mention that pNaKtide has been proven to be specific and effective in blocking the NKA/Src receptor signaling in vivo (2226), and our control experiments showed that pNaKtide could cross the placental barrier. Moreover, this lack of pNaKtide effect on mouse embryogenesis appears to be consistent with a previous report demonstrating that neutralization of endogenous ouabain by injection of an anti-ouabain antibody did affect the kidney development of neonatal mice but did not affect their overall survival (20). On the basis of these, we concluded that the NKA/Src receptor function in the CBM mutant embryo was not the direct cause of lethality and set out to identify a hitherto unrecognized NKA CBM-dependent yet NKA-Src–independent underlying mechanism.

The loss of α1 CBM does not block blastocyst formation and gastrulation but results in the arrest of organ development in mice

Embryo implantation within mice occurs around embryonic day 4.5 (E4.5) (27), followed by gastrulation around E5.5 to E7.5 (28), when the simple embryo develops into an organized and patterned structure with three germ layers (29). Subsequently, organogenesis takes place at E8.0 and onward; the patterned embryo starts to develop its organ systems including the brain, heart, limbs, and spinal cord.

To further analyze and explore the molecular mechanisms of the CBM mutation in the embryonic development of mice, we harvested the fertilized eggs at E1.5, and cultured them in vitro. It has previously been demonstrated that α1 knockout results in the failure of blastocyst formation (13). In contrast, we found that eggs from mCBM heterozygous parents developed into morphologically normal blastocysts. These findings indicate that loss of the CBM does not affect the molecular mechanisms necessary for blastocyst formation. Thus, a loss of functional α1 CBM and complete knockout of α1 NKA both result in embryonic lethality but differ by their specific mechanisms. Knockout of α1 NKA inevitably causes the loss of NKA enzymatic function, which is incompatible with life (13), and results in the failure of blastocyst formation in mice. In contrast, our in vitro data indicate that a loss of the CBM does not cause any notable alteration in NKA enzymatic activity, which is supported by the observation that mCBM mice are still capable of producing morphologically normal blastocysts. Consequently, CBM role in development appears to be critical at a developmental stage beyond blastocyst stage, and we further set out to identify this stage.

To this end, we collected and genotyped embryos or yolk sacs from mCBM heterozygous mice at different days of gestation. We first dissected 31 embryos at E12.5 from three different mice (Fig. 3A). Reabsorption and empty deciduae were observed in six implantation sites with only the mother’s genotype detectable. At E9.5, we were able to dissect a total of 303 embryos. Sixty-four of them were mCBM homozygous (21%), 71 were wild-type (23%), and 168 were mCBM heterozygous (55%) (Fig. 3A).

To further analyze the embryonic developmental defects, we examined mCBM embryos at E7.5, E8.5, and E9.5. The embryos looked similar between wild-type and mCBM homozygous mice at E7.5 and E8.5 under dissection microscopy. However, we found several severe morphological defects in homozygous embryos at E9.5 (Fig. 3, C and D). First, the overall size of embryos was considerably reduced in mCBM homozygous embryos (about 35% the size of the wild-type embryos). In addition, the observed effect of the CBM mutant on embryonic size was gene dose dependent, as the mCBM heterozygous embryos were significantly smaller than those of wild-type embryos but much bigger than the homozygous embryos. Second, most homozygous embryos did not “turn,” a process normally initiated at E8.5, suggesting that the loss of a functional CBM was responsible for a developmental arrest at an early stage of organogenesis. Last, the most severe morphological defects were observed in the heads of the mCBM homozygous embryos. In addition to the reduced size (about 25% of the size of wild-type embryos), we observed that mCBM homozygous embryos failed to close their cephalic neural folds (anterior neuropore) as indicated by the arrow in Fig. 3B. This phenotype more closely resembled wild-type embryos at E8.0 to E8.5, suggesting again that the loss of CBM arrested organogenesis in its early stages. On the other hand, all heterozygous embryos, although smaller than wild-type embryos, showed normal head morphology (Fig. 3B).

To follow up on the above observations, we collected and made histological sections of wild-type, heterozygous, and homozygous embryos at E9.5 (Fig. 3, C and D). Normally, formation and closure of the anterior neuropore occurs at E9.5 (Fig. 3D). In sharp contrast, mCBM homozygous embryos developed defects in neural closure. Specifically, failure of neural tube closure at the level of forebrain, midbrain, and hindbrain was prominent in homozygous embryos (Fig. 3D).

The loss of α1 CBM inhibits the expression of a cascade of transcriptional factors essential for neurogenesis

To further explore the molecular mechanism by which the loss of the CBM led to defects in organogenesis, we next conducted RNA sequencing analyses (RNAseq) in wild-type and mCBM homozygous embryos. More than 17,000 genes were read out in either mCBM homozygous or wild-type samples. Data analyses indicated that 214 and 208 genes from mCBM homozygous embryos were significantly down- and up-regulated, respectively (fig. S4). Among them, the expression of a cluster of transcriptional factors important for neurogenesis was significantly reduced. As depicted in Fig. 4A, the expression of neurogenin 1 and 2 (Ngn1/2), two basic helix-loop-helix (bHLH) transcriptional factors (30), was significantly down-regulated in homozygous embryos. Ngn1/2 are considered to be determination factors for neurogenesis, while members of the NeuroD family of bHLH work downstream to promote neuronal differentiation (31). We found that the expression of NeuroD1/4 was further reduced in mCBM homozygous embryos. As expected from these findings, the marker of neural stem cells nestin (Nes) and other genes related to neurogenesis including huntington-associated protein 1 (Hap1), nuclear receptor subfamily 2 group E members 1 (Nr2e1), and adhesion G protein (heterotrimeric guanine nucleotide–binding protein)–coupled receptor (Adgrb1) were all down-regulated in mCBM homozygous embryos (Fig. 4A). To verify these data, we performed reverse transcription quantitative PCR (RT-qPCR) analyses of both wild-type and mCBM homozygous embryos collected at E9.5. As depicted in Fig. 4 (B to D), the aforementioned transcriptional factors were all down-regulated in a cascade fashion. While a modest reduction was found with Ngn1/2, the expression of NeuroD1/4 was almost completely inhibited. To test whether the effects of the CBM mutation on the expression levels of these transcriptional factors were gene dose dependent, we also examined mRNA levels of Ngn1/2 and NeuroD1/4 in mCBM heterozygous embryos. As depicted in Fig. 4 (B and C), the expression of these genes followed the pattern found in homozygous embryos. The expression level in heterozygous embryos was significantly reduced compared to wild-type embryos but was much higher than that of mCBM homozygous embryos. These gene dosing–dependent cascade effects suggest that the α1 NKA is an important upstream regulator but not a determinant of neurogenesis like Ngn1/2 (32) or a key receptor mechanism like Wnt is.

Fig. 4 Expression of neurogenesis and neural stem cell markers in mCBM homozygous embryos at E9.5.

(A) RNAseq results of several neurogenesis and neural stem cell markers. Log2 ratio = 1 means twofold of change. *P < 0.05 compared to WT. (B and C) RT-qPCR analysis of selected gene expression in WT, heterozygous, and homozygous mCBM embryos at E9.5. (D) RT-qPCR analysis of neural stem cell marker gene expression in WT and homozygous mCBM E9.5 embryos. (E) RT-qPCR analysis of neurogenesis marker genes in WT and NKA α1+/− mouse E9.5 embryos. Quantitative data are presented as means ± SEM from at least six independent experiments. *P < 0.05, **P < 0.01 versus WT control.

As a control, we also assessed the expression of different isoforms of NKA and caveolin-1. As depicted in fig. S5, no changes were detected in the expression of the α1 isoform of the NKA. This is expected, as the mutations were only expressed on exon 4. Previous reports have demonstrated that, in addition to the α1 isoform, neurons also express the α3 isoform, while muscle and glial cells express the α2 isoform of the NKA (9). No difference was observed in the expression of α3, while the expression of α2 was too low to be measured. We were also unable to detect any change in the expression of caveolin-1.

Impact of reduced α1 expression on the embryonic development of mice

The total amount of protein recognized by the anti-NKA α1 antibody is unchanged in mCBM heterozygous mouse tissues compared to that of the wild type, albeit with changes in distribution in caveolar versus noncaveolar fractions. This indicates that the CBM mutant protein is fully expressed, as observed in cells (fig. S1), and further demonstrates that a reduction of enzymatic activity is not responsible for the observed phenotype in mCBM homozygous embryos. However, because the expression of wild-type α1 in mCBM heterozygous animals is most likely reduced, the phenotypic changes we observed in these mice could be due to the reduction of wild-type α1 expression rather than the expression of CBM mutant α1. To address this important issue, we collected embryos from α1 NKA heterozygous (α1+/−) mice and their littermate controls (33). In contrast to mCBM heterozygotes, reduction of α1 expression alone did not change the size of embryos (Fig. 3D), head morphology, or the expression of neuronal transcriptional factors (Fig. 4E). Because NKA α1 haploinsufficiency did not phenocopy mCBM heterozygosity, it was concluded that the mCBM allele was responsible for the observed changes.

The CBM is evolutionarily conserved

The CBM in NKA has a consensus sequence of “FCxxxFGGF” (fig. S6). To assess the generality of CBM-mediated regulation, we first turned to the conserveness of the CBM in animal NKA. A database search reveals that, like Wnt, the “mature” form of NKA (i.e., containing CBM, Na+/K+ binding sites, and β subunit) is absent in unicellular organisms but present in all multicellular organisms within animal kingdom (fig. S6). Further analysis of published data confirms the coevolutionary nature of the CBM and the binding sites for Na+ and K+ in the NKA. The first indication is from the analysis of single-cell organisms. No mature form of NKA is found in these organisms (fig. S6A). However, Salpingoeca rosetta, a marine eukaryote belonging to the Choanoflagellates class, undergoes a very primitive level of cell differentiation and specialization in their life cycle and expresses a putative NKA with several conserved motifs involved in the binding of Na+/K+. On the other hand, it contains no CBM (fig. S6) and there is also no evidence that it expresses a β subunit.

Second, as depicted in figs. S6 and S7, Caenorhabditis elegans, an example of a metazoan organism, expresses a mature form of NKA (eat-6) that contains binding sites for Na+ and K+ as well as the N-terminal CBM. It also expresses a couple of putative NKA such as catp-2 (34). However, they contain neither the CBM nor Na+ and K+ binding sites.

Third, although the “X” amino acids in the NKA CBM in invertebrates vary, only conserved substitutions occurred in this motif. This is in sharp contrast to many other membrane receptors/transducers such as Patched and Gα that also contain a consensus CBM (figs. S6 and S7). Within vertebrates, the CBM sequence “FCRQLFGGF” in NKA remains completely conserved across all species. Moreover, this sequence remains conserved in all isoforms of the α subunit except for the α4 isoform, which is exclusively expressed in sperm. The α4 isoform in some species still adapts the CBM sequence found in invertebrates (fig. S6). Moreover, of a total of nine α subunits found in zebrafish (35), five appear to be α1 homologs that, like the α4 isoform, contain both vertebrate and invertebrate CBM sequences.

Last, turning to the evolutionary aspect of the receptor NKA/Src complex, we found that the Src-binding NaKtide and Y260 sequences, in sharp contrast to the CBM, are only conserved in mammalian ATP1A1 (fig. S7). Therefore, the NKA/Src receptor may have evolved after the acquisition of the CBM, and hence is not a part of the fundamental regulation of animal organogenesis (fig. S3).

In short, the N-terminal CBM, like the binding sites for Na+ and K+, is conserved in all α subunits of NKA in animals, even after taking into consideration gene duplications and the generation of different isoforms or homologs. Thus, we postulate that this CBM must be evolutionally conserved to enable the NKA, in parallel with its enzymatic function, to serve an important role in the origination of multicellular organisms within the animal kingdom.

The loss of CBM in eat-6 results in the arrest of organogenesis in C. elegans

Organogenesis represents a unique feature of multicellular organisms. In considering the preceding findings, we reasoned that the loss of NKA CBM would also affect embryonic development in invertebrates such as C. elegans. To test our hypothesis, we used CRISPR-Cas9 to “knock in” the equivalent CBM double mutations of F75A and F78A in C. elegans NKA gene eat-6 (named as syb575) (fig. S8). Similar to the impact of the expression of CBM mutant α1 NKA in mice, no homozygous worms were produced, whereas the heterozygous worms hatched normally. Moreover, by using the gene balancer nT1, we confirmed that the F75A and F78A double mutations induced embryonic lethality in syb575 homozygotes secondary to L1 arrest (Fig. 5A). Furthermore, the observed larval arrest due to the loss of the eat-6 CBM was rescued by a transgene expressing a wild-type eat-6 complementary DNA (cDNA) through an extrachromosomal array (Fig. 5B). The lethality phenotype in syb575 mutants was different from those of the eat-6 mutants defective in enzymatic (transport) activity, because while the eat-6 mutants had growth defects, they were able to grow past the L1 stage (36). An exception to this was a cold-sensitive eat-6 (ad792) mutant with severely reduced transport activity, which exhibited L1 arrest at lower temperatures similarly to the syb575 mutant worms (36). Overall, those data suggest that both CBM-mediated signaling and ion transport activity by the NKA are essential to full-scale organogenesis in C. elegans.

Fig. 5 Effects of CBM mutation on C. elegans embryo development and human iPSC stemness.

(A) Heterozygous CBM mutant (mCBM) worms syb575/nT1 have GFP signals in pharynx (pointed with the arrowhead), while mCBM homozygous worms are GFP negative and arrested at larval stage (pointed with an arrow). (B) Rescue with a WT eat-6 gene showing a mCBM homozygous worm with a transgenic marker sur-5::GFP. Arrow points the somatic GFP signals. (C) Mutation of CBM–α1 NKA (F97A; F100A) results in reduced colony formation in human iPSC (mCBM iPSC). (D) RT-qPCR analysis of stem cell markers and primary germ layer markers in WT and mCBM iPSC. *P < 0.05 compared to WT. n = 7. Photo credit: Liquan Cai, Marshall Institute for Interdisciplinary Research at Marshall University.

In short, our data indicate that loss of the NKA CBM results in defective organogenesis in both mice and C. elegans. This, together with our finding that the NKA CBM is conserved in all NKA regardless of isoform or homolog, indicates that the NKA was originally evolved as a dual functional protein in multicellular organisms, and that it represents a primordial and common mechanism for regulating stem cell differentiation and early stage of organogenesis in animals.

Turning now to even more general features of the CBM in organogenesis, we searched for the plant plasma membrane H-ATPase that functions equivalently to the animal NKA. Like the NKA, the plant plasma membrane H-ATPase also contains a sequence motif at the first transmembrane segment that is in accordance with the consensus CBM. This motif is completely conserved from blue algae to land plants but does not exist within yeast and bacteria (fig. S6).

The loss of α1 CBM affects the differentiation status of human iPSCs

To assess the human relevance of our findings, we used CRISPR-Cas9 gene editing to generate the same mutations in human induced pluripotent stem cells (iPSCs) (fig. S9). As depicted in Fig. 5C, the expression of mutant CBM α1 reduced the colony formation ability of human iPSCs. Concomitantly, this was accompanied by a significant reduction in the expression of stemness markers (both Nanog and Oct4), and transcriptional factors controlling germ layer differentiation (gene MIXL and T for mesoderm, OTX2 and SOX1 for ectoderm, and GATA4 and SOX17 for endoderm) (Fig. 5D). These findings confirm an essential role of the NKA CBM in the regulation of stem cell differentiation and suggest the potential utility of targeting the NKA for improving tissue regeneration.

Expression of CBM mutant α1 NKA affects the dynamic nature of Wnt/β-catenin signaling

The canonical Wnt pathway is made of multiple components localized in the plasma membrane and cytosol (2, 3). Functionally, this pathway is critically important in animal organogenesis (2, 37). For example, it plays an essential role in the establishment of neurogenic niches and regulates the differentiation of neural stem cells into neuroblasts during organogenesis by regulating the expression of transcriptional factors Ngn and NeuroD (37, 38). Thus, we were prompted by the observed neural defects in mice to test whether the expression of the CBM mutant α1 NKA affects Wnt/β-catenin signaling.

In the first set of studies, we examined the cellular distribution of β-catenin in LW-mCBM cells. As depicted in Fig. 6A, confocal imaging analysis showed that β-catenin was distributed away from the plasma membrane in a vesicle-like form in LW-mCBM cells. To verify this finding, we fractionated the cell lysates as performed in Fig. 3B and observed that β-catenin, like Src and caveolin-1, moved from the low-density fractions to high-density fractions when compared to control cells (Fig. 6B). Control experiments showed no changes in the expression of E-cadherin, glycogen synthase kinase–3β (GSK-3β), LRP5/6 (Low-density lipoprotein receptor-related protein 5 and 6), and β-catenin in LW-mCBM cells (Fig. 6C).

Fig. 6 Effects of CBM mutation on Wnt/β-catenin signaling.

(A) β-Catenin staining of AAC-19 and LW-mCBM at basal level (n = 5). Blue arrow indicated β-catenin signal in the cytoplasm of cells. (B) Sucrose gradient fractionation of β-catenin in AAC-19 and LW-mCBM cells (n = 3). **P < 0.01. (C) Western blot analysis of Wnt/β-catenin signaling proteins in AAC-19, LX-α2, and LW-mCBM cells from at least six independent experiments. Two samples from each cell lines are presented. (D) Wnt3a induced TOPFlash luciferase report assay in AAC-19 and LW-mCBM (n = 8). ***P < 0.01. (E) Wnt3a induced expression of Wnt/β-catenin targeting genes (n = 8). **P < 0.01. (F) Wnt3a induced TOPFlash luciferase report assay in AAC-19, LX-α2, and LW-mCBM cells (n = 4). ***P < 0.01.

To test whether these changes in β-catenin distribution alter the function of canonical Wnt signaling, we conducted a TOPFlash luciferase activity assay (39). Cells were transiently transfected with the reporter plasmid, exposed to Wnt3a conditional medium, and then subjected to TOPFlash luciferase assays. As shown in Fig. 6D, while Wnt3a induced a greater than 35-fold increase in luciferase activity in AAC-19 cells, it only produced a fourfold increase in LW-mCBM cells, which equates to an approximate 90% reduction in the dynamics of Wnt activation. To further test the impact of the CBM mutation on Wnt signaling, we examined the effects of Wnt3a on the expression of Wnt target genes. Cells were exposed to Wnt3a for 6 hours and subjected to RT-qPCR analysis. As depicted in Fig. 6E, while Wnt3a increased the expression of c-Myc, Lef, and NKD1 expression in AAC-19 cells, it failed to do so in LW-mCBM cells.

Expression of α2 NKA rescues the dynamics of Wnt signaling

On the basis of the above observations, we reasoned that the NKA CBM might play an essential role in the dynamic regulation of Wnt signaling. We therefore analyzed Wnt signaling in our LX-α2 cell line. This cell line was made by the same strategy used for the generation of LW-mCBM cells, and it expresses essentially just the α2 isoform (40). We have observed that α2 NKA, like CBM mutant α1, maintains cellular pumping capacity but is unable to signal via Src like a wild-type α1 NKA (40). However, unlike CBM mutant α1, α2 does contain the same CBM at the N terminus (fig. S6). As depicted in Fig. 6F, expression of the α2 isoform produced a rescue of Wnt signaling dynamics when compared to that in LW-mCBM cells, which reinforces the idea that the NKA CBM is key to the dynamics of Wnt signaling. Like in LW-mCBM cells, no change in β-catenin expression was noted in LX-α2 cells. However, compared to LW-mCBM cells, caveolin-1 expression was decreased in LX-α2 cells, while ERK activity was increased (Fig. 6C). Together, these findings suggest that the conserved NKA CBM is essential for regulating Wnt signaling, which is independent of the pumping or CTS (ardiotonic steroid)–activated Src-dependent signaling transduction.

Evidence for transcriptional defects of Wnt signaling in homozygous CBM embryos

To see whether there is evidence of Wnt signaling defects in mCBM homozygous embryos, we examined the RNAseq data using a tool kit of pathway analysis. As depicted in fig. S10, Wnt signaling appears to be defective at the transcriptional level. First, the expression of one of the Wnt receptors [Frizzled homolog 5 (Fzd5)] and one of the Wnt ligands (Wnt7b) was down-regulated (fig. S10A). Second, the Wnt/β-catenin signaling inhibitor, secreted frizzled-related protein 5 (Sfrp5), was up-regulated in mCBM homozygous embryos. Third, the β-catenin “destruction complex” component adenomatosis polyposis coli (APC) was down-regulated in mCBM homozygous embryos. All these defects in Wnt signaling were confirmed by RT-qPCR analysis of both wild-type and mCBM homozygous embryos at E9.5 (fig. S10B). In addition, APC down-regulation was also observed at the protein level in mCBM iPSCs (fig. S10C). Last, the defect in Wnt signaling was further substantiated by the altered expression of Wnt downstream target genes. As shown in fig. S10B, the expression of Lef and NKD1 was significantly reduced in mCBM homozygous embryos. The expression of c-Myc was too low to be detected.

Together, these data provide strong support to the notion that the CBM is a key to the regulation of Wnt by the NKA. We hypothesize that this critical function of the NKA CBM may explain why the CBM is conserved in all four α subunit isoforms of the NKA. It is important to mention that the specific molecular defects in Wnt signaling that we have identified were tested in epithelial cells, a model we have previously used to characterize α1-specific signaling functions (16, 41). In view of the cell/tissue specificity of both NKA expression and subunit assemble (42) and Wnt signaling (13, 37), it is likely that this mechanism does not fully explain the Wnt signaling–related defects in embryogenesis.


The enzymatic function of NKA coordinates the transmembrane movement of Na+/K+, which is essential for the survival of individual animal cells. At the tissue/organ level, the ATP-powered transport of Na+/K+ by the NKA is required for neuronal firing, muscle contraction, and the formation and functionality of epithelia and endothelia. The NKA was found to be essential for forming septate junction in Drosophila melanogaster (43, 44) via a regulatory mechanism independent of its ion-pumping activity. Here, we reveal an additional fundamentally important role of NKA in the regulation of signal transduction through a separate functional domain (CBM) unrelated to its enzymatic activity.

Our findings raise the question of why NKA acquired the CBM in addition to its binding sites for Na+ and K+. One possible explanation for this is that the additional functionality in NKA (fulfilled by the CBM) evolved for the purpose of regulating stem cell differentiation and organogenesis in multicellular organisms. Two observations support this hypothesis. First, both Wnt and NKA are present in the first multicellular organisms within the animal kingdom and are evolutionally conserved ever since. Thus, it is likely that the NKA and Wnt work in concert to enable stem cell differentiation and organogenesis in animals. Second, while Wnt is key to the cellular programs of stemness and cell lineage specification (2), it does not directly participate in cell lineage–specific activities of newly differentiated cells. Instead, this particular function might be fulfilled by the NKA. Conceivably, the NKA could have been evolved, as exemplified by the mitochondrial cytochrome c in ATP generation, to bring together two seemingly unrelated processes (i.e., Wnt signaling regulation via the CBM and ion transport through Na+ and K+ binding) into one signaling circuitry, which is critical to the dynamic regulation of transcriptional factors that are required for organogenesis in a temporally and spatially organized manner. Needless to say, this hypothesis remains to be tested. In addition, other important signaling pathways such as Notch and Sonic Hedgehog may also be regulated by NKA.

It is also of interest to note the evolutionary conserveness of the CBM in the plant plasma membrane H-ATPase. Like its counterpart within the animal kingdom, the plasma membrane H-ATPase is essential for plant organogenesis (45). Unlike the NKA, the plasma membrane H-ATPase exists in single-celled organisms such as yeast, and their ion-pumping function is regulated by similar mechanisms (46). However, yeast, with no use for cellular machinery needed for organogenesis, does not contain the H-ATPase with conserved CBM. Moreover, we also observed that no CBM exists in the plasma membrane Ca-ATPase (fig. S6), both of which belong to the same type II P-type ATPase family as the NKA. While the Ca-ATPase is a more ancient protein than the NKA, as its expression can be found in unicellular organisms, the H/K-ATPase appeared later than the NKA, at some point during the development of vertebrates. Thus, we suggest that the NKA may have evolved from a P-ATPase of unicellular organisms via the gain of both the CBM and Na+/K+ binding sites. In contrast, the H/K-ATPase may have evolved from the NKA, losing not only the Na+ binding site but also the CBM.

We have shown a direct interaction between the NKA and caveolin-1 (8, 17), which has been independently confirmed (47). The loss of the CBM significantly reduced the interaction between NKA and caveolin-1 as revealed by multiple assays. In addition to caveolin-1, we and others have reported several signal transduction–related interactions (48). Of these, the potential interaction between α1 NKA and Src has attracted the most attention, especially in the past 10 years (7). While most studies indicated an important role of Src in CTS-activated signal transduction via α1 NKA, several publications have questioned whether α1 NKA interacts with Src directly to regulate Src functionality (49, 50). While this important difference remains to be experimentally addressed, we would like to point out the following facts. First, while we recognize the merit of using purified protein preparation to study protein interaction, it is important to recognize the limitation of using purified Src from bacterial expression system because they are heterogeneously phosphorylated. Second, we have reported multiple lines of evidence that support a direct interaction between α1 NKA and Src, including the identification of isoform-specific Src interaction, the mapping of potential Src-interacting sites in the α1 isoform, and the development of pNaKtide as Src inhibitor and receptor antagonist. These findings have substantially increased our understanding of α1 NKA/Src interaction in cell biology and animal physiology. It is important to mention that several groups not associated with us have successfully used pNaKtide to block ouabain and NKA signaling in vitro and in vivo (2326, 51). While our group and others continue to characterize the molecular basis and biological function of the NKA/Src receptor complex, we propound that the question of NKA/caveolin-1 interaction is a more pressing one in the context of this study. The role of CBM in caveolin-protein interaction and caveolae-related signaling is still debated (41, 52, 53).

Last, we conclude from these interesting findings that the NKA is not just an ion pump or a CBM-directed regulator but a critical multifunctional protein. This whole functionality underlies a hitherto unrecognized common mechanism essential for stem cell differentiation and organogenesis in multicellular organisms within the animal kingdom. Moreover, many recent studies also support the concept that the α1 NKA has acquired more functional motifs (e.g., Src-binding sites for the formation of NKA/Src receptor complex) during evolution. In addition, we have demonstrated that either knockdown of α1 NKA or the expression of an N-terminal fragment containing the CBM of the α1 subunit was sufficient to attenuate purinergic calcium signaling in renal epithelial cells (54). The α1 NKA is also found to be essential for CD36 and CD40 signaling in macrophages and renal epithelial cells (55, 56). Aside from the profound biological and fundamental implications, the previously unidentified NKA-mediated regulation of Wnt signaling through its N-terminal CBM may have substantial implications in our understanding of disease progression. The rapidly increasing appreciation of Wnt signaling in the pathogenesis of cancer and cardiovascular diseases (2, 3, 38) underlies the potential utility of NKA as a multidrug target (12, 22, 57, 58).


Experimental model

mCBM mice. Animal protocols were approved by the Marshall University Institutional Animal Care and Use Committee (IACUC) according to National Institutes of Health (NIH) guidelines. All mCBM heterozygous mice were backcrossed to C57BL6 for at least six generations. Female mCBM heterozygous mice (12 to 16 weeks old) were crossed with male mCBM heterozygous mice. Pregnant female mice at the indicated stage were humanely euthanized. Embryos at different ages were collected for morphology and gene analyses.

NKA α+/− mice. Animal protocols were approved by the Marshall University IACUC according to NIH guidelines. Female NKA α+/− mice were crossed with male NKA α+/− mice. Pregnant female mice at the indicated stage were humanely euthanized. Embryos at E9.5 were collected for morphology and gene analyses.

Cell lines. The parental LLC-PK1 cells were purchased from the American Type Culture Collection (ATCC). LLC-PK1, NKA knockdown (PY-17), α1 NKA–rescued (AAC-19), α1 NKA–mutant cell lines (LW-mCBM), and LX-α2 were all cultured in Dulbecco’s modified Eagle’s medium (DMEM) plus 10% fetal bovine serum (FBS) and 10% penicillin/streptomycin. After cells reached 95 to 100% confluence, they were serum-starved overnight and used for experiments unless indicated otherwise.

iPSC culture and generation of mutant human iPSCs by CRISPR-Cas9 genome editing human iPSCs were purchased from iXCells (catalog no. 30HU-002) and cultured with the TeSR-E8 Kit (STEMCELL Technologies Inc., catalog no. 05940). The DNA plasmids for Cas9–green fluorescent protein (GFP) (Addgene, no. 44719) and guide RNA (gRNA) cloning vector (Addgene, no. 41824) were ordered from Addgene. sgDNA (single-guide DNA) and ssODN (single stranded oligodeoxynucleotides) were designed according to the published protocol. The DNA sequences of primer oligos and their locations in the genome were included in fig. S9. The cloned Cas9-GFP vector inserted with sgDNA, along with ssODN, was transfected into human iPSCs with a 4D nucleofector device. Single-cell sorting and plating were later performed with fluorescence-activated cell sorting (FACS) flow cytometry, and the clones were selected and validated via genotyping PCR and DNA sequencing (fig. S9) (Table 1).

Table 1 Reagent and resource table.

View this table:

Generation of CBM mutation C. elegans and rescue of the mutant with wild-type eat-6. We used CRISPR-Cas9 to knock in the equivalent double mutations of the NKA α1 CBM mutation F75A and F78A into the C. elegans α1 gene eat-6 (syb575 allele) (fig. S8). Comparable to the impact of CBM mutant α1 NKA expression in mice, no live homozygous adults were obtained, whereas the heterozygous worms hatched normally. Moreover, by using the gene balancer nT1 [qIs51], we confirmed that the larva arrest occurred in syb575 homozygous animals because of L1 arrest (Fig. 5A). The defect of larval arrest by CBM mutation was rescued by a transgene expressing a wild-type eat-6 cDNA through an extrachromosomal array. The transgenic construct Peat-6::eat-6::unc-54 3′ untranslated region (SunyBiotech) was generated by fusion of 1963 base pairs (bp) upstream of the translation start of eat-6 with the open reading sequence eat-6 cDNA in the vector pPD49.78.

Method details

Immunoprecipitation and immunoblot analysis. Cells were washed with ice-cold phosphate-buffered saline (PBS) and then solubilized in modified radioimmunoprecipitation assay (RIPA) buffer containing 0.25% sodium deoxycholate, 1% NP-40, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 1 mM NaF, aprotinin (10 μg/ml), leupeptin (10 μg/ml), 150 mM NaCl, and 50 mM tris-HCl (pH 7.4). Cell lysates were centrifuged at 14,000 rpm for 15 min; supernatants were collected. After determining protein concentration by Lowry method, equal amounts of protein were loaded in each lane of freshly prepared SDS–polyacrylamide gel electrophoresis gel. Separated protein was transferred onto nitrocellulose membrane, blocked with milk or bovine serum albumin solution in a mixture of TBST (tris-buffered saline and Tween 20) for 1 hour, and then incubated with primary antibody in blocking solution for overnight. Membranes were washed three times with TBST, incubated with secondary antibody in blocking solution for 1 hour, and washed again. Proteins were visualized with enhanced chemiluminescence reagent (Thermo Fisher Scientific) and autoradiography films. Intensity of bands was quantified with ImageJ software (NIH). Immunoprecipitation was performed by adding 8 μg of anti–caveolin-1 (Cell Signaling, catalog no. 3267) antibody to 500 μg of cell lysate (1 μg/μl concentration). After overnight incubation at 4°C in a rotating shaker, protein G–agarose beads were added for 1 hour. The beads were washed three times with RIPA buffer, and proteins were dissociated from beads by incubating the beads with sample loading buffer in a 55°C shaking waterbath. Proteins were then subjected to Western blot as described above.

Immunostaining. Cells were grown on coverslips until 80 to 90% confluence and then fixed with prechilled methanol for 15 min. The fixed cells were blocked with an Image-iT FX signal enhancer (Invitrogen) and incubated with anti–α1 NKA antibody (Millipore, catalog no. 05-369) or anti–β-catenin (BD Biosciences, catalog no. 610153) overnight at 4°C, followed by incubation with Alexa Fluor 488–conjugated anti-mouse antibody. The stained cells on coverslips were washed, mounted, and then visualized using a Leica confocal SP5 microscope (Leica Microscopes, Wetzlar, Germany). The images were processed with the Leica (LAS/AF) suite (Wetzlar, Germany).

Cell growth assay. Cell growth assay was performed as previously described (59). Briefly, 20,000 cells per well were seeded in triplicate in 12-well plates in DMEM containing 10% FBS. Cells were serum-starved for at least 12 hours to achieve cell cycle synchronization. At indicated time points, cells were trypsinized, and the number of cells was counted with hemocytometer.

[3H]Ouabain binding assay. To measure the surface expression of the endogenous pig α1 NKA, [3H]ouabain binding assay was performed as described previously (16). Briefly, cells were cultured in 12-well plates until confluent and serum-starved overnight. Afterward, cells were incubated in K+-free Krebs solution [142.4 mM NaCl, 2.8 mM CaCl2, 0.6 mM NaH2PO4, 1.2 mM MgSO4, 10 mM glucose, 15 mM tris (pH 7.4)] for 15 min and then exposed to 200 nM [3H]ouabain for 30 min at 37°C. After incubation, the cells were washed three times with ice-cold K+ free Krebs solution, solubilized in 0.1 M NaOH and 0.2% SDS, and counted in a scintillation counter for [3H]ouabain. Nonspecific binding was measured in the presence of 1 mM unlabeled ouabain and subtracted from total binding. All counts were normalized to protein amount.

ATPase activity assay. Cells were harvested in Skou C buffer (30 mM histidine, 250 mM sucrose, 1 mM EDTA, pH 7.4) and briefly sonicated. After centrifugation (800g for 10 min), the post-nuclear fraction was further centrifuged (100,000g for 45 min) to get crude membrane. The crude membrane pellet was resuspended in Skou C buffer and treated with alamethicin (0.1 mg/mg of protein) for 10 min at room temperature. The preparation was then incubated in the buffer containing 50 mM tris (pH 7.4), 1 mM EGTA, 3 mM MgCl2, 25 mM KCl, 100 mM NaCl, 5 mM NaN3, and 2 mM ATP. Phosphate generated during the ATP hydrolysis was measured by BIOMOL Green Reagent (Enzo Life Sciences). Ouabain-sensitive NKA activities were calculated as the difference between the presence and absence of 1 mM ouabain. In kinetic studies, the enzymatic activity was measured as the function of varying concentrations of ouabain, Na+, and K+ as indicated.

Purification of caveolin-rich membrane fractions. Caveolin-rich membrane fraction was obtained by means of sucrose gradient fractionation as described previously (8). In brief, cells were washed with ice-cold PBS and collected into 2 ml of 500 mM sodium carbonate (pH 11.0). The cell lysates were then homogenized by Polytron tissue grinder, followed with sonication. The homogenate was then adjusted to 45% sucrose by addition of 2 ml of 90% sucrose prepared in MBS [MES (-”2-(N-morpholino)ethanesulfonic acidbuffered) saline] (25 mM MES and 0.15 M NaCl, pH 6.5) and placed at the bottom of an ultracentrifuge tube. The ultracentrifuge tubes were then loaded with 4 ml of 35% sucrose and 4 ml of 5% sucrose (both in MBS containing 250 mM sodium carbonate) and centrifuged at 39,000 rpm for 16 to 20 hours in an SW41 rotor (Beckman Coulter). A light-scattering band at the interface between the 5 and 35% sucrose gradients was observed. Eleven gradient fractions of 1 ml were collected from the top to the bottom of the centrifuge tube. Fractions 4 and 5 were combined and diluted and considered as caveolin-enriched fraction. Equal volumes of each fraction were analyzed by Western blot.

Site-directed mutagenesis and generation of mutant-rescued stable cell lines. Site-directed mutagenesis was done by QuikChange mutagenesis kit using pRc/CMV-α1 AACm1 vector as a template (41). The pRc/CMV-α1 AACm1 carries silent mutations that prevent the binding of α1-specific small interfering RNA (siRNA) to the transcript of rat α1 mRNA (41). The mutations were verified by DNA sequencing.

Mutant cell lines used in this work were derived from LLC-PK1 cells. The generation of α1 NKA knockdown PY-17 cells from LLC-PK1 was well described (41). PY-17 cells express about 8% of α1 NKA in comparison to that in LLC-PK1 cells and do not express other isoforms of NKA. They are used to generate wild-type and mutant-rescued cell lines. The generation and characterization of the α1-rescued AAC-19 cells and α2-rescued LX-α2 cells have been reported (16, 60). The expression level of wild-type as well as mutant rat α1 is comparable to that of pig α1 in LLC-PK1 cells. In addition, these cell lines exhibit ouabain-sensitive ATPase activity comparable to that in LLC-PK1 cells. Moreover, the expression of endogenous pig α1 in these cell lines is much lower than that in PY-17 cells, making it possible to study the functionality of mutant α1 NKA without much interference from the endogenous pig α1 (16, 41, 60). The same strategy was used to generate a stable cell line used in this work.

To generate the stable cell line (LW-mCBM), rat α1 NKA cDNA with two point mutations at F97A and F100A (numbered according to GenBank accession no. NM_01254) was transfected to PY-17 cells. Selection was initiated with 3 μM ouabain because endogenous pig NKA α1 in PY-17 cells is very sensitive to ouabain and only cells transfected by rat NKA α1 could survive. One week later, ouabain-resistant colonies were isolated and expanded into stable cell lines in the absence of ouabain.

Generation of mice expressing the mutant CBM α1 NKA. The F97A and F100A amino acid substitutions were introduced into the α1 isoform by 4-bp mutations in exon 4 by using PCR site-directed mutagenesis. Furthermore, three silent base pair substitutions were made to introduce restriction sites Pyu II and Nhe I. The LoxP-neomycin-LoxP cassette was cloned into intron 3 at a site 886 bp upstream of exon 4. The thymidine kinase gene was inserted downstream of the cloning sequence. The presence of the desired mutation in the targeting vector was verified by sequence analysis as well as Pyu II and Nhe I restriction digestion. The Duffy ES (embryonic stem) cell line was transfected with the targeting vector by electroporation, and successful homologous recombination was confirmed by Southern blot analysis. The 5′ probe was derived from intron 3, and the 3′ probe was derived from intron 4 of the α1 isoform gene. The presence of the Pyu II/Nhe I site and desired mutation in exon 4 was determined in successfully targeted ES cells by PCR using the P1 (5′-TGAGTCTTGTCTGCCACTGAG-3′) and P2 (5′-TCCAAGCCTCCCTGAGTTAAC-3′) primers, and the amplified exon 4 fragments were treated with either Pyu II and Nhe I restriction enzyme. Successfully targeted ES cells were subjected to the second electroporation with the Cre-recombinase encoding vector. Following the transfection, ES clones were duplicated and one set was grown in media supplemented with G418 (250 μg/ml). Genomic DNA was extracted from the neomycin-sensitive ES cells, digested with Pyu II or Nhe I, and analyzed by Southern blot using the 3′ and 5′ probes described above. The presence of the Pyu II and Nhe I sites, and therefore the presence of the desired mutations, in exon 4 was analyzed in targeted ES cells by PCR, as described above. Chimeric mice were generated via blastocyst injection of the positive ES cell clone and were bred to C57BL6 mice for two generations to establish a hybrid line. Genotyping was performed by allele-specific PCR analysis across the LoxP site in intron 3 of the genomic DNA. The presence of the mutation in heterozygous F/A mice was verified by sequence analysis of exon 4 using P1 and the P2 primers described above.

Isolation of mouse embryos. Pregnant mice at certain stages were euthanized by CO2 asphyxiation. Then, we opened the abdominal cavity and located the reproductive organs in the dorsal region of the body cavity: two uterine horns, the oviduct, and the ovaries. The uterine horn was removed by grasping the uterus below the oviduct and cut free along the mesometrium. After separating each embryo by cutting between implantation sites along uterine horn, decidua was exposed by peeling the surrounding muscle layer from each embryo. Then, we clipped off a portion of the exposed decidua at the apex (approximately 20% of the entire decidua tissue), which will expose the midventral or distal tip of the enclosed embryo. The embryo was then shelled out using the tips of forceps, and Reichart’s membrane was carefully removed and saved for genotyping.

TOPFlash luciferase assay. AAC-19, LW-mCBM, and LX-α2 plated on six-well plates were transfected with 1 μg of TOPFlash or FOPFlash plasmid and 100 ng of pRenilla as reporter control using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Eight hours after transfection, cells were changed to normal medium and incubated overnight. The second day morning, cells were stimulated with either Wnt3a conditional medium or control medium for 6 hours. The luciferase assay was performed using the dual luciferase assay reporter kit according to the manufacturer’s instructions (Promega).

RNAseq and RT-qPCR analysis. RNAseq analysis was performed by BGI Tech, USA. Briefly, total RNA was extracted from mouse embryos or cell lysates using RNeasy Mini Kit (Qiagen). mRNA was then fragmented with fragmentation buffer, and cDNAs were synthesized using them as templates. After agarose gel electrophoresis, suitable fragments were selected for PCR amplification as templates. During QC steps, Agilent 2100 Bioanalyzer and ABI StepOnePlus Real-Time PCR System were used for quantification and qualification of the sample library. The library was sequenced using HiSeq 2000 sequencer. Bioinformatic analysis was performed by deep analysis of gene expression. For RT-qPCR, after measuring the concentration of total RNA, the SuperScript III First-Strand Synthesis SuperMix for RT-qPCR (Thermo Fisher Scientific) was used for synthesizing first-strand cDNA. The cDNA from each sample was used as a template for the qPCR (SYBR Green) with Roche 480. All primers were synthesized by Integrated DNA Technologies. β-Actin was used as internal control.

Statistical analysis

Data were recorded as means ± SEM. Student’s t test were used to measure differences between two individual groups, and one-way analysis of variance (ANOVA) was used to measure differences between more than two groups. One-way ANOVA (Bartlett’s test) or paired t test was used to measure differences in the tissue array data, where appropriate. Paired t test followed by Wilcoxon signed-rank test was used to analyze gene expression data from TCGA database. Survival analysis was measured by log-rank survival test. P value less than 0.05 was considered as significant.

Correction (15 December 2020): In Fig. 3E, the second image on the top (Wild-type) and third image on the bottom erroneously contained duplicate images of Na/K-ATPase α1 (+/-) E9.5 embryos. The PDF and HTML (full text) versions have been updated with the correct figure.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: Funding: This work was supported by grants from: National Institutes of Health (NIH) Research Enhancement Award (R15) (R15 HL 145666); American Heart Association (AHA) Scientist Development Grant (#17SDG33661117); Brickstreet Foundation and the Huntington Foundation, which provide discretionary funds to the Joan C. Edwards School of Medicine. (These funds are both in the form of endowments that are held by Marshall University). Author contributions: Conceptualization: Z.X., X.W., J.X.X., L.C., G.-Z.Z., S.V.P., and J.I.S.; methodology: X.W., L.C., I.L., D.W., and G.-Z.Z.; investigation: X.W., L.C., X.C., J.W., Y.C., and J.Z.; writing (original draft): X.W., J.X.X., and Z.X.; writing (review and editing): Z.X., J.X.X., L.C., J.I.S., S.V.P., D.W., G.-Z.Z., and X.W.; funding acquisition: Z.X.; visualization: X.W. and Z.X. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article