Placental defects lead to embryonic lethality in mice lacking the Formin and PCP proteins Daam1 and Daam2

The actin cytoskeleton plays a central role in establishing cell polarity and shape during embryonic morphogenesis. Daam1, a member of the Formin family of actin cytoskeleton regulators, is a Dvl2-binding protein that functions in the Wnt/Planar Cell Polarity (PCP) pathway. To examine the role of the Daam proteins in mammalian development, we generated Daam-deficient mice by gene targeting and found that Daam1, but not Daam2, is necessary for fetal survival. Embryonic development of Daam1 mutants was delayed most likely due to functional defects in the labyrinthine layer of the placenta. Examination of Daam2 and Daam1/2 double mutants revealed that Daam1 and Daam2 are functionally redundant during placental development. Of note, neural tube closure defects (NTD), which are observed in several mammalian PCP mutants, are not observed in Wnt5a or Daam1 single mutants, but arise in Daam1;Wnt5a double mutants. These findings demonstrate a unique function for Daam genes in placental development and are consistent with a role for Daam1 in the Wnt/PCP pathway in mammals.


Introduction
The actin cytoskeleton plays a central role in the morphogenesis of the mammalian embryo by controlling cell shape, division, polarity and movement. Previous studies revealed that Rho-GTPases are important regulators of the actin cytoskeleton [1,2] and are essential for embryogenesis because loss of the Rho family members Rac1 or Cdc42 leads to early embryonic lethality [3,4]. One signaling pathway known to control the activity of Rho family proteins during mammalian development is the Wnt/Planar Cell Polarity (PCP) pathway [5]. Dishevelled2 (Dvl2) is a cytoplasmic phosphoprotein that has common roles in transducing Wnt signals. Dvl2 possesses three conserved functional domains, the N-terminal DIX domain, a central PDZ domain, and a C-terminal DEP domain (reviewed in [6,7]). The DIX domain is essential for transducing canonical Wnt signals through the Wnt/βcatenin pathway, whereas the PDZ and DEP domains function in the Wnt/PCP pathway [8][9][10][11][12]. Core components of the PCP pathway, which include the aforementioned Dvl, as well as the frizzled Wnt receptors, the transmembrane protein Van Gogh, the LIM-domain containing protein prickle, the G-protein coupled protein flamingo, and the ankyrin repeat-containing protein diego, were first identified in Drosophila where they were shown to control the direction of cells in the wing epidermis and eye. Establishment of PCP depends on the asymmetric localization of the PCP core components in these cells (Reviewed in [13]). Rho and Rac function downstream of these PCP core genes to control tissue polarity. In Xenopus and zebrafish, RhoA, Rac1 and Cdc42, along with the PCP pathway, control the convergence and extension (CE) movements responsible for the cell intercalations in the midline that drive the extension of the embryonic anterior-posterior (A-P) axis (reviewed in [5,[14][15][16][17]). Asymmetric localization of PCP core proteins was also observed in the cells undergoing the CE movements [18,19].
Mice with mutations in core PCP genes exhibit a common set of phenotypes, including defects in the cochlea, polarity of multiciliated cells, Left-Right (LR) axis formation, A-P axis elongation, neural tube closure, and cardiac outflow tract septation [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36]. Asymmetric localization of core PCP proteins in hair cells of the cochlea, multiciliated cells, and node cells has been reported. Disruption of these PCP proteins perturbed the orientation of the cells in the inner cochlea and multiciliated cells, or cilia position in the node in a fashion that is analogous to the PCP defects observed in the fly wing [22, 23, 27-29, 31-33, 37-39]. On the other hand, the mammalian A-P axis and neural tube closure defects are more similar to the CE movement defects observed when the PCP pathway is perturbed in Xenopus and zebrafish. These data suggest that the Wnt/PCP pathway has adapted to function in several different developmental contexts in diverse organisms.
These functional adaptations may have arisen from the regulation of different Rho GTPase effectors in a given cell type or tissue (reviewed in [40,41]). Daam1 is a member of the Formin subfamily of Rho GTPase effectors that was originally identified as a Dvl2 binding protein [42]. Formins contain a GTPase binding domain (GBD) at the N-terminus, and the Formin homology domains (FH) 1 and 2 at the C-terminus of the protein. Formins are auto-inhibited by interactions between the N-terminus diaphanous inhibitory domain (DID) and the C-terminus diaphanous autoregulatory domain (DAD). Binding of Rho GTPases to the GBD domain alleviates auto-inhibition, exposing the FH2 functional domain to increase nucleation of the actin cytoskeleton (reviewed in [43][44][45]). The N-terminal region of the Daam1 protein containing the GBD and DID domains has dominant negative activity, whereas the C-terminal region of Daam1, which contains the FH2 domain, is constitutively active [42,46]. In mammals, a second member of this highly conserved family, Daam2, exhibits both distinct and overlapping expression patterns with Daam1 during early mouse development [47]. Daam1 interacts with the PDZ and DEP domains of Dvl2 to control CE in Xenopus and zebrafish, consistent with a role of Daam1 in the Wnt/PCP pathway [42,46,[48][49][50].
It is unclear whether Drosophila DAAM participates in the establishment of PCP in the wing or eye, but it was demonstrated that DAAM regulates the actin cytoskeleton in the tracheal cuticle and the axonal growth cone [51,52]. Drosophila DAAM interacts genetically with RhoA and Rac GTPases [51,52], suggesting that Daam proteins have conserved functions as effectors of Rho GTPases. To determine whether Daam proteins function in the mammalian Wnt/PCP pathway, we generated loss-of-function (LOF) mutations in Daam1 and 2. We found that neither Daam1 nor Daam2 mutants display phenotypes characteristic of PCP mutants. However, genetic interactions between Daam1 and Wnt5a or Vangl2 are consistent with Daam1 having a role in the noncanonical Wnt pathway. Of note, the Daam1 mutant had abnormal connections between maternal blood sinuses and fetal blood vessels in the placenta, and Daam1/2 double mutants had a more severe phenotype. Our results demonstrate that Daam genes are essential genes with redundantly functions in placental development.
This study was carried out in compliance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by Frederick National Lab Animal Care and Use Committee (Proposal #17-408). Rodents were euthanized by CO 2 inhalation in accordance with the most recent AVMA Guidelines on Euthanasia. For the harvesting of viable embryonic tissue, the embryos were harvested, placed in cold PBS, cooled down on ice and decapitated before harvesting of fetal livers.

Western blot
Whole embryonic day 10.5 (E10.5) embryos were dissected in cold PBS. Placentas were dissected and separated from the maternal portion. These samples were subsequently homogenized on ice in TNE (50 mM Tris-HCl (pH 7.5), 0.15 M NaCl, 1 mM EDTA, and 1% NP-40 with protease inhibitors Complete Mini (Roche)), centrifuged, and the supernatants were used. Protein concentrations were measured using the Bio-Rad Protein Assay (Bio-Rad) according to the manufacturer's protocol. Proteins were separated using 4-15% Ready Gel (Bio-Rad), transferred to Immobilon transfer membrane (MILLIPORE), incubated with antibodies, and detected using the ECL Plus Western Blotting Detection System (GE Healthcare).

Immunocytochemistry
Immunocytochemistry and stress fiber quantification were performed as previously described [46]. Briefly, the plasmids were transfected into NIH3T3 cells using Polyfect reagent (Qiagen). The cells were fixed, permeabilized, and stained with anti-Myc antibody and phalloidin. For quantification of effects on actin fibers, a baseline of ten fibers/cell was used. Cells containing greater or less than ten fibers were scored as a normal or depleted stress fiber, respectively.

Xenopus embryo manipulation
Microinjection and manipulation of Xenopus embryos were performed as described previously [46]. Embryos with a curved and shortened axis were scored as mild defective embryos and embryos with an open neural tube and curved shortened axis were scored as severely defective embryos. The collective total number of injected embryos from all experiments is indicated below each bar.

Whole-mount immunohistochemistry
Embryos were fixed with 4%PFA/PBS and stained as previously described [56].

Isolation of primary mouse embryonic fibroblasts (MEFs)
MEFs were prepared from E12.5 embryos by an established procedure [57]. Cell proliferation was measured using <4-passage cells in 96 well plates using the CellTiter96 Aqueous One Solution Cell proliferation Assay (MTS assay) (Promega) according to the manufacturer's protocol.

Reverse Transcription and quantitative PCR (qPCR)
Total RNA was purified using TRIzol (Invitrogen) according to the manufacturer's recommendations. One microgram of RNA was treated with Dnase1 and first strand cDNA was synthesized using oligo dT and superscript III reverse transcriptase (Invitrogen). cDNA was quantitated by qPCR using the CFX96 real-Time PCR Detection System (Bio-Rad) and Fast Start Universal SYBR Green Master (Roche). The specificity of all primers was monitored by electrophoresis of the amplicons on agarose gels. The mean expression values obtained for each gene were normalized to GAPDH (ΔΔC (t) method) and to the expression of Daam1 in the lowest expressing sample. The following gene specific oligos were used.

In situ hybridization
Section in situ hybridization was performed as previously described [47,58]. Unless indicated otherwise, at least 4 mutant embryos were examined for each probe, and all yielded similar results.

Histology
Embryos and placentas were dissected in ice-cold PBS and fixed with 4% PFA/PBS overnight at 4˚C and transferred to 70% ethanol/saline for paraffin sections or 20% sucrose/PBS for frozen sections. Paraffin and frozen sections were cut at 8-μm and 10-μm thickness, respectively. Sections were stained with Hematoxylin and Eosin (H&E) following standard protocols.

Plasmids
Daam1 vectors were constructed by standard restriction digestion or PCR amplification cloning techniques, and subcloned into the pCS2+MT vector. The plasmids designated as Daam1 FL, Daam1 N, and Daam1 Δ contain cDNA fragments that correspond to a.a.1-1077, 1-420, and 1-148 of the mouse Daam1 protein, respectively.

GST pull-down assay
Myc-tagged proteins used in GST pull-down assays were generated by TNT Quick-Coupled Transcription/Translation systems (Promega). GST pull-down assays were performed as described previously [46].

Imaging
Embryos and sections were photographed with a Leica MZFLIII stereoscope (Leica) or a Zeiss AxioPlan2, equipped with a Zeiss Axiocam HR digital camera, and Zeiss Axiovision version 4.5 imaging software (Zeiss).

Statistical analysis
Student t-test and chi-square test were performed using excel software (Microsoft). Wilcoxon Rank-sum test was performed using R software (The R Foundation).

Characterization of Daam1 mutant mice
To address Daam1 function in vivo, we generated the Daam1 null allele (Daam1 Δ ) from Daam1 conditional knockout mice. Heterozygous Daam1 Δ/+ mice are viable, fertile, and have no obvious phenotypes. Homozygous Daam1 Δ/Δ offspring arising from Daam1 Δ/+ intercrosses were not found at birth (Table 1). Western blot analysis of whole embryo lysates using an antibody against the C-terminus of Daam1 confirmed that Daam1 Δ/Δ embryos did not express the full-length Daam1 protein, and that heterozygotes expressed roughly half the amount (S1A We therefore concluded that the Daam1 Δ allele is functionally null. Analysis of Daam1 Δ/Δ embryos demonstrated that embryo viability declined after E12.5, and live embryos were no longer observed by E16.5 (Table 1). E12.5-14.5 Daam1 Δ/Δ embryos were small and pale compared with wild-type and Daam1 Δ/+ littermates ( Fig 1A, 1B, 1D and 1E). Reduced liver mass was apparent in whole Daam1 Δ/Δ embryos by E14.5 ( Fig 1C and 1F), but reduced numbers of total fetal liver cells were quantified as early as E12.5 (2.2-fold reduction) and E13.5 (4-fold reduction) compared with wild-type and Daam1 Δ/+ littermates. The differentiation status of hematopoietic cells of the mutant fetal liver was assessed by flow cytometry. Daam1 Δ/Δ embryos exhibited normal differentiation of the T cell, B cell, and myeloid cell lineages, however erythrocyte differentiation was perturbed. Using CD71 and Ter119 expression as markers of erythrocyte differentiation [59], we found that Daam1 Δ/Δ embryos had fewer differentiated, mature erythrocyte cells, than wild-type and Daam1 Δ/+ embryos at E12.5 ( Fig 1G). The reduction of mature erythrocytes may have been caused by fewer erythrocyte progenitors or an inability to differentiate. To examine these possibilities, we performed an erythroid burst-forming units (BFU-E) colony formation assay. Daam1 Δ/Δ and control embryos had similar numbers of erythrocyte progenitors (Fig 1H). To examine their differentiation capacity in vivo, fetal liver cells from wild-type and mutant embryos were transplanted into lethally irradiated wild-type hosts. Engraftment of Daam1 Δ/Δ fetal liver cells leading to the viability of recipient hosts was observed (S1 Table), demonstrating that Daam1 Δ/Δ hematopoietic progenitor cells retain the ability to differentiate into functional erythrocytes. These results suggest that the reduced viability of Daam1 Δ/Δ embryos was not due to intrinsic defects in erythropoiesis or fetal liver function.

Delayed development of Daam1 Δ/Δ mutants due to placental defects
As Daam1 Δ/Δ embryos were smaller than their littermates, we considered the possibility that Daam1 might regulates cell proliferation. MEFs were established from E12.5 embryos, and cell proliferation was analyzed by MTS assay. No difference was observed in the number of Daam1 Δ/Δ MEFs compared with wild-type and Daam1 Δ/+ MEFs (Fig 1I), indicating that Daam1 does not play a major role in cell proliferation.
In addition to the reduced size of Daam1 Δ/Δ embryos, we also observed edema, which is indicative of a circulatory defect (Fig 1E). Circulatory defects can arise from defects in the vasculature, the heart, or the placenta. As the vasculature had developed normally in Daam1 ΔNeo/ ΔNeo embryos (S3 Fig), we considered a role of Daam1 in heart and placental function. As reported in a previous paper [53], myocardial-specific Daam1 conditional KO exhibits myocardial maturation defects, but does not cause embryonic lethality. Analysis of Daam1 protein expression in the placenta by Western blotting confirmed that Daam1 was indeed expressed in the placenta, at a similar level to that observed in embryos, and was undetectable in the Daam1 Δ/Δ placenta (Fig 2A). In situ hybridization of sectioned placenta revealed that Daam1 mRNA was highly expressed in the placenta, although restricted to the labyrinthine layer ( Fig  2B). The labyrinthine layer plays an important role in placental function, because it is where the direct exchange of nutrition and gases between the maternal and fetal blood supplies occurs [60]. Defects in the function of the labyrinthine layer lead to placental insufficiency, which may account for the fetal growth retardation and death. Histological analyses revealed that the spongiotrophoblast layer in Daam1 Δ/Δ placentas were of a similar size and thickness as controls, but the labyrinthine layer in Daam1 Δ/Δ placentas was slightly thinner than in wildtype. (Fig 2C-2F).
Marker analyses using the trophoblast giant cell marker Hand1, the endothelial trophoblast marker Plf, and the spongiotrophoblast marker Tpbpa, revealed no differences in their expression between control and Daam1 Δ/Δ placentas (Fig 2G-2L) suggesting that cell fates were unchanged in Daam1 Δ/Δ placentas. A detailed analysis of the labyrinthine layers revealed thin and refined maternal blood sinuses adjacent to fetal blood vessels in control placentas, however, maternal blood sinuses were enlarged and not entwined with fetal blood vessels in the Daam1 Δ/Δ labyrinthine layer ( Fig 2C'and 2D'). These results suggest that nutritional deficiency caused by reduced placental function was responsible for the delayed development of Daam1 mutants.
To assess whether Daam1 has an intrinsic function in the embryo, we used Meox2-cre to inactivate Daam1 specifically in the epiblast, thereby enabling wild-type function in the placenta. As Daam1 and Meox2 are located on the same chromosome, mutant alleles (Daam1 Flox/ Δ ; Meox2-cre) on different sister chromosomes were rarely inherited together ( Table 2). Notably, Daam1 Flox/Δ ; Meox2-cre mice were viable for >1 year and lacked overt phenotypes ( Table 2). These results are consistent with the absence of intrinsic function by Daam1 during embryonic development, however, the mosaic expression of Meox2-cre in the epiblast [61] remains a caveat. Indeed, Daam1 RNA and protein remain detectable in varying amounts in adult organs, depending on the tissue (ex. 1.3-29.6% of Daam1 RNA level observed in control tissues (S4A and S4B Fig)). Nevertheless, these results together suggest that the embryo lethality of Daam1 Δ/Δ mutants is due to the requirement of Daam1 in the placenta.

Daam1 and Daam2 are functionally redundant in the placenta
Despite the broad pattern of embryonic expression [47], removal of Daam1 only resulted in placental defects. To address the possibility that Daam2 functionally compensates for the loss of Daam1 in the embryo, we used Daam2 LacZ mice. Examination of β-galactosidase activity in Daam2 heterozygous (Daam2 LacZ/+ ) embryos revealed a pattern similar to that of endogenous Daam2 mRNA expression (S5A Fig, and [47]). Daam2 LacZ/LacZ mice were also viable and fertile and had no obvious phenotype.

Discussion
We investigated the role of Daam1 in mammalian development and discovered that Daam1 plays an essential role in murine placentation. The placenta functions as a fetomaternal organ during gestation, enabling nutrient uptake, waste elimination, and gas exchange between the mother and developing conceptus. Most rodents and primates (including humans) have hemochorial or labyrinthine placentas, in which maternal blood directly contacts the fetal chorionic trophoblast. The close apposition of the vascularized uterus and trophoblast in the labyrinth constitutes the major site of maternal and fetal exchange. We demonstrated that Daam1 is expressed in the trophoblast cells of the labyrinthine layer, but not in the embryonic mesoderm-derived vascular cells, and that Daam1 Δ/Δ placentas undergo normal trophoblast cell differentiation but exhibit structural defects in the labyrinthine layer. As gas and nutrient exchange relies upon diffusion across the placental membrane (reviewed in [64,65]), we suggest that the increased separation between maternal blood sinuses and fetal blood vessels observed in the Daam1 Δ/Δ labyrinthine layer (Fig 2D') is sufficient to impair the efficiency of the maternal/fetal exchange, and is therefore likely to result in fetal death. The observation that the embryo-specific deletion of Daam1 did not impair fetal development is consistent with a requisite role for Daam1 in placental trophoblast cells. How does Daam1 control the maternal blood sinus and fetal blood vessel connection? One possible explanation is that Daam1 controls trophoblast branching morphogenesis in the labyrinthine layer which is essential for the vascularization of the labyrinth, as the Wnt/PCP genes have been reported to be involved in branching morphogenesis of other epithelial tissues [66][67][68][69]. Another possibility is that Daam1 controls the maternal blood sinus and fetal blood vessel connection by stabilizing cellcell junctions. Daam1 interacts with the E-cadherin-β-catenin-α-catenin complex at the lateral membrane contact sites to stabilize the interactions between epithelial cells [70]. Similar molecular mechanisms might operate at the maternal blood sinus and fetal blood vessel connection. Indeed, the absence of E-cadherin or β-catenin in the placenta also causes vascularization defects in the labyrinthine layers [71,72]. Embryos lacking both Daam1 and Daam2 died earlier than single Daam1 mutants, but they did not have additional embryonic phenotypes. Daam1 Δ/Δ , Daam2 LacZ/LacZ placentas had more marked vascularization defects than Daam1 Δ/Δ placentas suggesting that the earlier embryo lethality arises from the more severe placental defects. mDia1, a paralog of Daam1/2, was reported to interact with RhoA, B, and/or C to control actin assembly, similar to Daam1 [73,74]: reviewed in [45,75]. mDia1 deficient mice also had no developmental defects [76], raising the possibility that these family members have redundant functions. Future studies are needed to address whether mDia1, Daam1 and Daam2 compensate for one another.
Our data demonstrating that the reduced Daam1 gene dosage worsened the vaginal atresia phenotype in Vangl2 Lpt/+ mice and caused NTDs in Wnt5a -/embryos. Most of the NTDs reported in the PCP core factors mutant mice were craniorachischisis, complete open neural tube [20,21,28,29,[34][35][36]62], however the NTD observed in Daam1 Δ/ Δ ;Wnt5a -/embryos was opened only at the hindbrain. Similar partial NTDs were observed compound heterozygous of PCP core factor mutants [35] and Dvl2 single null mice [21] with low penetrance. It is likely that the partial NTDs are mild phenotype of NTDs caused by the PCP pathway aberration. The NTD observed in Daam1 Δ/ Δ ;Wnt5a -/embryos suggested that Daam1 functioning, at least in part, in the Wnt/PCP pathway. This is supported by a report demonstrating a key role for Daam1 in the Wnt/PCP pathway during chick neural tube closure [77]. However, neither Daam1 Δ/Δ nor Daam1 Δ/Δ , Daam2 LacZ/LacZ mouse embryos exhibited neural tube closure defects and the early embryonic lethality observed in Daam1 Δ/Δ , Daam2 LacZ/LacZ embryos precludes our ability to fully evaluate the role of the Daam1 gene in the PCP pathway. A report of hypomorphic Daam1 mutants, which survive to later developmental stages, described cardiac outflow tract septation defects that are commonly found in PCP mutants [78]. As severe circulation defects from placental abnormality in Daam1 Δ/Δ and Daam1 Δ/Δ , Daam2 LacZ/LacZ mouse embryos caused midgestational lethality, we were unable to assess heart developmental defects in Daam1 and Daam2 null mutants. Our previous study of cardiomyocyte-specific conditional Daam1 knockouts revealed indispensable roles of Daam1 in cardiac development [53]. We did not observe heart defects in Daam1 Flox/Δ ; Meox2-cre mice presumably due to leaky expression of Daam1 caused by the mosaic expression of Meox2-cre. The low frequency of surviving Daam1 Flox/Δ ; Meox2-cre adult mice (3 out of 83 mice) compared with the frequency of Daam1 Flox/Δ ; Meox2-cre embryos (3 out of 37 embryos) suggest that some mutants did not survive to adulthood.
In addition to the neural tube and cardiac outflow tract, the Wnt/PCP pathway was reported to regulate tube formation in the kidney, gut, and female reproductive tract [25, [79][80][81][82][83][84]. Tissuespecific inactivation of Daam1 and Daam2 in these organs will provide an opportunity to analyze the role of Daam proteins in these PCP-dependent tissues at later developmental stages.
A role for the PCP pathway in placental development has not been previously described. However, mice deficient in Wnt2 or the Wnt receptor Fz5 were found to have vascularization defects in the labyrinthine layer of the placenta similar to the Daam1/2 mutant phenotype [85,86]. Wnt2 is expressed in embryo-derived allantoic mesoderm cells, whereas Fz5, like Daam1, is expressed in the extraembryonic labyrinth layer. During vascularization, interactions between trophoblast and embryonic mesoderm cells are important for forming capillary structures. We speculate that Wnt ligands provided by the allantois stimulate trophoblast cells in a paracrine manner to regulate capillary formation. Of note, Wnt2 is known to regulate the Wnt/β-catenin pathway in the developing lung [87], and Daam2 participates in the dorsal patterning of the spinal cord via the Wnt/β-catenin pathway [88]. Together, these results suggest that Daam1 and Daam2 could function in different Wnt pathways in a tissue-dependent manner. Another possibility is that Daam1 and Daam2 function in different signaling pathways to control placental development. Studies on Daam1 and Daam2 functions in cardiomyocytes and endothelial cells [53,89] suggested that Daam1 has roles in the activation of the Src-Akt-Gsk3β pathway. Indeed, mice deficient in Akt1 also exhibit vascularization defects in the labyrinthine layer of the placenta that are similar to the Daam1/2 mutant phenotype [90].
We demonstrated that Daam1 and Daam2 play essential roles in placental vascularization and the establishment of the maternal/fetal blood supply. Feto-placental problems are proposed to be one of the major causes of pre-eclampsia, intrauterine growth restriction and miscarriage [91]. Further analysis of the mechanisms through which Daam functions in the placenta may improve our understanding of placental pathologies.