Cables2 Is a Novel Smad2-Regulatory Factor Essential for Early Embryonic Development in Mice

CDK5 and Abl enzyme substrate 2 (Cables2), a member of the Cables family that has a C-terminal cyclin box-like domain, is widely expressed in adult mouse tissues. However, the physiological role of Cables2 in vivo is unknown. We show here that Cables2-deficiency causes post-gastrulation embryonic lethality in mice. The mutant embryos progress to gastrulation, but then arrest, and fail to grow. Analysis of gene expression patterns reveals that formation of the anterior visceral endoderm and the primitive streak is impaired in Cables2-deficient embryos. Tetraploid complementation analyses support the critical requirement of Cables2 in both the epiblast and visceral endoderm for progression of embryogenesis. In addition, we show that Cables2 physically interacts with a key mediator of the canonical Nodal pathway, Smad2, and augments its transcriptional activity. These findings provide novel insights into the essential role of Cables2 in the early embryonic development in mice.

physiological role of Cables2 in vivo is unknown. We show here that Cables2-deficiency 48 causes post-gastrulation embryonic lethality in mice. The mutant embryos progress to 49 gastrulation, but then arrest, and fail to grow. Analysis of gene expression patterns reveals 50 that formation of the anterior visceral endoderm and the primitive streak is impaired in 51 Cables2-deficient embryos. Tetraploid complementation analyses support the critical 52 requirement of Cables2 in both the epiblast and visceral endoderm for progression of 53 embryogenesis. In addition, we show that Cables2 physically interacts with a key mediator of 54 the canonical Nodal pathway, Smad2, and augments its transcriptional activity. These 55 findings provide novel insights into the essential role of Cables2 in the early embryonic 56 development in mice. 57

INTRODUCTION
12 catenin/TCF-mediated transcription in vitro with an almost two-fold increase in relative 253 TOP/FOP luciferase activity ( Figure 4A). Moreover, co-IP using N-terminal FLAG-tagged 254 Cables2 (FLAG-Cables2)-transfected 293FT cell lysates with or without exogenous β-catenin 255 indicated that β-catenin was present in the precipitated complexes with Cables2 ( Figure 4B  256 and Figure 4-figure supplement 1). These data suggest that Cables2 can physically 257 associated with β-catenin and increases its transcriptional activity at Wnt-responsive genes.  These results, in conjunction with the AVE analyses, support our notion that loss of Cables2 275 13 also impairs distal/anterior axis formation and suggest that Cables2 is involved in 276 reinforcement of Nodal/Smad2 signalling in the AVE. 277 We next investigated whether Cables2 enhanced the transcriptional activity of Smad2 by 278 luciferase reporter assay with the ARE-luc vector, which expresses firefly luciferase in a 279 Smad2-and FAST1-dependent manner. We co-transfected 293FT cells with the ARE-luc, 280 FAST1 expression vectors and a constitutive-active mutant form of TGF-β receptor together 281 with an empty vector or FLAG-Cables2 expression vector. The data indicated that forced 282 expression of Cables2 facilitated Smad2 activity in 293FT cells ( Figure 5K). Intriguingly,283 Cables2 had no effect on luciferase activity of the Smad3/4-specific reporter vector 284 (CAGA)9-luc (Dennler et al., 1998) ( Figure 5K), suggesting that Cables2 does not function 285 as an activator for Smad3. Luciferase assays also revealed that Cables1 had no effect on the 286 transcriptional activities of both Smad2 and Smad3 ( Figure 5K). These findings suggest that 287 the promoting activity on Smad2 function is a unique property of Cables2 rather than a 288 conserved function of the Cables family. We further conducted co-IP experiments of Cables2 289 with exogenous and endogenous Smad2 and found that Cables2 physically interacted with 290  showed GO term enrichments related to "nervous system development" and "negative 324 regulation of cell proliferation" and KEGG pathway enrichments related to several signalling 325 pathway including "p53 signalling pathway" among 122 out of 125 upregulated genes in 326 Cables2-deficient EpiLCs (Figure 6-figure supplement 3A and B). While, 52 out of 59 327 downregulated genes in Cables2-deficient EpiLCs represented GO terms related to 328 "proximal/distal pattern formation", "positive regulation of cell proliferation", and "male 329 gonad development" and KEGG pathway related to "signalling pathways regulating 330 pluripotency of stem cells" (Figure 6-figure supplement 4A and B). WISH data showed that Nanog was expressed in the posterior region of wild-type embryos at 336 E6.5 ( Figure 6C). In contrast, Cables2-deficient embryos showed only a low level of Nanog 337 gene expression over the whole epiblast region at E6.5 ( Figure 6D). On the other hand, the 338 pluripotency marker, Oct4, was expressed normally in Cables2 -/mutants at E6.5 ( Figure 6E  Cables2 -/embryos showed no elevation of Nanog-GFP expression at the PS ( Figure 6G and 343 H). Nevertheless, there was no difference in GFP expression between embryos of the same 344 litter at E5.5 ( Figure 6I and J), suggesting no detectable phenotype of Nanog before PS 345 formation. Overall, loss of Cables2 leads to downregulated Nanog expression at the start of 346 gastrulation. 347

Requirement of Cables2 in both epiblast and VE for the proper gastrulation in mice 348
To determine whether Cables2 is required in the VE, the epiblast, or both, chimera analysis 349 was performed using tetraploid wild-type embryos and Cables2 -/-ESCs. In tetraploid 350 complementation chimera has the advantage that the host tetraploid embryos can only 351 contribute to primitive endoderm derivatives and trophoblast compartment of the placenta, . We also demonstrated that Cables2 interacted with Smad2 in nuclear and 476 cytoplasm (Figure 5M and N). These results suggest that Cables2 functions as a nuclear 477 cofactor for Smad2 although the mechanism underlying enhancement of Tween 20) and digested with proteinase K (1 mg/mL) at 55°C for 2 hours, inactivated at 95 520 °C for 5 minutes, and then subjected to PCR. For paraffin slides, embryos were genotyped 521 using tissue picked from sections and digested directly with proteinase K (2 mg/mL) in PBS. 522 For others experiments, after collecting data, the whole embryos were used for genotyping. 523 Genotyping PCR was performed with AmpliTag Gold 360 Master Mix (Thermo Fisher 524 Scientific K.K., Tokyo, Japan) using the following primers: Cables2 D3-1: 5'-525 ACTGCAGAAGCTGGAGGAAA-3'; Cables2 D3-2: 5'-TCAAGGTGTCTGCCCTATCC-3'; 526 Cables2 D3-3: 5'-AGGGGATCCGCTGTAAGTCT-3'. 527

Indirect immunofluorescence assay (IFA) 709
After 24 or 48 hours transfection, cells were washed twice with PBS and then fixed with 4% 710 paraformaldehyde in PBS for 10 minutes. Permeabilization of cell membranes were done with 711 0.1% Triton X-100 in PBS for 20 minutes or methanol for 5 minutes. After blocking with 10% 712 goat serum or Superblock blocking buffer (Thermo Fisher Scientific) for 30 minutes, cells were 713 incubated overnight at 4°C with mouse anti-FLAG antibody. Then, the cells were washed with 714 PBS and incubated with Alexa Fluor-conjugated anti-mouse IgG antibody. Fluoresence signals 715 were detected using a BZ-X700 fluorescent microscope (Keyence Co., Ltd., Osaka, Japan). 716

Histology, EdU, and TUNEL assay 717
Mouse uteri including the decidua were collected and fixed in 4% paraformaldehyde in PBS. 718 Subsequently, paraffin blocks were made by dehydration in ethanol, clearing in xylene, and 719 embedding in paraffin. Embryo sections 5 µm thick were cut (Microm HM 335E; Thermo 720 Fisher Scientific) and placed on glass slides (Matsunami Glass Ind., Ltd., Osaka, Japan). For 721 haematoxylin-eosin (HE) staining, slides were deparaffinized and rehydrated through an 722 ethanol series, and then stained with HE. 723 To label the proliferating embryonic cells, pregnant mice at E6.5 were injected 724 intraperitoneally with 5-ethynyl-2'-deoxyuridine (EdU) at 200 µL/mouse and sacrificed 4 725 hours later. Embryos were embedded in paraffin blocks, and sections were refixed in 4% 726 paraformaldehyde and permeabilized in 0.5% Triton X-100/PBS. EdU assay was performed 727 with a Click-iT Plus EdU Imaging Kit (Thermo Fisher Scientific) and TUNEL assay was 728 performed with a Click-iT Plus TUNEL Assay for In situ Apoptosis Detection kit (Thermo 729 Fisher Scientific) according to the manufacturer's protocol. As the final step, embryo sections 730 were co-stained with Hoechst 33342, observed under a microscope (BZ-X700; Keyence), and 731 cell number was counted using ImageJ software. 732