Associationofdysfunctionalsynapsedefective1(SYDE1 ...download.xuebalib.com/xuebalib.com.41392.pdf · Biosciences, Mountain View, CA, USA) to generate pCDH-SYDE1-WT and pCDH-SYDE2-WT

Journal of PathologyJ Pathol 2017; 241: 324–336Published online 15 December 2016 in Wiley Online Library(wileyonlinelibrary.com) DOI: 10.1002/path.4835

ORIGINAL PAPER

Association of dysfunctional synapse defective 1 (SYDE1)with restricted fetal growth – SYDE1 regulates placental cellmigration and invasionHsiao-Fan Lo,1 Ching-Yen Tsai,2 Chie-Pein Chen,3 Liang-Jie Wang,4 Yun-Shien Lee,5 Chia-Yu Chen,3 Chung-TiangLiang,6 Mei-Leng Cheong7 and Hungwen Chen1,4*

1 Graduate Institute of Biochemical Sciences, National Taiwan University, Taipei 106, Taiwan2 Institute of Molecular Biology, Academia Sinica, Nankang, Taipei 115, Taiwan3 Division of High Risk Pregnancy, Mackay Memorial Hospital, Taipei 104, Taiwan4 Institute of Biological Chemistry, Academia Sinica, Nankang, Taipei 115, Taiwan5 Department of Biotechnology, Ming Chuan University, Tao-Yuan, Taiwan6 National Laboratory Animal Center, Taipei, Taiwan7 Department of Obstetrics and Gynecology, Cathay General Hospital, Taipei 106, Taiwan

*Correspondence to: H Chen, PhD, Institute of Biological Chemistry, Academia Sinica, Nankang, Taipei 115, Taiwan. E-mail:[email protected]

AbstractThe transcription factor glial cells missing 1 (GCM1) regulates trophoblast differentiation and function duringplacentation. Decreased GCM1 expression is associated with pre-eclampsia, suggesting that abnormal expressionof GCM1 target genes may contribute to the pathogenesis of pregnancy complications. Here we identified anovel GCM1 target gene, synapse defective 1 (SYDE1), which encodes a RhoGAP that is highly expressed inhuman placenta, and demonstrated that SYDE1 promotes cytoskeletal remodelling and cell migration and invasion.Importantly, genetic ablation of murine Syde1 results in small fetuses and placentas with aberrant phenotypes inthe placental–yolk sac barrier, maternal–trophoblast interface, and placental vascularization. Microarray analysisrevealed altered expression of renin-1, angiotensin I converting enzyme 2, angiotensin II type 1a receptor,and membrane metalloendopeptidase of the renin–angiotensin system in Syde1-knockout placenta, which maycompensate for the vascular defects to maintain normal blood pressure. As pregnancy proceeds, growth restrictionof the Syde1−/− fetuses and placentas continues, with elevated expression of the Syde1 homologue Syde2 inplacenta. Syde2 may compensate for the loss of Syde1 function because SYDE2, but not the GAP-dead SYDE2mutant, reverses migration and invasion activities of SYDE1-knockdown JAR trophoblast cells. Clinically, we furtherdetected decreased SYDE1 expression in preterm and term IUGR placentas compared with gestational age-matchedcontrols. Our study suggests a novel mechanism for GCM1 and SYDE1 in regulation of trophoblast cell migrationand invasion during placental development and that decreased SYDE1 expression is associated with IUGR.Copyright © 2016 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Keywords: SYDE1; GCM1; cell migration/invasion; IUGR, placenta

Received 18 May 2016; Revised 26 September 2016; Accepted 16 October 2016

No conflicts of interest were declared.

Introduction

The placenta is a temporary organ specialized for hor-mone production and maternal–fetal transportation ofnutrients and gases during pregnancy. During placentaldevelopment, trophoblast cells at the anterior end of thevilli that anchor the decidua proliferate and migrate toinvade the decidua. These migratory and invasive tro-phoblast cells, termed extravillous trophoblast (EVT)cells, may remodel spiral arteries to facilitate bloodcirculation between the mother and the fetus. Humanpregnancy complications are frequently associated withplacental disorders leading to maternal and neonatalmorbidity and mortality. Maternal and fetal symptoms

are found in pre-eclampsia (PE) and intrauterine growthretardation (IUGR) including hypertension, proteinuria,and poor fetal growth. The aetiology of PE and IUGRremains elusive because of their multifactorial nature;however, defective trophoblast invasion resulting in poorplacental development is believed to be a major cul-prit. Accordingly, a two-stage disorder theory has beenproposed whereby abnormal placental development andthereby insufficient placental perfusion produce factors(e.g. anti-angiogenic factors) resulting in clinical mani-festations of PE [1,2].

The human transcription factor glial cells miss-ing 1 (GCM1) is expressed primarily in the villouscytotrophoblast (CTB), syncytiotrophoblast (STB), and

Copyright © 2016 Pathological Society of Great Britain and Ireland. J Pathol 2017; 241: 324–336Published by John Wiley & Sons, Ltd. www.pathsoc.org www.thejournalofpathology.com

SYDE1 regulates placental development 325

EVT cells [3,4]. GCM1 is a master regulator of tro-phoblast differentiation and functionality [5–7]. It maytransactivate expression of the syncytin gene (ERVW-1)to facilitate the differentiation of the multinucleatedSTB layer which covers the outer surface of placentalvilli; hCGβ gene (CGB1) expression for the synthesisof hCG, which is essential for corpus luteum mainte-nance [8–10]; and HTRA4 gene expression to promotetrophoblast cell invasion [11].

The Rho family of small GTPases comprises abranch of the Ras superfamily and functions as binaryswitches between active GTP-bound and inactiveGDP-bound states. The active GTP-bound Rho GTPasesinteract with various downstream effectors, such asp21-activated kinase (PAK) and Rho-associated kinase(ROCK), to regulate many cellular activities includingcell cycle progression, actin polymerization, membranetrafficking, and cell migration [12,13]. Specifically,RhoA, Cdc42, and Rac1 are known to regulate therearrangements of the membrane-associated actincytoskeleton. Rho GTPase activity can be modulated byRho guanine nucleotide-exchange factors (RhoGEFs),which promote the exchange of GDP for GTP; RhoGTPase-activating proteins (RhoGAPs), which enhanceintrinsic GTPase activity; and Rho GDP-dissociationinhibitors (RhoGDIs), which modulate GTPase activ-ity by inhibiting nucleotide exchange and membraneassociation [14]. RhoGAPs contain a conserved GAPdomain of approximately 140 amino acids. A con-served arginine residue in the GAP domain is insertedinto the active site of Rho GTPases to stabilize aconformation required for GTP hydrolysis. Further-more, the Cdc42/Rac1 downstream effector PAK mayphosphorylate RhoGEFs to negatively regulate Rhoactivity [15].

A variety of factors have been reported to regulatetrophoblast cell migration and invasion. For example,matrix metalloproteinases and Wnt proteins have beenreported to regulate trophoblast cell migration and inva-sion [16–18]. Change in integrin expression, termedintegrin switching, by up-regulation of α1β1 and α5β1integrin and loss of E-cadherin expression have beenshown in EVT cells to facilitate their migration andinvasion [19,20]. Rac1, Cdc42, and ROCK are involvedin prostaglandin E2-stimulated trophoblast cell migra-tion, and RhoGDI2 inhibits trophoblast cell migration bysuppression of Rac1 activity [21–23]. Decreased Rac1activities have been observed in pre-eclamptic placentasand genetic ablation of mouse Rock2 leads to placentaldysfunction and IUGR [24–26].

We reason that the functional roles of GCM1 inthe regulation of placental cell migration and invasionare multifaceted through transactivation of differenttarget genes. Here, we identify SYDE1 (synapse defec-tive Rho GTPase homologue 1), which encodes aRhoGAP, as a novel GCM1 target gene in ChIP-chipexperiments. We demonstrate that SYDE1 promotesplacental cell migration and invasion and is crucial forplacental development. Importantly, retarded placentaland fetal growth with a defective spongiotrophoblast

layer, compromised vascularization, and an abnor-mal maternal–trophoblast interface are noted in theSyde1 homozygous knockout (KO) placenta. Similarly,decreased SYDE1 expression is observed in humanIUGR placentas. We further demonstrated that com-ponents of the renin–angiotensin system (RAS) and aSyde1 homologue, Syde2, are differentially expressedin Syde1-KO placenta, which might contribute to nor-mal neonatal delivery in Syde1-KO mothers. Our studyidentifies SYDE1 as a novel regulator for trophoblastmigration and invasion and supports abnormal SYDE1expression being associated with IUGR.

Materials and methods

Plasmid constructsDNA fragments encoding wild-type human SYDE1and SYDE2 with a C-terminal HA tag were clonedinto a pCDH lentiviral expression vector (SystemBiosciences, Mountain View, CA, USA) to generatepCDH-SYDE1-WT and pCDH-SYDE2-WT expres-sion plasmids, respectively. An arginine-to-alaninemutation was introduced into the conserved arginineresidue at the GAP active site of SYDE1 (R436)and SYDE2 (R854) to generate pCDH-SYDE1-MTand pCDH-SYDE2-MT expression plasmids for theGAP enzyme-dead SYDE1 and SYDE2 mutants. Agenomic fragment containing nucleotides from −500to +138 relative to the transcriptional start site of theSYDE1 gene was cloned into pGL3-E1b to generate thepGL3-E1b(−500/+138) reporter plasmid. Four potentialGCM1-binding sites (GBS1–4) in the above-mentionedSYDE1 promoter region were mutated individually orin combination by site-directed mutagenesis to generatepGL3-E1bMT reporter plasmids (MT1, −2, −3, −4, andMT3+ 4).

Cell culture, transfection, and lentivirustransductionCell culture, preparation of human EVT and CTB cells,and establishment of cells stably expressing scramble orGCM1 shRNA have been described previously [10,11].For transient expression, cells were transfected withthe indicated reporter and expression plasmids usingLipofectamine 2000 reagent (Invitrogen, Carlsbad, CA,USA). Luciferase assays were performed as describedpreviously [9]. JAR cells stably expressing SYDE1-WT,SYDE1-MT, SYDE2-WT or SYDE2-MT were estab-lished with lentiviruses harbouring the above-describedpCDH expression constructs.

Chromatin immunoprecipitation (ChIP) analysisChromatin immunoprecipitation-on-chip (ChIP-chip)experiments performed in BeWo31 cells stably express-ing HA-GCM1 have been described previously [11]. Tostudy the association between GCM1 and the SYDE1


326 H-F Lo et al

promoter, BeWo or CTB cells were subjected to ChIPassay using rabbit anti-GCM1 Ab. The precipitatedprotein–DNA complexes were analysed by PCR withspecific primers for a specific region containing theGBS sequences in the SYDE1 promoter. The primersequences were 5’-GCTGCGATTTGGAGACAGA-3’and 5’-TGCTGGCAGATGTAGTTGC-3’.

Reverse transcription-quantitative PCR (RT-qPCR)RNA isolation, reverse transcription, and qPCRwere performed as described previously [10]. Thesequences of the primers used are given in the sup-plementary material, Supplementary methods andmaterials.

Immunohistochemistry and immunofluorescencemicroscopyExpression of GCM1 and SYDE1 in human placenta,co-expression of GCM1 and SYDE1 in CTB and EVTcells, and analysis of F-actin filament distribution andfocal adhesion were performed as described in the Sup-plementary methods and materials (see supplementarymaterial).

Cell migration, cell invasion, and Rho GTPaseactivation assaysJAR cells stably expressing empty pCDH vector,SYDE1 or SYDE1-MT were seeded into fibronectin-and Matrigel-coated Transwells (BD Biosciences,San Jose, CA, USA), respectively. Quantification ofmigrated or invasive cells has been described previously[11]. In a separate experiment, SYDE1-knockdownJAR cells were transduced with lentiviruses expressingSYDE2-WT or SYDE2-MT and then subjected to cellmigration assays. Activation of Rho, Rac1, and Cdc42was measured in JAR cells stably expressing emptypCDH vector, SYDE1 or SYDE1-MT with commercialkits (EMD Millipore, Billerica, MA, USA) accordingto the manufacturer’s instructions.

Gene targeting and phenotypic analysisFor construction of the targeting vector, neomycinresistance gene, loxP, and FRT sites were insertedinto designated sites in the Syde1 gene of a mouse(C57BL/6 J) BAC clone via recombination (GeneBridges, Heidelberg, Germany). The targeting vectorwas linearized and electroporated into C57BL/6 J-Tyr/J(C2J) embryonic stem cells to generate Syde1 condi-tional knockout mice, which were further bred with EIIaCre recombinase-expressing (EIIa-Cre) mice for gen-eration of Syde1 complete knockout (Syde1-KO) mice.Mouse breeding was conducted at the Academia SinicaSPF animal facility according to the IACUC-approvedprotocol (RMiRbGPIBCCH2011015). Phenotypic anal-yses of Syde1-KO mice were performed as describedin the Supplementary methods and materials (seesupplementary material).

SYDE1 expression in IUGR placentasPlacental tissue biopsies were collected after deliv-ery; all experiments were approved by the IRB ofMackay Memorial Hospital, Taipei, Taiwan. The diag-nosis of IUGR was made according to the criteriaof fetal birth weight being in the bottom tenth per-centile with abnormal umbilical artery Doppler wave-forms and reduced amniotic fluid volumes (amnioticfluid index≤ 5). Deliveries from spontaneous-labourpreterm and term pregnancies with grossly normal pla-centas matching the closest gestational age of IUGRwere selected as controls. To study SYDE1 expression inIUGR placentas, preterm or term normal and IUGR pla-cental tissue sections were immunostained with SYDE1(HPA013328, 1:200 dilution; Sigma-Aldrich, St Louis,MO, USA) and cytokeratin 7 (CK7, MAB3226, 1:200dilution; EMD Millipore, Darmstadt, Germany) primaryAbs and then incubated with a secondary Ab conju-gated with AlexaFluor 488 or 568 for immunofluores-cence microscopy. In a separate experiment, pretermor term normal and IUGR placental tissue biopsieswere subjected to immunoblotting analysis with SYDE1(1:1000 dilution), SYDE2 (HPA027138, 1:1000 dilu-tion; Sigma-Aldrich), and GCM1 [11] (1:1000 dilution)antibodies. Densitometric analysis of immunoblot bandintensities was performed using ImageJ software.

Gene expression profilingWild-type and Syde1-KO placentas at E14.5 were col-lected for RNA purification for microarray analysis withthe Affymetrix (Santa Clara, CA, USA) mouse genome430 2.0 array according to the manufacturer’s protocol.A stringent analysis was performed to select genes thatwere expressed in Syde1-KO placentas at a higher levelthan in wild-type counterparts (>2-fold with p< 0.005),giving a total of 160 genes. We performed pathway anal-ysis of the 160 genes with DAVID bioinformatics toolsbased on the KEGG PATHWAY database.

Expression of RAS genes in wild-type and Syde1-KOplacentas was studied by RT-qPCR and immunoblot-ting analyses. For immunoblotting analysis, renin-1 Ab(sc-133145, 1:2000 dilution) was obtained from SantaCruz Biotechnology (Santa Cruz, CA, USA); Agtr1a(ab124505, 1:1000 dilution) and Ace2 (ab108252,1:1000 dilution) Abs were from Abcam (Cambridge,MA, USA); and Mme Ab (10302-1-AP, 1:1000 dilution)was from Proteintech (Rosemont, IL, USA).

All raw microarray data have been deposited inNCBI’s Gene Expression Omnibus (GEO) and areaccessible through GEO Series accession numberGSE81468 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE81468).

Statistical analysisStatistical analysis of the data was performed usingStudent’s t-test or the Mann–Whitney test. Statisticalsignificance was classified as *p< 0.05, **p< 0.01 or***p< 0.001.


http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE81468http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE81468


Figure 1. GCM1 regulates SYDE1 gene expression. (A) Identification of SYDE1 as a GCM1 target gene. BeWo31 cells expressing HA-GCM1were subjected to ChIP-chip analysis for GCM1 target genes using HA Ab (positive) or normal mouse IgG (control). Note that chip signalswere detected in the SYDE1 promoter region. (B) Regulation of SYDE1 expression by GCM1. BeWo and CTB cells were subjected to ChIPanalysis using GCM1 Ab or normal rabbit IgG (upper-left panel). The immunopurified GCM1-associated genomic fragments were subjectedto PCR amplification with a primer set for the SYDE1 promoter region. BeWo (middle and upper-right panels) or purified term EVT (lowerpanel) cells expressing scramble (Scr) or GCM1 shRNA were harvested for RT-qPCR analysis of SYDE1 mRNA level or immunoblottinganalysis of SYDE1 and GCM1 proteins. (C) Identification of GCM1-response elements in the SYDE1 promoter. Four tentative GCM-bindingsites (GBS1–4) are listed in the −500 to +138 region of the SYDE1 promoter. Sequences of wild-type GBS1–4 and their mutants (MT1–4)are listed, with the nucleotides essential for GCM1 binding being mutated and highlighted. 293 T cells were transfected with the indicatedreporter plasmid plus or minus pHA-GCM1. At 48 h post-transfection, cells were harvested for luciferase analysis. Mean values and the SDobtained from three independent experiments are presented. Fold induction was calculated relative to the activity of each reporter plasmidalone. (D) Microarray analysis of SYDE1 mRNA in human tissues. The data set was retrieved from the NCBI GEO database (GDS3113). Notethat SYDE1 is highly expressed in placenta.

Results

SYDE1 is a GCM1 target gene

The SYDE1 RhoGAP was identified as a GCM1 targetgene by ChIP-chip analysis in trophoblast-like BeWo31cells stably expressing HA-tagged GCM1 with HA Ab.Specifically, HA-GCM1 is associated with a genomicregion upstream of the first exon of SYDE1 (Figure 1A).We further confirmed that GCM1 associates with theSYDE1 promoter region by ChIP analysis in BeWoand primary human cytotrophoblast (CTB) cells withGCM1 Ab and that the levels of SYDE1 protein andtranscript are decreased in the GCM1-knockdown BeWocells (Figure 1B, upper panel). The GCM1-response

elements in the SYDE1 promoter were characterized byluciferase reporter assay. Four tentative GCM1-bindingsites (GBS1–4) were identified in the SYDE1 pro-moter and mutated individually or in combinationfor transient expression experiments in 293 T cellstransfected with or without the pHA-GCM1 expres-sion plasmid. As shown in Figure 1C, mutagenesis ofGBS3 or GBS4 or both sites significantly decreasedthe stimulatory effect of HA-GCM1 on the luciferaseactivity directed by the SYDE1 promoter, suggestingthat GBS3 and GBS4 are key response elements forGCM1 to regulate SYDE1 expression. Microarraydata deposited in NCBI’s Gene Expression Omnibus(GEO, GDS3113) indicated that SYDE1 is highlyexpressed in human placenta (Figure 1D). These results


328 H-F Lo et al

Figure 2. Expression of SYDE1 in human placenta. (A) Immunohistochemistry of SYDE1 and GCM1. Sections of first-trimester (8 weeks’gestational age, GA8) and term placentas were immunostained with SYDE1, GCM1 or CK7 Ab and then ImmPRESS reagent containingperoxidase-labelled anti-rabbit IgG or anti-mouse IgG. Antigenic detection was visualized with the chromogenic substrate DAB. Note thatSYDE1 and GCM1 are expressed in the EVT cells of first-trimester and term placentas. (B) Co-expression of SYDE1 and GCM1 in trophoblastcells. Purified CTB and EVT cells were stained with GCM1 and SYDE1 Abs and secondary Abs conjugated with AlexaFluor 488 (for GCM1) or568 (for SYDE1). Nuclei were stained with DAPI. Cells were visualized by confocal microscopy.

collectively suggested that GCM1 may be responsiblefor the high level of SYDE1 mRNA expression inplacenta.

We next examined the expression patterns of GCM1and SYDE1 in human placenta by immunohisto-chemistry. Both GCM1 and SYDE1 were detected inthe first-trimester trophoblast cells of villus and cellcolumns, and strong SYDE1 signals were also detectedin the term EVT cells at the maternal–fetal interface(Figure 2A). Correspondingly, GCM1 knockdownalso decreased SYDE1 expression in term EVT cells(Figure 1B, lower panel). Co-expression of GCM1and SYDE1 in purified term CTB and EVT cells wasfurther confirmed by immunofluorescence microscopy(Figure 2B).

SYDE1 regulates placental cell migrationand invasionBecause GCM1 and SYDE1 are expressed in migratoryEVT cells, we now tested whether SYDE1 is a GCM1downstream effector regulating trophoblastic migra-tion and invasion. We established trophoblast-like JARcells stably expressing SYDE1-WT or SYDE1-MT anddemonstrated that cell migration and invasion are signif-icantly enhanced for JAR cells expressing SYDE1-WT(Figure 3A–C). We also performed gap closure migra-tion assays to confirm that SYDE1 promotes JAR cellmigration in a GAP-dependent manner and that this pos-itive effect is not due to a differential growth rate inJAR cells expressing SYDE1-WT or SYDE1-MT (seesupplementary material, Figure S1). Moreover, SYDE1



Figure 3. SYDE1 stimulates placental cell migration and invasion. (A) Schematic representation of SYDE1 and SYDE2 domain structures. JARcells stably expressing empty pCDH vector, wild-type HA-tagged SYDE1 (SYDE1-WT) or GAP enzyme-dead mutant SYDE1 (SYDE1-MT) weresubjected to immunoblotting with SYDE1, HA, or β-actin Ab. (B, C) SYDE1 promotes JAR cell migration and invasion. The pCDH, SYDE1-WT,and SYDE1-MT JAR cells were plated into fibronectin-coated (B) and Matrigel-coated (C) Transwells for analysis of migration and invasionactivities, respectively. Mean values and the SD obtained from three independent experiments are presented. (D) Regulation of cytoskeletalremodelling by SYDE1. The pCDH, SYDE1-WT, and SYDE1-MT JAR cells were treated with or without 20 μM IPA-3 for 30 min and thensubjected to immunofluorescence staining using paxillin Ab and TRITC-conjugated phalloidin, respectively. Nuclei were stained with DAPI.Enlarged images of the boxed regions are shown. Arrows and arrowheads indicate focal adhesion and filopodial protrusion, respectively. (E)Regulation of Rho GTPase activity by SYDE1. The pCDH, SYDE1-WT, and SYDE1-MT JAR cells were harvested for analysis of Rho, Cdc42, andRac1 activity by pull-down assays with GST-Rhotekin-RBD or GST-PAK1-PBD beads and then immunoblotting with Rho, Rac1, and Cdc42Abs. In a separate experiment, scramble control and SYDE1-knockdown JAR cells were also subjected to Rho GTPase activity assays.


330 H-F Lo et al

knockdown decreased cell migration and invasion in pri-mary human CTB cells (see supplementary material,Figure S2). Our results suggested that SYDE1 may facil-itate placental cell migration and invasion through itsGAP activity.

Regulation of cytoskeletal remodelling by SYDE1Subsequently, we tested whether SYDE1 modulatescell migration through cytoskeletal remodelling bystaining focal adhesions and stress fibres in mock,SYDE1-WT-expressing, and SYDE1-MT-expressingJAR cells using anti-paxillin Ab and TRITC-conjugatedphalloidin, respectively (for details, see supplemen-tary material, Supplementary materials and methods).Compared with mock and SYDE1-MT-expressingJAR cells, focal adhesions were barely detected, stressfibres were disorganized, and more filopodia weredetected in the SYDE1-WT-expressing JAR cells,supporting SYDE1 facilitating cytoskeletal remod-elling for cell migration (Figure 3D). Interestingly,treatment of SYDE1-WT-expressing JAR cells withthe PAK inhibitor IPA-3 maintained focal adhesionsand stress fibres (Figure 3D). We went on to measureRho, Cdc42, and Rac1 activities by pull-down assayand revealed a decrease in Rho activity and activationof Cdc42 and Rac1 in the SYDE1-WT-expressing JARcells (Figure 3E, left panel). Correspondingly, Rhoactivity was increased and Cdc42 and Rac1 activitieswere decreased in the SYDE1-knockdown (SYDE1-KD)JAR cells (Figure 3E, right panel). We also examinedthe effect of IPA-3 on SYDE1-regulated Rho activity.Because IPA-3 released the suppression of RhoGEF byPAK-1 to enhance Rho activity, the Rho-GTP level wasnot different between mock and SYDE1-WT-expressingJAR cells. SYDE1-MT failed to convert Rho-GTP toRho-GDP; therefore, the Rho-GTP level was furtherelevated in the SYDE1-MT-expressing JAR cells byIPA-3 (see supplementary material, Figure S3A). Inthis manner, IPA-3 inhibited SYDE1-WT-expressingJAR cell migration (see supplementary material, FigureS3B). Taken together, these results suggested thatSYDE1 regulates cytoskeletal remodelling in placentalcells through inhibition of Rho and activation of Cdc42and Rac1.

Fetal growth restriction in Syde1-knockout miceWe performed gene targeting to establish Syde1+/loxp

mouse ES cells for generation of Syde1 conditionalknockout mice, which were subsequently crossedwith EIIa-Cre mice to generate Syde1−/− knockout(Syde1-KO) mice (see supplementary material, FigureS4). Gross morphological analysis revealed that theweights of Syde1-KO embryos and placentas are sig-nificantly lower at gestational days E11.5 to E18.5(Figure 4A, B). However, wild-type and Syde1-KOnewborns were delivered normally and no significantdifference in fertility between the two adult groupswas noted (see supplementary material, Figure S5).

While pathological examination did not reveal signifi-cant morphological differences between wild-type andSyde1-KO placentas at E11.5 (Figure 4C, left panel),cystic dysplasia was noted in the spongiotrophoblastlayer of the E14.5 Syde1-KO placenta (Figure 4C,middle panel). Quantitation further supported signifi-cant increases of cystic dysplasia in E14.5 and E18.5Syde1-KO placentas (p< 0.05; Figure 4C, bottom-leftpanel). In addition, spongiotrophoblast dysplasia in thelabyrinthine layer was detected in the E14.5 Syde1-KOplacenta, which was corroborated with staining of theTpbpa marker (Figure 4C, middle and bottom-rightpanels). At E18.5, apoptosis was observed in thespongiotrophoblast layer of Syde1-KO placenta withelevated caspase-3 activity (Figure 4C, right panel).

The placental vascular structures were examinedby immunostaining of the CD34 endothelial cellmarker. Weaker CD34 signals were detected in thelabyrinthine layer of E11.5 Syde1-KO placenta, andnon-vascularized areas (CD34-negative) significantlydeveloped in the labyrinthine layer of E14.5 andE18.5 Syde1-KO placentas compared with the regu-larly distributed CD34-positive cells in the wild-typeplacentas (Figure 4D). Abnormal vascularization inthe labyrinthine layer of Syde1-KO placenta was alsocorroborated by immunostaining of α-smooth muscleactin, which is expressed in vascular smooth musclecells (see supplementary material, Figure S6).

We further studied placental permeability by admin-istration of Evans Blue dye into pregnant wild-typeand Syde1-KO mice at 11.5 days post-coitum (dpc). Asignificant increase in the passage of Evans Blue dyeto the Syde1-KO placenta and yolk sac was observed,suggesting that the formation of a barrier structurebetween the fetal and maternal blood circulation inthe placenta may be compromised by Syde1 knockout(Figure 4E). The ultrastructure of the labyrinthine layerin wild-type and Syde1-KO placentas was compared byelectron microscopy, which revealed increased vacuolesin both STB layers I and II (Figure 4F). Finally, primarySyde1-KO trophoblast cells exhibited decreased inva-sion activity compared with their wild-type counterparts(see supplementary material, Figure S7). Therefore, it islikely that Syde1 deficiency compromises trophoblasticinvasion during placental development with a concomi-tant placental dysfunction.

Decreased expression of SYDE1 in IUGR placentasBecause Syde1 knockout results in fetal growth restric-tion, we surveyed SYDE1 expression in human IUGRplacentas by immunofluorescence microscopy. With theCK7 trophoblast marker as a reference, the stainingsignals of SYDE1 protein were lower in IUGR pla-centas than in normal placentas in three independentstudies (Figure 5A, cases 1–3). We also measured theSYDE1 and GCM1 protein levels by immunoblottingof preterm and term IUGR placentas and their gesta-tional age-matched controls (see supplementary mate-rial, Table S1 for demographic data of the patients). As



Figure 4. Phenotypic analysis of Syde1-knockout mouse. (A, B) Syde1 knockout leads to growth restriction of the embryo and placenta.Wild-type and Syde1-KO (knockout) embryos and placentas at E11.5, E14.5, and E18.5 were collected for weight analysis. (C) Sections ofwild-type and Syde1-KO placentas at different gestational ages (E11.5, E14.5, and E18.5) were subjected to H&E stain. Sections of E14.5and E18.5 Syde1-KO placentas were also immunostained with Tpbpa and cleaved caspase-3 Abs, respectively. Asterisk indicates cysticdysplasia in the spongiotrophoblast (SpT) layer of E14.5 Syde1-KO placenta. Arrows indicate spongiotrophoblast dysplasia. Arrowheadsindicate apoptosis in the spongiotrophoblast (SpT) layer of E18.5 Syde1-KO placenta, where Lab is the labyrinthine layer and TGC is atrophoblast giant cell. Quantitation of cystic areas in the spongiotrophoblast layer at E14.5 and E18.5 and spongiotrophoblast dysplasia(Tpbpa-positive areas) in the labyrinthine layer at E14.5 is presented. (D) Decreased vascularization in the Syde1-KO placenta. Sections ofwild-type and Syde1-KO placentas at E11.5, E14.5, and E18.5 were immunostained with CD34 Ab. Asterisk indicates a non-vascularizedarea in the labyrinthine layer. Quantitation of CD34-positive areas in the labyrinthine layer at different gestational ages is presented. (E)Defective placental barrier in the Syde1-KO placenta. Wild-type and Syde1-KO pregnant mice at 11.5 dpc were injected intraperitoneallywith Evans Blue. The passage of Evans Blue molecules into the placenta and yolk sac was quantitated as described in the Supplementarymethods and materials. (F) Ultrastructural analysis of the wild-type and Syde1-KO labyrinthine layers at E11.5 by transmission electronmicroscopy. Note accumulated vacuoles in the Syde1-KO STB layers I and II. MBS=maternal blood space; FBS= fetal blood space.


332 H-F Lo et al

Figure 5. SYDE1 expression is decreased in human IUGR placentas. (A) Immunofluorescence microscopy of SYDE1 in term normal and IUGRplacentas. Placental sections were subjected to immunostaining with SYDE1 and CK7 Abs. Sections were then stained with a secondaryAb labelled with AlexaFluor 568 (for SYDE1) or AlexaFluor 488 (for CK7). Nuclei were stained by DAPI. Insets are images of CK7 staining.(B) Decreased SYDE1 and SYDE2 expression in preterm and term IUGR placentas. Preterm (35 or 36 weeks) and term (37 or 38 weeks)IUGR and normal placentas were subjected to immunoblotting analysis with SYDE1, SYDE2, GCM1, and GAPDH Abs, respectively. Thenumbers underneath indicate the protein band intensities normalized against GAPDH. Note that the protein levels of SYDE1 and SYDE2 aresignificantly decreased in IUGR placentas (lower panel).



shown in Figure 5B, SYDE1 expression was decreasedin preterm and term IUGR placentas, whereas no sig-nificant difference in GCM1 expression was detectedbetween normal and IUGR placentas.

Differential expression of RAS and Syde2in Syde1-knockout placentaGiven that abnormal vascular structure was observed inSyde1-KO placenta, we failed to detect abnormalities inthe blood pressure of Syde1-KO pregnant mice (data notshown). To further examine the effects of Syde1 knock-out on placental physiology, we compared the geneexpression profiles of E14.5 wild-type and Syde1-KOplacentas by microarray analysis. A total of 160 geneswere identified to be differentially expressed betweenthe wild-type and the Syde1-KO placentas (>2-foldwith p< 0.005) and subjected to pathway analysis withthe KEGG PATHWAY database. Only the RAS path-way was of statistical significance (p= 0.0002), con-taining six RAS-related probe sets for genes encodingmembrane metalloendopeptidase (Mme), angiotensinII type 1a receptor (Agtr1a), angiotensin I convertingenzyme 2 (Ace2), and renin-1 (Figure 6A). We con-firmed the microarray results by RT-qPCR and demon-strated that the Renin1 (Ren1) transcript level wasdecreased and the Ace2, Agtr1a, and Mme transcriptlevels were increased in E14.5 Syde1-KO placentas(Figure 6A). By immunoblotting, a significant decreasein renin-1 precursor and mature renin-1 protein levelswas noted in E11.5, E14.5, and E18.5 Syde1-KO placen-tas (Figure 6B). An increased Ace2 or Mme protein levelwas detected in E14.5 or E18.5 Syde1-KO placentas,whereas increased Agtr1a protein levels were detectedin E14.5 and E18.5 Syde1-KO placentas (Figure 6B).Therefore, the activity of the placental RAS pathwaywas altered in Syde1-KO mice in response to a dysfunc-tional Syde1 condition.

Although Syde1 knockout causes defective placentaldevelopment in early gestational stages (E11.5 andE14.5), less severe phenotypes were observed in thelater gestational stage (E18.5). We were curious abouta possible functional redundancy between Syde1 andSyde2, which is a Syde1 homologue. By immunohis-tochemistry, Syde1 was detected in giant cells and thespongiotrophoblast and labyrinthine layers at E11.5 andE14.5 and barely detected at E18.5, whereas Syde2 wasdetected in the spongiotrophoblast layer at E11.5 andthe spongiotrophoblast layer and glycogen cells at E14.5and E18.5 (see supplementary material, Figure S8A).As pregnancy proceeded, the Syde1 and Syde2 tran-script levels were decreased and increased, respectively,and Syde2 expression was further enhanced in E14.5and E18.5 Syde1-KO placentas (see supplementarymaterial, Figure S8B). A significant increase of Syde2protein expression was also observed in Syde1-KOplacentas (Figure 6C). Introduction of SYDE2-WT, butnot SYDE2-MT, into SYDE1-knockdown (KD) JARcells significantly enhanced the migration activity ofSYDE1-KD JAR cells (Figure 6D). It is very likely that

Syde2 plays a role in the development of Syde1-KOplacenta during pregnancy and therefore contributesto normal neonatal delivery in Syde1-KO pregnantmice. Clinically, SYDE2 expression was significantlydecreased in IUGR placentas (Figure 5B).

Discussion

In the present study, we identified SYDE1 as a novelGCM1 target gene, which encodes a RhoGAP, and pro-vided evidence to support SYDE1 regulating placentalcell migration and invasion. Specifically, we identifiedkey GCM1-response elements, GBS3 and GBS4, inthe SYDE1 promoter based on ChIP-chip analysis andreporter assay. We demonstrated that SYDE1 is highlyexpressed in the human placenta based on data miningof the NCBI GEO database, which is also corroboratedby transcriptome studies by RNA-seq and microarrayanalyses of placenta and seven other types of tissues[27,28], and that SYDE1 and GCM1 are co-expressedin placental trophoblast cells by immunohistochemistryand confocal microscopy. Subsequently, we revealedthat SYDE1 facilitates placental cell migration andinvasion through enhancement of the cytoskeletal rear-rangement by down-regulation of Rho activity andup-regulation of Cdc42 and Rac1 activities.

That SYDE1 plays a crucial role in placental develop-ment is further supported by Syde1 transgenic studies,which showed significantly lower weights of Syde1-KOembryo and placenta at different gestational stages.This observed growth restriction may be attributed todefective placental development based on the followinglines of evidence. First, cystic dysplasia and apoptosiswere detected in the spongiotrophoblast layer of E14.5and E18.5 Syde1-KO placentas. Second, decreasedvascularization in the labyrinthine layer was observedin the E11.5 Syde1-KO placenta and non-vascularizedareas were also noted in the E14.5 and E18.5 Syde1-KOplacentas. Third, the barrier structure between fetaland maternal blood circulation was breached in theSyde1-KO placenta such that injected Evans Bluemolecules passed into the placenta and yolk sac.Fourth, at the cellular level, the invasiveness of primarySyde1-KO trophoblast cells was significantly reduced.We speculated that such impaired trophoblastic invasionmay undermine the infrastructure and functions ofplacenta, leading to growth restriction.

The RAS is crucial for regulation of blood pres-sure and electrolyte homeostasis. Recent studies havesupported a local RAS in placenta and decidua thatis critical to regulating blood flow in utero-placentalcirculation during pregnancy [29,30]. Indeed, aberrantexpression of placental RAS components is found inpregnancy complications such that elevated expres-sion of placental angiotensinogen, angiotensin II,and AGTR1 is detected in PE and reduced placentalAGTR1 expression in IUGR [31,32]. Additionally,cross-mating female transgenic mice expressing human


334 H-F Lo et al

Figure 6. Differential expression of placental RAS and Syde2 in Syde1-KO mice. (A) Gene expression profiling in wild-type and Syde1-KOplacentas. Transcripts were purified from E14.5 wild-type and Syde1-KO placentas for microarray analysis for identification of differentiallyexpressed genes associated with the RAS pathway. Down-regulation of Renin1 (Ren1) gene expression and up-regulation of Ace2, Agtr1a,and Mme gene expression in the Syde1-KO placenta were confirmed by RT-qPCR analysis. A schematic representation of regulation ofvasoconstriction and vasodilation by RAS is provided. (B) Immunoblotting analysis of RAS proteins in wild-type and Syde1-KO placentasat different gestational stages. A representative set of results is shown. Quantification of band signals was performed by densitometry. (C)Elevation of Syde2 expression in the Syde1-KO placenta. Wild-type and Syde1-KO placentas at different gestational stages were subjectedto immunoblotting with SYDE2 Ab. The levels of Syde2 protein were quantified by densitometry. (D) SYDE2 reverses the migration activityof SYDE1-knockdown cells. JAR cells expressing scramble shRNA (Scr) or SYDE1 shRNA (E1 KD) were transduced with lentivirus harbouringan empty pCDH, wild-type, or GAP enzyme-dead SYDE2 (E2 MT) expression cassette. Cells were plated into fibronectin-coated Transwellsfor migration analysis. Mean values and the SD obtained from three independent experiments are presented in A–D.



angiotensinogen with male transgenic mice expressinghuman renin leads to elevated systolic blood pressure,placental defects, and fetal growth restriction, whichare ameliorated when the Agtr1a gene is also deletedin the pregnant transgenic mice [33,34]. Because thehuman Renin transgene is expressed in the placenta, thismay activate RAS through cleavage of angiotensinogenand activation of Agtr1a to cause hypertension. Unlikethe above-mentioned RAS transgenic studies, theblood pressure was not significantly different betweenwild-type and Syde1-KO pregnant mice. Instead, wedetected down-regulation of renin-1 expression andup-regulation of Ace2, Mme, and Agtr1a expressionin Syde1-KO placenta based on microarray analy-sis (Figure 6A). We speculated that the differentialexpression of RAS genes in the placentas of Syde1-KOpregnant mice might be a local response to counteractthe expected elevation of blood pressure caused byplacental defects. In this scenario, decreased renin-1expression may reduce the vasoconstrictive angiotensinII level, whereas elevated Ace2 and Mme expressionmay stimulate the vasodilative angiotensin-(1–7) levelto maintain normal blood pressure in Syde1-KO preg-nant mice. We also detected increased expression of theSyde1 homologue, Syde2, in the Syde1-KO placentaand demonstrated that SYDE2 is able to reverse themigration activity of SYDE1-knockdown JAR placentalcells. Therefore, it is feasible to speculate that increasedSyde2 expression may help to maintain pregnancy inSyde1-KO mice, due to the functional redundancy ofSyde1 and Syde2. Further study is warranted to test thishypothesis by knocking out Syde2 in Syde1-KO mice.

A recent microarray analysis by Trifonova et al [35]has revealed a modest increase in SYDE1 expression inPE. In contrast, decreased SYDE1 and SYDE2 proteinlevels were observed in IUGR placentas in the presentstudy. Because both SYDE1 and SYDE2 modulate pla-cental cell migration, it is highly possible that com-promised SYDE1 and SYDE2 functions contribute tothe development of IUGR. We did not observe statisti-cally significant changes in the GCM1 protein level inIUGR as reported by Bainbridge et al [36]. Neverthe-less, decreased expression of the GCM1 target genes,syncytin-1 and syncytin-2, has also been reported inIUGR [37]. Therefore, the GCM1 protein level may notreflect overall GCM1 activity in IUGR placentas. Sev-eral cellular factors are known to interact with GCM1to regulate its activity at the post-translational level.Caspase-14, p45NF-E2, and GATA3 suppress GCM1activity, whereas CBP and RACK1 enhance GCM1activity [4,38–41]. It will be interesting to investigatethe roles of the aforementioned factors in the regula-tion of GCM1 activity in IUGR. Wentzel et al [42] haverecently reported that mouse Syde1 is involved in theregulation of synaptic vesicle docking at the active zone.By electron microscopy, we found accumulated vac-uoles in STB layers I and II of Syde1-KO placenta. Itwill also be intriguing to further investigate a potentialrole for Syde1 in modulating vesicular trafficking.

Acknowledgements

This work was supported by grants (to HC) fromthe Ministry of Science and Technology (grant No103-2311-B-001-024-MY3), the National HealthResearch Institutes (grant No NHRI-EX100-10049SI),and Academia Sinica, Taiwan.

Author contribution statement

HFL and LJW performed experiments and analysed thedata. CYT, CTL, and YSL were involved in the gener-ation and analysis of knockout mice. CPC, CYC, andMLC were involved in the analysis of human placentalbiopsies. HC designed the study and wrote the paper.

References1. Roberts JM, Hubel CA. The two stage model of preeclampsia:

variations on the theme. Placenta 2009; 30(suppl A): S32–S37.2. Huppertz B. Placental origins of preeclampsia: challenging the cur-

rent hypothesis. Hypertension 2008; 51: 970–975.3. Nait-Oumesmar B, Copperman AB, Lazzarini RA. Placental expres-

sion and chromosomal localization of the human Gcm 1 gene. J His-tochem Cytochem 2000; 48: 915–922.

4. Chiu YH, Chen H. GATA3 inhibits GCM1 activity and trophoblastcell invasion. Sci Rep 2016; 6: 21630.

5. Anson-Cartwright L, Dawson K, Holmyard D, et al. The glial cellsmissing-1 protein is essential for branching morphogenesis in thechorioallantoic placenta. Nature Genet 2000; 25: 311–314.

6. Schreiber J, Riethmacher-Sonnenberg E, Riethmacher D, et al. Pla-cental failure in mice lacking the mammalian homolog of glial cellsmissing, GCMa. Mol Cell Biol 2000; 20: 2466–2474.

7. Chang M, Mukherjea D, Gobble RM, et al. Glial cell missing 1regulates placental growth factor (PGF) gene transcription in humantrophoblast. Biol Reprod 2008; 78: 841–851.

8. Yu C, Shen K, Lin M, et al. GCMa regulates the syncytin-mediatedtrophoblastic fusion. J Biol Chem 2002; 277: 50062–50068.

9. Liang CY, Wang LJ, Chen CP, et al. GCM1 regulation of the expres-sion of syncytin 2 and its cognate receptor MFSD2A in human pla-centa. Biol Reprod 2010; 83: 387–395.

10. Cheong ML, Wang LJ, Chuang PY, et al. A positive feedback loopbetween glial cells missing 1 and human chorionic gonadotropin(hCG) regulates placental hCGβ expression and cell differentiation.Mol Cell Biol 2016; 36: 197–209.

11. Wang LJ, Cheong ML, Lee YS, et al. High-temperature require-ment protein A4 (HtrA4) suppresses the fusogenic activity ofsyncytin-1 and promotes trophoblast invasion. Mol Cell Biol 2012;32: 3707–3717.

12. Kaibuchi K, Kuroda S, Amano M. Regulation of the cytoskeleton andcell adhesion by the Rho family GTPases in mammalian cells. AnnuRev Biochem 1999; 68: 459–486.

13. Hall A. Rho family GTPases. Biochem Soc Trans 2012; 40:1378–1382.

14. Cherfils J, Zeghouf M. Regulation of small GTPases by GEFs, GAPs,and GDIs. Physiol Rev 2013; 93: 269–309.

15. Alberts AS, Qin H, Carr HS, et al. PAK1 negatively regulates theactivity of the Rho exchange factor NET1. J Biol Chem 2005; 280:12152–12161.

16. Gellersen B, Reimann K, Samalecos A, et al. Invasiveness of humanendometrial stromal cells is promoted by decidualization and bytrophoblast-derived signals. Hum Reprod 2010; 25: 862–873.


336 H-F Lo et al

17. Onogi A, Naruse K, Sado T, et al. Hypoxia inhibits invasion ofextravillous trophoblast cells through reduction of matrix metallo-proteinase (MMP)-2 activation in the early first trimester of humanpregnancy. Placenta 2011; 32: 665–670.

18. Pollheimer J, Loregger T, Sonderegger S, et al. Activation of thecanonical wingless/T-cell factor signaling pathway promotes inva-sive differentiation of human trophoblast. Am J Pathol 2006; 168:1134–1147.

19. Damsky CH, Librach C, Lim KH, et al. Integrin switching regulatesnormal trophoblast invasion. Development 1994; 120: 3657–3666.

20. Zhou Y, Fisher SJ, Janatpour M, et al. Human cytotrophoblasts adopta vascular phenotype as they differentiate. A strategy for successfulendovascular invasion? J Clin Invest 1997; 99: 2139–2151.

21. Nicola C, Lala PK, Chakraborty C. Prostaglandin E2-mediatedmigration of human trophoblast requires RAC1 and CDC42. BiolReprod 2008; 78: 976–982.

22. Nicola C, Chirpac A, Lala PK, et al. Roles of Rho guanosine5’-triphosphatase A, Rho kinases, and extracellular signal reg-ulated kinase (1/2) in prostaglandin E2-mediated migration offirst-trimester human extravillous trophoblast. Endocrinology 2008;149: 1243–1251.

23. Liu S, Cui H, Li Q, et al. RhoGDI2 is expressed in human tro-phoblasts and involved in their migration by inhibiting the activationof RAC1. Biol Reprod 2014; 90: 88.

24. Fan M, Xu Y, Hong F, et al. Rac1/beta-catenin signalling pathwaycontributes to trophoblast cell invasion by targeting Snail and MMP9.Cell Physiol Biochem 2016; 38: 1319–1332.

25. Hannke-Lohmann A, Pildner von Steinburg S, Dehne K, et al. Down-regulation of a mitogen-activated protein kinase signaling pathway inthe placentas of women with preeclampsia. Obstet Gynecol 2000; 96:582–587.

26. Thumkeo D, Keel J, Ishizaki T, et al. Targeted disruption of themouse rho-associated kinase 2 gene results in intrauterine growthretardation and fetal death. Mol Cell Biol 2003; 23: 5043–5055.

27. Saben J, Zhong Y, McKelvey S, et al. A comprehensive analysis ofthe human placenta transcriptome. Placenta 2014; 35: 125–131.

28. Dezso Z, Nikolsky Y, Sviridov E, et al. A comprehensive functionalanalysis of tissue specificity of human gene expression. BMC Biol2008; 6: 49.

29. Irani RA, Xia Y. The functional role of the renin–angiotensin systemin pregnancy and preeclampsia. Placenta 2008; 29: 763–771.

30. Verdonk K, Visser W, Van Den Meiracker AH, et al. The renin–angiotensin–aldosterone system in pre-eclampsia: the delicate bal-ance between good and bad. Clin Sci 2014; 126: 537–544.

31. Li X, Shams M, Zhu J, et al. Cellular localization of AT1 receptormRNA and protein in normal placenta and its reduced expression inintrauterine growth restriction. Angiotensin II stimulates the releaseof vasorelaxants. J Clin Invest 1998; 101: 442–454.

32. Anton L, Merrill DC, Neves LA, et al. Activation of local chorionicvilli angiotensin II levels but not angiotensin (1–7) in preeclampsia.Hypertension 2008; 51: 1066–1072.

33. Saito T, Ishida J, Takimoto-Ohnishi E, et al. An essential rolefor angiotensin II type 1a receptor in pregnancy-associated hyper-tension with intrauterine growth retardation. FASEB J 2004; 18:388–390.

34. Takimoto E, Ishida J, Sugiyama F, et al. Hypertension induced inpregnant mice by placental renin and maternal angiotensinogen.Science 1996; 274: 995–998.

35. Trifonova EA, Gabidulina TV, Ershov NI, et al. Analysis of the pla-cental tissue transcriptome of normal and preeclampsia complicatedpregnancies. Acta Naturae 2014; 6: 71–83.

36. Bainbridge SA, Minhas A, Whiteley KJ, et al. Effects of reducedGcm1 expression on trophoblast morphology, fetoplacental vascu-larity, and pregnancy outcomes in mice. Hypertension 2012; 59:732–739.

37. Ruebner M, Strissel PL, Langbein M, et al. Impaired cell fusion anddifferentiation in placentae from patients with intrauterine growthrestriction correlate with reduced levels of HERV envelope genes.J Mol Med 2010; 88: 1143–1156.

38. Kashif M, Hellwig A, Kolleker A, et al. p45NF-E2 represses Gcm1in trophoblast cells to regulate syncytium formation, placentalvascularization and embryonic growth. Development 2011; 138:2235–2247.

39. Wu YH, Lo HF, Chen SH, et al. Caspase-14 suppresses GCM1acetylation and inhibits placental cell differentiation. FASEB J 2013;27: 2818–2828.

40. Chang CW, Chuang HC, Yu C, et al. Stimulation of GCMa transcrip-tional activity by cyclic AMP/protein kinase A signaling is attributedto CBP-mediated acetylation of GCMa. Mol Cell Biol 2005; 25:8401–8414.

41. Wang CC, Lo HF, Lin SY, et al. RACK1 (receptor for acti-vated C-kinase 1) interacts with FBW2 (F-box and WD-repeatdomain-containing 2) to up-regulate GCM1 (glial cell missing 1)stability and placental cell migration and invasion. Biochem J 2013;453: 201–208.

42. Wentzel C, Sommer JE, Nair R, et al. mSYD1A, a mammaliansynapse-defective-1 protein, regulates synaptogenic signaling andvesicle docking. Neuron 2013; 78: 1012–1023.

SUPPLEMENTARY MATERIAL ONLINESupplementary materials and methods

Supplementary figure legends

Figure S1. Regulation of placental cell migration by SYDE1

Figure S2. SYDE1 regulates migration and invasion of primary human trophoblast cells

Figure S3. SYDE1 regulates Rho GTPase activity

Figure S4. Generation of Syde1-knockout mice

Figure S5. Growth curves of Syde1-KO and wild-type mice

Figure S6. Decreased vascularization in the Syde1-KO placenta

Figure S7. Syde1 regulates mouse trophoblast invasion

Figure S8. Expression of Syde1 and Syde2 in placenta

Table S1. Demographic data of patients


本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/

Documents

Associationofdysfunctionalsynapsedefective1(SYDE1 ...download.xuebalib.com/xuebalib.com.41392.pdf · Biosciences, Mountain View, CA, USA) to generate pCDH-SYDE1-WT and pCDH-SYDE2-WT