Bves expression during avian embryogenesis


  • Megan E. Osler,

    1. The Stahlman Cardiovascular Research Laboratories, Program for Developmental Biology, and Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee
    Search for more papers by this author
  • David M. Bader

    Corresponding author
    1. The Stahlman Cardiovascular Research Laboratories, Program for Developmental Biology, and Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee
    • Stahlman Cardiovascular Laboratories, Program for Developmental Biology, Department of Medicine, 346 PRB, Vanderbilt University Medical Center, Nashville, TN 37232-6300
    Search for more papers by this author


Bves (blood vessel/epicardial substance) is a transmembrane protein postulated to play a role in cell adhesion. While it is clear that Bves and gene products of the same family are expressed in adult striated muscle cells, the distribution of these proteins during development has not been critically examined. An understanding of the expression pattern of Bves is essential for a determination of protein function and its role in embryogenesis. In this study, we present an expression analysis of Bves during chick gastrulation and germ layer formation. Our data show that Bves is expressed in epithelia of all three germ layers early in development. Furthermore, Bves protein is observed in epithelial tissues during organogenesis, specifically the developing epidermis, the gut endoderm, and the epicardium of the heart. These data support the hypothesis that Bves may play a role in cell adhesion and movement of epithelia during early embryogenesis. Developmental Dynamics 229:658–667, 2004. © 2004 Wiley-Liss, Inc.


Formation of epithelia is a critical and recurring event during development. Development of the entire avian embryo begins with the epithelial epiblast. All cells within the organism, whether they are adhesive or freely migratory at later stages, are derived from this epithelium. Each of the three germ layers, the ectoderm, mesoderm, and endoderm, begin as epithelia also (Duband et al., 1988). Of interest, embryonic cells may exhibit intermediate epithelial and mesenchymal phenotypes during their differentiative program. For example, skeletal muscle precursors are first part of the epiblast, then convert to a mesenchymal phenotype during gastrulation, form an epithelial somite, and again convert to a freely migratory, myogenic cell type before finally fusing to differentiate into skeletal muscle cells (Pourquie et al., 1995; Streit and Stern, 1999; Buckingham et al., 2003). This example is one of many that demonstrate how epithelia play important and varied roles in the generation of embryonic structure.

Many molecules regulate formation, morphogenesis, and function of embryonic epithelia. Proteins found in cell–cell junctions, including cadherins (Hatta and Takeichi, 1986), integrins (Yi et al., 1993), and desmoplakins (Runswick et al., 2001), often play critical roles during embryogenesis. Mutation of many of these gene products leads to abnormal embryonic development and disease (Allen et al., 1996). Discovering new molecules that regulate the formation and function of embryonic epithelia is important for a better understanding of development.

In a search to identify new molecules involved in heart development, we performed a subtractive hybridization and discovered a novel transcript we named bves (Reese et al., 1999). Andree et al. (2000) independently discovered an identical gene that they termed popeye. (To clarify, the accepted nomenclature for this gene product is bves (NCBI), whereas the gene family, described below, is termed popeye. The present study focuses on the bves/pop1a gene and gene product, and thus we will refer to the gene product as Bves.) The popeye gene family has been described for several different species, and the number of gene family members varies in a species-specific manner. Andree et al. (2000) were the first group to determine that three genes were present in mouse and two genes were detected in chicken and that differential splicing is responsible for variant mRNA transcripts. Subsequently, a single gene was discovered in Drosophila (Lin et al., 2002), while the human genome contains three popeye genes (NCBI). Multiple expressed sequence tags (ESTs) have been reported in Xenopus, Danio, and other experimental models but the number of genes in these species has not been reported (NCBI).

Results from functional studies of bves have been published. Our previous work demonstrates that Bves confers adhesive properties to nonadherent L-cells cells upon transfection of cloned chicken bves cDNA (Wada et al., 2001). In the same study, we have shown that Bves is trafficked to points of cell–cell contact early in the process of cell adhesion. Additionally, Andree et al. (2002) have demonstrated that skeletal myoblasts from pop1-null mice exhibit impaired differentiation, and they suggested that impedance of cell adhesion may play a role in this process. From these studies, we hypothesize that Bves is a cell–cell adhesion molecule.

Controversy remains, however, concerning the expression of the Bves gene product. Our original report on this protein in the developing chick embryo demonstrated antibody reactivity to Bves in the epicardium, epicardium-derived cardiac mesenchyme, and smooth muscle of coronary arteries (Reese et al., 1999). Subsequent generation of mono- and polyclonal antibodies in our laboratory detected mouse and chicken Bves in cardiac and skeletal muscle as well as additional epithelial cells types (Wada et al., 2001, 2003). Conversely, DiAngelo et al. (2001) showed high levels of expression in cardiac muscle using a monoclonal antibody to chicken Bves, but could not detect the protein in the epicardium. In addition, Andree et al. (2000) reported high levels of bves/pop1 mRNA expression by reverse transcriptase-polymerase chain reaction (RT-PCR) and in situ hybridization in developing mouse embryos in cardiac and skeletal muscle. Close examination of data from their publication (specifically, Fig. 4G; Andree et al., 2000) reveals that the transcript is also present in the epithelium of the gut. Furthermore, Andree and coauthors detected bves/pop1 mRNA by PCR analyses in several tissue types during embryogenesis (Fig. 3B; Andree et al., 2000), which did not match the presented RNA in situ hybridization analyses (Fig. 4; Andree et al., 2000). Similarly, Hitz et al. (2002) reports high levels of Xbves/Pop1 mRNA restricted to the frog heart. However, numerous ESTs have been isolated from the unfertilized egg and early embryo (NCBI), and in situ studies from our laboratory demonstrate the presence of the bves transcript in the early Xenopus embryo.

Figure 4.

Bves protein expression during germ layer formation. A–C: An embryo at Hamburger and Hamilton stage HH8 is labeled with B846. At the termination of neural tube closure, Bves protein is seen expressed in the ectoderm, endoderm, and the epithelium that forms the lumen of the neural tube (NT). Mesodermal labeling by anti-Bves antisera at this stage is observed (arrow in B). B: A higher power of the NT illustrates the apical/lateral distribution of Bves, like ZO-1, in the epithelial structures of the ectoderm. C: Higher magnification of the endoderm reveals Bves staining at cell membranes, colocalizing with ZO-1 (C, merged image). D: B846 labeling of an HH11 embryo undergoing somitogenesis reveals Bves protein expression at the ectoderm/endoderm layers, the neural tube, the epithelial somites, and epithelialized intermediate and lateral mesoderm. D′: D033 labeling of a serial section is shown. D: Higher magnification of epithelial somites allows an analysis of B846 staining with that of ZO-1. Both antibodies label the apical region of the NT epithelia and the somitic epithelia and colocalize as observed in the merge (D, far right). E: The intermediate (arrow) and lateral plate mesoderm are both epithelial in nature and express Bves, as detected by B846. Colocalization with ZO-1 is observed with a merged image in E.

Figure 3.

Detection of Bves during gastrulation. A: Hamburger and Hamilton stage HH4 embryos labeled with B846 show Bves expression in the epiblast (arrow). This posterior section of the embryo has not undergone gastrulation. B: D033 labeling of the same HH4 embryo reveals a similar pattern of staining in the epiblast (arrow). Arrows denote epiblast in A and B. Higher power images of an HH6 embryo undergoing gastrulation allow the observation of distinct Bves localization. C: B846 labels Bves in the epiblast/definitive ectoderm at cell–cell borders, as indicated by colocalization with ZO-1. D: Regions of the hypoblast/developing endoderm also express Bves at cell membranes, as detected by B846 antisera. E: At the primitive streak, B846 detects Bves apically in cells that are involuting, a pattern similar to that of ZO-1. Arrow denotes loss of membrane staining of ingressing cells.

Based on data generated from our laboratory, our central hypothesis predicts that Bves plays a role in cell–cell adhesion in developing epithelia and muscle. Nevertheless, because controversy exists concerning the distribution of Bves protein in epithelia of the developing embryo, the progress toward determination of its basic function is impaired. Clearly, additional studies are necessary to indisputably determine the timing and distribution of Bves during embryogenesis. Therefore, in the present study, we focus on elucidating the expression of Bves during gastrulation and germ layer formation, because these processes represent fundamental movements of epithelia during development. Our data show that Bves is expressed in the gastrulating embryo and in all three germ layers during early development. Furthermore, Bves protein is detected in some but not all epithelial structures derived from endoderm, mesoderm, and ectoderm. Taken together, the distribution of Bves, as shown by our studies and those of others, suggests that this gene product not only has a function in muscle development but also in the developing chicken embryo through its involvement in epithelial morphogenesis.


Characterization of Antisera Against Bves

Two polyclonal antibodies (DO33 and B846) were previously generated against the C-terminus of chicken Bves but characterization of their binding activities has not been published thus far. Here, we are presenting a series of experiments to establish the reactivity of these antisera for Bves. Antibody binding of B846 to chicken Bves protein was initially established by using Western blot analysis of Flag-tagged chicken bves transfected into COS-7 cells (Fig. 1A). B846 binds a protein with the same mobility as anti-flag and does not react with nontransfected cells. Detection of Bves on a Western blot by D033 was presented previously (Reese et al., 1999). Specificity of B846 and D033 was further demonstrated by immunofluorescence microscopy as both antisera colocalize with anti-Flag in COS-7 cells transfected with Flag-tagged bves (Fig. 1B,B′). Peptide competition studies extinguish antiserum reactivity, further establishing the specificity of these reagents (Reese et al., 1999).

Figure 1.

Characterization of anti-Bves polyclonal antibodies. A: Specificity of the B846 antisera is determined by Western blot. B846 recognizes a ∼52-kDa band from COS cell lysate transfected with a Flag-tagged chicken Bves construct (lane 2). Minor bands are unglycosylated forms of Bves (Knight et al., 2003). No reactivity is detected with untransfected COS lysate (lane 1). Anti-Flag antibody recognizes a band of similar mobility (lane 3). B,B′: Specificity of B846 (B) and D033 (B′) is demonstrated by detection of transfected chicken Bves in COS cells processed for immunofluorescence. Bves antisera is labeled using Alexa 488 (green) secondary antibody and anti-Flag antibody staining is represented by Alexa 568 (red) secondary antibody. This merged image shows colocalization between Bves and Flag. Cell nuclei are stained blue with DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride). C,D: The Bves staining pattern in at cell–cell borders in immortalized cell lines is detected with B846 antisera in EMC (C) and HCA-7 (D) epithelial cells. E: Bves mRNA is detected in the HCA-7 clonal intestinal epithelial cell as shown by reverse transcriptase (RT) -polymerase chain reaction analysis, which reveals the presence of a 600-bp product. F: A reaction in the absence of RT reveals no product. Epitopes of D033 (red) and B846 (blue) antisera are shown within the Bves protein sequence. Putative hydrophobic/transmembrane domains are depicted in green.

At present, the expression of bves in epithelial cells is controversial. Because clonal cell lines permit analysis of pure cell populations without contaminating cells from primary tissue sources, we characterized anti-Bves reactivity in various epithelial cell lines. Since avian epithelial cell lines do not exist at present (see ATCC, 2003), we turned to mammalian cells to obtain pure cell populations. As seen in Figure 1, Bves protein was detected with B846 at the membrane of epicardial–mesothelial cells (EMC) and human colonic adenocarcinoma (HCA-7) cells (Fig. 1C and D, respectively). This antibody reactivity can be competed with the antigenic peptide (Reese et al., 1999). RT-PCR analyses confirmed the presence of bves mRNA in these cell lines (Wada et al., 2003; Fig. 1E). Bves protein was also observed in MDCK cells, 4T-1 mouse mammary cells, human corneal epithelial cells, and A549 human lung epithelial cells (Dr. Lynn Matrisian, Vanderbilt University) with B846 antiserum (unpublished data). Additionally, several new monoclonal antibodies directed against mouse Bves react with mouse epithelia, as well as heart and skeletal muscle, and various mammalian cell lines (Smith and Bader, manuscript in preparation).

Finally, an analysis of antisera reactivity with different vertebrates determined that D033 detects Bves in the avian class, while B846 recognizes Bves protein in a variety of species, including avian and mammalian classes. An antiserum against amphibian Bves has been generated and shows the same general pattern of localization presented here (Nesset, Wright, and Bader, in review). These results demonstrate that the B846 and D033 polyclonal antibodies identify the Bves protein and will be useful tools for studying the expression patterns of Bves.

The location of the antibody epitopes, D033 and B846, within the full-length chick Bves protein sequence is shown in Figure 1F. All experimental evidence thus far suggests that these epitopes are specific to the Bves/pop1 isoform. Extensive blast searches do not reveal these epitopes in other pop gene products.

bves Is Expressed During Early Development

bves is most highly expressed in cardiac myocytes after the formation of a definitive heart tube heart has fully developed (Reese and Bader, 1999; Andree et al., 2000). Andree et al. (2000) generated RT-PCR products from several tissues in addition to the heart, and numerous ESTs from nonmuscle sources were previously reported in a variety of organisms (NCBI). However, bves mRNA was not detected at high levels in developing chick embryos (Hamburger and Hamilton stage [HH] 11–HH18) by using whole-mount in situ hybridization (Andree et al., 2000). Therefore, we used RT-PCR analysis to determine whether bves mRNA is detectable in the early chick embryo before heart tube formation. As seen in Figure 2A, a bves RT-PCR product was observed in HH 6 whole embryo preparations at the predicted mobility. At this time, germ layer differentiation is occurring anteriorly, while gastrulation continues posteriorly. As expected, the Bves transcript was also detected in embryos with a developing primordial heart tube (Fig. 2A).

Figure 2.

Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of early chick embryos. Total RNA was extracted from whole chick embryos at Hamburger and Hamilton stages HH6, HH24, and HH31 using Trizol reagent. A: RT-PCR was performed using primers specific to chicken Bves. Bves mRNA was detected in both HH6 and HH24. RT- reactions for HH6 and HH24 show no contamination of genomic DNA. Glyceraldehyde-3-phosphte dehydrogenase (GAPDH) controls for HH6 and HH24 verify input of RNA (C in A). B: RT-PCR on HH31 embryonic tissue detects chicken Bves in total RNA from heart, gut, and skin. GAPDH primers serve as a positive control (C in B) and the negative control RT- reaction shows no product. A cytokeratin transcript was detected in the skin as expected (CK in B).

Further RT-PCR analyses were conducted to determine whether bves could also be detected in extracardiac tissues after the heart is formed, specifically in the ectoderm/skin and gut, as we were particularly interested in the development of epithelial structures in these two organ systems. Expression of bves transcript was observed in the ectoderm/skin and the developing gut (small intestine) at HH stage 31 (Fig. 2B). As expected, bves was detected in the developing heart (Fig. 2B). These data indicate that bves message is present in a variety of embryonic tissues, in addition to the heart and that investigation of the protein expression at the cell and tissue levels is warranted.

Localization of Bves Protein During Gastrulation

The gastrulating embryo is a complex structure comprised of both epithelial and mesenchymal elements. After initial cleavage stages, the avian embryo becomes a 2-layered blastoderm with a dorsal epiblast and a ventral hypoblast. The embryo proper is derived entirely from the epithelial epiblast/definitive ectoderm. Primitive streak formation occurs upon the thickening of the epiblast, marking the initial ingression of mesendodermal precursors. The cells migrating into the blastocoele undergo an epithelial-to-mesenchymal transition. A specific population of these ingressing cells are programmed to become endoderm and revert to an epithelial phenotype as they replace hypoblast cells (Lawson and Schoenwolf, 2003). Some of the remaining mesodermal precursors will remain mesenchymal, while others will convert back to epithelia upon formation of specific structures, such as somites (Pourquie et al., 1995). Immunocytochemical methods using both antisera were used to localize Bves protein in the gastrulating embryo at stage 4. At this time, Bves is detected in the epiblast/forming ectoderm (Fig. 3A,B, see arrows). At higher magnification, Bves is seen at the cell membrane in the epithelium of the epiblast/ectoderm (Fig. 3C). Colocalization with ZO-1, a tight junction marker, demonstrated cell membrane deposition of Bves in the epithelial epiblast/ectoderm in these specimens. Bves is also detected at cell–cell borders in the ventral region of the gastrulating chick embryo (Fig. 3D, see arrow). At present, we cannot determine whether this staining is hypoblast or newly forming endoderm. Of interest, Bves protein is distributed apically/laterally in cells of the primitive streak and colocalizes with ZO-1 (a section posterior to Henson's node in a stage 6 embryo is shown, Fig. 3E). As cells ingress as individuals or in groups to form endoderm and mesoderm, Bves protein is eliminated or greatly reduced at the cell surface (Fig. 3E, see arrow). The reactivity seen in these cells is above background staining. Still, it should be noted that newly gastrulated cells do not traffic Bves protein to their cell surface when they are in a mesenchymal state. This reduction of membrane staining of Bves is similar to that observed during epithelial-to-mesenchymal transition in vivo (Reese et al., 1999; Wada et al., 2001) or in vitro (Wada et al., 2003).

Localization of Bves During Early Germ Layer Differentiation

Development of the chick continues with the differentiation of the three definitive germ layers. Ectoderm and endoderm primarily remain epithelia throughout development with specific exceptions, for example neural crest cells (Selleck and Bronner-Fraser, 1996). The mesoderm, however, undergoes various epithelial and mesenchymal phases depending on the structure to be generated. Because one of our central hypotheses is that Bves plays a general role in cell–cell adhesion in epithelia, we sought to determine whether Bves is present in epithelial elements of the three germ layers. As seen in Figure 4, Bves is clearly present at membranes in cells of the definitive ectoderm (A and B) of an HH8 embryo. After gastrulation is initiated, the formation and closure of the neural tube proceeds anteriorly to posteriorly. Bves is expressed at the luminal surface of the developing neural tube, which was once continuous with the apical surface of the ectoderm. It is also interesting to note that the connection between the surface ectoderm and the developing neural tube remains positive for Bves and ZO-1 for a brief period before complete loss of contact between these structures (Fig. 4B, see arrow). Bves staining is also observed in the endoderm of the open foregut (Fig. 4C). As previously noted, Bves protein was not detected at high levels in newly gastrulated mesoderm, especially at the cell membrane (Fig. 3E, see arrow). Because avian cells gastrulate individually or in small groups, elimination of Bves from the cell surface was not unexpected.

We next examined the expression of Bves protein in a HH11 embryo as mesodermal cells form various epithelial structures. Findings in Figure 4 exemplify a recurring pattern of Bves deposition observed during the differentiation of mesoderm, namely that Bves is present at the apical/lateral regions of mesodermally derived epithelia, similar to the ZO-1 pattern. After neurulation is under way, the paraxial mesoderm begins to organize into epithelial somites. During this process of mesodermal epithelialization, B846 detects Bves along the inner margin of the somites (Fig. 4D,E), which is the apical surface of this epithelium. Also, the epithelial cells of the somites become attached by tight junctions (Bellairs, 1979), which can be observed by ZO-1 staining (Fig. 4E). D033 labels Bves in the somite also, but the distribution within the epithelium is not as highly localized (Fig. 4D′). The intermediate and lateral plate mesoderm are also formed at this time by the clustering of cells into epithelial structures. Interestingly, when these mesodermal structures were examined with anti-Bves antisera, the protein was observed at the apical/lateral regions (facing inward toward the coelomic cavity; Fig. 4F, arrow). Bves shares the same labeling pattern as tight junction protein ZO-1 in the intermediate and lateral plate mesoderm (Fig. 4F). Additionally, the notochord is positive for Bves (data not shown). Thus, the initial epithelial structures formed by the partitioning of mesoderm are positive for Bves. However, B846 antisera do not detect Bves protein at high levels in mesodermally derived mesenchyme. As development proceeds, mesodermal derivatives such as cardiac, smooth, and skeletal muscle cells are positive for anti-Bves antisera.

Bves Is Expressed in the Epithelia of the Developing Skin and Gut

From the data presented above, it is apparent that Bves, like other adhesive proteins, is expressed in epithelial structures of the developing organism. Because Bves is initially detected in both ectoderm and endoderm, we wished to determine the pattern of Bves expression in differentiating skin, an ectodermal derivative, and gut epithelium, which is originates as endoderm.

The outermost surface cell layer of the developing embryo, the periderm, arises from an initial single layer of ectoderm. This one cell layer quickly becomes two layers, the outer being the temporary periderm, and the inner being the cuboidal germinal layer that gives rise to the stratified epidermis. The cells of the periderm die, keratinize, and are eventually exfoliated from the underlying stratified epidermis several days before hatching (Isokawa et al., 1996). In the developing skin, the epidermis that surrounds the embryo expresses Bves and the protein appears to be restricted to cell–cell borders (Fig. 5A–C). The developing dermis is negative when labeled with anti-Bves antibodies and remains so throughout development.

Figure 5.

Bves distribution in the chick epidermis. A,C: B846 (A) and D033 (C) anti-Bves antisera exclusively label the epidermis (e) of Hamburger and Hamilton stage HH31 chick embryos in cross sections of the skin. The dermis (d) is negative. B: At higher power, epidermis labeled with B846 detects Bves protein at the borders of cells of the epidermal epithelium. D: An epidermal culture, extracted from HH31 embryo and grown on a glass slide, was fixed and labeled with B846 antisera. This image shows Bves at cell membranes in vitro, a pattern consistent with the pattern observed in vivo (B).

We have established an epidermal in vitro culture system to study Bves protein expression and function in more detail. After growing patches of epidermis for several days in culture, antibodies to Bves detect the protein at the cell membranes, similar to the observed pattern in vivo (Fig. 5D). These data suggest that Bves protein retains its expression in the ectoderm as it differentiates to form the epidermis.

The endoderm of the chick embryo forms the mucosa/epithelium of the gut tube and undergoes dynamic changes throughout gut morphogenesis (reviewed in Roberts, 2000). Because Bves expression in the early stage endoderm was pronounced, we sought to determine whether Bves protein could be detected in the gut epithelium. As shown in Figure 6, Bves is expressed throughout gut development. Closure of the gut tube occurs around HH20–HH22 and at this stage, the gut epithelium forms a simple tube (Fig. 6A). The developing gut of the chick embryo undergoes proliferation and morphogenesis up until hatching. As the primordial villi begin to take shape, the resulting gut epithelium is more undulated and the epithelial surface area is enlarged (Fig. 6B–D). Protein expression analysis of Bves in a HH41 gut reveals that Bves is restricted to the gut epithelial cells, specifically to the apical/lateral region of cell–cell borders (Fig. 6E). Mesenchyme derived from lateral splanchnic mesoderm that surrounds the gastric mucosa is negative. In addition, Bves is expressed in the serosal epithelium (Fig. 6A, see arrow) which is a mesothelium, similar to the epicardium that surrounds the developing heart.

Figure 6.

Bves is detected throughout development of the endodermally derived gut. A–E: Chick embryos were labeled with B846 at various stages of gut development: day 7 (Hamburger and Hamilton stage [HH] 30–HH31; A), day 8 (HH33–HH34; B), day 11 (HH37; C), day 13 (HH39; D), day 15 (HH41; E). B846 antisera labels the epithelium of the gut throughout gut morphogenesis and extension of the villi. The serosa, the outer epithelial covering of the gut, is also positive for Bves (arrow in A). D033 staining of the gut is much weaker and not shown.

Differences in Bves Localization With DO33 and B846

While the protein expression patterns of Bves revealed by DO33 and B846 are in general agreement, specific variations in antigen labeling exist. These differences are best exemplified by the reaction of these two sera with sections through the developing heart. As previously reported (Reese et al., 1999), DO33 labels Bves in the proepicardial organ (PEO), epicardium, and delaminated mesenchyme of the heart at stage 18 (Fig. 7A,B). Reaction of D033 antisera with the myocardium is minimal or reduced. In contrast, B846 detects Bves in the PEO, epicardium, and delaminated mesenchyme but also stains myocardial cells, as visualized by colocalization with anti-desmin (Fig. 7C). Bves localizes to cell–cell borders of the epicardial epithelium (Fig. 7D, see arrow). Transverse sections of HH31 embryos also reveal subtle differences between the two antisera. Both B846 and D033 label the epidermis (Fig. 7C,D). D033 strongly labels somites but is not reactive with the developing neural tube (Fig. 7E,G), whereas B846 selectively labels the ependyma (Fig. 7F,H). B846 does not label somites at this stage, in contrast to what was observed in stage HH11 embryos (Fig. 4D). At this point in development, somites are no longer epithelial in nature and migrate as individual cells due to their commitment to the muscle-cell lineage.

Figure 7.

Differing labeling patterns of B846 and D033. Transverse sections of Hamburger and Hamilton stage HH18 chick embryos show the proepicardial organ (PEO) migration over the heart. A: The PEO is positive for Bves, as detected in red by D033, while anti-desmin labels the myocardium in green. B: The epicardium is formed (arrow) as the Bves-positive PEO migrates from the liver primordium over the surface of the myocardium, a process observed at higher power. C: Labeling of a serial section with B846 detects Bves protein (red) in the migrating PEO/epicardium as well as the heart tube, which is also labeled with anti-desmin in green. D: View of a section through a HH31 heart at high magnification is consistent with a previous report that Bves protein (green) localizes to the cell–cell borders of the epicardium (e), as labeled with B846. Bves is expressed in the epicardium (arrow) and colocalizes with MF20 (red) in cardiac myocytes (m). B846 and D033 affinity differences are observed in cross-sections of developing chick (HH31). E,F: D033 labels epidermis and developing somites (E), whereas B846 detects Bves in the epidermis and the developing ependyma (F). An inset shows a magnification of the ependymal staining in F. G,H: Overlays of E and E with DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride) are seen in G and H, respectively.

Additionally, in epithelial cells that are reactive with both antisera, such as the epicardium (Fig. 7A–D) and the epidermis (Fig. 5), B846 has a more distinct membrane staining pattern, while D033 has a more diffuse signal. There are several explanations for this variation. First, it is possible that antigen presentation in different cell types varies slightly for the two sera. This variation could be the result of the interaction of Bves with accessory proteins in a tissue-specific manner. In addition, multiple splice variations of these genes have been identified. Thus, it is possible that immunochemical reagents, whether they be mono- or polyclonal antibodies, recognize variant isoforms of a gene family to different degrees. For example, we have generated monoclonal antibodies for mouse Bves that recognize variant subcellular distribution patterns. While the present study has clearly demonstrated the expression during the early phases of embryogenesis and suggest a possible function at this time, further analyses with these new tools will aid in delineating the expression of all gene family members.


Previous studies have demonstrated that members of the popeye gene family are highly expressed in cardiac and skeletal myocytes during embryogenesis (Andree et al., 2000; DiAngelo et al., 2001). Although our initial findings were consistent with these publications, we also detected Bves in epicardial cells in vivo and in vitro (Reese et al., 1999; Wada et al., 2001, 2003). Still, several pieces of evidence including RT-PCR, database searches, and in situ hybridization suggested that Bves may be more widely expressed (Andree et al., 2000; Wada et al., 2001). In the present study, we are presenting data showing Bves is expressed during early embryogenesis when epithelial movements are critical for the initial differentiative events.

We detect bves transcripts by using RT-PCR technology during early chick embryogenesis and in many mammalian epithelial cell lines (Wada et al., 2003). As previously noted, the whole-mount studies of Andree et al. (2000) also demonstrate popeye gene expression in the gut epithelium, a cell type highly reactive with our antisera. In addition, EST database analyses reveal that transcripts of the popeye gene family are widely distributed in various cell types at different stages of vertebrate development. The observed discrepancy in expression patterns between data from Andree et al. (2000) and the findings from our laboratory (Wada et al., 2003) could be due to a low level of bves gene expression in those epithelia that produce the transcripts. In addition, the lifetime of the mRNA message in epithelia is unknown and could potentially be short-lived, impairing its detection. In this study, we clearly demonstrate that Bves protein is present in many embryonic epithelia during development. Thus, variable sensitivity of assay systems may account for the apparent variation in mRNA and protein distribution.

Results of the present study demonstrate that Bves is expressed in epithelial cells of the epiblast as well as the ectoderm, mesoderm, and endoderm. While these embryonic epithelial structures are highly dynamic during development, the necessity of cell adhesion during morphogenesis is critical to proper shaping of these layers. This finding is consistent with previous data suggesting an adhesive role for this protein (Wada et al., 2001). The presence of Bves in the early embryo suggests that this protein may have a conserved function in epithelial structures, specifically during gastrulation and germ layer differentiation. Our expression studies demonstrate that Bves is present in epithelia undergoing dynamic rearrangements such as the epicardium, epiblast, cavitation of the lateral plate, formation of the epithelial somite, and tubulogenesis in the neuroectoderm and intermediate mesoderm. Later, after gross movements and reshaping of these embryonic structures is completed, Bves expression may be down-regulated in epithelia but remain highly expressed in the heart and skeletal muscles. If Bves indeed functions as an adhesion molecule between actively moving cells, its expression in embryonic epithelia and contracting muscle may indicate this conserved function. Thus, the present data lead us to the hypothesis that Bves regulates adhesion in various types of epithelial morphogenesis.

As seen in this study, Bves protein is confined to specific regions of developing epithelia. Specifically, Bves is localized to the apical/lateral regions of epiblast, neural tube, somites, and newly formed intermediate, somatic, and splanchnic mesoderms. Bves in these regions suggests that the protein plays a role in the generation or maintenance of cell polarity. This hypothesis is supported by previous data demonstrating that Bves is one of the first adhesion proteins to traffic to points of cell–cell contact in forming epithelia in vitro (Wada et al., 2001). The present study lays the groundwork for future analyses to discover the role of Bves in epithelia and in early embryogenesis.



Two antisera against chick Bves have been reported previously (Reese et al., 1999; Wada et al., 2001, 2003). Antiserum DO33 reacts with only with avian species (Reese et al., 1999). Antiserum B846 reacts with avian and mammalian Bves. While the specificity of these antisera has been reported, further characterization is given here to document their reactivities. We recently have developed a bank of monoclonal antibodies against mouse Bves that do not cross-react with the chicken protein. Other monoclonal antibodies used include ZO-1 (Zymed), Desmin (Sigma), Flag (Sigma); MF20 hybridoma supernatant (Hybridoma Bank). ZO-1, a well-characterized component of the tight junction, is marker of epithelial junctions to which the localization of Bves is compared. Secondary antibodies were conjugated to Alexa 488 and Alexa 568 (Molecular Probes). Samples were counterstained with DAPI (4′,6-diamidine-2-phenylidole-dihy-drochloride; Roche) to visualize nuclei. Reagents were used at manufacturer-recommended dilutions.


Fertilized White Leghorn eggs were obtained from Spafus/Trelow Farms, and all animal protocols have been approved by Vanderbilt University. Embryos were staged using the standard staging protocols (Hamburger, 1951).

Cloning of Chick bves and Cell Culture Studies

Chick bves was cloned previously as a cDNA product of a subtractive library screen (Reese et al., 1999). A full-length chick bves with a 3′ Flag tag was generated by using PCR and cloned in frame into pCIneo at the SalI/NotI sites (Promega). Primer sequences are as follows: 5′ SalI - AGAGCTAGCGTCGACTTCAAGAT-GGACACTACGGCA and 3′ NotI/flag TACATATGCGGCCGCCTACTTGTCATCGTCGTCCTTGTAGTCAGGC AGCCGCTGCAGCTC (Aya Wada, personal communication). COS7 fibroblast cells were transiently transfected with pCIneo/bves (or positive and negative control plasmids) by using the Fugene6 Transfection reagent (Roche). EMCs were obtained from Hoda Eid (Eid et al., 1994) and grown as previously described (Wada et al., 2001, 2003). HCA cells, a human intestinal carcinoma line and COS7 cells were obtained from Dr. Robert Coffey (Vanderbilt University) and ATCC, respectively. As previously described, cell lines were grown in DMEM and passaged upon reaching confluency (Wada et al., 2001).

Generation of Chick Epidermal Cultures

Epidermal cultures were generated as follows: skin (epidermis and dermis) was removed from chick embryos staged to HH30–HH32 and rinsed in phosphate buffered saline. The skin was incubated in Dispase II (Invitrogen) for 5 min at 37°C, then transferred to media (described below) for recovery. The epidermis was separated from the dermis, transferred to chamber slides (Lab-Tek), and fragmented with forceps to create small clumps. Epidermal patches were grown for 24–48 hr in M199 medium (Cellgro) with the following additives: 10% tryptose (Invitrogen), penicillin/streptomycin (Cellgro), 5% fetal bovine serum, and 1% chick serum. Epidermal patches were subsequently processed for immunofluorescence as described below.

Immunochemical Analyses

Standard methods for mono- and polyclonal antibodies were used for tissue sections and cultured cells (Bader et al., 1982; Reese et al., 1999; Wada et al., 2003). Tissue sections and cell cultures were fixed 10 min in 70% methanol for anti-Bves antibodies. Reactivity is reduced with paraformaldehyde or formalin fixation. Secondary antibodies (Alexa) were used at manufacturer's specifications. Negative controls included peptide competition, no primary antibody application, and nonimmune antibody treatment. All of these preparations revealed no reactivity and are not shown. Western blotting experiments were conducted using standard protocols (Ausubel et al., 2002). Anti-mouse and anti-rabbit alkaline phosphatase-conjugated secondary antibodies were purchased from Sigma and used at a 1:10,000 dilution.

RT-PCR Analyses

Human HCA-7 cells grown to confluency were washed once with serum-free medium and harvested in Trizol reagent (Invitrogen) using 1 ml per two plates of cells. RNA was extracted following the standard protocol (Invitrogen). Whole chick embryos and tissues were isolated under sterile conditions, and RNA was extracted with the Trizol reagent system. Approximately 50 ng of RNA template was used for each reaction. RT-PCR was performed by using the Titanium One-Step kit (BD Biosciences Clontech). The RT step was performed for 60 min at 50°C. PCR was performed by using the following conditions: 94°C for 5 min, 40 cycles of 94°C for 30 sec, an annealing step for 30 sec at primer-specific temperatures (listed in Table 1), and 68°C for 3 min. Final extension was 68°C for 2 min. The primer sets used for RT-PCR reactions are listed in the Table 1. Chick Bves primers are specific to the bves/pop1a message and do not amplify other splice variants.

Table 1. Sequences of Primers Used for RT-PCR
Gene product5′ sequence3′ sequenceSize (bp)Topt (°C)


We thank members of the laboratory for critical reading and helpful comments.