Main

Cardiac morphogenesis is a complex process that requires a number of coordinated events, including cell fate determination, proliferation, differentiation, cell migration, and cell-to-cell communication. Many molecules are known to play a role in this process, including the transforming growth factor-βs, retinoic acid receptors, and transcription factors such as msx-1 and GATA-4 (reviewed inRefs. 1–3). Appropriate signaling mechanisms are necessary to mediate the changes induced by these molecules during normal embryogenesis. To date, there is limited information regarding signal transduction molecules involved in cardiac development(4). This study examines the expression pattern of one signal transduction molecule, PP2A, in the developing rat heart.

PP2A is one of the major serine/threonine phosphatases responsible for dephosphorylating many essential proteins, including cell surface receptors, metabolic enzymes, ion channels, cytosolic protein kinases, and transcription factors(5). Such dephosphorylation is required for cell-cycle regulation, cell transformation, and cell fate determination(5, 6), suggesting a role for PP2A in normal embryogenesis. A number of recent findings support this hypothesis. Studies in Drosophila(7) and Xenopus(8) have shown that various subunits of PP2A are developmentally regulated in whole embryo extracts. Furthermore, interruption of a PP2A subunit by p-element insertion in mutant Drosophila causes abnormalities in the larval brain, including mitotic defects(9). A different PP2A subunit mutation results in partial duplication of wing imaginal discs, indicating a role for PP2A in cell fate determination(10). In contrast, the role of PP2A in mammalian development is unexplored.

PP2A usually exists as a heterotrimer consisting of a highly conserved 36-kD catalytic subunit, a tightly associated 65-kD structural subunit, and a highly variable regulatory subunit of 54, 55, 72, or 74 kD. This third, regulatory subunit is thought to target the enzyme to specific substrates(11). At least two isoforms of the catalytic subunit have been identified, α and β, which are encoded by separate single genes(12). Although the two isoforms have similar tissue distributions by Northern analysis, the α isoform predominates with expression levels about 10 times higher than the β in most mammalian tissues(13).

We previously identified a 2.4-kb clone of the α catalytic subunit of protein phosphatase 2A (PP2Aα) in the embryonic rat heart(14). We now describe the mRNA and protein distribution of PP2Aα at various stages of cardiac development. The high level of expression and specific cellular localization of PP2Aα suggest that this important second messenger is involved in normal cardiac differentiation and development.

METHODS

All studies were approved by the Animal Research Committee at the University of Virginia.

Northern analysis. Total RNA was extracted from Sprague-Dawley rat hearts at various developmental stages (Hilltop Laboratory Animals, Inc., Scottsdale, PA) by the method of Chomczynski(15) using Tri-Reagent (MRC, Cincinnati, Ohio). Ten micrograms of total RNA were loaded into each lane and electrophoresed under denaturing conditions in a 4-morpholinepropanesulfonic acid/formaldehyde 1.2% agarose gel with transfer to a charged nylon membrane (Zetaprobe; Bio-Rad, Hercules, CA) by capillary action in 20 × SSPE. The membranes were then cross-linked by UV light exposure. A 2.4-kb PP2Aα cDNA probe was labeled by random priming with [α-32P]dCTP (Pharmacia Biotech, Uppsala, Sweden). Hybridization and washing were done at 65°C according to the method of Church and Gilbert(16). Specific hybridization was detected by a PhosphorImager (Molecular Dynamics, Sunny-vale, CA). The membrane was reprobed for the control gene GAPDH as previously reported(17) to verify equal loading and gene specific expression.

The template for the probe was previously cloned in our laboratory from a 14-d embryonic rat heart cDNA library(13). This 2.4-kb clone is 100% homologous within the coding region to a previously described 1.7-kb rat heart PP2Aα(18), and has less than 20% sequence similarity to other members of the serine/threonine protein phosphatase family.

Western analysis. Western blotting was performed by a method described previously with minor modifications(19). Fetal hearts at various gestational ages were dissected and stored at-80°C. After thawing, they were homogenized in ice-cold 50 mM Tris-HCl buffer (pH 7.4) containing 0.1 mM EDTA, 0.1 mM EGTA, 0.1% 2-mercaptoethanol, 1 mM phenylmethylonfonyl fluoride, 2 mM leupeptin, and 1 mM pepstatin A (Sigma, St. Louis, MO). The homogenates were centrifuged at 1000 × g for 10 min at 4°C, and the pellets were discarded. The heart homogenates(50 μg each) and purified PP2Aα protein (1 μg, a gift of Dr. David Brautigan, University of Virginia) were separated on a 7.5% SDS-PAGE gel followed by transfer to a nitrocellulose membrane (Bio-Rad). Membranes were stained with Ponceau-S to verify equal loading and transfer. Then membranes were blocked with buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 2% BSA, and 0.1% Tween 20, for 1 h at room temperature. The membranes were incubated with an anti-PP2A antibody at 1:1000 dilution for 1 h at room temperature. We used a polyclonal rabbit antibody against the carboxy-terminal region of PP2A (amino acids 295-309). This antibody is completely specific for PP2A and does not react with other protein phosphatases(20). The membranes were washed six times with PBS, incubated for 1 h with anti-rabbit IgG antibodies conjugated with horseradish peroxidase (Bio-Rad) at room temperature, and washed six more times with PBS. Immunoreactive PP2Aα protein was detected by enhanced chemiluminescence (ECL System, Amersham Corp., Arlington Heights, IL) and film autoradiography. Quantitative differences in protein levels were determined by densitometry (Personal Densitometer and ImageQuant software, Molecular Dynamics). Protein levels were determined from three Western blots with different protein samples and compared statistically using one-way ANOVA with Bonnferoni correction. Significance was determined as p < 0.05.

Probe preparation for in situ hybridization. Digoxigenin-labeled sense and antisense RNA probes were synthesized with T4 and T7 RNA polymerase (RNA Labeling Kit, Boehringer Mannheim, Indianapolis, IN) from a template consisting of an 840-bp EcoRI/SstI restriction fragment of the 2.4-kb rat heart PP2Aα coding region(bp 70-910) subcloned into Bluescript plasmid vector (Stratagene). Subsequently the probes were hydrolyzed for 35 min at 60°C in hydrolysis buffer (80 mM NaHCO3, 120 mM Na2CO3) to give a final length between 150 and 250 bp, confirmed by Northern analysis.

Whole mount in situ hybridization. Sprague-Dawley rat embryos(12-d gestation) were dissected from anesthetized time-dated pregnant females, and fixed overnight in 4% paraformaldehyde at 4°C. Embryos were washed in 70% ethanol and stored at -20°C. Just before hybridization, they were rehydrated and washed in PBS + 0.3% Triton X-100 (PBT). They were digested in proteinase K (1 μg/mL) at 37°C for 30 min, refixed in 4% paraformaldehyde for 30 min, and rewashed in PBT. Digoxigenin-labeled PP2Aα cRNA probes as prepared above were denatured and added to HB at a final concentration of 300 ng/mL. HB contained 50% formamide, 5× SSC, 1 × Denhardt's solution, 10% dextran sulfate, 0.25 mg/ml yeast tRNA, and 0.4 mg/mL denatured salmon sperm DNA. Hybridization took place overnight at 60°C. The probe was removed by washing in 2 × SSC, at 60°C, then RNase A (20 μg/mL) for 30 min at 37°C, followed by 0.2× SSC + 0.3% Triton X-100 at 50°C. Embryos were then incubated at 4°C overnight in anti-digoxigenin antibody (Boehringer Mannheim) diluted 1:2000 in PBT with 10% heat-inactivated sheep serum (Gibco BRL, Gaithersburg, MD). Excess antibody was removed, and color development took place in NBT/X-phosphate color substrate solution (DIG Nucleic Acid Detection Kit, Boehringer Mannheim) with 0.3% Triton X-100, for 30-90 min. Development was stopped by washing in PBS, and the embryos were photographed. At least four embryos were used for each probe (sense and antisense).

Tissue preparation of frozen sections. Embryos of various gestational ages were removed from anesthetized time-dated pregnant Sprague-Dawley rats and rapidly placed into PBS. The earlier embryos (12-16-d gestation) were fixed whole in 4% paraformaldehyde in PBS at 4°C. The hearts were rapidly dissected from older embryos (18-20-d gestation) and from newborn animals (<24 h of age) before immersion in paraformaldehyde. Fixation took place for 90 min to 2 h, depending on the thickness of the specimen. After fixation, the specimens were dehydrated in an increasing gradient of sucrose in PBS, then embedded in a 1:1 solution containing OCT compound (Miles Inc., Elkhart, IN) and 20% sucrose in PBS, and quickly frozen by immersion in liquid nitrogen. Sections (2-4-μm) were cut (at least 12 sections or 6 slides for each gestational age) and thaw-mounted onto precleaned Superfrost Plus slides (Fisher Scientific Inc., Springfield, NJ). The slides were used for both in situ hybridization and immunohistochemistry.

In situ hybridization of frozen sections. Frozen tissue sections prepared as above were washed in PBT, digested in proteinase K (1μg/mL) for 10 min at 37°C, refixed in 4% paraformaldehyde for 2 min, and acetylated for 10 min in 0.25% acetic anhydride, 100 mM triethanolamine HCl, and 0.09% NaCl (Sigma). Sections were then dehydrated in an ethanol series, delipidated in chloroform for 15 min, washed again in ethanol, and air-dried. Probes were denatured and added to HB at a final concentration of 1μg/mL. Hybridization took place overnight at 37°C in a humid chamber. The probe was removed by washing in 2 × SSC, followed by RNase A (20μg/mL) at 37°C for 30 min, and then 0.2 × SSC at 42°C. Slides were incubated overnight at 4°C in anti-digoxigenin antibody diluted 1:2000 in PBS/10% heat-inactivated sheep serum solution, and washed again. Finally, slides were incubated in the NBT/X-phosphate color substrate solution(DIG Nucleic Acid Detection Kit, Boehringer Mannheim) plus 0.3% Triton X-100, in the dark for 24 h. Color solution was removed by washing in PBS, and the slides were mounted in aqueous medium (Geltol, Lipshaw, Detroit, MI) and photographed using an Olympus Vanox AHBS3 bright field microscope.

Immunohistochemistry. The immunohistochemical method employed has been previously described(19), with slight modification. After preincubation with 20% horse serum (Sigma) for 15 min, tissue cryostat sections were washed (twice for 10 min each, in PBS) and incubated at 4°C overnight with the same carboxy-terminal PP2A antibody used in Western analysis(20) at a 1:100 dilution. Control slides were incubated with PBS alone or with PBS and nonspecific rabbit IgG. After unbound primary antibody was washed off with PBS, sections were incubated for 1 h with a biotinylated anti-rabbit antibody (Amersham Corp.) at a 1:250 dilution, followed by incubation in avidin-biotin-horseradish peroxidase complex, 1:50 dilution, for 45 min(Vector Laboratories Inc., Burlingame, CA). Peroxidase activity was visualized by a color reaction using diaminobenzidine, 0.5 mg/ml (Sigma), as the substrate. The slides were mounted and examined under an Olympus Vanox AHBS3 bright field microscope.

RESULTS

Northern and Western analysis. Northern analysis of PP2Aα expression in the developing heart probed with the 2.4-kb cDNA is shown in Figure 1. The PP2Aα gene is highly expressed in the developing heart, and as expected, two major transcripts were detected(8, 13, 21). At all developmental stages examined, the predominant 1.8-kb and less abundant 2.4-kb transcript were equally present. PP2Aα mRNA levels of both transcripts appear relatively unchanged in the embryonic, newborn, and adult rat heart when compared with the control gene GAPDH.

Figure 1
figure 1

Expression of PP2Aα mRNA during heart development. Shown is a representative Northern blot performed using total RNA(10 μg per lane) extracted from embryonic d 14 (E14), newborn(NB), and adult (AD) rat hearts. The membrane was probed with a 2.4-kb PP2Aα full-length cDNA. Two predominant transcripts were identified at each developmental stage, a more abundant 1.8-kb transcript and a less abundant 2.4-kb transcript. The same membrane was subsequently hybridized with the control gene GAPDH. Molecular size markers are indicated on the left.

Full size image

Analysis of PP2Aα protein levels was then performed by Western blotting with a specific polyclonal antibody to the C-terminal region of the PP2A catalytic subunit (Fig. 2, top). This antibody specifically detected the purified 36-kDa PP2Aα control protein(lane 1) and a single protein band of the same size in all of the heart homogenates examined (lanes 2-6). Data from three Western blots were averaged and analyzed (Fig. 2, bottom). As shown, PP2A protein levels remained relatively constant from embryonic d 14 through term, but decreased significantly by 50% in the adult rat heart. Therefore, PP2A mRNA is highly expressed in the developing heart with the highest protein levels in the embryonic and newborn heart.

Figure 2
figure 2

PP2A protein expression during heart development.(Top) Representative Western blot using total protein (50 μg per lane) from rat hearts on embryonic d 14-20 (E14-E20), and from newborn (NB) rat hearts. The left lane (control) contained 1 μg of purified PP2Aα protein. An anti-PP2A C-terminal polyclonal rabbit antibody was used for detection. A single 36-kD protein band was identified at each gestational age. Protein molecular weight markers are indicated on the left. (Bottom) PP2A protein levels at various gestational ages are expressed as a percentage of adult level. Densitometric data from three different Western blots (mean ± SD) are shown. *E14-NB vs. adult, p < 0.05.

Full size image

In situ hybridization. To localize PP2Aα mRNA expression in the embryo and to compare expression in the heart with that of other tissues, in situ hybridization was performed in whole mount embryos at 12 d of gestation using a digoxigenin-labeled riboprobe from the PP2Aα coding region. Specific hybridization signal was detected in the heart, limb buds, tail, and parts of the brain (Fig. 3, top, Antisense). PP2Aα was regionally expressed in the developing heart with relatively abundant expresion in the ventricle and less in the walls of the atrium and truncus arteriosus. Strikingly, there was a gradient of expression from the positive outflow tract and truncus arteriosus to the aortic sac, the aortic arches, and the dorsal aorta, which had no detectable PP2Aα hybridization (Fig. 3, top and bottom, Antisense). The positive hybridization signal was not due to probe entrapment, as no signal was detected in control embryos hybridized with the noncomplimentary sense probe (Fig. 3, Sense).

Figure 3
figure 3

In situ hybridization of PP2Aα mRNA in whole mount rat embryos at 12 d of gestation. Digoxigenin-labelled sense and antisense RNA probes were synthesized from an 840 bp restriction fragment from the coding region of the 2.4 kb PP2Aα cDNA. A purple color indicates specific hybridization to PP2Aα mRNA.(Top) As viewed from the left side of the embryo, positive hybridization signal is detected in the tail, the lower limb bud, parts of the brain, and in the heart. No specific hybridization is seen with the sense control probe. (Bottom) Higher magnification of the heart is shown from the right side. Specific hybridization is detected in the walls of the atrium, the ventricle, and the truncus arteriosus. No hybridization is seen in the aortic sac. H, head; TL, tail; LB, limb bud;V, ventricle; A, atrial walls; TA, truncus arteriosus; and AS, aortic sac.

Full size image

To further localize cardiac PP2Aα expression, cross-sectional in situ hybridization was also performed on embryonic hearts throughout the last half of gestation (Fig. 4). At 12 d of gestation, PP2Aα mRNA was detected in cells throughout the heart, with the highest levels in the pericardium, the epicardium, the atrial wall, and in the endothelial cells that line the endocardial cushions (Fig. 4A). Note that the endocardial cushions are relatively acellular at this stage, being comprised primarily of extracellular matrix material. The lowest levels of PP2Aα were detected in the spongy myocardium. No hybridization was detected with the control sense probe in a serial section (Fig. 4B). At 15 d of gestation, PP2Aα was detected in the pericardium, the endocardial endothelium, myocardial cells throughout the compact and spongy layers, and in mesenchymal cells now populating the endocardial cushions(Fig. 4,C and D, arrowheads). At 16 d, expression persisted in some of these mesenchymal cells, and was particularly intense in the endothelial cells at the tips of the developing atrioventricular valve leaflets (Fig. 4E). By 20 d of gestation, the heart is structurally complete. At this time PP2Aα expression was predominantly localized to endothelial cells lining the trabeculations of the atrium and ventricle (Fig. 4F), with minimal diffuse hybridization in the underlying myocardium.

Figure 4
figure 4

In situ hybridization of PP2Aα mRNA during heart development. Frozen tissue sections of embryonic rat hearts are shown at various developmental stages from embryonic day (e)12-20. The same digoxigenin-labeled sense and antisense probes were used as in the whole mount experiments in Figure 3. A purple precipitate indicates specific PP2Aα mRNA expression(arrowheads). Sections are hybridized with the antisense probe (A, C, D, E, and F) or a control sense probe (B). Note the absense of specific hybridization signal in the representative sense control section (B). A and B are serial sections of an e12 heart. Arrowheads in A represent PP2Aα expressing pericardial, epicardial, and endothelial cells. C is a section of an e15 heart. D shows a higher magnification of the endocardial cushion region shown in C. Arrowheads in C and D point to PP2Aα expressing cells in the endocardium, myocardium, and endocardial cushions. E shows the developing atrioventricular valve leaflets of an e16 heart at high magnification with hybridization to endothelial cells lining the leaflets and underlying mesenchymal cells(arrowheads). F shows an e20 heart at low and high magnification(inset) with hybridization to endothelial cells (arrowheads). A, atrium; V, ventricle; P, pericardium; EC, endocardial cushions; S, interventricular septum; TA, truncus arteriosus.

Full size image

To verify specificity of the whole mount in situ hybridization in the brain and tail, these regions were also examined by cross-sectional in situ hybridization at 13-18 d of gestation. PP2Aα mRNA was highly expressed in the somites of the tail and in the columnar epithelial cells lining the ventricles of the brain. There was no detectable hybridization using the sense probe (data not shown).

Immunohistochemistry. To determine whether PP2Aα mRNA expression correlated with protein expression, we examined PP2Aα protein in the heart by immunohistochemistry (Fig. 5,A-D). At 15 d of gestation, PP2Aα protein was present in the same cell population that expressed mRNA (compare serial sections in Figs. 4C and 5A). This includes the pericardium, epicardium, endocardial endothelium, and myocardial cells in the compact and spongy layers. Transformed mesenchymal cells in the endocardial cushions also contained abundant PP2Aα protein, demonstrating both nuclear and cytoplasmic staining (Fig. 5B). At 20 d of gestation PP2Aα protein is concentrated in endothelial cells (Fig. 5C, inset), but is also present in the underlying myocardium, indicated by the diffuse brown staining (Fig. 5C). This is similar to the pattern of mRNA expression (compare serial sections of Figs. 4F and 5D). A control section, incubated without primary antibody, is shown in Figure 5D. Note the lack of nonspecific background staining.

Figure 5
figure 5

PP2A protein expression during heart development by immunocytochemistry. Frozen tissue sections of rat hearts from embryonic day(e) 15-20 were stained with an anti-PP2A C-terminal polyclonal rabbit antibody (A-C). The presence of PP2A protein is indicated by a brown color. A representative control section, shown in D was incubated without the primary antibody. A and B show the e15 heart at low and high magnification, respectively. Arrowheads illustrate PP2A containing ventricular and endocardial cushion cells with nuclear and cytoplasmic staining. C and D are serial sections of the e20 heart. The insets in C and D show the e20 heart at higher magnification. Arrowheads point to PP2A-containing endothelial cells. A, atrium; V, ventricle; EC, endocardial cushions; S, interventricular septum.

Full size image

At term, both PP2Aα mRNA and protein mapped specifically to cells in the epicardium and endocardium (Fig. 6,A and B) with very little mRNA or protein expression in the underlying myocardium. Interestingly, PP2Aα mRNA, but not protein, was expressed in the muscular walls of large coronary arteries (Figs. 6,C and D).

Figure 6
figure 6

PP2Aα mRNA and protein expression in the newborn rat heart. A and B demonstrate serial sections of newborn atrium, and C and D demonstrate serial sections of newborn ventricle containing a large coronary artery. Sections A and C were hybridized with the same antisense in situ probe as in Figure 3 and 4. Sections B and D were stained with the same anti-PP2A antibody used in Figures 2 and 5. Arrowheads in A and B indicate PP2Aα mRNA and protein expressing endothelial and epicardial cells. Arrowheads in C indicate hybridization to smooth muscle of a coronary artery.

Full size image

DISCUSSION

Cardiac morphogenesis requires the coordinate regulation of many developmental events. A number of primary effector molecules are involved including members of the transforming growth factor-β family, connexin-43, retinoic acid receptors, homeobox genes such as msx-1 and msx-2, and other transcription factors, such as GATA-4 and helix-loop-helix proteins(1–3). Clearly, appropriate signaling molecules must be present to mediate the interaction of signals from the extracellular matrix, cell surface receptors, and transcription factors. Many of these molecules are regulated by coordinate phosphorylation/dephosphorylation by protein kinases and phosphatases. To date the role of protein phosphatases in heart development is unknown.

PP2A is a highly conserved phosphatase responsible for dephosphorylating serine and threonine residues in all eukaryotes. PP2A is essential for a variety of cellular functions, including cell cycle regulation, cell transformation, and cell fate determination. It generally functions as a“growth suppressor”(6). We previously reported finding high levels of PP2Aα mRNA in the embryonic rat heart(14), and in the present study have determined PP2Aα expression at various stages of cardiac development.

Using a cDNA probe for the α-catalytic subunit of PP2A, Northern analysis demonstrated that PP2Aα mRNA levels were not grossly altered in the developing heart at the ages examined. What this Northern blot did demonstrate was that multiple PP2Aα transcript lengths exist in the rat heart, similar to that seen in Xenopus(8), Drosophilia(7), other adult rat tissues(13), and malignant cell lines(21). Genomic studies have demonstrated that PP2Aα is a single gene in the rat(12), and there are at least two possible polyadenylation sites within the sequence(12, 21). Therefore, these two transcripts represent either alternate splice products or different polyadenylation sites within the PP2Aα gene.

Although we did not detect obvious developmental differences in PP2A mRNA by Northern analysis, we did detect a significant 50% decrease in protein level in the adult heart by Western blotting. This suggests the possibility that post-transcriptional regulation is different in the fully mature heart compared with the developing heart.

To further examine the developmental expression of PP2Aα we performed in situ hybridization with whole mount 12-d gestation embryos. High levels of mRNA expression were found in the heart, portions of the brain, the limb bud, and the tail. Using Northern analysis of adult tissues, other investigators have similarly shown preferential expression of PP2Aα in certain organs. For example, mRNA levels are elevated in adult rat brain(13), pig brain and heart(13), and Xenopus heart and ovary(8), when compared to other adult tissues. This is the first demonstration of tissue specificity in the embryo. The gradient of PP2Aα expression from the ventricular outflow tract to the aortic arches suggests regional specificity for the action of PP2Aα in outflow tract formation.

We further localized PP2Aα within the developing heart using cross-sectional in situ hybridization and immunohistochemistry. PP2Aα mRNA was initially abundant in the pericardium, epicardium, and cardiac endothelial cells (e12), which all have different embryoloic origins. As development progressed (e15-16), PP2Aα was also highly expressed in mesenchymal cells of the endocardial cushions, and in individual myocytes throughout the myocardium. As the heart assumed its mature phenotype, there was less expression in the underlying myocardium, but both mRNA and protein remained abundant in the endothelial and epicardial cells. Although PP2A has been thought to be a ubiquitous enzyme, necessary for basic cell function, we have clearly shown that the catalytic subunit is not uniformly expressed in all embryonic cells. It is likely that small amounts of PP2A are present in all cells, but our results demonstrate marked up-regulation in certain cell populations within the developing embryo.

In the endocardial cushions, both the endothelial and newly transformed mesenchymal cells expressed high levels of PP2Aα. Of particular interest is that, at 16 d of gestation, the highest levels occurred at the tips of the endocardial cushions, in endothelial cells that were not transformed. Perhaps the role of PP2Aα here is to limit epithelial-mesenchymal transformation, or inhibit cell migration, keeping such processes in check. It is interesting that other places where important epithelial-mesenchymal interactions occur in the embryo, such as the brain and the limb buds, also expressed high PP2Aα levels by whole mount in situ hybridization.

There was excellent agreement between protein and mRNA expression in all of the tissues examined except in the walls of large coronary arteries, where we detected message, but no protein. It is unlikely that the message is transcribed but not translated. Another explaination may be that PP2Aα protein is present, but phosphorylated on Tyr307 within the carboxyterminal antibody binding region. Such phosphorylation is known to impair antibody binding to PP2A, and has been shown to occur in intact cells in response to growth factors or cell transformation(22). Therefore, the coronary artery results shown in Figure 6 may reflect physiologic regulation of PP2Aα itself.

The question then becomes what target proteins are dephosphorylated by PP2Aα during cardiac morphogenesis? Multiple targets are likely, as PP2Aα is expressed in both the nucleus and cytoplasm of cells. Furthermore, activity of the catalytic subunit is nonspecific, and substrate specificity is provided by the associated regulatory subunits. Our antibody and in situ probes would have detected any holoenzyme complex containing the α isoform of the catalytic subunit regardless of which regulatory subunits were present.

We can speculate on a number of possible PP2Aα targets during cardiac development. For example, connexin-43 gap junction communication is impaired in vitro by phosphorylation on serine residues and is restored by PP2A(23). Whether or not this occurs during embryogenesis is unknown, but it is especially interesting in light of the cardiac malformations seen in the connexin-43 knockout mouse(24) and the recent link between connexin-43 mutations in humans and congenital heart disease(25). Another example is the retinoblastoma protein, a nuclear “pocket protein” involved in cell-cycle regulation. The under- or unphosphorylated form has antiproliferative activity and is predominant in G1, whereas the hyperphosphorylated form is predominant during G1/S phase transition(26). This suggests that PP2A may keep proliferation in check via the retinoblastoma protein. Growth factor receptors are yet another possible target for PP2A. The epidermal growth factor receptor, for example, loses its affinity for its ligand when phosphorylated on serine/threonine residues(27). PP2A might therefore promote the mitogenic response to growth factor stimulation. These examples suggest that PP2A may be a growth promoter or inhibitor, depending on its molecular target. Its striking predominance in endothelium also suggests a possible role in transduction of mechanical signals such as shear forces. It should be emphasized that there are no data implicating any specific function for PP2A in development at this time. The pattern of expression shown here suggests a number of interesting possibilities. Further studies are necessary to define the role of protein phosphatases in embryogenesis.