Main

During embryonic development normal cardiac structure and function is established through coordinated gene expression in response to developmental and environmental cues. During midgestation, the embryonic heart must accomplish several simultaneous goals including rapid growth, morphogenesis, and contraction. This requires continued cellular proliferation concurrent with myocyte differentiation. It is during this period of development that the embryonic circulation becomes necessary to sustain viability of the embryo. Identification of genes whose expression is required and specifically enhanced at this critical stage of embryonic cardiac development will contribute to our understanding of the molecular regulation of cardiogenesis and cardiac disease.

Between Theiler stages 16 (E10.5) and 19 (E11.5), critical developmental events occur in the heart and in the great vessels(1). Within the common atrial chamber of the heart, increased evidence of septation is noted with the growth of the primum septum toward the atrioventricular bulbar cushion. In the ventricular chamber, significant progress in the development of the muscular and membranous components of the interventricular septum is evidenced along with a dramatic increase in the degree of trabeculation. In the outflow tract, further differentiation of the aorticopulmonary spiral septum occurs in preparation for separation into two distinct outflow tracts that will be completed by stage 21 (E13). The later event coincides with cardiac neural crest migration and colonization of the cardiac outflow tract(2,3). The lack of information regarding the regulation of these significant events between Theiler stages 16 and 19 led us to focus our attention on the study of cardiac genes regulated during this period.

Recently, an expanding number of introduced or identified mutations resulting in midgestation lethality have been discovered that result in a common cardiac phenotype characterized by a "thinned myocardium" and signs of heart failure including pericardial and peripheral edema. Examples include the Splotch and Patch mutations as well as mutations in the neurofibromatosis type 1 gene (Nfl) and various retinoic acid receptor genes(4–8). In each case, the myocardium appears poorly developed though trabeculae composed of relatively mature cardiomyocytes are present. The abnormal myocardium is evident by E11.5 and embryonic lethality results by E13.5. We have sought to more fully understand the molecular nature of these defects by defining gene programs required for midgestation cardiac maturation.

Here, we report the identification of several genes with activated or enhanced expression between E10.5 and E11.5 of mouse cardiac development. One of these, p57Kip2, encodes a cyclin-dependent kinase inhibitor that is implicated in cardiomyocyte differentiation. In Splotch embryos, p57Kip2 expression is expanded in the myocardium, suggesting that premature cardiomyocyte exit from the cell cycle and precocious differentiation may account for the "thinned myocardium" phenotype.

METHODS

Tissue collection. E10.5 and E11.5 embryos were collected by standard methods in ice-cold PBS from anesthetized and euthanized pregnant CD-1 mice (Charles River Laboratories, Southbridge, MA). The day of vaginal plugging was designated as E0.5. Dissected mice embryos had to fulfill the morphologic criteria for gestational age as described by Theiler (stages 16-17 for E10.5 embryos and 19 for E11.5 embryos, respectively)(1). The embryonic heart was dissected from the pericardial sac and divided from the body of the embryo at the level of the common arterial trunk. The whole embryonic heart was harvested. Approximately 300 hearts were collected and pooled at E10.5 and 200 at E11.5. The tissue was rapidly frozen by immersion in liquid nitrogen and stored up to 2 wk before RNA preparation. All protocols involving animals were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania.

Total RNA extraction and isolation of poly(A)+ RNA. After pulverization of the collected embryonic cardiac tissue in a glass dounce homogenizer, total RNA was extracted using TRIzol reagent (GIBCO, Life Technologies, Grand Island, NY). Polyadenylated RNA (mRNA) was then prepared by affinity to oligo-dT cellulose (Oligotex, Qiagen, Valencia, CA) with a yield of 1.1-1.3%. All RNA samples were kept dissolved in 2% diethyl pyrocarbonate-treated deionized H2O at -80°C. We collected 4.5 µg of poly(A)+ RNA from 300 E10.5 hearts, and 4.7 µg of poly(A)+ RNA from 200 E11.5 hearts.

Generation of subtracted E11.5-E10.5 cDNA. The subtractive hybridization strategy used was designed to identify transcripts present at increased levels at E11.5 compared with E10.5 in the heart. We used SSH to increase our chances of isolating rare as well as high copy number transcripts(9–11). Single-stranded cDNA was synthesized from 1 µg of poly(A)+ RNA using the PCR-Select cDNA Subtraction Kit (Clontech, Palo Alto, CA) with an oligo-dT primer and Superscript II reverse transcriptase (GIBCO, Life Technologies). Second-strand synthesis was performed in the presence of DNA polymerase I, Escherichia coli DNA ligase, and RNAse H. Double-stranded cDNA was treated with T4 polymerase and digested with RsaI. SSH was carried out as recommended (Clontech) using E10.5 cDNA and E11.5 cDNA after addition of appropriate linkers to allow for subsequent PCR amplification.

Two rounds of subtractive hybridization were performed. In the first, a 10-fold excess of E10.5 cDNA was added to E11.5 cDNA. The mixtures were then heat denatured (1.5 min at 98°C) and allowed to reanneal for 12 h at 68°C. After the first hybridization, the two primary hybridization samples were combined without denaturing, and an additional 10-fold excess of fresh denatured E10.5 cDNA was added. The sample was allowed to hybridize for an additional 24 h at 68°C. After a brief preincubation (5 min, 75°C), two PCR amplifications were performed. In the primary PCR amplification the subtracted double-strand cDNA templates were amplified using a program of 45 cycles each of 94°C for 30 s, 65°C for 30 s, and 72°C for 90 s. In the secondary PCR, nested primers were used under the same conditions for 25 cycles. For both PCR amplifications, three different template concentrations were used for optimization of reaction efficiency. PCR mixtures were prepared using Advantage Klen Taq Polymerase Mix (Clontech) containing TaqStart Antibody for hot start PCR. The PCR products were analyzed by 2% agarose TAE gel electrophoresis. The final PCR products were purified (QIAquick PCR purification kit, Qiagen), eluted in 50 µL of deionized H2O, and stored at -20°C. All PCR and hybridization reactions were performed in an MJ Research thermal cycler. A portion of the purified PCR products was subsequently cloned directly into PCRII using TA cloning (Invitrogen, Carisbad, CA).

Analysis of the subtracted cDNA clones. Clones harboring inserts were sequenced by automated fluorescence sequencing at the Sequencing Core Facility of the University of Pennsylvania. Nucleic acid homology searches of the GenBank, EMBL, and dbEST databases were performed using the BLAST program through e-mail servers at the National Center for Biotechnology Information (National Institutes of Health, Bethesda, MD).

Northern dot-blot analysis. Clone-specific DNA probes were prepared from 5 to 10 µg of miniprep plasmid DNA that was subjected to EcoRI restriction and agarose gel insert purification (Qiaex gel extraction kit, Qiagen). Twenty-five nanograms of purified product was used for radioactive labeling with [32P]dCTP. Unincorporated label was removed before hybridization using a Micro Bio-Spin chromatography column (Bio-Rad, Hercules, CA). Aliquots of 0.04 µg of poly(A)+ RNA from E10.5 and E11.5 embryonic hearts were denatured and blotted in 10 × SSC onto nylon membranes (Gene Screen, Dupont, Boston, MA) using a Bio-dot blotting apparatus (Bio-Rad). The filters were hybridized with 25 ng of the 32P-labeled cDNA probes (specific activity 0.76-1.9 × 109 cpm/µg) at 42°C for 12 h. A 32P-labeled mouse β-actin probe (Ambion, Austin, TX) was used as control. The filters were washed in 0.1× SSC, 0.1% × SDS at 45°C and then subjected to autoradiography at -80°C for 1 d to 3 wk.

In situ hybridization. The radioactive in situ hybridization protocol used in this study has been described elsewhere(12). E10.5 and E11.5 whole mouse embryos were collected, fixed overnight in 4% PFA at 4°C, dehydrated through an ethanol series, and embedded in paraffin. The embedded embryos were sectioned at 8 µm and placed onto slides (Fisher Superfrost Plus, Pittsburgh, PA). After rehydration, pretreatment, and prehybridization, samples were allowed to hybridize overnight at 50°C with 35S-labeled probes. For preparation of sense and antisense probes, plasmids containing cDNA inserts were linearized with BamH1 or Not1, and cRNA 35S-labeled probes were transcribed using T7 or Sp6 RNA polymerases (Promega, Madison, WI). After hybridization and washing, dried slides were exposed to autoradiography film for 1-3 d and subsequently coated with photographic emulsion (Kodak NTB2, Rochester, NY). We determined the length of emulsion exposure from 4 to 14 d based on the intensity of the signal on autoradiography film. Slides were developed and counter-stained with 2 µg/mL Hoechst 33258 nucleic dye (Sigma Chemical Co., St. Louis, MO). In all cases, sense probes gave no signal over background.

Reverse transcription-PCR. RNA was isolated from E9.5 microdissected hearts derived from crosses of heterozygous Splotch mice. The genotype of each embryo was determined from genomic DNA isolated from extra-embryonic membranes using a PCR-based method, as described previously(12). Reverse transcription was performed using random hexamers and Superscript II reverse transcriptase as directed (GIBCO, BRL). PCR amplification was performed using p 57Kip2 or G6PDH-specific primers for 30 cycles (94°C for 1 min, 60°C for 1 min, 72°C for 1 min) and analyzed by ethidium bromide-stained agarose gel electrophoresis.

Electron microscopy. Embryonic hearts were excised and fixed in 0.1 M cacodylate buffer with 2% glutaraldehyde for 2-4 h at 4°C, washed in 0.1 M cacodylate buffer, fixed again in 0.1 M cacodylate buffer with 1% osmium tetroxide for 60 min at 4°C, washed in 0.1 M cacodylate buffer and twice with distilled water, and stained with 2% aqueous uranyl acetate for 30 min at 4°C. Samples were then washed in distilled H2O, dehydrated through an ethanol series, and embedded in epoxy. Ultrathin (100 nm) sections were prepared and analyzed using a JEOL 100CX transmission electron microscope by the Bio-medical Imaging Core Facility at the University of Pennsylvania. A low-power image was used as a guide for the analysis of high-magnification images that formed a composite of an entire cross section at the mid-ventricular level of wild-type and Splotch E9.5 hearts. Every cell contributing to the cross sections was analyzed by planimetry to determine cytoplasmic area (total cellular area minus nuclear area), and myofibril area. Individual cells that contained myofibrils with two consecutive Z-lines and those with intercalated disks were scored and expressed as the percent of the total number of cells analyzed, which was greater than 100 in each case.

RESULTS

To identify genes with enhanced or activated expression between E10.5 and E11.5, we performed a subtractive hybridization experiment using cDNA derived from E10.5 and E11.5 mouse hearts. We used SSH to improve the representation of low copy number transcripts(9–11). We used excess E10.5 cDNA so that transcripts present at increased frequency in the E11.5 pool would be identified. After subtraction, 48 individual bacterial clones containing inserts were sequenced and found to represent 21 different genes. The average insert size was 253 bp, approximating the statistically expected 256-bp size after digestion with a four-base cutting restriction enzyme. Homologous sequences were identified by BLAST searches of GenBank, EMBL, and dbEST databases. Of the 21 unique sequences, 16 corresponded to known genes. Included among these were three genes identified as fetal hemoglobin, likely from contaminating red cells, and one identified as 28s rRNA. These genes were not further examined. Five cDNA clones with no significant homology to any known sequences in the GenBank or EMBL databases were isolated. Four of these five matched ESTs, and in two cases the longer EST sequences corresponded to known genes. The fifth unknown clone represented a novel sequence. The expression of individual genes in E10.5 and E11.5 hearts was determined by Northern dot-blot analysis. Differential expression was noted in nine of the 17 clones (Fig. 1 and Table 1). The exposure times for the dot-blot autoradiograms shown in Figure 1 ranged from 1 to 20 d, suggesting that differentially expressed cDNA of varying abundance were identified by SSH, consistent with previous reports(9–11). Control blots were performed in parallel using β-actin as a probe to confirm equal loading of E10.5 and E11.5 mRNA (Fig. 1J). A summary of the differentially expressed cDNA clones is shown in Table 1. The known genes are classified by functional category according to Adams et al.(13). Three of these cardiac genes encode contractile or structural proteins (α-tropomyosin, cardiac troponin T, and cardiac α-actin); two correspond to cell-cycle regulatory genes (cyclin B1 and p57Kip2); one represents a gene involved in cellular metabolism (cytochrome c oxidase III); and one is the transcription factor YB-1. One fragment (clone 35) is identical to mouse EST encoding unknown functions, and one is a novel sequence (clone 74).

Figure 1
figure 1

Northern dot-blot analysis of differentially expressed subtracted clones. cDNA inserts from each clone were radioactively labeled and hybridized to dot blots containing 0.04 µg of poly(A)+ RNA from cardiac tissue of E10.5 (left) and E11.5 (right) embryos. A, Cardiac troponin T; B, cardiac α-actin; C, α-tropomyosin; D, YB-1; E, p57Kip2; F, cyclin B1; G, cytochrome c oxidase; H, clone 35; I, clone 74; J, control (mouse β-actin); K, laminin receptor 67-kD; and L, NADH ubiquinone reductase. K and L represent examples of isolated clones that were expressed equally at both gestational days. The exposure times were 1 d (A-F, J-L), 3 d (G), and 20 d (H, I).

Full size image
Table 1 Differentially expressed genes in mouse heart at E11.5 compared with E10.5
Full size table

In situ hybridization was performed for all differentially expressed clones and was consistent with the results of the dot-blot screening. For example, cardiac troponin T was not detected at E10.5 by dot blotting (Fig. 1A) or by in situ hybridization (Fig. 2A). However, by E11.5, transcripts were detected by both methods (Figs. 1A, 2B) with greatly enhanced expression in both atrial and ventricular myocardium (Fig. 2B).

Figure 2
figure 2

In situ hybridization analysis of differentially expressed clones. mRNA expression at E10.5 and E11.5 of differentially expressed genes were analyzed. Cardiac troponin T (A, B), clone 35 (C, D), and p57Kip2 (E, F, G) expression at E10.5 (A, C, E) and E11.5 (B, D, F-I) is shown in transverse (A-D, H, I) or sagittal (E-G) sections. Troponin T message was undetectable at E10.5 (A) and present at E11.5 (B) in the myocardium of the heart (ht). Expression of clone 35 was not detected at E10.5 (C), but was evident in the outflow tract (ot) of the heart, the neural tube (nt), the dorsal root ganglia (drg), the somite (so), and the nasal epithelium (ne) at E11.5 (D). p57Kip2 was present at E10.5 (E) and E11.5 (F) in the heart, but up-regulated at E11.5 (see also Fig. 1E). A higher magnification of the sagittal section of the heart in F is shown in G, to emphasize the greater expression of p57Kip2 in the inner trabecular zone compared with the peripheral compact zone.

Full size image

We identified two novel genes with enhanced expression at E11.5. Clone 74 contained a short (93 bp) insert with no homology to known sequences. It was expressed at low levels in the heart by both dot blot (Fig. 1I) and in situ hybridization (not shown). A second novel gene was represented by clone 35. Both Northern blot analysis (Fig. 1H) and in situ hybridization (Fig. 2C) revealed weak or no expression at E10.5. By E11.5 (Fig. 2D), expression was evident in the outflow tract of the heart, in the neural tube, the nasal epithelium, the dorsal root ganglia, and the somites. Clone 35 included a 309-bp open-reading frame encoding a putative peptide conserved through millions of years of evolution (Fig. 3). The 103-amino acid putative peptide is 75% identical to a protein of unknown function in the puffer-fish (Fugu rubripes). Construction of a cDNA contig using overlapping murine EST sequences revealed a high level of predicted amino acid sequence conservation throughout the length of the protein (Fig. 3). Highly homologous human EST sequences, including those derived from human fetal heart (AA733158), predict the existence of a very similar human protein (97% amino acid identity; Fig. 3). A related protein with ∼40% amino acid identity has been reported in C. elegans (accession No. CELR12C12.6). Expression of this gene in the outflow region of the heart is of particular interest as outflow tract septation is just beginning at E11.5 and gene programs regulating this process are poorly understood. Further description and analysis of these novel genes will be provided elsewhere.

Figure 3
figure 3

Clone 35 represents a novel gene encoding a putative protein conserved through evolution. The 438-bp cDNA designated clone 35 contains an open-reading frame encoding 103 amino acids (underlined portion of murine sequence). Overlapping murine ESTs were used to predict the remainder of the murine sequence shown (accession Nos. AA427236, C89394, AA242342). Homologous human ESTs predict a very similar human protein (accession Nos. AI076446, AI034167, AA733158, AA121948, AA985583, AA824395, AA661638, AA401153, W32784, AA297477). The predicted peptide shares high sequence homology to a putative protein in Fugu rubripes (accession No. AF026198). Positions of amino acid identity to the predicted murine sequence are shown with a dash (-) and amino acid differences are given in small letters. Gaps to improve sequence alignment are shown with dots.

Full size image

One of the genes that emerged from our screen had a particularly interesting pattern of expression in the developing heart. p57Kip2, encoding a cyclin-dependent kinase inhibitor(14), was expressed at low levels at E10.5 and at much higher levels by E11.5 (Fig. 1E). In situ hybridization analysis revealed differential expression within the ventricular myocardium (Fig. 2, F and G). It was poorly expressed in the outer compact zone that is composed of replicating cardiomyoblasts, whereas it was highly expressed in the forming trabeculae, which consist of nonreplicating, more fully differentiated cardiomyocytes. The increased size of the trabecular layer at E11.5 compared with E10.5 and the concomitant increased level of p57Kip2 expression account for the isolation of this clone in our screen.

To assess the possible functional role of p57Kip2 in cardiac development, we have examined relative expression in a mouse model with developmental myocardial defects. Homozygous Splotch embryos die by E13.5 owing to cardiac malformations that include outflow tract disorders related to neural crest deficiency(15). In addition, these embryos have a "thinned myocardium" (Fig. 4B) in which trabeculation is evident but overall wall thickness is reduced. This defect becomes obvious at the histologic level by E12.5. We examined the expression of p57Kip2 in homozygous E11.5 Splotch embryos and wild-type litter mates (Fig. 4, C and D). We noted expansion of p57Kip2 expression in the Splotch hearts such that nearly the entire myocardial wall was included. No significant outer compact zone devoid of p57Kip2 expression could be identified. Unlike wild-type hearts, nearly all of the cardiomyocytes in Splotch hearts expressed p57Kip2 by E11.5.

Figure 4
figure 4

Abnormal myocardium in Splotch mutant embryos. Hematoxylin and eosin-stained sections of E12.5 wild-type (A) and Splotch (B) right ventricle reveals an abnormally thinned myocardium in Splotch embryos. Note that trabeculation is present and that the outer compact zone is deficient (black arrows). p57Kip2 expression in transverse sections of wild-type (WT) and Splotch (Sp) E11.5 hearts is shown (C, D). The probe, exposure, and processing conditions were identical for the control and Splotch samples. Note expanded expression of p57Kip2 in the Splotch heart compared with wild-type such that a compact zone devoid of p57Kip2 expression (white arrow, C) is not present.

Full size image

This result suggested to us that the thinned myocardium of Splotch embryos might be related to precocious exit from the cell cycle and myocyte differentiation within the compact layer of mutant hearts. Therefore, we examined the expression of p57Kip2 in wild-type and Splotch hearts earlier, at E9.5, by RT-PCR (Fig. 5). We noted more abundant p57Kip2 transcripts in Splotch hearts whereas the control G6PDH transcripts were detected at comparable levels. We next examined the state of myocardial differentiation in Splotch and wild-type hearts by electron microscopy. We analyzed all of the cells contained within several cross sections taken at the mid-ventricular level and determined the degree of cardiomyocyte differentiation by counting mature sarcomeres (containing more than two consecutive Z-bands), intercalated discs, and the percent of the cytoplasm occupied by myofibrils. As has been previously reported(16), we noted more mature sarcomeres in the inner trabecular layer of wild-type hearts compared with the outer compact zone. As early as E9.5, there was a significant increase in the number of mature sarcomeres in Splotch compared with wild-type hearts (Table 2, Fig. 6). This was evident in both the subepicardial region (25% versus 10% of cells with mature sarcomeres) and in the trabecular layer (68% versus 15%). We also noted a significant increase in the number of cells with intercalated discs in Splotch compared with wild-type hearts (60% versus 20%) whereas the percent of the cytoplasm occupied by myofibrils was similar (8.9% versus 7.2%, an insignificant difference). These results are consistent with a more advanced state of cardiomyocyte maturation in Splotch embryo hearts although the overall amount of myofibrilogenesis is similar.

Figure 5
figure 5

RT-PCR analysis of Splotch and wild-type E9.5 hearts reveals increased expression of p57Kip2 in Splotch hearts. Total RNA was isolated from E9.5 microdissected hearts from Splotch and wild-type litter mates. p57Kip2 and G6PDH (control) expression was determined by RT-PCR. Enhanced expression of p57Kip2 is evident in Splotch hearts.

Full size image
Table 2 Increased maturation of contractile apparatus in Splotch embryos as determined by electron microscopy
Full size table
Figure 6
figure 6

Electron microscopic analysis reveals early maturation of Splotch embryo myocardium. Subepicardial regions of E9.5 wild-type (A) and Splotch (B) myocardium show significant differences in relative maturation of contractile apparatus by electron microscopy. Small bundles of myofibrils are present in wild-type myocardium, whereas more mature sarcomeres with consecutive Z-lines (black arrows, B) are present in Splotch myocardium. Many more intercalated disks are also evident (white arrowhead, B). Quantification of electron microscopic analysis is shown in Table 2.

Full size image

DISCUSSION

SSH using RNA from microdissected embryonic hearts was used in this study to identify known and novel genes that participate in mid-cardiac embryonic development. Our results undoubtedly represent only a sample of the genes with enhanced expression between E10.5 and E11.5. We concentrated on those with the most dramatic and unambiguous increase in expression. Three of the differentially expressed cardiac genes encode contractile components of the sarcomere (α-tropomyosin, cardiac troponin T, cardiac α-actin) not previously known to be regulated during this time of development. Mutations in the genes encoding α-tropomyosin and cardiac troponin T are responsible for some forms of hypertrophic cardiomyopathy suggesting that these genes are important in maintaining force generation(17,18). The expression of α-cardiac actin, the mature actin isoform, is associated with changes in cardiac function. It is down-regulated in the adult heart in response to hypertrophic stimuli, whereas the fetal isoform of α-skeletal actin is up-regulated(19,20). α-Cardiac actin expression is also decreased in association with anthracycline-induced cardiomyopathy(21). A defect in force transmission secondary to actin dysfunction has been suggested as a cause of idiopathic dilated cardiomyopathy after the identification of cardiac actin mutations in some patients with this disease(22). It is likely that enhanced expression of these sarcomeric proteins during the period studied is associated with an increase in contractile function, though further studies will be needed to confirm this contention. Consistent with this hypothesis, we found enhanced expression of cytochrome c oxidase III, the terminal enzyme of the respiratory cycle, during the same period. Decreased expression of cytochrome c oxidase has been reported in conditions such as heart failure and cardiac aging(23), indicative of its significance in maintaining adequate energy supply to the myocardium.

One of the most interesting genes with enhanced expression at this time is the gene encoding p57Kip2, a cyclin-dependent kinase inhibitor. This message was confined to the trabecular zone of wild-type E11.5 myocardium and is likely to be involved in cell-cycle arrest and differentiation of cardiomyocytes. A related gene, p21, has been implicated in myoD-dependent differentiation of skeletal muscle(24). We have also noted p57Kip2 expression in the CNS confined to the intermediate zone where maturing neurons stop dividing after migrating from the ventricular zone. Inactivation of the imprinted p57Kip2 gene in the mouse results in diffuse abnormalities of cellular proliferation and apoptosis reminiscent of patients with Beckwith-Wiedemann syndrome(25,26). Cardiac defects have not been reported in association with this condition, probably because of compensation by other cyclin-dependent kinase inhibitors including p27(27). We believe that p57Kip2 nevertheless is likely to be a critical regulator of embryonic myocardial cell differentiation.

This possibility is strengthened by the expression pattern of p57Kip2 in the myocardium of Splotch embryos. The Splotch myocardium is thinned and displays abnormalities of excitation-contraction coupling(28). Our results indicate expanded expression of p57Kip2 throughout the ventricular wall of Splotch embryos. This is associated with more mature cardiomyocytes and contractile structures and a relative deficiency of the compact layer in which replicating cardiomyoblasts required for continued growth of the embryonic heart reside. Each of these results is consistent with an early suppression of cardiomyocyte replication by p57Kip2.

In summary, we have identified several genes with enhanced expression in the embryonic heart at E11.5 compared with E10.5. Several of these encode contractile proteins consistent with increased functional cardiac demand during this time in development. In addition, expression of p57Kip2 is enhanced during this period associated with maturation of cardiomyocytes and expansion of trabeculation. p57Kip2 expression is normally restricted to the trabecular layer at E11.5, but expression is expanded to include nearly the entire thickness of the myocardium in Splotch embryos, which display precocious cardiomyocyte differentiation. These results suggest that p57Kip2 may play an important role regulating the balance between proliferation and differentiation in the embryonic heart.