INTRODUCTION

During cardiogenesis a repertoire of cardiac and muscle specific genes are activated. At the same time, genes characteristic of an embryonic phenotype are down-regulated (1, 2, 3, 4, 5). Under conditions such as ischemia, hypertrophy, and injury, some of these embryonic genes are reexpressed (6, 7, 8, 9). Analyses of the activation of cardiac genes (e.g. rat myosin light chain 2, slow/cardiac troponin C) have resulted in the identification of cis elements and trans factors determining cardiac specific gene expression (10, 11). The transcriptional mechanism(s) regulating embryonic gene down-regulation during development or reexpression in response to stressors have not been characterized.

Creatine kinase, via the reversible phosphorylation of ADP and creatine, is critical for the maintenance and regulation of cellular energy stores in highly oxidative tissues such as the heart (12). The brain creatine kinase (BCK) gene is the predominant cytoplasmic creatine kinase expressed in the embryonic heart, is down-regulated to low but easily detectable levels in the adult heart, and is reinduced in the adult in response to stimuli producing ischemia, hypertrophy, or failure (6, 13, 14, 15). We and others have demonstrated that BCK gene expression is transcriptionally regulated via sequence located within the 5-flanking region in C₂C₁₂, HeLa, and neuroblastoma cells (16, 17). We recently described the presence of two negative regulators in this region, one which functions as a silencer (17). That report also suggested that sequence located between 92 and +80 contained the elements necessary for regulating BCK expression during C₂C₁₂ differentiation. Subsequent reports have localized the critical region to a small sequence within the first exon (18, 19). These data suggest that BCK is transcriptionally regulated during cardiac development and imply that the regulatory elements are likely contained within the 5-flanking region. Thus, analyzing the transcriptional regulation of BCK in the heart may provide insight into the mechanisms governing embryonic gene expression during cardiogenesis. Accordingly, the transcriptional regulation of BCK in the heart was analyzed.

EXPERIMENTAL PROCEDURES

Primary neonatal cardiocytes were isolated by standard methods that result in 90% pure cardiocytes (20). One- to 2-day-old Sprague-Dawley rats were killed, and their hearts were quickly rinsed in ice-cold PBS.¹ The hearts were then minced in a 0.2% pancreatin-containing solution. Following a 15-min incubation at 37 °C with continuous stirring, the supernatant was removed and discarded. An additional 10 ml of pancreatin solution was added, and the process was repeated. The process was repeated five to six additional times with retention of the supernatant following each 15-min time interval. The digested cells contained in the supernatant were quickly pelleted following each step and resuspended in PC-1 media (Hycor) supplemented with 10% fetal calf serum, L-glutamine, and antibiotics. All cells were then plated in large flasks for 1.5 h to allow fibroblast fallout. The cell-containing solution was plated on Primaria (Falcon) 60-mm dishes (20). The number of dishes plated was approximately one per heart isolated and the volume 3 ml per dish. This resulted in 5 × 10⁵ to 1 × 10⁶ viable cells per dish 24 h following plating.

Ten- to 12-week-old Sprague-Dawley rats weighing approximately 250 g were sedated by injection of 0.37 mg/g body weight of chloral hydrate intraperitoneally. Once sedated, the heart was delivered to the exterior of the chest wall via a left mid-chest wall incision, grasped, and a solution containing 15 µg of pSV2CAT (to control for transfection efficiency) and 50 µg of experimental DNA injected into the left lateral wall and apex of the ventricle. The chest wall was closed, and the rats were allowed to recover (10).² Five days later, hearts were removed for isolation of nuclear proteins and determination of luciferase (LUC) and chloramphenicol acetyltransferase (CAT) activity (21).

pMSVGal, pSV₂CAT, and pSVOCAT have been described elsewhere (15, 17). The constructs identified as BCK1100CAT, BCK388CAT, and BCK92CAT have been previously reported (17). Constructs labeled BCK+57CAT, BCK+41CAT, BCK+25CAT, BCK+17CAT, BCK+10CAT, and BCK+1CAT were prepared using convenient restriction enzyme sites within the first exon and/or using oligonucleotides as described previously (22). Briefly, BCK promoter fragments with the same 5 end at 92 and different 3 ends at +57, +41, +25, +17, +10, and +1 were placed in the sense orientation upstream of the CAT reporter gene. The BCK388LUC, BCK92LUC, BCK+57LUC, BCK+41LUC, BCK+25LUC, and BCK+1LUC constructs were prepared by subcloning the BCK promoter fragments upstream of the luciferase reporter gene of pXP₂LUC (23) (generously donated by Muthu Periasamy). For analysis of the element as an enhancer, a double-stranded oligonucleotide corresponding to bases within the first exon from +25 to +57 was subcloned into the SmaI site upstream of the hamster sarcoma virus thymidine kinase promoter of pT109LUC, a luciferase reporter vector designed for testing enhancer function (ATCC no. 37584) (23, 24). This construct, which contains two copies of the element in the sense orientation, is termed Seam/pT109. As a control for nonspecific enhancer effects of a DNA fragment, the 32-base pair polylinker region of pXP₂LUC from the BamHI site to the XhoI site was isolated, the 5 overhangs were blunt ended using the Klenow fragment, and the fragment was subcloned into the SmaI site of pT109LUC. This construct was termed Poly/pT109. Orientation and copy number for all constructs was determined by the dideoxynucleotide sequencing method (Sequenase, U. S. Biochemical Corp.) (25).

Mutated constructs were prepared using the approach noted above for preparing deletion constructs (22). The protocol used is as follows. Mutations were made in the University of Cincinnati core oligonucleotide synthesis laboratory. Mutant oligonucleotides are shown in Scheme 1 in the sense orientation 5 to 3 with the mutated nucleotides capitalized and in bold type. The name of the resulting mutant construct is listed to the left of the oligo used in its preparation. Mutated oligonucleotides were combined with the oligonucleotide corresponding to the 5 end of the insert in a solution containing dNTPs, magnesium chloride, buffer, Amplitaq DNA polymerase (Perkin-Elmer), and template (BCK92CAT). Thermocycler parameters were empirically optimized, and the product was identified on a polyacrylamide gel according to the manufacturer's instructions (26). This product was subcloned into the TA cloning vector (Invitrogen). Mutations were confirmed by sequence analysis. The remainder of the product was also sequenced to ensure that no additional mutations were present. Inserts were released by HindIII digestion and subcloned into HindIII-digested and dephosphorylated pSVOCAT or pXP₂LUC. Orientation and copy number were determined by differential polymerase chain reaction analysis and confirmed by sequencing.

Scheme 1

EMSAs were performed using a 1/2 × tris borate-EDTA, 5% polyacrylamide gel electrophoresed for 2.5 h at 160 mV at 4 °C (27). Typically 10 µg of nuclear protein were incubated with 2 µg of poly(dI-dC), 10,000 cpm of radiolabeled probe in an aqueous solution with a final concentration of 20 mM Hepes, pH 7.9, 1.5 mM MgCl₂, 1 mM EDTA, 60 mM KCl, 1 mM DTT, and 5% glycerol. For competition experiments, an indicated molar excess amount of unlabeled oligonucleotide or DNA fragment was added to the solution and incubated for 10 min prior to the addition of the labeled probe. Unlabeled competitors were prepared from plasmid DNA fragments or annealed oligonucleotides, native and mutated. Probes were end- or fill in-labeled with ³²P using standard methods (28). Oligonucleotides used in EMSAs as probes and competitors were prepared in the University of Cincinnati core oligonucleotide synthesis laboratory and are listed in Scheme 2 5 to 3.

Scheme 2

Plated cells were scraped in PBS and pelleted at 1850 × g for 10 min, and the pellet was resuspended in a volume of hypotonic buffer (10 mM Hepes, pH 7.9, 1.5 mM magnesium chloride, 10 mM potassium chloride, 0.2 mM PMSF, and 0.5 mM DTT) five times the packed cellular volume (27). Cells were immediately pelleted at 1850 × g for 5 min, and the pellet was resuspended in hypotonic buffer three times the original packed cellular volume. Following 10-15 min of swelling on ice, cells were transferred to a glass Dounce homogenizer and homogenized with 25-30 strokes of a type B pestle. Isolated, intact nuclei were confirmed by trypan blue staining. Nuclei were pelleted by centrifuging for 15 min at 3300 × g. Nuclei were then resuspended in one-half packed nuclear volume of low salt buffer (20 mM Hepes, pH 7.9, 25% glycerol, 1.5 mM magnesium chloride, 0.02 M potassium chloride, 0.2 mM EDTA, 0.2 mM PMSF, and 0.5 mM DTT). In a dropwise fashion, a high salt buffer of an equal volume was then added (20 mM Hepes, pH 7.9, 5% glycerol, 1.5 mM magnesium chloride, 1.2 M potassium chloride, 0.2 mM EDTA, 0.2 mM PMSF, and 0.5 mM DTT). Nuclei were allowed to extract for 30-45 min on a shaker at 0-4 °C. The extracted nuclei were pelleted by centrifuging for 30 min at 25,000 × g. The nuclear protein-containing supernatant was then dialyzed, the precipitated proteins were pelleted by centrifugation for 5 min in a microcentrifuge, and the remaining nuclear extract proteins (supernatant) were quantified (29).

Whole tissue extracts were isolated as follows. Isolated hearts were placed in ice-cold PBS, rinsed twice, and then minced in whole heart homogenization buffer (25 mM glycylglycine, 15 mM MgSO₄, 4 mM EGTA, pH 8.0, 1 mM DTT) (20, 21, 30).² The minced solution was homogenized with 10-15 strokes of a motor-driven Teflon homogenizer. The homogenate was removed and pelleted at 6000 × g for 10 min at +4 °C in a microcentrifuge. The supernatant was kept for LUC and CAT assays (see below). The nuclei-containing pellet was resuspended in hypotonic buffer. The subsequent steps were as described above for cellular nuclear extracts.

Twenty-four hours following isolation of neonatal rat cardiomyocytes, transient transfection assays were performed (31). Twenty µg (15 µg of experimental and 5 µg of pMSVGal) of DNA and 50 µl of Lipofectin were resuspended in 1.5 ml of Opti-MEM (Life Technologies, Inc.) medium and incubated for 15 min. pMSVGal was cotransfected with experimental constructs as an internal control for transfection efficiency. Following the 15-min incubation, PC1 medium (Hycor) was removed, the cardiocytes were washed twice with sterile PBS, and 1.5 ml of Opti-MEM medium were added. The DNA-containing Opti-MEM solution was then applied. Following 5 h of incubation, the Opti-MEM/DNA solution was removed, and 3 ml of PC1 medium were added. Forty-eight hours later, cells were harvested by standard methods, and protein extracts were assayed for -galactosidase and LUC or CAT activity (17). CAT and LUC activity were normalized for transfection efficiency using relative -galactosidase activity (17). The percentage of CAT conversion was between 0.5 and 35%, which was in the linear range. Luciferase activity varied between 290 and 97,112 relative light units (RLU) which was also in the linear range. For in vivo competition, 2 µg of BCK92CAT was cotransfected with 16 µg of pTZ19 or pTZ19/ExI and 2 µg of pMSVGal. For testing of the enhancer element, 5 µg of pMSVGal was co-transfected with 4 µg of Poly/pT109 or Seam/pT109.

Five days following injection of DNA-containing solutions, hearts were isolated and treated as for isolating nuclear extracts (see above). Once the supernatant was obtained, it was transferred to a fresh tube and quantified. One hundred microliters of the supernatant were heated to 60 °C for 7 min, cooled immediately on ice, and spun for 5 min in a microcentrifuge, and 30 µl were used in a standard CAT assay (17). Triton was added to 200 µl of the supernatant for a final Triton concentration of 0.27%. Luciferase activity was determined on 10 µl of this solution by standard assay (Promega). Luciferase activity of individual constructs was normalized for initial volume isolated and for transfection efficiency using pSV2CAT driven CAT expression. LUC activity was between 205 and 17,901 RLU, which was in the linear range.

RESULTS

Developmental end points in this investigation were defined as fetal (primary neonatal rat heart cultures) and adult (adult rat heart in vivo). These tissues had previously been well characterized with high level and low level endogenous BCK expression, respectively (15, 17, 32). To begin to identify the regulatory elements responsible for BCK gene expression early in cardiac development, a series of promoter deletion constructs were transiently transfected into freshly isolated neonatal rat cardiac myocytes (Fig. 1). Removal of 5-flanking DNA from between 1100 to 388 and 388 to 92 base pairs upstream of the transcription start site led to increased levels of gene expression, suggesting the presence of two separate negative elements. These same negative elements were previously identified as functional in C₂C₁₂ cells (17). That study also suggested that sequence between 92 and +80 contained the elements necessary for driving differentiation responsive expression. Subsequent analyses have determined that the critical region is contained within the first exon and that interactions with specific nuclear factors confer adequate and differentiation responsive levels of expression in C₂C₁₂ cells (18, 19). Accordingly, transfection analyses in neonatal cardiocytes also investigated the importance of the first exon. As shown in Fig. 1, removal of the first (untranslated) exon markedly reduced CAT activity, implicating the first exon as critical for BCK expression in cardiomyocytes. To isolate the specific region responsible, a series of promoter constructs with the same 5 end at 92 and differing 3 ends were prepared. Deletion of 3 sequence from +80 to +57 had no effect on expression. Removal of an additional 16 base pairs (to +41) had the same impact as removal of the entire first exon: CAT activity markedly decreased. This suggests that in neonatal rat cardiomyocytes, cis elements critical for BCK expression are located within the first exon. The element is likely between +41 and +57. However, disruption in the middle of an element by deletion at +41 cannot be excluded.

To identify the regulatory elements mediating BCK expression later in development, a series of promoter deletion constructs were injected into the apex of the adult rat left ventricle. These chimeric constructs contained BCK promoter fragments upstream of the LUC reporter gene. Luciferase was used as a reporter gene because transfection efficiency is very low by this method and because endogenous BCK is expressed at low levels in the adult heart. To ensure that changing the reporter gene would not alter expression, transient transfections using the LUC constructs in neonatal cardiocytes were performed. These analyses confirmed that expression was not reporter-dependent. In vivo analyses were then initiated and are summarized in Table I. The longest construct produced luciferase activity of 1783 RLU. This amount is significantly above background. Deletion of 5 base pairs to 92 led to a substantial increase in expression, suggesting that at least one of the negative elements functional in vitro in neonatal cardiocytes also functions in vivo in adult cardiac tissue. Investigation of the importance of the first exon was then performed. Deletion of 3 flanking DNA from +80 to +57 had no effect. Removal of 3 sequence to +41 significantly decreased expression to 1814 RLU. Thus, cis elements identified as important for transfected BCK expression in neonatal cardiocytes in vitro are also functional in the adult heart in vivo.

Table I. Promoter analysis of human BCK gene in vivo in adult rat heart Luciferase activity is quantified as RLU. Values shown are mean ± standard error of the mean. n represents the number of injection experiments performed. At least four plasmid preparations were used for each construct. Construct RLUa n BCK388LUC 1,783  ± 506* 10 BCK92LUC 13,189  ± 889*§¶ 12 BCK+57LUC 14,627  ± 1,146*§¶ 7 BCK+41LUC 1,814  ± 662* 8 BCK+25LUC 1,746  ± 402* 7 BCK+1LUC 1,635  ± 590* 12 pXP2LUC 215  ± 13 14 pSV2CAT 220  ± 21 4 a  All constructs are significantly greater than promoterless pXP2LUC (* = p < 0.05). BCK92LUC and BCK+57LUC are both significantly greater than the other constructs (¶ = p < 0.05). There is no difference between BCK92LUC and BCK+57LUC (§ = p > 0.05). There is also no difference among the less well expressed constructs ( = p > 0.05).

To begin to determine the specific region within the first exon that was responsible and to assess if DNA-nuclear protein interactions were involved, EMSAs using a series of oligonucleotides corresponding to different and overlapping regions of the first exon were performed. Deletion analysis suggested that sequence 3 of +41, particularly between +41 and +57, was important. However, disruption of a binding site by deletion at +41 could not be excluded. Accordingly, the first oligonucleotide prepared spanned sequence on either side of +41 (+25 to +57). This probe was termed ``seam'' oligo. EMSAs were performed using the seam probe with nuclear proteins isolated from neonatal cardiocytes and adult hearts. An easily detectable complex is seen using neonatal cardiocyte extracts (Fig. 2a). A very faint band of similar mobility is seen with extracts from the adult rat heart. This band was easily and specifically competed by unlabeled seam (+25 to +57). The amount of shifted band was significantly increased in neonatal cardiocyte nuclear extracts relative to adult nuclear extracts. The relative abundance of the interacting factor correlates with the level of endogenous BCK gene expression (high in neonates, low in adults), suggesting that this factor may be involved in the developmental regulation of BCK in the heart.

The +25 to +57 element does not contain any previously described nuclear factor binding sites. However, the portion of the element between +41 and +57 does contain a sequence which could be construed as an inverted Egr-1 site (33). Consequently, a commercially available oligonucleotide containing the Egr-1 consensus binding site was used as a competitor. As seen in Fig. 2a, the Egr-1 element is not capable of competing the interacting factor in either neonatal cardiocytes or adult hearts. An inverted Sp-1 site could also be derived from this cytosine rich region (34). Therefore, an EMSA competition assay using an oligonucleotide containing the Sp-1 binding site was performed. As seen in Fig. 2b, Sp-1 does not compete the shifted complex in adult rat heart extracts. The Sp-1 binding site oligonucleotide also does not compete the shifted band when using neonatal cardiocyte nuclear extracts. Thus, the identified complex does not represent Egr-1 or Sp-1.

To begin to localize the binding site within the seam probe, competition EMSAs using oligonucleotides corresponding to the 5 (+25 to +41) and 3 (+41 to +57) halves of the seam probe were performed (Fig. 3). While the 5 half competed somewhat, the 3 half competed nearly as effectively as unlabeled seam, suggesting that the major site of binding is between +41 and +57. Competition with an oligonucleotide corresponding to base pairs +1 to +27 (oligo 3) had no affect, showing that the interacting factor was specific for the seam region of the first exon. Interestingly, competition with an oligonucleotide-containing sequence corresponding to the 3 half and also extending to +68 (oligo 4, +41 to +68) appeared to be as effective as unlabeled seam. Although this may be due to an additional binding site provided by the added base pairs, because the +57 to +68 region does not have any functional significance (deletion from +80 to +57 had no affect on transfected expression) it more likely is because these appended bases serve to nonspecifically stabilize the binding of the factor. Despite the apparent localization of binding to +41 to +57, EMSAs using this region as a probe failed to demonstrate a shifted complex. Hence, all subsequent EMSAs used the full-length seam probe.

Prior work had shown that the seam sequence (+25 to +57) was also capable of producing shifted complexes using nuclear extracts from C₂C₁₂ myoblasts and myotubes (18). Accordingly, an EMSA was performed using nuclear extracts from C₂C₁₂ myoblasts and myotubes and neonatal cardiocytes (Fig. 4). This EMSA shows that the complex identified using neonatal cardiocyte nuclear extracts migrates to the same position in the gel as that identified using C₂C₁₂ extracts. This suggests that a similar factor is present in both cardiac and skeletal muscle and is capable of interacting with this region.

To assess whether regions of the first exon other than +25 to +57 were capable of interacting with additional nuclear factors, additional EMSAs using oligonucleotides corresponding to separate elements were performed. Interestingly, an EMSA using an oligonucleotide corresponding to +1 to +27 (oligo 3) shifted a single band in C₂C₁₂ myotubes> myoblasts but not in neonatal or adult cardiocytes (Fig. 5). The identified band likely represents a different trans factor due to sequence discrepancy and the inability of oligo 3 to compete the band shifted with the seam probe (Fig. 3). Thus, this factor is present in skeletal but not cardiac nuclear extracts and may represent a skeletal muscle specific factor.

Transient transfection and EMSA analyses supported the presence of a trans factor interacting with a sequence from +25 to +57, with highest affinity for a sequence between +41 and +57, that positively influenced expression. To determine if single-base pair mutations could affect the ability of the element to influence expression, a series of mutant constructs were prepared and transfected into neonatal cardiocytes. All constructs extended from 92 to +57 and contained single-base pair mutations at either +32, +36, +40, or at both +47 and +53 (termed BCK+57+32CAT, BCK+ 57+36CAT, BCK+57+40CAT, and BCK+57+47,+53CAT, respectively). As shown in Fig. 6, mutation of bases +47 and +53 resulted in a marked reduction in expression. This affect was not seen with mutation of the other base pairs. These data imply that bases +47 and +53 are critical for the +25 to +57 element to induce expression.

To determine if the effect was due to modification of the binding site, EMSA was performed using an oligonucleotide containing the nonfunctional mutation as a competitor. Fig. 3a shows that while the 3 half effectively competes the shifted band, the 3 half containing mutations at +47 and +53 (F) competes poorly. This suggests that these bases are important for proper binding of the interacting factor. These data imply that mutating +47 and +53 alters binding of the interacting transcription factor, preventing factor mediated inducibility of expression.

Analyses demonstrated that the +25 to +57 region bound a nuclear factor and that removal of the element or mutation of specific bases within the binding site of the interacting factor prevented the element from positively influencing transfected BCK promoter-driven reporter gene expression. To determine if the putative positive regulatory element could induce expression of a heterologous promoter, the element (+25 to +57) was placed upstream of the hamster sarcoma virus thymidine kinase promoter of pT109LUC, a LUC reporter plasmid designed for testing enhancer elements. As seen in Table II, this construct (Seam/pT109), which contains two copies of the element in the sense orientation, markedly induces expression. This affect was not seen with a nonspecific 32-base pair fragment (Poly/pT109). These data confirm the functional independence of the +25 to +57 element and imply that it acts as an enhancer.

Table II. Analyses of +25 to +57 (seam) element in heterologous promoter Luciferase activity is quantified as RLU. Values shown are mean ± standard error of the mean. n is the number of transfection experiments, all performed in duplicate. Two separate plasmids of each experimental construct were used. Construct RLUa n pT109LUC 1315  ± 283 4 Poly/pT109 1533  ± 322 4 Seam/pT109 64,541  ± 2264* 4 a  Two copies of the element in the sense orientation upstream of the HSV TK promoter in pT109luc (Seam/pT109) significantly induces expression (* = p < 0.05 relative to Poly/pT109 and pT109). There is no significant difference between pT109 and Poly/pT109.

DISCUSSION

These experiments sought to identify the regulatory elements mediating brain creatine kinase gene expression in cardiac cells during development. In vitro and in vivo promoter deletion analyses, EMSAs, and site-directed mutagenesis show that a nuclear factor binds and mediates the expression of the brain creatine kinase gene in the heart through an enhancer element at +25 to +57 via sequence between +41 and +57. The binding site of this factor involves nucleotides +47 and/or +53. This factor may also determine developmental stage specific expression in the heart as the abundance of the factor correlates with the level of endogenous and transiently transfected BCK expression. These data also show that a factor present in C₂C₁₂ nuclear extracts but not cardiocytes is capable of binding to the first exon between +1 and +27. The functional importance of this factor is unknown but may determine skeletal muscle specificity.

The enhancer element is interesting in that the critical area identified, particularly the pivotal base pairs, is not a previously described transcription factor recognition sequence. Though the element does contain sequence that could be considered either an inverted Sp-1 or Egr-1 site, EMSAs clearly show that the factor interacting with this region does not represent either of these factors. It is also intriguing that the element binds a similar size factor in skeletal and cardiac muscle cells. However, it is unlikely that this factor functions similarly in both muscle cell types. This is based on transient transfection assays in C₂C₁₂ cells which showed that removal of +25 to +57 sequence had no effect on expression and EMSAs which show subtle but substantive differences between cardiac and skeletal muscle derived extracts (18, 19) (Figs. 3 and 4). Taken together, these data suggest that this region is capable of binding factors in a tissue and developmentally specific fashion and implies that this region may be involved in conferring tissue and developmental stage specificity. This is reminiscent of the DNA-protein interactions governing the rat cardiac myosin light chain-2 (MLC-2) gene in which MLC-2 cardiac muscle-specific expression is regulated by the combination of a ubiquitous factor (HF-1a) and a muscle-specific factor (HF-1b) via their interactions with an HF-1a factor-specific sequence and a MEF-2-like site, respectively (11, 35, 36). Other cardiac genes also manifest tissue specific expression through novel DNA-protein interactions. For example, distinct transcriptional regulatory pathways in cardiac and skeletal muscle utilizing novel sequence and factors mediate slow/cardiac troponin C gene expression (10). The BCK gene may have a similar regulatory schema. For example, the presence of a skeletal muscle specific factor may be important in determining tissue-specific expression (see Fig. 5).

Except for the 3-nontranslated region translational control of BCK expression in U937 cells, endogenous BCK expression is primarily transcriptionally mediated, and in C₂C₁₂, HeLa, and neuroblastoma cells, is dependent upon the sequence within the 5-flanking region and the untranslated first exon (17, 18, 19, 37, 38, 39). The information presented within this report is consistent with these previous reports and indicate that BCK expression in the developing heart is also under transcriptional control. The region mediating BCK expression in the heart, not previously been described, functions as an enhancer and interacts with a factor whose abundance correlates with the level of endogenous BCK expression. These data implicate the interacting factor as responsible for mediating BCK expression in the heart. The identity of this factor has not yet been determined but likely represents either a described factor interacting with a new sequence or a novel nuclear factor.