INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE FUNCTIONS OF SERINE/THREONINE KINASES in cardiac gene expression and hypertrophy have been studied extensively, in particular those of protein kinase C, protein kinase A, and members of the mitogen-activated protein (MAP) kinase pathways (37). However, the functions of serine/threonine phosphatases in these processes are less well understood. Recently, calcineurin (protein phosphatase 2B) has been shown to play a role in the induction of cardiac gene expression and hypertrophy through a mechanism that involves dephosphorylation of the transcription factor NF-AT3 (nuclear factor of activated T cells-3) (24, 40). Other serine/threonine phosphatases are also involved in cardiac function. This became apparent when it was discovered that okadaic acid, an inhibitor of protein phosphatase 1, protein phosphatase 2A (PP2A), and related serine/threonine phosphatases, induces a shortening in contraction time of isolated cardiac muscle (13). The effects of okadaic acid are similar to those induced by -adrenergic receptor agonists. It has been suggested that they are mediated by increased phosphorylation of the cardiac-specific proteins phospholamban and troponin I (27, 39). Because okadaic acid inhibits several serine/threonine phosphatases (6, 7), it is difficult to correlate the biological effects of okadaic acid with the inhibition of a specific phosphatase. However, there is biochemical evidence that protein phosphatase 1 is mainly responsible for the dephosphorylation of phospholamban (18), whereas dephosphorylation of troponin I might be carried out by PP2A (25). PP2A is also involved in regulating ATP-sensitive potassium channels, as demonstrated by single channel recordings on inside-out membrane patches from ventricular myocytes (16). Furthermore, PP2A may modulate MAP kinase pathways, which are activated in response to treatment with phorbol ester, endothelin-1, or phenylephrine (3, 30, 43). As shown for cardiomyocytes (4) and other systems (38, 45), PP2A can either enhance or reduce signaling through the MAP kinase pathway.

PP2A exists in cells as two forms: the heterodimeric core enzyme, composed of the catalytic subunit C and the regulatory subunit A, and the heterotrimeric holoenzyme, composed of the core enzyme and a second regulatory subunit, subunit B. The A subunit polypeptide consists of 15 nonidentical repeats. The B subunit binds to repeats 1-10, and the C subunit binds to repeats 11-15 of the A subunit (Fig. 1, see Ref. 26 for review). The core enzymes and holoenzymes are expressed in approximately equal quantities (15), and they differ in substrate specificity (26). B subunits fall into three families, designated B (10, 20, 29, 46), B' (or B56) (8, 22, 23, 41, 42), and B'' (5), which are unrelated by primary sequence. The B' family consists of numerous isoforms and splice variants that determine the specificity and intracellular localization of the holoenzyme (22, 42). In addition, there is tissue-specific expression of B' subunits (8, 22, 23, 42). Importantly, B56- and B56-subunits are preferentially expressed in heart and skeletal muscle, suggesting that they direct the holoenzyme to a specific subcellular location in these tissues and/or toward a specific substrate.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Model of protein phosphatase 2A (PP2A) and location of epitopes on the A subunit. The B subunits and T antigens bind to the NH2-terminus, and the C subunit binds to the COOH-terminus of the A subunit (34, 35). The regions where the epitopes of the 6F9 and 5H4 antibodies are located are indicated by horizontal bars. Figure modified from Ref. 15.

We have previously demonstrated that expression of the A5 subunit, a NH₂-terminal mutant of the A subunit of PP2A with a deletion of repeat 5 [which binds the catalytic C subunit but none of the regulatory B subunits (Fig. 1)], causes an increase in core enzymes and a decrease in holoenzymes in tissue culture cells (32). As a result of this change in the core enzyme-to-holoenzyme ratio, the Tat protein-stimulated transcription from the human immunodeficiency virus-1 (HIV-1) long-terminal repeat and HIV-1 virus production are strongly inhibited. On the basis of this result, we asked whether expression of the mutant A5 subunit in animals would alter PP2A activity and provide information about the function of PP2A. Here we show that mice expressing the mutant A5 subunit under the combined control of the chicken -actin promoter and the cytomegalovirus (CMV) enhancer, yielding high gene expression in muscle tissue, have an increased heart weight-to-body weight ratio early in life and develop a form of dilated cardiomyopathy.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DNA construction. PCR was used to engineer a sequence encoding the nine amino acid epitope tag EEEEYMPME (EE) at the 3' end of the human A subunit-cDNA open reading frame (33). The construct was subcloned into the ecdysone-responsive expression plasmid pIND (Invitrogen) using EcoR I and Xho I. This vector was termed pINDA5^EE. To create an EcoR I site at the end of the A5^EE sequence, the vector pINDA5^EE was digested with Xho I, treated with Klenow, and blunt-end ligated to an EcoR I linker. The EcoR I-EcoR I insert of A5^EE was ligated into the EcoR I site of the pCAGGS vector (28) resulting in pCAGGSA5^EE (Fig. 2). The pCAGGS vector containing the CMV enhancer, the chicken -actin promoter, and the rabbit -globin polyadenylation signal was used because this vector yields high gene expression in muscle tissue (19).

View larger version (9K): [in this window] [in a new window]   Fig. 2.   Construct containing mutant A5EE subunit (the A subunit with repeat 5 deleted) used to generate transgenic mice. This construct was generated by inserting the glutamic acid-glutamic acid epitope (EE)-tagged A5 (A5EE) mutant into the vector pCAGGS, which contains the cytomegalovirus (CMV) enhancer, the chicken -actin promoter, and the rabbit -globin polyadenylation signal.

Generating transgenic mice. The vector pCAGGSA5^EE was cut with Ssp I and BamH I. The 3.7-kb fragment (shown in Fig. 2) was purified from the gel using a Qiagen kit and then concentrated with an Elutip (Schleicher & Schuell). The DNA was ethanol precipitated and dissolved in 5 mM Tris (pH 7.5) and 0.1 mM EDTA at a concentration of 2.0 µg/ml.

Analyzing mouse organs for expression of A5^EE protein. The mice were euthanized by CO₂ inhalation for 30 s. The different tissues were removed, frozen on dry ice, and weighed. Protein was extracted by grinding the defrosting tissue with a pestle and then grinding with 10 µl of SDS sample buffer [2% SDS, 5% -mercaptoethanol, 5% glycerol, 10 mM Tris (pH 6.8), 10 mM dithiothreitol, and 0.025% bromophenol blue] per milligram of tissue. An 18-gauge needle was used to further dissociate the tissue, and the lysate was then boiled for 5 min, diluted, and boiled again. For protein determination, aliquots of solubilized tissues were precipitated with trichloroacetic acid. Protein (10-20 µg) was analyzed on a 10% acrylamide gel and transferred to a polyvinylidene difluoride membrane (Immobilon-P, Millipore). Western blotting was done using the 6G3 antibody (15).

Heart weight-to-body weight ratios and histology. Mice were euthanized using CO₂ inhalation as described in Analyzing mouse organs for expression of A5^EE protein, and the mice were weighed. Their hearts were removed, blotted on filter paper, and weighed. For histology, hearts were fixed in 10% buffered Formalin for at least 48 h, cut longitudinally or transversely, dehydrated, and infiltrated with paraffin. Sections (5 µm) were cut after embedding the hearts in paraffin. Sections were floated onto slides and dried overnight. The sections were stained with hematoxylin and eosin or with Masson trichrome stain.

Immunoprecipitation of PP2A. Mice were euthanized by 30 s of CO₂ inhalation, and their hearts were removed and weighed. Sections of 100-150 mg of heart were frozen in liquid nitrogen and ground with a mortar and pestle and lysis buffer [50 mM Tris · HCl (pH 8.0), 150 mM NaCl, 3 mM MgCl₂, 1 mM Pefabloc SC (Boehringer), 50 µM leupeptin (Sigma), 0.2 mg/ml soybean trypsin inhibitor (Calbiochem), 1 µg/ml aprotinin, and 1 mM dithiothreitol]. The lysate was spun at 16,000× g at 4^oC, and the supernatant protein concentration was determined with Pierce's Coomassie Plus Assay. Over 95% of PP2A was solubilized under these conditions, as determined by Western blotting. Addition of 0.5% Triton X-100 to the lysis buffer did not increase the amount of PP2A in the supernatant. For immunoprecipitation, 300 µg of lysate was incubated for 2.0 h at 4^oC with 360 µg of anti-EE, 300 µg of 6F9, or 200 µg of 5H4 monoclonal antibodies coupled to Gamma Bind Plus sepharose beads (Pharmacia) (15). After centrifugation, we incubated the supernatant with the next antibody. The immunoprecipitates were washed three times with 1 ml of wash buffer (above lysis buffer with 0.5% Triton X-100 without Pefabloc, aprotinin, and soybean trypsin inhibitor) and boiled in SDS sample buffer. One-half of the immunoprecipitate was loaded on a 10% high-bis gel (19 × 16 cm). The protein was transferred to the polyvinylidene difluoride membrane, and antibody detection was carried out with rat monoclonal anti-A subunit (6G3), rabbit anti-B-subunit, and mouse monoclonal anti-C subunit antibodies, as described (32).

Culturing and immunostaining of cardiomyocytes and heart fibroblasts. For the preparation of fibroblasts, hearts from 4-mo-old animals were rinsed in cold PBS, the atria and aorta were removed, and the ventricles were cut into small pieces and placed into 37^oC buffer containing 18.3 mM HEPES, 0.8 mM MgSO₄, 1.0 mM NaH₂PO₄, 116.4 mM NaCl, 5.5 mM D-glucose, and 5.4 mM KCl. The hearts were digested six times for 20 min at 37^oC with 100 U/ml collagenase type 2 (Worthington Biochemicals) and 0.6 mg/ml pancreatin (GIBCO) in the same buffer (5.6 ml/heart). The cells dissociated from the last five digestions were pooled and plated in DMEM, 10% FCS, and 1% penicillin-streptomycin solution. Cells were transferred once and plated on tissue culture plates or coverslips. Cells were fixed for 7 min with 4% p-formaldehyde and permeabilized with 1% NP-40/PBS for 1 min. The cells were blocked for 30 min in 1% BSA/PBS, incubated for 1 h with anti-glutamic acid-glutamic acid (EE) tag antibody (50 µg/ml of 1% BSA/PBS), washed with PBS, incubated for 1 h with anti-mouse horseradish peroxidase (1:250 Jackson Immuno Research), washed with PBS, and incubated with diaminobenzidine substrate (Vector). Cardiomyocytes were prepared from five 1-day-old mouse hearts, as described above except that 0.3 ml of enzyme solution (0.8 mg/ml pancreatin and 0.55 mg/ml collagenase type 2) was used per heart. The myoblasts were plated into media containing 13.9 g/l DMEM, 3.02 g/l medium 199, 3.4 g/l sodium bicarbonate, 10% horse serum, and 5% fetal calf serum (11, 14). The reason why cardiomyocytes were prepared from 1- to 5-day-old mice and not from adults was because adult cardiomyocytes are difficult to isolate and to maintain in culture for immunostaining.

Phosphatase assays. Phosphorylase A and Rb peptide were used as substrates and labeled as described (32). Heart extracts were made as described for Immunoprecipitation of PP2A, and the protein contents of the heart extracts were immediately determined and adjusted to be equal. Extracts containing equal amounts of protein were diluted 1:40 in 0.1% BSA, 50 mM Tris · HCl (pH 7.5), 0.1 mM EGTA, and 0.1% 2-mercaptoethanol. Extracts (5 µl) containing 0.6 µg of total protein were assayed in the presence of protein phospatase 1 inhibitor 2 as described (32). The reaction time was 10 min. The assay was linear for at least 10 min and directly dependent on the amount of extract used. The amounts of catalytic C subunit in the wild-type and I1 heart extracts used for the assay were identical as determined by Western blotting.

Transthoracic echocardiography. Echocardiography was performed in intact anesthetized mice (2.5% Avertin, 14 µl/g ip) using an Apogee CX echocardiograph (Interspec-ATL, Bothell, WA) as previously described (31). Wild-type and homozygous I1-transgenic mice were studied at age 4-7 mo. The average heart weight-to-body weight ratios of the I1-transgenic and wild-type mice were 6.7 and 5.0, respectively (data not shown). The operator who performed and measured the echocardiograms was blinded to the genotype of the animals. The following parameters were measured: left ventricular end-diastolic dimension (LVEDD), left-ventricular systolic dimension (LVESD), percent fractional shortening [calculated as (LVEDD  LVESD)/LVEDD × 100], septal wall thickness, posterior wall thickness, and heart rate.

Arterial blood pressure. Mice were anesthetized with ketamine (100 mg/kg) and xylazine (5 mg/kg ip). A polyethylene-50 flame-stretched, fluid-filled catheter was introduced through a cervical incision into the carotid artery and attached to a modified P50 Statham transducer. Systolic and diastolic blood pressure were recorded.

Northern blot analysis. Two- to three-month-old wild-type and homozygous I1-transgenic mice were euthanized for 30 s with CO₂ inhalation. The hearts were removed, frozen quickly in liquid nitrogen, and kept on dry ice until homogenized. To avoid RNA degradation, the heart weight was not determined. Total RNA was isolated using the RNeasy Midi Kit (Qiagen) following the manufacturor's instructions. The RNA concentration was determined by measuring optical density at 260 nm. RNA (10 µg) was loaded onto a 1% agarose gel under denaturing conditions, as described by Qiagen. The RNA was transferred onto a Magna Graph nylon membrane (MSI) overnight with 20× saline-sodium citrate buffer and hybridized with probes. Myosin heavy chain- (-MHC) and myosin heavy chain- (-MHC) oligonucleotides were end labeled with deoxynucleotide transferase. Hybridization was carried out overnight at 65^oC. The mouse -MHC oligonucleotide had the sequence AACGTTTATGTTTATTGTGGATTGGCCACAGCGAGGGTCT derived from the 3' untranslated region that does not show homology to the -form. The oligonucleotide from rat -MHC was purchased from Calbiochem (cat. no. ON366, sequence unknown). Atrial natriuretic factor, calsequestrin, sarco(endo)plasmic reticulum Ca²⁺-ATPase, phospholamban, and the ryanodine receptor RNAs were labeled with Amersham's multiprime kit. Hybridization was carried out overnight at 42^oC, as described by He et al. (9).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Expression of A5^EE subunit in muscle tissue of transgenic mice. As previously described (11), transgene expression controlled by the CMV enhancer and the chicken -actin promoter combined, as in the vector pCAGGS (28), is very high in muscle tissue and low or undetectable in other tissues. We chose this promoter because the endogenous level of the A subunit of PP2A is very high (~0.1% of total protein) in all cells and tissues that have been studied (15, 17, 36). Therefore, a dominant negative mutant of the A subunit, such as the A5 subunit (35), can only compete with the endogenous A subunit if the mutant subunit is expressed as high or higher than the endogenous protein. We found two transgenic founder mice that expressed a high amount of the A5^EE protein in the heart, skeletal muscle, and smooth muscle (stomach). These mice were bred to generate the lines I1 and I4. As shown in Fig. 3, the approximate ratios of the A5^EE subunit to the endogenous A subunit for the I1 line are 4:1 for heart, 8:1 for skeletal muscle, and 1:1 for stomach. The approximate ratios for the I4 line are 1:1 for heart, 1:2 for skeletal muscle, and 1:8 for stomach. Note that the amounts of the A subunit and the A5^EE subunit cannot be compared between lanes because unequal amounts of protein were loaded. In addition, to optimize visualization and quantitation of the ratio of the wild-type A subunit to the A5^EE subunit, different exposure times were chosen for different lanes. Importantly, expression of the A5^EE subunit had no significant effect on the level of the endogenous A subunit in the heart (data not shown). Only trace amounts of the A5^EE subunit were found in the kidney, liver, and brain (data not shown). During the first year after these two mouse lines were established, the expression of the A5^EE subunit dropped. In hearts from I1-transgenic mice, which were used for the majority of our experiments, the ratio of the A5^EE subunit to the endogenous A subunit reached a level of 2:1. The A5^EE and endogenous A subunits were evenly distributed between the ventricles and atria of I1- and I4-transgenic mouse hearts (data not shown). Importantly, the A5^EE subunit was only expressed in cardiomyocytes but not in heart fibroblasts, as shown in Fig. 4.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Expression of mutant A5EE protein in transgenic mice. Western blot of wild-type A subunit and mutant A5EE subunit in the heart (He), skeletal muscle (Sk), and stomach (St) from two independent lines of trangenic mice, I1 and I4. Monoclonal antibody 6G3 was used as described in MATERIALS AND METHODS.

View larger version (67K): [in this window] [in a new window]   Fig. 4.   Expression of A5EE subunit in cardiomyocytes but not in heart fibroblasts. Cardiomyocytes were prepared from 1-day-old mice, and heart fibroblasts were prepared from 4-mo-old mice. The cells were fixed with paraformaldehyde and stained with antibodies against the EE tag, as described in MATERIALS AND METHODS. Only cardiomyocytes from I1-transgenic mice (top right) were stained with the antibodies. I1, I1-trangenic mice; WT, wild-type mice.

Transgenic expression of A5^EE subunit causes increase in heart weight-to-body weight ratio. We discovered that the hearts of transgenic animals were significantly larger than normal hearts and that there was no sign of interstitial fibrosis (Fig. 5). Furthermore, the heart weights and heart weight-to-body weight ratios of the I1-transgenic mice were already increased on day 1 after birth and continued to be high throughout life. These data are presented in Figs. 6 and 7, showing the values of individual animals in the different age groups (plus signs, closed triangles, and open squares for I1-transgenic, I4-transgenic, and wild-type mice, respectively), and in Table 1, presenting the average values. Similar results to the I1-transgenic mice were obtained for the I4-transgenic mice, although the increase in heart weight and heart weight-to-body weight ratio was generally smaller than for I1-transgenic mice. This latter finding is consistent with the observation that I4-transgenic mice expressed less A5^EE subunit than I1-transgenic mice. Heterozygous I1- and I4-transgenic mice had lower heart weight-to-body weight ratios than the corresponding homozygous mice (as shown in Fig. 8), further indicating that the heart weight and heart weight-to-body weight ratio correlate with the level of A5^EE subunit expression. These results strongly suggest that the A5^EE subunit is the cause for the increased heart weight of the A5^EE-transgenic mice. There was no apparent sex difference in heart weight and heart weight-to-body weight ratio. Interestingly, starting at 3-4 mo of age, I1-transgenic mice with extremely large hearts and heart weight-to-body weight ratios (between two- and fivefold above the norm) were observed. These mice represented ~25% of all animals in the 7- to 12-mo age group. They were characterized by general weakness, slow movements, and increased respiratory rate, and they died within a few weeks after showing these symptoms. A summary of all data is presented in Fig. 7 and Table 1. Note that the P values for the heart weights and the heart weight-to-body weight ratios of I1-transgenic mice in the 3- to 4-, 5- to 6-, and 7- to 12-mo-old age groups are skewed by the mice with extremely high ratios. For example, the increased ratio of 3- to 4-mo-old mice appears statistically nonsignificant due to the presence of two animals with extreme ratios, although the increase is highly significant, as illustrated in Fig. 6.

View larger version (74K): [in this window] [in a new window]   Fig. 5.   Increase in heart size of A5EE-transgenic mice and lack of interstitial fibrosis. Hearts from a normal (A) and a A5EE-transgenic mouse (B) were sectioned transversely and stained with Masson trichrome stain. The scale is 2 mm.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Increase in heart weight of A5EE-transgenic mice on postnatal day 1 (A) and throughout life. Body weights and heart weights of WT (), A5EE I1-trangenic (+), and A5EE I4-transgenic () mice were determined at different ages [10-day-old mice (B), 1- to 2-mo-old mice (C), 3- to 4-mo-old mice (D), 5- to 6-mo old mice (E), and 7- to 12-mo-old mice (F)]. G and H: comparison of body weight to heart weight in male (G) and female (H) WT, I1-, and I4-transgenic mice. I: comparison of body weight to heart weight in all of the mice. This figure illustrates the general increase in heart weight of transgenic compared with WT animals and the appearance of animals with extremely large hearts beginning at 3-4 mo of age. It also illustrates the general tendency of transgenic hearts toward a slightly decreased body weight compared with WT hearts.

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Increase in heart weight-to-body weight ratio of A5EE-transgenic mice. The heart weight-to-body weight ratio was plotted for the different age groups. This figure illustrates the general increase in the ratio for transgenic animals and the extremely high ratio for a fraction of older animals. Also note that the ratio is constant for WT mice at all ages. , WT; +, I1-transgenic mice; , I4-transgenic mice.

View this table: [in this window] [in a new window]   Table 1.   Increase in heart weight and heart weight-to-body weight ratio in A5-transgenic mice

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Homozygous A5EE-transgenic mice have a larger heart weight-to-body weight ratio than heterozygous mice. The P values (means ± SD) for homozygous (Homo) I1- and I4- and heterozygous (Hetero) I1- and I4-transgenic mice compared with WT mice are 0.0005, 0.0054, 0.0011, and 0.24. The P value for I1 Homo- compared with I4 Homo-transgenic mice is 0.0178; for I1 Hetero compared with I4 Hetero, 0.0077; for I1 Homo compared with I1 Hetero, 0.0062; and for I4 Homo compared with I4 Hetero, 0.0235.

Heart function of A5^EE-transgenic mice is impaired, and blood pressure is unchanged. Echocardiography was used to evaluate cardiac chamber size and function in vivo in wild-type and homozygous A5^EE-transgenic mice. As shown in Table 2, LVEDDs for I1-transgenic mice were significantly greater compared with wild-type mice (4.5 vs. 3.5 mm, respectively), indicating chamber enlargement of the I1-transgenic mouse hearts. Furthermore, a significant increase in LVESD was found in the I1-transgenic mouse hearts compared with wild-type mouse hearts (3.5 vs. 2.2 mm, respectively). Calculated fractional shortening (end-diastolic dimension  end-systolic dimension × 100%/end-diastolic dimension), a measure of systolic function, indicates depressed cardiac performance in vivo in the I1-transgenic mice (21%) compared with wild-type mice (36.5%). Another important result from the echocardiographic analysis was that the thickness of the septum and of the left ventricular posterior wall was significantly reduced in transgenic mice. Taken together, these data of chamber enlargement, reduced systolic function, and wall thinning are consistent with a phenotype of dilated cardiomyopathy.

View this table: [in this window] [in a new window]   Table 2.   Left ventricle function is impaired in A5-transgenic mice

A5^EE subunit binds C subunit but not B subunit: similar abundance of core enzyme and holoenzyme in hearts. The underlying assumptions on which the present work is based are that 1) the A5^EE mutant subunit competes with the endogenous A subunit for binding of C subunit, 2) the A5^EE subunit does not bind B subunits, 3) the ratio of core enzyme to holoenzyme increases in A5^EE-expressing tissue, and 4) the phosphatase activity changes due to the increase in the core enzyme-to-holoenzyme ratio. To test these assumptions, we carried out immunoprecipitations with normal and transgenic heart extracts, using monoclonal antibodies against the EE tag (anti-EE) and monoclonal antibodies 5H4 and 6F9 directed against the wild-type A subunit. These latter antibodies distinguish between the core enzyme and holoenzyme. Whereas 5H4 recognizes the core enzyme but not the holoenzyme, 6F9 precipitates both forms (Fig. 1) (15). To achieve quantitative precipitations, two consecutive precipitations were carried out with each antibody. The amounts of the A, B, and C subunits in immunoprecipitates were determined by Western blotting with antibodies against the individual subunits as described in MATERIALS AND METHODS.

View larger version (62K): [in this window] [in a new window]   Fig. 9.   Abundance of holozyme and core enzyme in heart tissue of WT and A5EE-transgenic mice. Lanes 1-6, extracts from I1-transgenic mouse hearts were incubated sequentially with antibodies against the EE tag precipitating A5EE-C core enzyme (lanes 1 and 2), 5H4 precipitating WT core enzyme (AC) (lanes 3 and 4), and 6F9 precipitating holoenzyme (ABC) (lanes 5 and 6). Extract from WT hearts (lanes 7-10) and extract from I1-transgenic mouse hearts (lanes 11-14) were incubated sequentially with 5H4 (lanes 7, 8, 11, and 12) and 6F9 (lanes 9, 10, 13, and 14). To deplete the extract of the different forms of PP2A, each antibody was used twice. The precipitates were analyzed by Western blotting as described in MATERIALS AND METHODS. The positions of the A, B, and C subunits are indicated on the right.

Reduction of holoenzyme activity in transgenic hearts. To determine whether a decrease in holoenzyme activity and an increase in core enzyme activity in fact occurs as a result of A5^EE subunit expression, we used phosphorylase a, phosphorylated by phosphorylase kinase, and Rb peptide, phosphorylated by cdk1 kinase/cyclin B, as substrates. The holoenzyme is 100-fold more active than the core enzyme in dephosphorylating Rb peptide (1), whereas the core enzyme activity toward phosphorylase a is ~2.5-fold higher than that of the holoenzyme (44). As shown in Table 3, the phosphatase activity toward the Rb peptide was reduced by 29% in extracts from I1-transgenic mouse hearts compared with extracts from control hearts, whereas the activity toward phosphorylase a increased by 24%. This change in activity is reflected in a drop in the ratio of the Rb peptide phosphatase activity to the phosphorylase a phosphatase activity from 3.7 to 2.1. Because the activity toward the Rb peptide comes almost exclusively from the holoenzyme, these results indicate a 29% drop in the level of the holoenzyme; and, because holoenzyme represents approximately two-thirds (66%) of the total PP2A in the wild-type heart extract (Fig. 9), a 29% drop would result in a final amount of 42% holoenzyme in I1-transgenic mouse heart extract. Furthermore, because the decrease in holoenzyme leads to a corresponding increase in the core enzyme, the level of core enzyme is expected to rise from ~33% in wild-type heart extract (Fig. 9) to 58% in I1-transgenic mouse heart extract. Because the core enzyme is 2.5 times more active than the holoenzyme, one would expect a 25% increase toward phosphorylase a, which is very close to the experimental value of 24%. Thus the results obtained from phosphatase assays confirm those obtained by immunoprecipitation.

View this table: [in this window] [in a new window]   Table 3.   Altered phosphatase activity in heart extracts from A5-transgenic mice

Increased -MHC gene expression in A5-transgenic hearts. A characteristic of cardiac hypertrophy is the reactivation of genes whose expression is normally restricted to fetal development, e.g., the -MHC and the atrial natriuretic factor gene. On the other hand, the transcription of genes encoding sarco(endo)plasmic reticulum Ca²⁺-ATPase, phospholamban, and -MHC has been reported to be suppressed in hypertrophic hearts (see Refs. 2 and 14 for review). As demonstrated by Northern blotting, the level of -MHC transcripts in homozygous A5-transgenic hearts was significantly increased, whereas no change in atrial natriuretic factor, sarco(endo)plasmic reticulum Ca²⁺-ATPase, phospholamban, and -MHC expression was observed (Fig. 10A). The increase in -MHC transcripts in A5-transgenic hearts was fourfold compared with control animals (Fig. 10B). Elevated levels of -MHC were also observed in transgenic mice that developed cardiac hypertrophy and ventricular dilation due to overexpression of the calcium-dependent phosphatase calcineurin in the heart (24). We also found no change in calsequestrin and ryanodine receptor gene expression.

View larger version (25K): [in this window] [in a new window]   Fig. 10.   Cardiac gene expression in I1-transgenic mice. A: Northern blot showing cardiac expression in both I1-transgenic and WT mouse hearts. RNA was isolated and hybridized with specific probes as described under MATERIALS AND METHODS. B: quantitative comparison of transcripts from I1-transgenic and WT mice. WT gene expression was set to 100%. Numbers (means ± SD) in arbitrary units were obtained from 3 animals. MHC, myosin heavy chain; CSQ, calsequestrin; SERCA, sarco(endo)plasmic reticulum CA2+-ATPase; PLB, phospholamban; RYR2, ryanodine receptor 2; ANF, atrial natriuretic factor. The measured values (in arbitrary units) for WT mice are the following: -MHC, 4.6 ± 1.3; ANF, 260 ± 79; CSQ, 19.4 ± 1.6; -MHC, 310 ± 49.6; SERCA, 141.9 ± 8.5; PLB, 793.8 ± 50.9; RYR2, 1.6 ± 0.4. The measured values for I1-transgenic mice are the following: -MHC, 18.2 ± 6.3; ANF, 350 ± 170; CSQ, 21.2 ± 3.8; -MHC, 324.5 ± 95.5; SERCA, 120.4 ± 20.6; PLB, 705.6 ± 101.8; and RYR2 1.3 ± 0.7. Note that the level of expression varies greatly between the different genes. The exposure times shown in A were chosen to best visualize the WT and I1 signals in each case.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In this report, we have shown that transgenic expression of a dominant negative mutant (A5) of the regulatory A subunit of PP2A, which is defective in binding regulatory B subunits but normal in binding the catalytic C subunit, leads to a significant increase in heart weight and heart weight-to-body weight ratio. Previous experiments with tissue culture cells (32) have demonstrated that expression of the A5 subunit causes an increase in the amount of the core enzyme and a decrease in the amount of the holoenzyme, resulting in a change in PP2A activity. In the heart, expression of the A5 subunit also caused an increase in the ratio of the core enzyme to the holoenzyme and a change in activity, and this change is responsible for the observed increase in heart weight and the depressed cardiac function. At present, we do not know what causes the dramatic increase in heart weight at older age. It is conceivable that this is due to genetic or environmental factors in addition to expression of the A5 subunit. The increase in heart weight was proportional to the expression of the A5 subunit, because I4-transgenic mice, expressing less of the A5 subunit than I1-transgenic mice, had smaller hearts than the I1-transgenic mice. Furthermore, heterozygous I1-trangenic mice had smaller hearts than homozygous mice. These observations further indicate that the A5 subunit is the actual cause for the observed phenotype. Echocardiographic analyses revealed that transgenic hearts were significantly dilated, the walls of the hearts were thinner, and the capacity of the hearts to contract was considerably reduced. These findings are consistent with a phenotype of dilated cardiomyopathy. The A5-transgenic hearts show only a modest increase in -MHC expression and normal atrial natiuretic factor expression. These findings are compatible with the echocardiography results showing no increase in wall thickness. In addition, no significant interstitial fibrosis occurred.

We have shown that the heart weight of the A5-transgenic animals was significantly increased at birth. Because up to this time in development the heart grows by an increase in the number of cardiomyocytes, this finding might suggest that expression of the A5 subunit stimulates cardiomyocyte proliferation (hyperplasia) during fetal development. Alternatively, the cardiomyocytes in A5-transgenic mice may be abnormally large at birth, whereas the number of cardiomyocytes is unchanged. Transgenic overexpression of the c-myc proto-oncogene also causes cardiac enlargement, which in this case is due to myocyte hyperplasia during fetal development (12).

We have demonstrated that the core enzyme and holoenzyme are both similarly abundant in the heart, as shown previously for bovine hearts (25) and for several lines of cultured fibroblasts (15). This was achieved by quantitative immunoprecipitation of the core enzyme and holoenzyme with specific monoclonal antibodies. The conditions of extract preparation and immunoprecipitation were chosen such that proteolysis and dissociation of the holoenzyme into the core enzyme and B subunit were minimized. Our data contradict the commonly held view that the holoenzyme is the only form of PP2A in intact cells and that the core enzyme is an artifact of enzyme purification (for discussion, see Ref. 15). A simple calculation shows that the change in the amounts of the core enzyme and holoenzyme, which would result from a twofold A5 subunit overexpression, are relatively small (in the order of 30%) and fall within close range of the observed values of 20-30%. If core enzymes were not initially present in cardiomyocytes, the expected increase in core enzyme activity due to the formation of the A5-C subunit would be dramatic (50-fold). That this formation is, in fact, only 24% further supports the notion that the core enzyme is as similarly abundant as the holoenzyme. Our findings are important because the core enzyme differs significantly in substrate specificity from the holoenzyme. In addition, the core enzyme may serve as a reservoir for the binding of various regulatory B subunits. Besides the core enzyme and B-subunit-containing holoenzyme, at least three additional heart-specific forms of the holoenzyme can be inferred from the fact that three additional B subunits are expressed in the heart. The B56- and B56-subunits (members of the B' family) are preferentially expressed in the heart and may have cytosolic and nuclear functions, respectively (22, 42). The 74-kDa B'' subunit is highly expressed in the heart and brain (5). It is likely that these B subunits do not exist as free monomers but are bound to excess core enzymes forming distinct holoenzymes. The phenotype of A5-transgenic mice may be attributed either to an increase in core enzymes or a decrease in one or several forms of holoenzyme. Because the A5 mutant protein does not bind any form of the B subunit, we assume that transgenic expression of the A5 subunit caused a decrease in all forms of holoenzyme. Hence, it is not possible to link the phenotype of the A5-transgenic mice to the decrease of a particular form of the holoenzyme. To identify the heart-specific substrates of the core enzyme and the various holoenzymes is an important goal. One candidate is MAP kinase, a known substrate of PP2A (38), which is involved in the hypertrophic response of cardiomyocytes to a variety of factors and stimuli. Interestingly, there is an inverse relationship between increasing MAP kinase activity and decreasing PP2A phosphatase activity in cultured cardiomyocytes upon treatment with phorbol ester, an inducer of protein kinase C and of hypertrophic response (4).

It is intriguing that the genes encoding the B56- and B56-subunits have been mapped to regions on the human chromosome linked to heart disease (21). The B56-subunit gene is located on chromosome 1q41, a region linked to rippling muscle disease and ventricular cardiomyopathy, and the gene for the B56-subunit is located on 3p21, a region linked to a form of familial cardiomyopathy. However, mutations in the B56- or B56-subunit genes have not been reported in patients with heart disease. It is conceivable that mutations of the A subunit, which impair binding of the B-, B56-, B56-subunits, or the 74-kDa B'' subunit to the core enzyme, are genetically linked to heart disease, in particular because a large number of point mutations in the A subunit abolish binding of specific B subunits (33). We recently constructed mutants that do not bind B' (B56) subunits but bind B- and B'' subunits normally (33). The expression of these mutants in the heart is expected to shed light on the function of B' subunits, including the B56- and B56-subunits.