Cardiac Hypertrophic and Developmental Regulation of the β-Tubulin Multigene Family*

Increased microtubule density, through viscous loading of active myofilaments, causes contractile dysfunction of hypertrophied and failing pressure-overloaded myocardium, which is normalized by microtubule depolymerization. We have found this to be based on augmented tubulin synthesis and microtubule stability. We show here that increased tubulin synthesis is accounted for by marked transcriptional up-regulation of the β1- and β2-tubulin isoforms, that hypertrophic regulation of these genes recapitulates their developmental regulation, and that the greater proportion of β1-tubulin protein may have a causative role in the microtubule stabilization found in cardiac hypertrophy.

Increased microtubule density, through viscous loading of active myofilaments, causes contractile dysfunction of hypertrophied and failing pressure-overloaded myocardium, which is normalized by microtubule depolymerization. We have found this to be based on augmented tubulin synthesis and microtubule stability. We show here that increased tubulin synthesis is accounted for by marked transcriptional up-regulation of the ␤1and ␤2-tubulin isoforms, that hypertrophic regulation of these genes recapitulates their developmental regulation, and that the greater proportion of ␤1-tubulin protein may have a causative role in the microtubule stabilization found in cardiac hypertrophy.
When under pathological circumstances the heart is forced to eject blood against an increased impedance, the terminally differentiated cardiac muscle cell, or cardiocyte, responds by hypertrophic growth (1). The resultant increase in muscle mass constitutes the basic compensatory cardiac response to sustained hemodynamic overloading, but this initial compensation is frequently vitiated by a progressive decline in cardiocyte contractile function (2), so that congestive heart failure ensues.
We have found that this cardiocyte contractile defect is caused by increased density of the cellular microtubule network (3), which imposes a viscous load on the shortening sarcomeres during contraction (4). Thus, microtubule depolymerization in hypertrophied cardiocytes restores normal cellular contractile function, and induced microtubule hyperpolymerization in normal cardiocytes causes these cells to exhibit the same contractile abnormality found in hypertrophied cells (3).
The ␣␤-tubulin heterodimer-microtubule system is in a dynamic steady state. Therefore, in attempting to uncover the cause of increased microtubule density in hypertrophied cardiocytes, we focused on increased tubulin synthesis (5) and thus microtubule formation as well as on increased stability of the microtubules once formed. With respect to the latter, we have indeed found marked stabilization of the microtubule network in hypertrophied cardiocytes associated with a substantial increase in the predominant microtubule-associated protein of the heart, MAP4 1 (6). Although recent data (7,8) put into question the previously accepted role of MAP4 in microtubule assembly and stability for some cell types, the muscle-specific variant of MAP4 appears to play a role in striated muscle (9). Given that MAPs and the expressed proteins of the ␤-tubulin multigene family exhibit coordinate developmental regulation (10) and that the latter may via their isoform-variable carboxyl-terminal domain confer differing MAP binding affinity and microtubule stability after assembly (11), the question of whether there is differential regulation of the members of the ␤-tubulin multigene family during cardiac hypertrophy assumed pivotal importance.

EXPERIMENTAL PROCEDURES
Right Ventricular Pressure Overloading-Pressure overload hypertrophy of the feline right ventricle (RV) was created (12) by placement of a 2.9-mm inner diameter band around the proximal pulmonary artery. Because the RV mass increase stabilizes by 2 weeks after a step increase in load (5), at 2 weeks after surgery intravascular pressures were measured in these and in control cats; values in the systemic circulation were the same for both groups. The heart was then removed, RV mass was determined, and Ca 2ϩ -tolerant quiescent cardiocytes were isolated enzymatically from the RV and left ventricle (LV) separately (13). All operative procedures were carried out under full surgical anesthesia; all procedures and the care of the cats were in accordance with institutional guidelines. At 2 weeks, RV systolic pressure was doubled, and there was a 59% increase in the ratio of RV to body weight; the mass of the normally loaded same animal control LV was unchanged.
Western Blotting-Peptide synthesis and coupling were performed as described (14) with minor modifications. The peptides underlined in Fig. 1 were synthesized, purified, keyhole limpet hemocyanin-conjugated via glutaraldehyde cross-linking, and injected into rabbits. Sera were monitored on slot blots using bovine serum albumin-conjugated peptides until high antibody activity was achieved; the IgG fractions were then purified using covalently coupled peptide columns. BamHI fragments of the human h␤1, mouse m␤2, and human h␤2 (14) genes, which encode the carboxyl-terminal 100 amino acids of ␤1-, ␤2-, and ␤4-tubulin as noted here, respectively, were ligated into the pET-28 histidine-tagged expression vector (Novagen). Cultures were induced to express for 1.5 h with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. Cells were then spun down, resuspended in binding buffer, and lysed with a French press. The resultant fusion proteins, after isolation on a Ni 2ϩ -chelation column (Novagen), were used to verify the monospecificity of each of the site-directed polyclonal antibodies. Because of potentially differing antibody affinities, aliquots of the purified fusion proteins for ␤1-, ␤2-, and ␤4tubulin were resolved on SDS-PAGE and silver-stained; loading was adjusted to produce equal tubulin band intensities. For the subsequent immunoblots, fusion protein loading normalized in this manner was then used as an internal standard for analysis of the tubulin present in the samples, where the ratio of integrated optical density for the single band for each sample to that for each fusion protein was estimated.
Northern Blotting-To generate isoform-specific cDNA probes, the 3Ј-untranslated region of each ␤-tubulin isoform gene was cloned from the cat. ␤1and ␤4-tubulin were cloned from a cat cDNA library (Strat-agene) by PCR using oligonucleotides specific to the conserved region of each ␤-tubulin isoform and a poly(T) oligonucleotide as primers. ␤2tubulin was cloned by reverse transcriptase-PCR from total RNA isolated from cat brain using a poly(T) primer for reverse transcription; oligonucleotides from the conserved region of ␤-tubulin and a poly(T) primer were then used for PCR amplication of the cDNA products, which were cloned into the pT7Blue vector (Novagen) and sequenced (Sequenase, U. S. Biochemical Corp.). Oligonucleotides that specifically amplified the 3Ј-untranslated region of each gene were used to generate 32 P-radiolabeled PCR probes. To estimate ␤-tubulin isoform mRNA levels, total RNA was extracted from frozen RV and LV samples from the same hearts (15). The RNA samples were dissolved in 500 l of diethylpyrocarbonate-treated water and quantified spectrophotometrically at a 260/280 nm extinction coefficient, and to assess RNA quality and to assure equal loading of RV and LV RNA for Northern blots, 3-5-g RNA samples were stained with ethidium bromide and run on 1% agarose check gels. We then electrophoresed 5-7-g RNA samples on denaturing 2% formaldehyde, 1% agarose gels followed by 1.5 h of pressure-driven blotting to a nylon membrane (Hybond-N, Amersham Pharmacia Biotech). The RNA was immobilized on the nylon membrane by UV cross-linking (Stratalinker, Stratagene), and the membrane was prehybridized for 4 h at 42°C in a solution containing 50% (v/v) deionized formamide, 0.2% (w/v) Ficoll, 0.02% (w/v) polyvinylpyrrolidone, 5ϫ SSC, 10 mM MOPS, pH 7.0, 2 mM EDTA, 100 g/ml denatured salmon sperm DNA, and 0.2% (w/v) SDS. The membrane was hybridized for 16 h at 42°C in a solution containing a 32 P-radiolabeled probe (0.5-1.0 ϫ 10 6 cpm/ml) for each ␤-tubulin isoform. The Northern blots were washed 3 times in 2ϫ SSC, 0.1% SDS for 1.5 h at 42°C followed by a wash in 0.2ϫ SSC, 0.1% SDS for 20 min at 42°C and processed for autoradiography. Each blot was normalized by stripping and reprobing with a glyceraldehyde-3-phosphate dehydrogenase probe PCR-generated from feline glyceraldehyde-3-phosphate dehydrogenase cDNA.
mRNA Stability-Hearts were removed from control cats and from cats 2 weeks after RV pressure overloading; cardiocytes isolated enzymatically (13) from the RV and LV separately were incubated at 37°C in 1.8 mM Ca 2ϩ mitogen-free M-199 medium at pH 7.4. The cardiocytes were either untreated or exposed to 5 g/ml actinomycin D for 0, 4, 8, or 18 h. To test the effect of an acute increase in the cytosolic concentration of tubulin heterodimers on tubulin mRNA stability, further control and hypertrophied cardiocytes were simultaneously exposed to both 5 g/ml actinomycin D and 10 M colchicine for 0 or 4 h. Total RNA was then extracted from the cells for Northern blot analysis using the three isoform-specific ␤-tubulin probes or a probe that recognizes the isoformcommon region of feline ␤-tubulin mRNA (5).
␤-Tubulin Isoform Fractionation and Localization-For the immunoblots, a fresh 100-mg tissue specimen from the RV and LV of each cat was homogenized in 2 ml of hypotonic Tris buffer (20 mM Tris, pH 7.4, 20 mM ␤-glycerophosphate, 1 mM dithiothreitol, 10 g/ml leupeptin, 2 g/ml pepstatin, 10 g/ml aprotinin, 0.5 mM phenylmethylsulfonyl fluoride) and centrifuged at 15,000 ϫ g at 4°C for 15 min. The supernatant was mixed with an equal volume of SDS-sample buffer and saved as the cold-extractable fraction. The pellet was extracted in 2% Triton buffer (100 mM Tris, pH 7.4, and 2% Triton X-100) with the same protease inhibitors and centrifuged at 15,000 ϫ g at 4°C for 5 min; no tubulin was detected in the supernatant. The pellet was added to SDS-sample buffer and boiled for 5 min. The SDS-solubilized pellet and the coldextractable fraction were resolved on 12.5% SDS-PAGE and immunoblotted with a 1:10,000 dilution of the ␤1-, ␤2-, or ␤4-tubulin-specific antibodies. For double label immunofluorescence confocal micrographs, freshly isolated RV and LV cardiocytes in 37°C modified M-199 medium (10 mM Hepes, pH 7.4, 250 M Ca 2ϩ ) were sedimented onto FIG. 1. Carboxyl-terminal domains of the five vertebrate ␤-tubulin polypeptides, the sequences within these peptides used for antibody generation, and validation of the specificity of three of these antibodies. Peptides corresponding to the underlined region of each ␤-tubulin isoform shown in the upper panel, which uses the one-letter amino acid code, were used to prepare site-directed polyclonal antibodies. Fusion proteins corresponding to the ␤1, ␤2, and ␤4 isoforms were then used, as shown in the lower panel, to determine antibody specificity. These fusion proteins, with concentrations adjusted to produce comparable staining intensities, were loaded onto each lane of the 12.5% SDS-PAGE before immunoblotting and visualization by enhanced chemiluminescence (DuPont). Each isoform-specific antibody bound strongly to the corresponding fusion protein, but there was no detectable antibody cross-reactivity.
FIG. 2. Immunoblots of ␤-tubulin isoforms in RV and LV myocardium from a control cat and a RV pressure-overloaded cat 2 weeks after pulmonary artery banding. Both free (lanes 1 and 3) and polymerized (lanes 2 and 4) tubulin fractions, prepared and validated as before (6), were probed with the ␤1-, ␤2-, or ␤4-tubulin-specific antibodies or, for total ␤-tubulin, with an antibody denoted as "common ␤," which recognizes an epitope common to all ␤-tubulin isoforms (DM1B, Amersham Pharmacia Biotech). Equal proportions of the free and polymerized samples were loaded onto the two lanes for each ventricle, and an equal amount of protein as determined by a bicinchoninic acid assay (Pierce) was loaded for the RV and LV samples. The LV pressure of both cats was normal and equivalent.
laminin-coated coverslips for 45 min and then, after 1 h of exposure to either 37 or 8°C, extracted for 1 min in 1% Triton X-100 in microtubule stabilization buffer (16), washed three times in the same buffer, and fixed for 30 min with 3.7% formaldehyde, all at 25°C. After blocking with 10% donkey serum in 0.1 M glycine, the cardiocytes were incubated overnight at 4°C both with a 1:200 dilution of our rabbit ␤1-tubulinspecific antibody and with a 1:500 dilution of a mouse monoclonal antibody (B-5-1-2, Sigma), which recognizes all native ␣-tubulin isoforms (17), followed by both Cy3-labeled anti-rabbit IgG and Cy5-labeled anti-mouse IgG (Jackson ImmunoResearch) secondary antibodies. Micrographs were acquired as single 0.7-m confocal sections taken at the level of the nuclei (LSM GB-200, Olympus).
Cardiac Developmental Expression of ␤-Tubulin-Pregnant and 1-, 20-, and 90-day-old postpartum Sprague-Dawley rats were anesthetized with sodium pentobarbital (50 mg/kg, intraperitoneally). The 15day-old embryonal and 1-day-old neonatal rats were decapitated, the hearts were removed, and the atria and great vessels were trimmed away. About 30 embryonal hearts and ϳ50 mg of ventricular myocardium from the other stages were homogenized in 500 l of lysis buffer (10 mM Tris, 0.5 mM dithiothreitol, 1 mM sodium vanadate, 1% sodium dodecyl sulfate, pH 7.4), boiled for 5 min, and centrifuged at 16,000 ϫ g at room temperature for 10 min; the supernatants were saved. For the subsequent 12.5% SDS-PAGE, an equal amount of protein (25 g) as determined by a bicinchoninic acid assay (Pierce) was processed with SDS-sample buffer and loaded for each sample; immunoblotting was done with the same antibodies as those specified in Fig. 1. Three samples from each stage were studied with confirmatory results.

RESULTS AND DISCUSSION
The five ␤-tubulin isoform proteins whose expression was examined in this model, classified according to Cleveland (14,18), are shown in Fig. 1. Site-directed polyclonal antibodies were generated against each of these ␤-tubulin isoforms using synthetic peptides having the sequences underlined in Fig. 1. After preparing the appropriate fusion proteins, the specificity of the peptide column-purified ␤1, ␤2, and ␤4 antibodies was validated, as also shown in Fig. 1. No reactivity of the ␤3 and ␤5 antibodies with homogenates from either normal or hypertrophied myocardium was detected despite strong reactivity of each antibody with the respective bovine serum albumin-conjugated peptide.
Immunoblots of normal feline hearts showed that ␤4-tubulin expression is greatly preponderant (see below). As shown in Fig. 2, by taking advantage of the much higher affinity of the ␤1and ␤2as opposed to the ␤4and common (all isoforms) FIG. 3. Summary of ␤-tubulin isoform immunoblots during development of RV hypertrophy. For this semiquantification, fusion proteins over a concentration range, which with the corresponding antibody produced a linear densitometric relationship with protein concentration, were loaded onto 12.5% SDS-PAGE gels along with cardiac samples whose protein concentration had been adjusted to produce a densitometry signal within this same linear range. Statistical comparisons were by two-way ANOVA (analysis of variance) and a means comparison contrast, where n is the number of cats at each of the four time points. *, p Ͻ 0.01 for difference from the value for free tubulin at matched time points; †, p Ͻ 0.01 for difference from the initial control value within a group.

FIG. 4. Time course for expression of ␤-tubulin isoform transcripts during development of cardiac hypertrophy.
Northern blots of each ␤-tubulin isoform in myocardium from control and RV pressure-overloaded cats were prepared from total RNA isolated from the RV and LV at the indicated times after pulmonary artery banding. The blots were probed with ␤-tubulin isoform-specific cDNA probes. Equal RV versus LV loading was confirmed by stripping and reprobing with a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe. Similar results were obtained in two additional cats at each time point.
␤-tubulin antibodies and by varying blot exposure times, it was possible to visualize all four classes of ␤-tubulin in the same homogenates, where the samples from the control and RV hypertrophy hearts were treated identically. In the hypertrophied RV, despite the marked increase in both free and polymerized isoform-common ␤-tubulin, there was little change in ␤4-tubulin. Rather, the increased ␤-tubulin was accounted for by marked increases in free and polymerized ␤1and ␤2-tubulin. Densitometric analysis of immunoblots from the RVs and LVs of two further control and RV hypertrophy cats, where the total tubulin protein fraction (6) was assayed and equal amounts of fusion proteins specific to each ␤-tubulin isoform were loaded along with the myocardial samples, showed that in both ventricles from control cats and in the normally loaded LV from RV hypertrophy cats, ␤1-, ␤2-, and ␤4-tubulin were 4, 1, and 95%, respectively, of total ␤-tubulin; in the hypertrophied RVs, ␤1-, ␤2-, and ␤4-tubulin were 20, 4, and 76% of total ␤-tubulin. Of singular interest, as shown by the immunoblot summary data in Fig. 3, solely for ␤1-tubulin in the hypertrophied RV there was a disproportionate increase in the microtubule-assembled pool. That is, at 2 weeks of RV pressure overloading the increase of ␤1-tubulin in the microtubule-assembled pool was about twice that in the unassembled heterodimer pool. Although functional significance has been ascribed to differential ␤-tubulin isoform expression in the male germ line of Drosophila (19,20) and has been inferred from heterogeneous cellular isoform distributions (e.g. Ref. 21), the present data represent the first such direct evidence in vertebrates, and this is in a context of potential consequence to an important human disease state.
When the time course of these changes was examined, where the pressure-overloaded RV was compared with the normally loaded LV in five cats at each time of 2 days, 1 week, and 2 weeks after an increase in RV load, by 1 week there were increases in both the polymerized and free ␤-tubulin fractions, with the greater increase being in the polymerized fraction. This increase was accounted for in its entirety by increases in The two cytosolic lanes were loaded equally as were the two cytoskeletal lanes; the loading ratio of the soluble versus cold-stable fractions was constant. The same result was obtained in the hearts from four further RV pressure-overloaded cats. The right panels show double label immunofluorescence confocal micrographs of cardiocytes isolated from the RV and LV of a feline heart 2 weeks after RV pressure overloading. Prior to fixation, the two upper cells were maintained at 37°C and the two lower cells at 8°C for 1 h; the latter condition causes selective depolymerization of labile microtubules (6). These cells were double-stained for ␤1-tubulin (red) and for ␣-tubulin (green), where ␤1and ␣-tubulin primary antibodies were followed by species-specific fluorochrome-conjugated secondary antibodies; areas of coincident decoration by both primary antibodies range in color from orange to yellow. The inset at the lower right corner of each micrograph is a magnified view of a segment of a single microtubule. Cardiocyte preincubation with the ␤1-tubulin peptide abolished ␤1-tubulin labeling. When this protocol was repeated using the ␤2-tubulin antibody, very little microtubule decoration was apparent, and it was not obviously selective for cold-stable microtubules (data not shown). the ␤1 and ␤2 isoforms. This pattern was found in additional studies to be maintained for up to 6 months after the hypertrophic response was complete (data not shown).
Expression of ␤-tubulin isoform transcripts was then examined by Northern blot analysis. Fig. 4, where the alternate polyadenylation site of ␤1-tubulin (22) produces two transcripts, shows that the pattern of ␤-tubulin isoform expression on the mRNA level mimics that seen on the protein level. That is, the amount of ␤4-tubulin mRNA was equivalent in the RV and LV of control cats and changed very little in either ventricle during the development of RV hypertrophy. There was, however, a striking and persistent up-regulation of ␤1and of ␤2-tubulin mRNA in the hypertrophying RV.
To gain some insight into the relationship of mRNA levels to protein levels for these ␤-tubulin isoforms, we examined mRNA stability in cardiocytes from the hypertrophied RV and the control LV of pulmonary artery-banded cats. Because it has not proved possible to isolate transcriptionally active nuclei from these cells, we used actinomycin D to inhibit RNA polymerase in fresh primary cultures of these cells and then measured the rate of decline of mRNA levels for the ␤-tubulin isoforms and for total ␤-tubulin by Northern analysis. Lanes 1-4 of Fig. 5 show that mRNA stability estimated in this manner is quite similar in hypertrophied RV versus control LV both for the three ␤-tubulin isoforms examined and for isoform-common total ␤-tubulin. The same result was obtained in the RV versus LV of normal cats (data not shown). Thus, increased ␤1and ␤2-tubulin protein in hypertrophied myocardium results from increased transcription of these genes.
The concurrent up-regulation of ␤-tubulin on both the protein and the message levels, which we have found to persist indefinitely in myocardium hypertrophying in response to a pressure overload (5), appears to contravene the co-translational negative feedback control that tubulin exerts on its own rate of synthesis (23) via reduced mRNA stability. That is, the stability of ribosome-bound ␤-tubulin mRNA is controlled by co-translational binding of either ␤-tubulin itself or an intermediary factor to the amino-terminal ␤-tubulin tetrapeptide as it emerges from the ribosome. This binding then activates an RNase or causes ribosomal stalling with the result in either case being accelerated ␤-tubulin mRNA degradation, such that the net effect is an inverse relation between ␤-tubulin protein concentration and ␤-tubulin mRNA half-life. A breakdown of this regulatory control of tubulin synthesis, which might selectively affect different ␤-tubulin gene products, could explain increased microtubule density in hypertrophied cardiocytes. Thus, we used colchicine to acutely increase the concentration of tubulin heterodimers in control and hypertrophied cells and observed the effect of this intervention on tubulin mRNA stability. Lanes 5 and 6 of Fig. 5 show that to an equivalent degree for both control and hypertrophied cardiocytes, colchicine-induced microtubule depolymerization accelerates the rate of mRNA degradation for all three ␤-tubulin isoforms and for total ␤-tubulin. Further, in intact cats given 1 mg/kg colchicine intravenously with or without 2 or 7 days of RV hypertrophy, where ␣and ␤-tubulin mRNAs are markedly increased in hypertrophied RVs (5), ␤-tubulin mRNA levels had decreased equivalently 4 h later in normal and hypertrophied RVs and in the control LVs to 25.5 Ϯ 1.7% of the respective control values for cats not given colchicine. Thus, these data demonstrate that FIG. 7. Summary of ␤-tubulin isoform distribution in cold-stable versus cold-labile microtubules. These ratios were obtained by densitometric quantification of the two fractions for each isoform when run on 12.5% SDS-PAGE gels and immunoblotted with the corresponding antibodies. Statistical comparisons were by two-way ANOVA and a means comparison contrast, where n is the number of control or RV hypertrophy cats (RVH). *, p Ͻ 0.01 for difference from the ventriclespecific value for the other isoforms; †, p Ͻ 0.01 for between ventricle difference for a given isoform. the co-translational regulatory mechanism for controlling tubulin mRNA stability is intact in normal and hypertrophied terminally differentiated cardiocytes, such that, again, there is an authentic increase in the transcription of the ␤1and ␤2tubulin genes in cardiac hypertrophy. However, they also strongly suggest that such co-translational control is exerted as a rate-dependent rather than a concentration-dependent function of the cytosolic concentration of tubulin heterodimers, a finding having general rather than cardiocyte-restricted implications.
A central goal of this study was to determine whether if up-regulation of specific ␤-tubulin isoforms during cardiac hypertrophy was discovered, such changes in gene expression have functional significance. That is, the cardinal alterations of the extramyofilament cytoskeleton of hypertrophied cardiocytes, in terms of inducing contractile dysfunction, are interrelated increases in the quantity and stability of the microtubule network. If there were increased expression of specific ␤-tubulin genes, the protein products would directly account for the persistent increases in both free and polymerized tubulin. However, if one or more of the up-regulated isoforms conferred greater stability on the microtubules once assembled, this would contribute to increased density of the microtubule network via a mechanism independent of augmentation of the heterodimer pool. Thus, because stable microtubules are resistant to cold-induced depolymerization (6), we measured ␤1-, ␤2-, and ␤4-tubulin in the cold-stable cytoskeletal fraction of normal and hypertrophied myocardium and cardiocytes. The left panels of Fig. 6 show that although the proportion of ␤2or ␤4-tubulin in this fraction is very low for normal or hypertrophied myocardium, a significant proportion of ␤1-tubulin is found in the cold-stable cytoskeletal fraction, and this is more pronounced for the hypertrophied RV. The right panels of Fig.  6 show that the microtubule array of the hypertrophied cardiocyte is more cold-stable than that of the control cardiocyte, that microtubules of the hypertrophied cardiocyte incorporate more ␤1-tubulin than those of the control cardiocyte, and that this latter finding is especially pronounced in the cold-stable microtubules of the hypertrophied cell. These findings are consistent with densitometric analysis of tubulin isoforms, because the data in Fig. 7 show first in normal RVs and LVs a greater proportion of ␤1-tubulin in the cold-stable microtubule fraction and second a selective further shift solely of ␤1-tubulin to this fraction in the hypertrophied RV. Nonetheless, there is not necessarily an exact correspondence between microtubule cold stability and microtubule stability in vivo, such that these correlative data do not constitute proof that ␤1-tubulin-enriched microtubules are more stable in the intact cardiocyte of the heart in situ. We are therefore testing this point directly via adenovirus-mediated overexpression of ␤1-tubulin in isolated cardiocytes and via cardiac-targeted ␤1-tubulin overexpression in transgenic mice.
Because many protein isoforms normally expressed in the developing heart and then down-regulated in the adult heart are re-expressed after hemodynamic hypertrophic stimulation and because the specificity of these isoform switches may become important to understanding transcriptional regulation during cardiac hypertrophy, we examined cardiac developmental regulation of the ␤-tubulin multigene family in hearts extirpated from a developmentally timed series of rats that were subjected to immunoblot analysis of ␤1-, ␤2-, and ␤4-tubulin as well as total isoform-common ␤-tubulin. Fig. 8 shows that between embryonic day 15 and postpartum day 90 there is a rather modest decrease in total ␤-tubulin and a very minor increase in ␤4-tubulin. However, both ␤1and ␤2-tubulin peak at postpartum day 1 and decline to very low levels in the adult heart.
In common with the questionable functional significance of many other cardiac isoform switches wherein hypertrophy recapitulates phylogeny (1), these data do not imply a role for altered ␤-tubulin isoform expression in the generative processes either of cardiac hypertrophy or of cardiac development. Rather, the functional significance of the hypertrophic cardiac ␤-tubulin isoform switch described here is presently of known consequence only in terms of the resultant disordered contractile function (3). However, hypertrophic expression of the ␤-tubulin multigene family clearly does recapitulate its developmental expression, which, again, may provide eventual insight into ␤-tubulin transcriptional control mechanisms common to these two phases of cardiocyte growth. Apart from the intrinsic interest of these observations, the major impetus for this study was to ascertain their basis in terms of the augmented tubulin quantity (5) and microtubule stability (6) found in the pressure overload-hypertrophied heart. Although our finding of MAP4 up-regulation in cardiac hypertrophy (6) may well be important to the latter phenomenon, it does not directly explain the former. Thus, the possibility that increased expression of one or more members of the ␤-tubulin multigene family might explain both the greater quantity of tubulin and the greater microtubule stability, either directly via differing intrinsic properties or indirectly via differing MAP4 affinities (11), was quite intriguing. The data in this study indeed show that whereas expression of the predominant cardiac ␤-tubulin isoform is but little affected, there is marked up-regulation of two ordinarily minor cardiac ␤-tubulin isoforms. In addition to the possibility that this explains augmented tubulin quantity, selective localization of the ␤1-tubulin isoform to stable microtubules may also explain augmented microtubule stability. Finally, the fact that hypertrophic regulation of ␤-tubulin mimics its developmental regulation may provide the eventual insight required to understand transcriptional regulation of the ␤-tubulin gene family during cardiac hypertrophy.