INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

MICROTUBULES ARE REQUIRED for a variety of functions that are common to most cells, as well as functions that are specific to particular cell types or developmental stages. The common functions include secretion, organization of the cytoplasm, and cell division (13). Specific functions include neurite stability and outgrowth (29), and particular stages of skeletal muscle myogenesis (1, 4, 48). The underlying mechanisms for these microtubule (MT)-mediated functions have not been fully elucidated.

The MT array in most cell types is composed of a large number of dynamic MTs and a small subset of drug- and cold-stable MTs (41). Specifically, most cellular MTs "turn over" rapidly, exchanging MT-bound subunits with cytosolic subunits by end-dependent mechanisms (54). By these mechanisms (termed "dynamic instability" and "treadmilling") the cellular MT array is continually reorganized over time. Among other possibilities, the constant remodeling of the cytoskeleton allows cells to either initiate motility or assume new shapes rapidly. In addition, whereas most MTs (the dynamic MTs) are quickly depolymerized by high concentrations of MT-disrupting drugs such as nocodazole (8), some are resistant to drug-induced disassembly (5) and have been characterized as stable MTs. Because MTs contribute to a variety of cellular functions, it is probable that these two populations comprise functionally distinct arrays. Finally, it has been demonstrated that low concentrations of MT antagonists (in the nanomolar range) can suppress MT dynamics (growth and shrinkage) without generating MTs that are resistant to high concentrations of depolymerizing drugs. In those studies, the process of MT dynamics itself was shown to be important in several MT-mediated functions, including growth cone advance (47) and fibroblast motility (25, 55). Thus both MT dynamics and the stabilization of a subset may impart specificity to MT function.

MTs in most cell types are subject to one or more posttranslational modifications, including the reversible removal of the COOH-terminal tyrosine residue from -tubulin (detyrosination). Because these modifications occur slowly relative to the turnover of most MTs, the older (and more stable) MTs accumulate more modified subunits than do the dynamic MTs. Detyrosination of MTs increases the binding affinity of the plus-end-directed motor kinesin (24), which may in turn mediate the selective association between MTs and intermediate filaments (10). It has been speculated that other kinesin-mediated processes, such as organelle placement and motility, may be directed or enhanced by MT detyrosination (19). Although recent studies have demonstrated the presence of a large population of stable, detyrosinated MTs (commonly called "Glu" MTs due to the glutamic acid residue which serves as the new COOH terminus of -tubulin in detyrosinated MTs) in both neonatal (56, 57) and adult (39) heart cells, their functional importance remains unresolved.

MTs also contribute to many functions in heart cells, including secretion and receptor recycling (23, 26, 42), organelle placement (14, 15, 36), sarcomere mechanics (45), calcium channel activity (12), and beating rate (22, 39). When neonatal heart cells are isolated and then cultured on a variety of substrates, they will begin to beat spontaneously at a rate of ~60-100 beats/min. When the MTs in these cells are completely depolymerized with either colchicine or nocodazole, the beating rate increases by 40-80% (18, 22), suggesting that MTs exert a negative regulatory effect. To date, no information is available concerning the contribution of MT stabilization or modification to this process.

When adult feline hearts are induced to hypertrophy via pressure overload, the cells that are isolated from those hearts possess increased levels of tubulin, MTs, and MT associating protein-4 (MAP4), which are correlated with heart cell enlargement and altered contractile mechanics (39, 49). Furthermore, the MTs in those cells are more resistant to drug-induced depolymerization and are detyrosinated extensively. When all of the MTs in those cells are depolymerized with colchicine, the values for sarcomere mechanics return to normal, suggesting that the increased density of MTs is responsible for the deterioration of contractile function. Thus the results from the feline model suggest that, in addition to their role in regulating the beating rate in normal cells, some aspect of MT physiology (dynamics, stabilization of a subset, or detyrosination) may also contribute to the deterioration of heart cell function that accompanies the progression of certain cardiomyopathies.

It is important to note that although both beating rate and various aspects of sarcomere mechanics have been measured in cells from hypertrophying hearts (3, 50) they represent distinct processes. Although contraction velocity, the extent of contraction, duration of relaxation, etc., represent individual parameters of each contraction cycle, the spontaneous rate of beating represents the frequency of these processes as well as others, which include calcium dynamics and the activity of membrane ion channels (27, 28, 35).

In this study, we examined the influence of the relevant MT properties (dynamics, stabilization of a subset, or detyrosination) on the beating rate directly using neonatal rat heart cells as our model and found that the stabilized MT subset was sufficient to maintain the normal rate. In addition, we found that the direct introduction of excess tubulin into cells depressed the beating rate by one-half, demonstrating directly an inverse relationship between tubulin and MT polymer levels and one aspect of cardiomyocyte function.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. Nocodazole and the Tyr-tubulin-specific antibodies (1A-2) were purchased from Sigma Chemical (St. Louis, MO). Taxol was obtained from the Drug Synthesis and Chemistry Branch of the National Cancer Institute (Bethesda, MD). Texas Red-labeled n-hydroxysuccinimidyl (NHS) and Texas Red-labeled dextran were purchased from Molecular Probes (Eugene, OR). DME and other culture supplies were obtained from Fisher Scientific (Pittsburgh, PA).

Animals and cell preparations. Timed-pregnant, Sprague-Dawley rats were obtained from the National Cancer Institute Animal Program (Frederick, MD) and were fed ad libitum until needed. Cardiomyocytes were isolated from the ventricles according to the procedure provided by Worthington (Neonatal Cardiomyocyte Isolation System, Worthington Biochemical; Freehold, NJ), and were grown on glass coverslips (6-7 in a common 10-cm dish) at 37°C in a 95% air-5% CO₂ humidified atmosphere. For some experiments, the coverslips were precoated with calf serum overnight to enhance the attachment of cells.

Beating rate assay. At the beginning of the study, cells on coverslips were counted for two 60-s periods to determine the consistency of the beating rate and our counting method. Thereafter, beats were routinely counted for 30-60 s. For the nocodazole and taxol experiments, cells that were grown on glass coverslips for 4-7 days were transferred to a holding dish and placed on a stage-mounted, warmed culture chamber (Medical Systems; Greenvale, NY) that was blanketed with 5% CO₂. Then 15-30 cells were chosen at random and cell beats were visually counted. The cells were then treated with a drug (nocodazole or taxol) for 2 h and returned to the microscope stage. Again, cells to be counted were chosen at random. After similar numbers of cells were counted, the drug-containing medium was washed out, and the cells were incubated for another 2 h and counted a third time. No drug washout experiments were performed for cells treated with taxol. Often, the entire dish of coverslips were treated with the drug, and different coverslips were used for each cell count. Significant differences were assessed (by one-tailed, unpaired t-test, ANOVA, and several post hoc tests) using the StatView statistics program (version 5). All of the post hoc tests used (Fisher's protected least-significant difference test, Scheffé, Bonferroni-Dunn, Dunnett, and Tukey-Kramer) were consistent in assigning statistical significance to the results.

Protein preparations. The Glu tubulin antibodies were prepared by SynPep (Dublin, CA) using a peptide corresponding to the seven most COOH-terminal amino acids of Glu tubulin as the immunogen. The crude antiserum was then affinity purified in our laboratory after coupling a 15 amino acid, COOH-terminal peptide (containing an NH₂-terminal cysteine residue) to a sulfhydryl-reactive matrix (Sulfolink, Pierce Chemical; Rockford, IL). The flow-through and eluted fractions were concentrated through centrifugal filters (Millipore; Bedford, MA), aliquoted, and frozen for future use. The antibodies were characterized by competitive ELISA, using Tyr peptides as the competitor. The cells were microinjected with either a 3-6 mg/ml solution of Glu tubulin antibodies or a 10-30 mg/ml solution of the flow-through fraction, together with a Texas Red-conjugated marker (dextran or BSA).

MT protein (tubulin plus MAPs) was prepared from bovine brain essentially as described earlier (30). Tubulin was purified from the MT protein by DEAE chromatography (52) and was subsequently labeled with NHS-biotin as described (20). Tubulin was labeled with Texas Red according to a protocol developed by John Peloquin (University of Wisconsin) (33). Texas Red-labeled BSA was labeled using the same procedure. Cells were microinjected with a 4-8 mg/ml solution of tubulin.

Microinjections and immunofluorescence. For the antibody-injection experiments, cells were first incubated with 33 µM nocodazole for 1 h, to depolymerize all MTs and maximize the ability of the injected antibodies to bind (on drug removal) to Glu MTs. For all of the injection experiments, cells were switched from DME plus 10% (vol/vol) calf serum to calcium- and magnesium-free Hanks' balanced salt solution immediately before injection to reduce their sensitivity to the procedure. The cells were kept in this medium for as little time as possible (~10-20 min) to minimize problems of detachment from the substratum. Cells were microinjected using a constant-flow microinjection apparatus, such that ~10% of the cell volume was injected (58). After injection, the cells were quickly returned to normal culture medium and incubated overnight, to allow an appropriate interval for recovery. After measuring the beating rates for both the injected cells and neighboring, uninjected cells, they were quickly washed twice in warm PHEM buffer (60 mM PIPES, pH 6.95, 25 mM HEPES, 10 mM EGTA, and 2 mM MgCl₂) (40) and then fixed in ice-cold methanol. Fixed cells from all experiments were treated with 5% (vol/vol) normal goat serum in PBS to block nonspecific binding of the antibodies and then incubated with antibodies specific for either Glu or Tyr tubulin. Injected tubulin was detected with either anti-biotin antibodies (Sigma) or fluorescein-labeled streptavidin (Jackson Immunoresearch; Malvern, PA). Secondary antibodies were conjugated to either Texas Red, Alexa 488, or aminomethylcoumarin acetate (AMCA) (Molecular Probes). Micrographs were obtained using a Zeiss Axiovert microscope and hypersensitized Technical Pan (type 2415) film.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Characterization of spontaneous beating of neonatal myocytes. We prepared heart cells from neonatal rats using a standard protocol (Tissue Dissociation Guide, Worthington Biochemical). Cells were plated at low density, so that after 4-7 days of incubation most were observed as either single cells or as parts of small groups, with few intervening fibroblasts. Because wide variations in the beating rates of neonatal cells have been observed in other studies (18, 21, 31), we measured the average rate under a variety of conditions. Cells that were incubated for 3-11 days after their initial plating beat at similar rates. However, at either end of this time range many cells beat irregularly, with many beats involving only a part of the cell. Because most cells that were incubated for 4-7 days beat at steady rates, we used those cultures for all of our experiments.

Because the design of the microinjection experiments favored the use of single cells, we next compared the beating rates of isolated, single cells with cells in groups of 2-10. We found no significant differences in the mean beating rates among the groups, allowing us to use individual cells for microinjection and either individuals or small groups for all other experiments. In addition, we compared the beating rates from cultures derived from whole hearts with cultures derived only from the ventricles. Again, no differences in beating rates were observed.

Selective depolymerization of MTs with nocodazole. Low (nanomolar) concentrations of MT antagonists, including nocodazole, vinblastine, and taxol, have been shown to suppress MT dynamics in other cell types without generating drug-stable MTs or eliciting MT rearrangements (60). Those studies also implicated a requirement for MT growth and/or shrinkage in various cell processes, including growth cone advance, cell motility, and cell division (17, 25, 47, 55). Furthermore, the application of higher (micromolar) drug concentrations selectively depolymerizes the dynamic MT array, leaving the stable subset (9). At very high concentrations (33 µM, for example), nocodazole will normally depolymerize all MTs within a short period of time. In this study, we manipulated the MTs in heart cells in all three ways, to investigate whether MT dynamics (growth and shrinkage of MTs), the stable MT subset, or a more general property of the entire array was important for the regulation of beating rate. We first examined a variety of drug concentrations and incubation intervals to determine the most appropriate treatments. We found that 50-100 nM nocodazole, when applied for 1-10 h, resulted in little, if any, reduction in the number of Tyr MTs present in both the myocytes and the non-myocytes (compare Fig. 1, D-F with A-C). In addition, the level of Glu immunostaining on MTs did not vary significantly in either cell type. When a more extensive drug treatment was given (1 µM for 1-4 h), the density of MTs was significantly reduced (Fig. 1, G-I), with the majority of remaining MTs staining brightly with the Glu antibodies. This selective depolymerization was reversible; within 2 h of drug washout, a normal-looking MT array was present (Fig. 1, J-L). When high drug concentrations were applied (33 µM for 1-4 h), all MTs were disassembled (Fig. 1, M-O). Again, the normal MT array returned within 2 h of drug wash-out (Fig. 1, P-R) with one exception. Some cells had not yet regenerated a large subset of Glu MTs. However, it is likely that within this time period a stable population had been regenerated that had not yet become significantly detyrosinated. From these experiments we chose conditions that would "freeze" dynamics, disassemble the dynamic array, or disassemble all MTs. We next examined the beating rates of cells whose MTs had been manipulated in these ways.

View larger version (115K): [in this window] [in a new window]   Fig. 1.   Differential depolymerization of microtubules (MTs) by nocodazole. Phase-contrast micrographs are shown on left, Tyr immunostaining reactivity is shown on middle, and Glu immunostaining fluorescence is shown on right. Untreated cells (A-C) display abundant MT array that immunostains with both Tyr and Glu antibodies. Treatment of cells with 50 nM nocodazole for 2 h (D-F) reveals no obvious diminution in MT staining, whereas treatment of cells with 1 µM nocodazole for 2 h (G-I) depolymerizes most Tyr MTs, leaving a drug-resistant subset that stains with the Glu antibodies. Two hours after release from the drug (J-L), normal MT array is observed. Treatment of cells with high concentrations of nocodazole (33 µM, M-O) depolymerizes all MTs, including both Tyr and Glu MTs. Two hours after drug washout (P-R), abundant Tyr MT array regrows, whereas generation of prominent Glu array often occurs later. Bar, 50 µm.

Normal beating rates are maintained after selective depolymerization of dynamic MT population. Beating rates were determined visually, after equilibrating individual coverslips of cells on a stage-mounted incubator for ~10-20 min. Initially, we counted the number of beats per cell twice, each over a 60-s interval, to determine the most economical and appropriate measuring period. We determined that single counts that encompassed 25-30 beats were adequate. The average ± SE beating rate for all untreated cells (from all experiments) was 93 ± 1.5 beats/min, whereas the rates for individual experiments varied somewhat. When cells were treated with 50 nM nocodazole for 2 h, the beating rate declined slightly from 117 ± 3.9 to 107 ± 5.0 beats/min (Fig. 2). After the drug-containing medium was replaced with fresh medium and the cells were incubated for 2 h, the beating rate increased to 127 ± 7.0 beats/min. The initial decrease was not significant, but the change from the drug-treated rate to the rescued rate was significant (P = 0.044). The rate after drug rescue was not significantly different from the rate determined before drug treatment (P = 0.83; labeled "Pre-" in Fig. 2), suggesting that the overall effect of low nocodazole treatments was at most marginal.

View larger version (35K): [in this window] [in a new window]   Fig. 2.   The stable MT subset confers negative regulation of beating rate on heart cells. Randomly chosen cells were counted before (Pre-) and during (Noc) drug treatment, as well as 2 h after drug washout (Rescue). For the 50 nM drug treatment, no difference between the predrug and the drug-treated groups was observed, whereas there was a small but significant difference (* P = 0.044) between drug-treated and rescued groups. No significant differences were observed among groups treated with 1 µM nocodazole. Cells treated with 33 µM nocodazole exhibited signficantly higher beating rates (** P < 0.001) than either predrug or rescued groups. Values shown are means ± SE. For the 50 nM treatments, n = 7 experiments (and heart preparations) with n = 119 cells (Pre-), 81 cells (Noc), and 60 cells (Rescue); for the 1 µM treatments, n = 4 experiments with n = 57 cells (Pre-), 64 cells (Noc), and 54 cells (Rescue); and for the 33 µM treatments, n = 5 experiments with n = 71 cells (Pre-), 62 cells (Noc), and 55 cells (Rescue). Differences were assessed using one-way ANOVA.

When myocytes were treated with 1 µM nocodazole for 2 h, the beating rate did not change (114 ± 5.1 vs. 113 ± 8.2 beats/min, Fig. 2), demonstrating that the presence of the dynamic MT array was not required to maintain the normal beating rate. The rate was similar 2 h after drug washout (106 + 6.2 beats/min), a time period sufficient to fully restore the normal complement of MTs (refer to Fig. 1).

It has been reported that the disassembly of all cardiomyocyte MTs increases the beating rate by ~40-80% (18, 22). For a direct comparison with the earlier results, beating myocytes were treated with high concentrations of nocodazole (33 µM) before measuring their rates (Fig. 2). Consistent with the earlier studies, we found that the beating rate increased significantly (P = <0.0001 over a 2-h period) from 115 ± 5.1 to 150 ± 5.6 beats/min, an increase of 30%. The effect was reversible, in that 2 h after drug washout the beating rate returned to normal values (120 ± 5.7 beats/min). In conclusion, although the "freezing" of MT dynamics did not affect beating rate, the disassembly of all MTs resulted in a mild increase, again demonstrating the negative regulatory role of MTs in this cell function. Furthermore, the stable MT subset was sufficient for this purpose, not requiring the dynamic array. We next tested the effect of stabilizing all MTs on beating rate.

Beating rate is unaffected by exogenous stabilization of MTs with taxol. Lampidis et al. (21) found that a taxol treatment of 10 µg/ml for 24 h reduced the beating rate of confluent, neonatal cells by ~35%, but shorter treatments (2 h) revealed no differences. We showed earlier (56) that a 4-h treatment with taxol stabilized myocyte MTs and that longer (18 h) incubations rearranged the MT array into thick bundles. In this report, we tested the ability of 10 µM taxol to decrease the beating rate, after both short (4 h) and long (~18 h) incubation periods (Fig. 3). In both cases, however, we noted no significant changes in the beating rates (compared with cells on other coverslips that were not treated with taxol), showing that at least moderately increased MT polymerization, stabilization, and rearrangement into bundles did not measurably affect the beating rate.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Exogenous stabilization of MTs with taxol does not affect heart cell beating rate. Cells on coverslips were incubated with either normal growth medium or medium plus 10 µM taxol and then incubated for either 4 or ~18 h. Both control (Con) and taxol-treated cells (Tax) were assayed for beating rate as described. Values (means ± SE) for taxol-treated cells were not significantly different from control-treated cells at either time period; n = 3 experiments (4-h treatment) with n = 44 cells (Con) and 45 cells (Tax); n = 4 experiments (18-h treatment) with n = 45 cells (Con) and 50 cells (Tax).

Detyrosination of stable cardiac MTs is not required for regulation of beating rate. Neonatal cardiomyocytes possess a large population of stable, detyrosinated MTs (56). It has been suggested that posttranslational modification may act to specify or enhance the functions of particular MTs (6). Therefore, we sought to determine the contribution of detyrosination to the regulation of beating rate. To accomplish this goal, we microinjected affinity-purified antibodies to Glu tubulin (plus a fluorescent marker) into heart cells, allowed them to recover from the injection process overnight, and then monitored their beating rates the next day. A previous study showed that antibodies to a different protein, tubulin tyrosine ligase, remained within the cytoplasm and were able to inhibit enzyme activity for many hours after injection (59), suggesting that the antibodies would remain undegraded during the overnight incubations allotted for these experiments. Indeed, cells that were immunostained after the completion of each experiment showed the decoration of a subset of MTs by the injected antibodies (data not shown). However, the beating rates of the Glu antibody-injected cells were indistinguishable from those of cells (on separate coverslips) that had been injected with the flow-through fraction (P = 0.998 in an unpaired t-test) and were very similar to uninjected cells found on the same coverslips (Fig. 4). We concluded that detyrosination did not contribute significantly to this MT-mediated function.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Acute elevation of intracellular tubulin concentration decreases heart cell beating rate, while masking the Glu MT population has no effect. Cells were microinjected with either affinity-purified antibodies to Glu tubulin or flow-through fraction from affinity column (left) or in other experiments, with either concentrated tubulin or buffer alone (right). No differences in beating rates were observed between flow-through-injected cells (F-th) and the antibody-injected cells (Ab; P = 0.998 using a one-tailed, unpaired t-test), or between treated and untreated cells (U). Tubulin-injected cells (Tub) exhibited a marked decrease in rate, from 66 ± 2.5 to 30 ± 1.5 beats/min (** P < 0.001 from ANOVA comparison of tubulin-injected, flow-through-injected, and uninjected cells), whereas buffer-injected cells (Buff) beat at similar rates as uninjected cells (U) in those experiments. Values shown are means + SE; n = 7 experiments (antibody injections), with n = 108 cells (U), 66 cells (F-th), and 33 cells (Ab); and n = 8 experiments (tubulin injections), with n = 112 cells (U), 59 cells (Buff), and 120 cells (Tub).

Acute elevation of intracellular tubulin concentration depresses beating rate directly. The deterioration of cardiac performance during the progression of cardiac hypertrophy, including all aspects of sarcomere contraction (velocity, extent, and relaxation) has been correlated with an increase in the tubulin concentration and the density of MTs within the cardiomyocytes (49). We tested the effect of increased tubulin and MT levels on the beating rate of otherwise healthy, beating heart cells. Tubulin was purified from brain tissue, labeled with either biotin or Texas Red, and then microinjected into heart cells. When biotin was the tubulin reporter, a fluorescent marker was also injected. Enough tubulin was injected to increase the estimated tubulin concentration approximately two- to threefold. As before, the cells were allowed to recover from the injection process overnight before assaying their beating rates. As a control, cells on different coverslips were injected with buffer plus the fluorescent marker. Whereas the beating rates of the control-injected cells did not differ from uninjected cells on the same coverslips, the beating rates from the tubulin-injected cells decreased by one-half (Fig. 4). Cells that were fixed and prepared for immunofluorescence after the assay showed the incorporation of the labeled tubulin into the MT network, without inducing a noticeable change in the distribution of MTs, including the type of bundling usually seen after extensive taxol treatment (data not shown). Thus we concluded that the acute elevation of the intracellular tubulin concentration has a direct, negative effect on beating rate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

It has been proposed that MTs degrade the sarcomeric performance of cardiomyocytes by imposing a viscous load on the cytoplasm, rather than by interfering with sarcomeric mechanics directly (45). This conclusion was reached by comparing the "cortical stiffness" and apparent viscosity of the cytoplasm in cells isolated from pressure-overloaded, ventricular myocardium with cells isolated from the opposite, normally loaded ventricle. Although the evidence for this conclusion was strong, it relied on a model (hypertrophic myocardium) whose cells had experienced many noncytoskeletal alterations [for example, changes in the complement of sarcoplasmic reticulum proteins (2)], as well as the observed increase in MTs. Because the beating rate of myocytes derived from hypertrophic hearts also has been shown to be lower than in normal heart cells (3), we measured that parameter in otherwise healthy myocytes within which the intracellular tubulin concentration had been immediately and dramatically increased. The results from this study show conclusively that the elevation of intracellular tubulin (and MT) levels is sufficient to reduce the beating frequency of the injected cells.

An earlier report estimated that the tubulin concentration was elevated approximately twofold in myocytes isolated from pressure-overloaded right ventricles (49). We attempted to "load" the neonatal myocytes with a similar amount of tubulin. We estimated that if the volume of a neonatal myocyte was similar to that of other cultured cells [~3 pl (11)], and if the amount of tubulin was two to three times higher in neonatal cells (57) than in adult rat cells [~0.4 pg/cell (38)], or 0.8-1.2 pg/cell, then injecting roughly 10% of the cell volume (~0.3 pl) with a 4 mg/ml solution of tubulin should introduce an extra 1.2 pg of tubulin, to roughly double the intracellular concentration. Thus the final intracellular tubulin concentrations achieved in this study were similar to those estimated to be present in the hypertrophied cells studied earlier.

We found that the stable MT subset was sufficient to dampen the rate of beating. It is somewhat surprising, however, to find that the increased stability of all MTs by taxol did not decrease the beating rate. It may be that only more significant increases in MT levels, i.e., that induced by either hypertrophy or the microinjection of exogenous tubulin, result in an observable decrease in rate. For example, it has been determined that roughly 25% of the total tubulin in adult rat cardiomyocytes (32) and heart tissue (61) is in polymer form, which corresponds to roughly 0.1 pg tubulin/cell. Furthermore, after the treatment of heart tissue with taxol (10 µM for 3 h, a treatment that alters the viscous properties of the tissue), the proportion that is in polymer form rises to ~40% of the total (61) or to about 0.16 pg/cell. Assuming no changes in cell volume over that period, these results suggest that the concentration of polymeric tubulin increases by ~60% after taxol treatment. In contrast, it has been determined that after pulmonary artery banding the hypertrophied right ventricular myocytes display an increase in polymeric tubulin of ~300% over their nonhypertrophied counterparts (50), demonstrating that hypertrophy induces a greater increase in MT mass than does taxol treatment. The proportions of cytosolic versus polymeric tubulin that exist after neonatal cardiomyocytes are treated with taxol are unknown. However, the observed lack of effect of taxol treatment could be explained if direct tubulin injection elevated the polymer level to a higher degree than did taxol. Also, we did not measure various aspects of contraction, such as the extent and velocity of shortening. It is quite possible that taxol might have affected those properties of contraction in our system. However, the results from this study are compatible with the report by Lampidis et al. (21), which showed that taxol affected the beating rate only after very long (24 h) treatments. It should also be noted that most if not all of the taxol treatments that have been performed have used cells possessing an intact MT cytoskeleton. Because taxol begins to act on MTs immediately after its application, the stabilization and extension of preexisting MTs would be expected to predominate over the formation of new MTs, possibly resulting in more subtle effects on beating rates. Finally, cells from pressure-overloaded, hypertrophic ventricles in the adult feline heart also possess an enhanced population of stable, Glu MTs (39). It seems logical (though it remains unproven) to assume that the stable cardiomyocyte MTs in the failing cat heart are sufficient to retard their beating rates as well.

Although the drug-stable population of MTs was sufficient for the regulation of the beating rate, neither the "quenching" of MT dynamics or the posttranslational detyrosination of MTs contributed significantly to this function. This result is consistent with the hypothesis that stabilized MTs regulate the beating rate by interfering mechanically with the process, in which case the overall bulk of the MT array would be more important than the modification of subunits on their surfaces. An a priori analysis could have envisioned detyrosination to be important if, for example, it served to enhance the binding of MTs to other cellular components (i.e., myofibrils, the desmin filament array, or the sarcolemma) and by that mechanism helped to "rigidify" the cytoplasm. However, cells injected with the Glu tubulin antibodies showed no obvious changes in the immunostaining pattern for either myosin or the rather disorganized array of desmin filaments (data not shown). Therefore, detyrosination must be at best just one component of a mechanism designed to integrate the cytoplasm of myocytes. Other contributors might include a protein such as plectin, which has been shown to interact both with MTs and microfilaments, with the latter mediated by myosin (43). Therefore, it is possible that the lack of functional Glu MTs might not decrease the structural rigidity of the cytoplasm significantly. Alternatively, detyrosination may be important only for other functions, such as receptor recycling (26), maintaining the position (15) or the activity (12) of the sarcoplasmic reticulum, secretion (23), or any other kinesin-mediated process (19).

Many of the apparently conflicting results on the role(s) of MTs in cardiomyocyte beating may be due to differences in the animal model being used [e.g., guinea pig vs. cat, neonatal vs. adult (53)], and the parameter being measured (beating rate vs. rate and extent of sarcomere shortening). For example, whereas the disassembly of all MTs in cultured, neonatal myocytes increases the beating rate and decreases the amplitude of contraction (18, 22), in adult cells there is no effect (44). In contrast, taxol-treated adult cells display the same alterations in sarcomere shortening as do the cells that have undergone hypertrophy, demonstrating that the maintenance of the normal dimer:polymer steady state is important for normal sarcomere mechanics. However, the precise relationship between the beating rate and individual parameters of contraction remains unclear. Although there are differences in the organization of MTs between neonatal and adult rat cardiomyocytes (7, 39, 56, 57), MT levels have been shown to rise during the onset of hypertrophy in both cases (16, 34), suggesting that the results from this study may be applicable to the reports that used adult cell models. Furthermore, alterations in MT organization after pressure-induced hypertrophy are similar in both ventricular chambers of the heart and are similar in canine and feline hearts (44). However, differences remain between the animal models tested. For example, the increased MT density in the adult rat model is transient (37), whereas MTs remain at high levels in the feline adult heart (16). In addition, MTs are not increased in the feline model if the increased load is due to an increase in volume instead of pressure. Finally, the role of MTs in the deterioration of heart cell function in tachycardia-induced dilated cardiomyopathy is minor (46), demonstrating that MTs may contribute little to some cardiac abnormalities, while exacerbating the dysfunction present in others. Thus considerable care must be exercised when the results obtained using different animal models and different assays of heart cell function are compared. Clearly, however, the MT component of the cardiomyocyte cytoskeleton can contribute significantly to the normal process of beating. One major goal for future work should be to further consolidate the results obtained using different animal models. Another goal will be to explore other contributions of MTs (or MT subsets) to normal heart cell physiology, including the organization and activity of plasma membrane channels (12, 51), to present a clearer picture of the role of the extra-myofibrillar cytoskeleton in this critical heart cell function.