INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CYTOCHROME-C OXIDASE (COX), the terminal enzyme complex of the mitochondrial electron transport chain, catalyzes the transfer of electrons from reduced cytochrome c to molecular oxygen and translocates protons across the mitochondrial inner membrane. The establishment of this electrochemical gradient across the membrane provides the driving force for ATP synthesis (7). Three-dimensional X-ray structural analysis reveals that the cytochrome-c oxidase enzymatic complex is composed of four subunits in Paracoccus denitrificans (16) and 13 subunits in the bovine heart (18, 40). Three of the thirteen subunits (I, II, and III) of mammalian cytochrome-c oxidase are encoded in mitochondrial DNA and constitute its catalytic core. The remaining 10 subunits (IV, Va, Vb, VIa, VIb, VIc, VIIa, VIIb, VIIc, and VIII) are encoded in nuclear DNA and are less well understood (7), although many functional and structural features of the homologous subunits have been analyzed in the yeast system (1, 5, 9, 12, 21, 27, 35, 37, 39, 43).

Isoforms of subunits VIa, VIIa, and VIII, designated heart (H) and liver (L), exist in mammals that may function to modulate COX activity under various conditions of cellular metabolism (17, 19, 36, 44). The H form of subunit VIa (as well as VIIa and VIII) is expressed only in striated muscle, whereas the L form is more widely expressed (17, 36). The function of these isoforms cannot be studied in the yeast system, because only subunit Va and Vb isoforms, corresponding to subunit IV in mammals, exist (4). The location and protein interactions of subunit VIa at the interface of the two cytochrome-c oxidase monomers in the dimeric structure suggest that subunit VIa is involved in dimerization and possibly in modulation of the enzymatic properties. Currently, little is known about the enzymatic properties of COX isoforms of subunits VIIa and VIII, but evidence suggests that isoforms of subunit VIa, modulate COX activity and show different nucleotide dependence (2, 29). For example, COX activity in heart, but not in liver, was stimulated by ADP and this stimulation was abolished with preincubation of heart COX with an antibody against subunit VIaH (2). These results indicate that COX VIaH may be required for mediating the tissue-specific allosteric effects of ADP and ATP on the enzyme in striated muscle.

The transcript levels of nuclear genes encoding for VIaH and VIaL during embryogenesis and postnatal development have been analyzed in bovine skeletal muscle (11) and in striated muscle of the mouse (26). We have previously shown that in striated muscle of the mouse, the predominant VIaL isoform in all fetal tissue was gradually replaced by the VIaH isoform in both cardiac and skeletal muscles, and this transition was not complete until the postnatal day 28 in the heart. These results suggested that transition of liver-type COX to heart-type COX might be essential for normal functioning of tissues with high aerobic metabolic demands.

The focus of this work was to determine the functional significance of subunit VIaH on the activity of COX and its effect on oxidative phosphorylation in the context of the intact working heart. This was accomplished using a gene-disruption strategy to create a murine line that lacked functional VIaH.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Generation of COXVIaH-deficient mice. The structure of the murine gene for VIaH has been previously described (41) (GenBank accession No. U34801) as has the developmental expression of the H and L subunits (26). Briefly, an expression cassette for neomycin phosphoglycerate kinase (PGK-neo, 1.7 kb) was inserted at the beginning of exon 2 at a SalI/EcoRV site, deleting 207 bp and truncating the 97 amino acid coding region of exon 2 at amino acid 28. A 1.8-kb herpes simplex virus-thymidine kinase cassette was inserted downstream of the 0.8-kb 3' flanking region of the gene, and the entire construct was linearized at a unique NotI site. The targeting vector consisted of 5.0-kb of 5' COXVIaH sequence, the neomycin cassette, a 0.8-kb fragment of 3' COXVIaH sequence, and a cassette encoding thymidine kinase drug resistance. Approximately 2.5× 10⁹ J1 embryonic stem cells (kindly provided by R. Jaenisch, Whitehead Institute), after growth to near confluence were electroporated in a final volume of 0.75 ml with 50 µg of the linearized construct and plated on mitotically arrested mouse embryonic fibroblasts. Positive-negative selection was carried out as described (38). Neomycin-resistant and gancyclovir-sensitive clones were expanded, and DNA was isolated by standard techniques. Authentically targeted clones (targeting frequency 1:9) were identified by hybridization with the 300-bp probe isolated downstream of the 3' flank (Fig. 1), expanded, and injected into day 3.5 blastocysts from naturally mated C57BL/6 mice. After injection, the blastocysts were implanted into pseudopregnant Swiss outbred females (6-10 per female), and Agouti offspring were mated to wild-type C57BL/6 partners. Progeny from these matings were screened for the presence of the mutated allele (Fig. 1), and heterozygous animals were crossed to generate mice deficient in VIaH, as assessed by Southern analysis (Fig. 1A). Correctly targeted mice were identified by digesting genomic DNA (isolated from tail biopsies) with SacI (unique site introduced by insertion of the PGK-neo cassette in exon 2) to generate a wild-type allele fragment (5 kb) and a mutant allele fragment (4 kb) identified using the 300-bp probe and Southern blot analysis.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Targeted disruption of the cytochrome-c oxidase subunit VIaH (COXVIaH) gene. A: the targeting vector was constructed by replacing exon 2 with a phosphoglycerate kinase (PGK) promoter driving the neomycin (neo) expression cassette. The PGK-neo cassette was flanked by a 5.0-kb 5' arm and a 0.8-kb 3' arm of homologous DNA. Exons are shown as filled boxes and the position of the DNA segment used as a probe for Southern blot analysis is shown as P and is located outside of the 3' arm of the targeting vector. B: Southern blot analysis using SacI digested genomic DNA and the 3' DNA probe. The wild-type allele is 5 kb and the mutated allele is 4 kb in size. C: Northern analysis to detect mRNA transcripts produced from the native and targeted COXVIaH alleles. Note COXVIaH transcripts are absent in RNA isolated from mutant (/) hearts and skeletal muscle and transcripts are reduced in the heterozygote (+/) compared with wild-type (+/+) striated muscle using the full length cDNA probe. Presence of -actin mRNA transcripts demonstrates the integrity of the RNA in all of the samples. HSV-TK, herpes simplex virus-thymidine kinase.

Northern analysis. Total RNA from mouse skeletal muscle and heart was prepared by directly homogenizing the tissues in 4 M guanidinium followed by CsCl density gradient centrifugation. RNA electrophoresis and filter hybridization were performed as described previously (24). Hybridization probes were synthesized by primer extension using sequence-specific primers and the entire COXVIaH cDNA template (22). Hybridization membranes were washed in 0.1× SSC/0.1% SDS at 60°C for 1 h (high stringency) to eliminate cross reactivity between the H and L isoforms (data not shown).

Electrophoretic techniques, isolation of complex IV subunits, and protein sequencing. Blue native-PAGE (BN-PAGE), second-dimension Tricine-SDS-PAGE, staining, and densitometric quantification were undertaken as previously described (31). Heart muscle (10 µg wet weight) was homogenized and solubilized by the addition of 6 µl 10% dodecylmaloside. After centrifugation at 100,000 g, the supernatant was applied to a 1-cm sample well for BN-PAGE (4-13% acrylamide gradient gel). The above protocol was scaled up 18-fold for sample preparation and protein sequence analysis. The visible blue bands after BN-PAGE, representing complex IV from wild-type and VIaH mutants, were excised from the gel and cut into six portions. These portions were loaded as a stack of six slices onto a Tricine-SDS gel (16% acrylamide containing 6 M of urea). After electrophoresis, the proteins were blotted onto Immobilon P and sequenced directly using a 473A protein sequencer (Applied BioSystems) (30).

Sequence analysis. Sequence analysis was performed as previously described (31). The protein sequences were: WT VIaH: ASAAKGDHGG; WT VIII-H: VSAKP; MUTANT VIaL: xxGAHGEEG (the two serines at the NH₂-terminus could not be identified); and mutant VIII-H: VSAKP.

Biochemical assays. Hearts were isolated, quickly weighed, and frozen between aluminum tongs precooled in liquid nitrogen. Myocardial tissue was pulverized in a mortar cooled in liquid nitrogen and extracted in perchloric acid. ATP concentration was determined by standard enzymatic assay using a commercially available kit (Sigma; St. Louis, MO). COX activity was determined at 25°C in (in mM): 120 NaCl, 1 EDTA, 20 HEPES, pH 7.5. After the addition of 40 µM reduced cytochrome c (horse heart), the reaction was initiated by the addition of solubilized samples. Samples were solubilized either by dodecylmaltoside under the conditions used for BN-PAGE (0.6 mg dodecylmaltoside per 10 mg heart muscle) or by 1.2 mg digitonin per 10 mg heart muscle, which produced identical results. The reaction was terminated by the addition of 5 mM NaCN.

Histochemistry and electron microscopy. Hearts from mutant and wild-type littermates were Langendorff perfused with Millonig's 10% buffered-Formalin and processed for paraffin embedding. With the use of a Leitz 1512 microtome, 4-µm-thick sections were obtained and stained with either hematoxylin and eosin for routine evaluation or Masson's trichrome for detection of collagen by using standard histologic techniques (33). Sections were examined under brightfield illumination using a Leitz Laborlux-S microscope. For transmission electron microscopy, hearts were perfusion fixed with 3% glutaraldehyde in phosphate buffer, postfixed in 1% osmium tetroxide and dehydrated in increasing concentrations of ethanol. Hearts were embedded in Spurr resin and sectioned with a diamond knife on a Reichert Jung Ultracut E microtome. Semithin (1 µm) sections were stained with toluidine blue and examined using light microscopy. Ultrathin sections (70 nm) were mounted on Formvar-coated copper oval slot grids, stained with uranyl acetate and lead citrate, and examined with a JEOL 1200EX II electron microscope at an accelerating voltage of 120 kV in the transmission mode.

Isolated working heart perfusion. Mice (25-30 g) were killed and the thoracic contents (washed immediately with 4°C perfusate) were excised and placed in 4°C perfusate. The heart and aorta were dissected free of extraneous tissue, and a fluid-filled cannula was threaded into the aorta and secured with a suture. Hearts were perfused in the Langendorff mode at 37.5°C with a (retrograde) perfusion pressure of 80 cmH₂O (60 mmHg). Mouse hearts require a retrograde perfusion pressure higher than 50 mmHg to avoid slowly developing cardiac dysfunction (14). A cannula was inserted into a small opening in the left atrial (LA) wall and secured circumferentially with a suture. The preparation was then converted to a working mode with LA pressure (LAP) of 15 cmH₂O and an anterograde perfusion pressure (APP) of 70 cmH₂O (52 mmHg). An APP of 70 cmH₂O was selected, because this was the highest APP tolerated by the COXVIaH/ hearts and because this pressure has been successfully utilized in working heart preparations described previously (3). Aortic pressure was monitored through a sidearm connector attached to the aorta cannula for measures of arterial pressure and heart rate (and for gating as needed) with a multichannel polygraph (Colbourne Instruments; Lehigh Valley, PA) for cardiac perfusions utilizing nuclear magnetic resonance (NMR) spectroscopy. The nonrecirculating medium was a modified Krebs-Henseleit bicarbonate buffer (95% O₂-5% CO₂; pH 7.4) containing (in mM) 118 NaCl, 4.7 KCl, 2.5 NaHCO₃, 1.2 MgSO₄, 5.5 glucose, 2.5 CaCl₂ and 0.5 Na₂EDTA. Coronary and aortic flows were measured by timed collections (over 5-min intervals); cardiac output is the sum of coronary and aortic flows. These collections were obtained by diverting either aortic flow or coronary vein effluent into a graduated cylinder. Stroke work was computed as mean aortic pressure multiplied by stroke volume times 0.0136 (14). A tiny capillary tube with a known volume phosphate standard was positioned along the length of the heart within the NMR tube to be used as a standard for quantitation of metabolites from the spectral data. Left ventricular performance curves were generated using the isolated working heart preparation. At a constant APP of 70 cmH₂O, the LAP was increased in a stepwise fashion from 15 to 20 to 25 cmH₂O and stroke work response was measured. One set of experiments was performed on the bench, and a second set was performed inside the vertical bore of the 400-MHz spectrometer during NMR imaging.

NMR imaging. Hearts were perfused within the vertical narrow bore of a Bruker 9.4 T/8.9-cm (gradient-enhanced) spectroscopy system. High resolution, spin-echo images were obtained using a standard imaging sequence modified to accommodate a hard 180-degree pulse and cardiac triggering. Images of a 1-mm slice were collected in sagittal and transaxial planes during end systole and end diastole with a repetition time of 350 ms, echo time of 9 ms, 90° flip angle, number of acquisitions of 2, matrix size of 256 × 256 and a field of view of 17 mm. Volumes were calculated using the volume calculation for a prolate ellipse with variables of length in the sagittal plane and minor and major axes in the transaxial view using National Institutes of Health Image software (SymanTec).

³¹P NMR spectroscopy. Hearts were perfused in the working mode in the vertical narrow-bore of a 14.1 T/5.1-cm Oxford magnet interfaced with a high resolution Varian Inova spectrometer equipped with a 10-mm broadband probe that can be tuned to ²³Na or ³¹P. The field homogeneity was optimized using the ²³Na resonance until a line width of 10-15 Hz was obtained. Partially saturated ³¹P NMR spectra were averaged over 20 min using a 45-degree pulse of 15 µs, an interpulse delay of 665 ms, a 12-kHz spectral width, and 4-K time domain data points. Resonance areas obtained under these conditions were corrected for T₁ relaxation by factors determined from a fully relaxed spectrum. For NMR determination of ATP concentration, peak areas, corrected for saturation effects, were compared with a known reference standard for calculation of absolute concentration. After NMR data collection, total myocardial ATP was also measured using standard biochemical methods described above to verify the precision of the NMR methodology. NMR spectra were analyzed using NUTS software (Acorn NMR; Palo Alto, CA). Free-induction decays were baseline corrected, exponentially multiplied with 60 Hz of line broadening, Fourier transformed, and phase corrected manually. Peak areas were calculated using the automated line-fit procedure.

Statistical analyses. The spectrometer operator was blinded to the genotypes of all animals throughout the spectral and imaging data analysis until the data were grouped for pooled statistical analysis. Data are reported as means ± SD unless otherwise noted. Resonance areas for metabolites were compared using the repeated-measures analysis of variance with Dunnett's test for comparison to control or with Tukey's test for multiple comparisons among all groups. Hemodynamic data are compared using paired t-tests or the repeated-measures analysis of variance with Tukey's test for multiple comparisons (for 3 or more sequential measures). All statistical analyses are performed using an automated software program (TexaSoft; Cedar Hill, TX). P < 0.05 is considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Generation of mice with a null mutation in COXVIaH. The murine COXVIaH gene was mutated by a replacement vector resulting in the deletion of two-thirds of exon 2 (Fig. 1A). Confirmation of a successfully targeted COXVIaH locus included an absence of full-length or truncated COXVIaH mRNA isolated from heart and skeletal muscle of the COXVIaH/ mouse (Fig. 1C). The presence of -actin mRNA, included as a control for RNA loading, demonstrated the integrity of the RNA in all the samples. Mice heterozygous or homozygous for the mutation demonstrated normal fecundity and a normal life span, and there was no evidence of embryonic loss of mutant mice (Table 1).

Lack of histochemical and ultrastructural changes in myocardial tissue from mutant mice. Sections of myocardial tissue stained with hematoxylin and eosin revealed no obvious structural differences between mutant and wild-type cardiac muscle. There were neither pericardial abnormalities nor increases in myocardial fibrosis in the mutant hearts (not shown). The ultrastructural evaluation of adult wild-type and mutant hearts revealed no qualitative differences in basic cellular ultrastructure including mitochondrial morphology, size, or number per cell (Fig. 2).

View larger version (112K): [in this window] [in a new window]   Fig. 2.   Ultrastructural assessment of VIaH+/+ and VIaH/adult cardiac ventricles. There are no structural abnormalities or differences noted in mitochondria content in the hearts of the VIaH/ mice (A) compared with the wild-type age- and sex-matched controls (B). Bar = 0.5 µm.

COX activity and subunit tissue content. The tissue contents of complexes I, III, and V (ATP synthase) in the knockout hearts were normal or slightly increased compared with wild-type levels (Fig. 3). However, the complex IV protein (COX) content (from densitometric quantification of two-dimensional gels) and COX activity was only 23% of wild-type levels. Direct protein sequence analysis of the subunit VIa-band in wild-type hearts revealed that this band was composed of two isoforms (about 80% VIaH and 20% VIaL, Fig. 4). However, the subunit VIa-band in the knockout mice was composed of exclusively the VIaL isoform. Because total complex IV from the knockout mice was reduced to 23%, the tissue content of complex IV comprising subunit VIaL seems to be unchanged in knockout and wild-type mice (i.e., there is no apparent compensation by induction of the biosynthesis of subunit VIaL in the mutant hearts). Protein sequence analysis of subunit VIII revealed that the heart isoform VIII-H was retained in the knockout mice and not replaced by VIII-L. A chimeric complex IV comprising subunits VIaL and VIII-H thus was generated by deletion of subunit VIaH. These results were further supported using a spectrophotometric enzymatic activity for COX in VIaH+/+ and VIaH/ hearts. COX enzymatic activity in the VIaH+/ heart was not significantly different compared with the wild-type heart.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   A: separation of native oxidative phosphorylation (OXPHOS) complexes from heart muscle by blue native-PAGE (BN-PAGE). Cardiac muscle (10 mg wet weight) from wild-type and COXVIaH knockout hearts were solubilized by dodecylmaltoside and separated by a linear 4-13% acrylamide gradient gel for BN-PAGE, as described previously (31). The native complexes bind Coomassie dye and are visible during the electrophoresis. The OXPHOS complexes from solubilized bovine heart mitochondria were used as molecular mass standards (Std). The apparent mass of complex IV from mice is ~50 kDa larger than that of the bovine complex. Complex IV from the VIaH mutant mice is minimally detectable indicating a considerable specific reduction of the complex IV protein amount. B: quantification of OXPHOS protein amounts and COX activity in the subunit VIaH mutant hearts. The protein amounts of OXPHOS complexes in the subunit VIaH mutant hearts were determined by densitometric analysis of the Coomassie staining intensities of subunits of complexes in two-dimensional gels (Fig. 4) and compared with the staining intensities of the wild-type subunits. Complex I, III, and V protein amounts were increased by 36, 51, and 45%, respectively, whereas complex IV protein amount and COX activity were reduced to 23%.

View larger version (58K): [in this window] [in a new window]   Fig. 4.   Two-dimensional resolution of the subunits of OXPHOS complexes. Lanes from the first dimension native gel (Fig. 3) were processed in a second dimension by Tricine-SDS-PAGE using 16% acrylamide gels containing 6 M urea as previously described (30). Comparison of the gels from (A) wild-type and (B) VIaH mutant hearts revealed a reduction of the complex IV protein amount in the mutant hearts and a concomitant loss of silver-staining of subunits I and III. *Protein sequencing of subunit VIa-bands revealed the presence of about 80% VIaH and 20% VIaL in the wild-type mice but exclusively VIaL in the mutant mice. Subunit VIII was retained as the heart form. The more intense silver-staining of subunit VIII in B, therefore, seems to be a silver-staining artifact. Other subunits of complex IV were assigned according to (31).

Myocardial performance. Isolated hearts from both VIaH+/+ and VIaH+/ mice consistently generated stroke work when perfused with an APP of 70 and LAP of 15 cmH₂O. The hemodynamic variables from hearts perfused under these standard conditions are presented in Table 2. Left ventricular performance curves were generated using the isolated working heart preparation in hearts from VIaH+/+ and VIaH+/ mice (n = 7 for each group; Fig. 5). Hearts from VIaH/ mice were unable to consistently generate stroke work in the isolated working heart under these conditions. Initial perfusion conditions required for the stabilization of these hearts were an APP of 50 and a LAP of 25 cmH₂O. Perfusion pressures were then adjusted slowly to an APP of 70 and a LAP of 15 cmH₂O, which VIaH/ hearts simply could not tolerate (n = 20). To evaluate the contribution of systolic dysfunction to the failure of the working heart preparation in the VIaH/ hearts, null mice underwent echocardiography, which revealed the presence of normal left ventricular systolic function [fractional shortening of wild-type controls = 0.57 ± 0.02 (n = 3); fractional shortening of COXVIaH mutants = 0.56 ± 0.03 (n = 3)]. The failure of stroke work to increase with increasing LAP in the VIaH+/ hearts was a consequence of either impaired systolic function (inability of the heart to eject increasing ventricular filling volume) or impaired ventricular filling (inability of the heart to increase ventricular volume). To distinguish between these two possibilities, systolic function and ventricular filling (diastolic function) were assessed in isolated working heart preparation using magnetic resonance imaging in hearts from VIaH+/+ and VIaH+/ mice. As a result of the high failure rate of the COXVIaH/ working heart perfusion, this group was omitted from these technically challenging imaging studies. Figure 6, A and B, shows typical gated magnetic resonance images of an isolated working mouse heart in transaxial views at end diastole and end systole. Ejection fraction determined from left ventricular end-diastolic and end-systolic volumes under basal conditions was not significantly different in hearts from VIaH+/+ and VIaH+/ mice (51 ± 1 vs. 51 ± 2%, respectively, n = 5) suggesting preserved systolic function in the COXVIaH+/ heart. To assess ventricular filling, left ventricular end-diastolic volumes were measured at baseline and after an increase in LAP. Under basal conditions, there were no significant differences in mean arterial pressure, cardiac output, stroke volume, heart rate, or left ventricular end-diastolic volumes in VIaH+/+ and VIaH+/ hearts (Table 2). Figure 6C reveals the relationship between left ventricular end-diastolic volume and LAP in hearts from VIaH+/+ and VIaH+/ mice (n = 7 in each group). Whereas left ventricular end-diastolic volume increased significantly with increasing LAP in hearts from VIaH+/+ mice, there was no significant change in left ventricular end-diastolic volume in hearts from VIaH+/ mice, suggesting the presence of impaired ventricular filling. As in the previous studies, stroke work increased with increasing LAP in hearts from VIaH+/+ mice, whereas it remained constant in hearts from VIaH+/ mice.

View this table: [in this window] [in a new window]   Table 2.   Hemodynamic data from working heart perfusions from VIaH+/+ and VIaH+/ mice

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Left ventricular performance curves in the isolated working heart. Stroke work (means ± SD) was calculated in hearts from VIaH+/+ (denoted by ) and VIaH +/ (denoted by ) mice as left atrial (LA) pressure was increased from 15 to 20 to 25 cmH2O. Stroke work did not significantly increase in hearts from mutant mice, whereas it increased significantly (*P < 0.05, 15 vs. 25 cmH2O) in the wild-type hearts.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   A: typical gated magnetic resonance images of the isolated working mouse heart in diastole (left) and systole (right). B: left ventricular end-diastolic volume (LVEDV, means ± SD) was determined from magnetic resonance images during diastole as LA pressure was increased from 15 to 20 to 25 cmH2O in hearts from VIaH+/+ (denoted by ) and VIaH +/ (denoted by ) mice. LVEDV did not significantly increase in hearts from mutant mice, whereas it increased significantly (*P < 0.05, 15 vs. 25 cmH2O) in the wild-type hearts.

Energy metabolite levels in myocardial tissue. The concentration of ATP was measured using both standard biochemical methods and ³¹P NMR spectroscopy with a reference standard. In Table 3, ATP concentrations are listed for wild-type, heterozygote, and homozygote mutant hearts. There are no differences in ATP concentration in the three groups measured by either method. Although ATP concentration is lower in work-performing hearts (measured by NMR and confirmed by standard biochemical methods), there were no differences among the three genotypes.

View this table: [in this window] [in a new window]   Table 3.   Myocardial ATP concentration from VIaH+/+, VIaH+/, and VIaH/ mice

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of COXVIaH deficiency on cardiac function. To examine the functional significance of the cytochrome-c oxidase subunit VIaH on oxidative energy production and myocardial performance, we generated COXVIaH-deficient mice. The predominant cardiac phenotype resulting from gene disruption of COXVIaH is impaired left ventricular filling or diastolic dysfunction under maximal cardiac load. Abnormal left ventricular diastolic function in the absence of systolic dysfunction is the etiology of congestive heart failure in as many as one-third of patients who carry this diagnosis (8, 34). Although diastolic dysfunction is a common clinical diagnosis, simple animal models (i.e., diastolic dysfunction in the absence of dramatic systemic confounders such as hypertension, diabetes, or ischemia) that can be used to study metabolism and mechanical function are uncommon.

Cellular mechanisms linking genotype to phenotype. A number of different etiologies for diastolic dysfunction have been described. Increased resistance to filling of ventricular chambers may be due to structural abnormalities such as constrictive pericarditis or myocardial fibrosis. Numerous cellular or metabolic causes of diastolic dysfunction have also been proposed. Impaired signaling or function of the sarcoplasmic reticulum ATP-dependent Ca²⁺ pump leading to abnormal intracellular Ca²⁺ homeostasis has been implicated in the setting of ischemia, hypertrophic cardiomyopathy, and chronic heart failure where systolic dysfunction is commonly present, as well (13, 25). Impaired restoration of diastolic Na⁺ levels has been implicated in postischemic hypertrophied hearts (23). Inhibition of glycolysis may cause diastolic dysfunction through a number of mechanisms: loss of glycolytically derived ATP, which is used preferentially for cation homeostasis (6, 28, 42); alterations in intracellular pH and P_i accumulation, which may influence the sensitivity of the myofilament to Ca²⁺ (15); and accumulation of sugar phosphates, which may have toxic effects on intracellular Ca²⁺ homeostasis (20). Given that nitric oxide has been shown to enhance left ventricular relaxation under normal and stressed conditions, from endogenous or exogenous sources respectively, inhibition of nitric oxide may also be linked to diastolic dysfunction via a cGMP-induced reduction in myofilament response to Ca²⁺ (10, 32).

Role of VIaH in COX function. Finally, what insight does this transgenic model provide regarding the role of subunit VIaH in modulating COX activity? Previous reports indicated that VIaH regulates COX activity via binding of ADP or ATP (2, 29). The present results using VIaH/ mice indicate that the deletion of subunit VIaH also has dramatic effects on assembly/stability of COX so that COX activity is reduced considerably due to low COX protein amounts. Additional experiments are in progress to further define the effects of this mutation on the assembly and stability of COX enzyme. We previously observed that the turnover number of COX increased from 120 s¹ to 200 s¹ and the K_m for cytochrome c dropped from 38 µM to 25 µM when the liver isoforms (VIaL, VIIIL) were replaced by the heart isoforms (VIaH, VIII-H) in the developing rat heart (31). Because VIa and VIII were exchanged on similar time scales, we could not discern whether subunit VIa or VIII, individually or collectively, were responsible for the altered enzymatic properties.