Physiological Genomics


Despite identical cardiac outputs, the right (RV) and left ventricle (LV) have very different embryological origins and resting workload. These differences suggest that the ventricles have different protein programming with regard to energy metabolism and contractile elements. The objective of this study was to determine the relative RV and LV protein expression levels, with an emphasis on energy metabolism. The RV and LV protein contents of the rabbit and porcine heart were determined with quantitative gel electrophoresis (2D-DIGE), mass spectrometry, and optical spectroscopy techniques. Surprisingly, the expression levels for more than 600 RV and LV proteins detected were similar. This included proteins many different compartments and metabolic pathways. In addition, no isoelectric shifts were detected in 2D-DIGE consistent with no differential posttranslational modifications in these proteins. Analysis of the RV and LV metabolic response to work revealed that the metabolic rate increases much faster with workload in the RV compared with LV. This implies that the generally lower metabolic stress of the RV actually approaches LV metabolic stress at maximum workloads. Thus, identical levels of energy conversion and mechanical elements in the RV and LV may be driven by the performance requirements at maximum workloads. In summary, the ventricles of the heart manage the differences in overall workload by modifying the amounts of cytosol, not its composition. The constant myocyte composition in the LV and RV implies that the ratio of energy metabolism and contractile elements may be optimal for the sustained cardiac contractile activity in the mammalian heart.

  • left ventricle
  • energy metabolism
  • proteomics
  • myocardial biochemistry and function

the right ventricle (RV) and left ventricle (LV) of the mammalian heart work in series to deliver oxygenated blood to the body. The workload of the heart can be determined by the product of afterload pressure and cardiac output (stroke volume·heart rate), with heart rate being the major determinant (5, 8). Although the RV and LV have identical cardiac outputs, their afterload pressures differ significantly. Afterload pressure in the RV is determined by the pulmonary vasculature (∼25 mmHg), whereas the LV is dominated by peripheral vasculature (∼120 mmHg). Thus, the workload ratio of the left to right heart approaches 5, as shown by previous studies (17, 21). Ventricular mass is an important normalization factor when comparing LV and RV parameters. The reported ratios of LV to RV mass include 3.4 (17) and 2.6 (21), which are not proportional to the fivefold difference in work. This discrepancy suggests that in addition to workload, there are other underlying differences between the ventricles that determine the ratio of LV/RV mass, including metabolic capacities and contractile properties. Consistent with this notion is the observation that oxygen consumption in the “resting” right heart is less than the left (15, 32), implying that the metabolic stress (ATP generation rate/maximum ATP generation rate) (9) of the RV is lower, based on a wet weight comparison. Additionally, given the different determinants of LV and RV afterload pressures, these two pumps respond very differently to systemic perturbations. One example involves alpha agonists, which decrease RV inotropy but increase inotropy of the LV (29). Additionally, oxidative capacity in the LV is apparently more sensitive to hypoxia than the RV (20). Mechanically, the RV is more compliant than the LV and does not rely on twisting to eject blood, which results in a more linear contraction pattern. Myocytes of the LV and RV also have different embryological origins that are regulated by different genetic pathways (31), which may also lead to differences in the differentiated cell types. Other differences between the LV and RV related to muscle structure and contraction, oxygen supply reserve, and the regulation of blood flow have been summarized in recent reviews (11, 32).

Given the noted differences in the left and right mammalian heart, the purpose of this study was to determine the relative protein expression levels and assess the posttranslational modification status of the LV and RV, with an emphasis on the enzymes associated with energy metabolism. This was performed using quantitative proteomic approaches, including difference in-gel electrophoresis, iTRAQ (isobaric tags for relative and absolute quantitation) quantitative protein labeling, coupled to mass spectrometry (MS), and optical spectroscopy. The molecular activity of select enzymes was also surveyed. These studies reveal that the metabolic protein programming and structural elements of the LV and RV are effectively identical, and provides insight into the relative sensitivity of these chambers in response to chronic metabolic challenges.


Collection of tissues.

All experiments were conducted in accordance with protocols approved by the NHLBI Animal Care and Use Committee. Hearts were harvested from anesthetized rabbits (2–4 kg, male and female) and pigs (30–40 kg, male and female) and transiently perfused with ice-cold saline to remove blood and excess free calcium. Each animal served as its own control, comparisons were only made between the left and right ventricles of a given animal, not between animals or species.

Histology preparation and staining.

Hearts were harvested from anesthetized rabbits, transiently perfused with ice-cold saline, and then immediately perfused, fixed with 10% buffered formalin, and embedded in paraffin. Whole rabbit hearts were serial sectioned at a thickness of 4 μm. The sections were stained with hematoxylin and eosin or Masson tri-chrome stains. Histology images were taken by using Leica macro- (M420) and microscopes (DMRXA) with a DC500 digital camera.

Tissue homogenization.

Fresh heart tissue from the LV and RV free walls was dissected of all fat and connective tissue on ice. A sample of heart muscle (∼500 mg) was placed in an equal weight of cold 0.1 M potassium-phosphate buffer, at pH 7.0. The tissue was then minced with scissors on ice. This mixture was homogenized using to 5 s treatments at maximum speeds with a tissue homogenizer (Virtis, Gardiner, NY), with the tissue on ice.

Cytochrome a,a3 and protein content.

Whole tissue cytochrome a,a3 content was determined spectrophotometrically using Triton X-100 solubilized tissue homogenates and cyanide as previously described (3). Approximately 0.5 ml samples of the homogenate (described above) were weighted and then dissolved in 2.5 ml of 2% (vol/vol) Triton X-100 in 0.1 M potassium-phosphate buffer, at pH 7.0. After mixing, the suspension was centrifuged to remove any residual connective tissue and solid material. The difference in absorbance at 605 nm between the oxidized and cyanide forms of the enzyme, corrected for baselines, were measured to determine cytochrome a,a3 content in the presence of myoglobin and hemoglobin. Total protein content was determined using the Bradford method (USB, Cleveland, OH), with bovine serum albumin as a protein reference.

Two-dimensional differential electrophoresis.

To compare the proteomes of the porcine and rabbit LV and RV, two-dimensional differential in-gel electrophoresis (2D-DIGE) was performed. LV and RV samples, suspended in DIGE lysis buffer [7 M urea, 2 M thiourea, 4% CHAPS (wt/vol), 15 mM Tris·HCl, pH 8.5], were labeled with CyDyes, separated by 2D gel electrophoresis, imaged, analyzed, and identified as previously described (14) (19). Specifically, 50 μg of LV and 50 μg of RV were labeled with Cy3 and Cy5 dyes, respectively. Each sample (25 μg) was mixed together and labeled with Cy2 to serve as an internal reference. After 30 min of labeling, the CyDye reaction was quenched with 10 mM lysine for 15 min before combining the three samples. Proteins were loaded onto 13 cm Immobiline DryStrip gels (pH 3–11 NL), and isoelectric focusing (Ettan IPGphor2 apparatus) was achieved by active rehydration at 30 V for 12 h, followed by stepwise application of 250, 500, 1,000, and 6,000 V for a total of 15,000 V-h.

Immobiline DryStrip gels were then equilibrated in 5 ml of SDS equilibration solution [50 mM Tris·HCl (pH 8.8), 6 M urea, 30% glycerol, and 2% SDS] for 10 min, first containing 100 mg of DTT and then 250 mg of iodoacetemide. Gel strips were applied to 8–16% Tris·HCl gels (Bio-Rad Laboratories, Hercules, CA), and electrophoresis was performed in a Criterion Cell (Bio-Rad), with SDS electrophoresis buffer [25 mM Tris·HCl (pH 8.3), 192 mM glycine, and 0.2% SDS] for ∼210 V-h. Gels were subsequently imaged using a Typhoon 9400 scanner at 100 μm resolution. After scanning, gels were stained with Coomassie blue and proteins were subsequently picked using an Ettan Spot Handling Workstation.

Protein identification from the gels was carried out using a MALDI-TOF/TOF instrument (4700 Proteomics Analyzer; Applied Biosystems, Foster City, CA) with MS/MS peptide analysis. At least two peptides were obtained for each protein using MS/MS, for which the spectra were searched against the NCBI protein database, using the MASCOT search algorithm (Matrix Science, Boston, MA). A positive protein identification was indicated by >95% confidence in the MASCOT algorithm.

Identical total protein of LV and RV were added to the gels, with the caveat that fat and connective tissue on the endocardium and epicardium were removed by blunt dissection.

iTRAQ processing.

We iTRAQ-labeled 100 μg of TCA-acetone cleaned homogenate according to manufacturer's instructions (Invitrogen, Carlsbad, CA). Isobaric tagging was performed from 113 to 116 for the RV samples and 117–121 for the LV samples. All iTRAQ data are presented as the ratio of RV proteins to LV proteins. After the labeling reaction was quenched, the resulting peptide mixtures were combined and dried until the final volume of 100 μl was achieved. The combined peptide digest was resuspended in 900 μl of 0.1% FA and desalted using Waters Oasis HLB 1 cm3 cartridges (Milford, MA) per the manufacturer's instructions using acetonitrile instead of methanol. Eluent was dried and resuspended with 100 μl strong cation exchange (SCX) buffer A (10 mM KH2PO4/25% acetonitrile, pH 3.0).

SCX chromatography.

An Agilent 1200 HPLC system (Agilent Technologies, Santa Clara, CA) was used to separate the peptides by their charge using a polysulfoethyl A column; 200 × 2.1 mm, particle size 5 μm, 200 Å (Poly LC, Columbia, MD). A 60 min linear ramp from 0%-40% buffer B (10 mM KH2PO4/500 mM KCl/25% ACN, pH 2.7) was used to separate the peptides. Column temperature was maintained at room temperature, and a flow rate of 200 μl/min throughout the run. Fractions were collected at 1 min intervals on a 96-well microtiter plate for a total of 60 fractions. The chromatographic peaks were monitored using the built in UV detector (214 nm) and fractions were combined for a final count of 23 fractions. Each fraction was desalted using Waters Oasis HLB 1 cm3 cartridges per the manufacturer's instructions using acetonitrile instead of methanol.

MS liquid chromatography-tandem mass spectrometry: Orbitrap Velos.

Liquid chromatography-tandem mass spectrometry (LC MS/MS) was performed using an Eksigent nanoLC-Ultra 1D plus system (Dublin, CA), coupled to an LTQ Orbitrap Velos mass spectrometer (Thermo Fisher Scientific, San Jose, CA). We loaded 10 μl of the peptide digest onto an Zorbax 300SB-C18 trap column (Agilent, Palo Alto, CA) at a flow rate of 6 μl/min for 6 min and then separated it on a reversed-phase PicoFrit analytical column (New Objective, Woburn, MA) using a 40 min linear gradient of 5–40% acetonitrile in 0.1% formic acid, at a flow rate of 250 nl/min. LTQ-Orbitrap Velos settings were as follows: spray voltage 1.5 kV; full MS mass range: m/z 300–2,000, operated in positive ion mode with data-dependent acquisition. A single full-scan MS in the Orbitrap (30,000 resolution, 300–2,000 m/z) was followed by six data-dependent MS2 scans for precursor ions above a threshold ion count of 10,000, using the HCD cell with the resolution set to 7,500 and 45% normal collision energy.

MS data analysis.

The raw files generated were analyzed for protein identification and quantification using Proteome Discoverer v1.2 software (Thermo Fisher Scientific, San Jose, CA) using Mascot search engine v.2.3. This resource was managed and weekly updated protein databases by the Center for Information Technology at the National Institutes of Health. The MS/MS spectra were searched against the Swiss-Prot Protein Knowledgebase (Sprot) combined pig, bovine, and human database (release 2011_06). The combined pig, bovine, and human database was used because the pig protein database is partially annotated. The enzyme parameters were limited to trypsin with a maximum mis-cleavages set to 2. The selection criteria: variable modifications; oxidation (M), deamidation (NQ), iTRAQ8plex (Y); fixed modifications, (MMTS) methyl methanethiosulfonate (C), NH2-terminal iTRAQ8plex, iTRAQ8plex (K); 10 ppm for precursor ions and 0.05 Da for fragment ions. The quantitative measurements were performed in the MS/MS scan using the iTRAQ reporter ions. The experimental bias was corrected by using the generated normalization factor (Proteome Discoverer v1.2) based on the protein median for each sample group. The automatic decoy database search option was selected on Mascot, and the high confidence peptides were only used for protein identification that contained a false discovery rate (FDR) <1% at the peptide level.

Briefly, every time a peptide sequence search is performed on a target database a random sequence of equal length is automatically generated and tested. The statistics for matches are calculated, and a peptide significance is generated; an in depth explanation can be found at Protein identifications with one or more unique peptides were selected for quantitative analysis. A single unique peptide hit was accepted if it passed the following criteria: expectation value (E-value) <1.7 × 10−2, which is a measure of the number of matches with scores equal to or better than the score values that are expected to occur only by chance. In addition, Mascot ion score is equal to or better than the identity high, which is a threshold that determines whether peptides are ranked as high confidence when you perform a decoy search to calculate the FDR (0.01). All mass spectra used in this study are publicly available at the Proteome Commons Tranche website (

The complete results from the iTRAQ analysis of the pig and rabbit heart can be found in Supplemental Tables S1 and S3, respectively.1

Aconitase and manganese superoxide dismutase activity assays.

Aconitase activity was determined by homogenizing fresh porcine heart tissue in buffer A [(in mM) 125 KCl, 20 HEPES, 15 NaCl, 5 MgCl2, 1 K2EDTA, 1 EGTA, 2 NaPO4, and 0.1 K-malate, pH 7.1] on ice. After centrifugation, the pellet was resuspended and sonicated in aconitase reaction buffer (50 mM Tris·HCl, pH 7.4, 5 mM sodium citrate, 0.6 mM MnCl2, 0.2 mM NADP+, and 1–2 units of isocitrate dehydrogenase). Total aconitase activity (cytosolic + mitochondrial) was measured by following the linear absorbance change at 340 nm at room temperature for 1 h. The linear potion of the curve was used for analysis, where 1 mU of aconitase activity was defined as the amount catalyzing the formation of 1 nmol of isocitrate per min. The activity of manganese superoxide dismutase (Mn-SOD) was measured spectroscopically using a commercially available assay kit (Trevigen, Gaithersburg, MD) on tissue protein homogenates. Mn-SOD, the mitochondrial form of the enzyme, was specifically assayed by using potassium-cyanide. Superoxide anions generated by the conversion of xanthine to uric acid and hydrogen peroxide by xanthine oxidase in turn convert NBT to NBT-diformazan, which absorbs light at 550 nm.


Morphology/protein content.

In preliminary studies we found that using simple, crude homogenates of RV contained large amounts of contaminating fat and connective tissue that interfered with protein identification. This was seen in our initial 2D-DIGE gels as well as our early iTRAQ trials. Thus, we carefully removed the epicardial fat and the visceral pericardium as well as the extensive endocardial connective tissue before analysis. With this treatment reproducible cellular proteins were extracted from the RV free wall.

To confirm this observation we performed histological studies on the LV and RV of the rabbit heart. Although this issue was most apparent in porcine heart, we performed histology studies on the rabbit so that the entire free wall could be observed in a single section. These results are presented in Fig. 1. The RV had a much larger fraction of connective tissue and fat compared with the LV in both rabbit and porcine (not shown) hearts. This supported our observations that, without careful removal of the fat and surface connective tissues, this noncardiomyocyte content could significantly contribute to the overall protein content of the RV and alter the direct comparison of left and right heart proteins as well as dry and wet weight values. The noncardiomyocyte content of the RV has important ramifications for comparing enzyme activities of the LV and RV as a function of weight.

Fig. 1.

Histological comparison of the left (A) and right ventricle (B) of the rabbit heart. Gross sections of the left (LV) and right (RV) ventricles with expanded regions from the epicardium and endocardium where the largest amounts of connective tissue and fat are present. The scale bars for LV and RV are 2 mm, and the scale bars for all the insets are 200 μm. Details of the staining procedures are presented in materials and methods.

Protein and cytochrome oxidase content.

Cytochrome a,a3 is a component of cytochrome oxidase. It reflects to the total capacity for respiratory activity since it is the enzyme that reduces oxygen to water in the oxidative phosphorylation reaction. The absolute content of cytochrome a,a3 can be directly determined in tissue extracts (3) and is useful in providing information on the functional mitochondrial content of the tissue. The concentration of cytochrome oxidase (complex IV) was not significantly different between the LV and RV in the rabbit and pig. In the rabbit the RV and LV contained 35.1 ± 0.6 vs. 33.9 ± 0.5 (n = 3) (nmol/g wet wt), while the pig RV and LV contained 31.7 ± 0.5 vs. 30.4 ± 0.3 (n = 6) (nmol/g wet wt). There was statistically more cytochrome oxidase/gram wet weight in the rabbit ventricles than the pig ventricles (P < 0.05).

In the current studies we used total protein content as the normalizing factor. Since previous physiological data generally used wet weight, the mg protein/g wet wt conversion factors were determined to aid comparisons. Removing the connective tissue and fat, the mg protein/g wet wt values were, for pig: LV 150 ± 4, RV 153 ± 8, and for rabbit: LV 156 ± 1, RV 178 ± 5. The effect of trimming on the content of noncardiomyocyte protein was directly assessed by measuring the ratio of cytochrome a,a3 content to total protein before and after trimming (Fig. 1C). The nmoles cytochrome a,a3/mg protein in the porcine LV was 0.20 ± 0.01 in untrimmed and 0.20 ± 0.01 in trimmed tissue (n = 3), suggesting that a majority of the total LV mass is composed of cardiomyocytes. However, in porcine RV the nmol cytochrome a,a3/mg protein was 0.13 ± 0.02 in untrimmed and 0.20 ± 0.00 in trimmed tissue. Since the cardiomyocyte content (i.e., cytochrome a,a3/mg protein) in RV and LV are nearly identical after trimming, these data suggest that a significant fraction (∼1/3) of the gross RV wet weight is composed of noncardiomyocyte connective tissue and fat. Therefore, unless noncardiomyocyte components are properly dissected, comparison of LV and RV parameters, where tissue weight or protein is used as the normalization factor, is likely to be problematic for the RV.

Two-dimensional differential electrophoresis.

A representative example from four 2D-DIGE studies comparing the pig or rabbit LV and RV is presented in Fig. 2, A and B, respectively. Surprisingly, the gels are almost universally yellow, indicating that the relative amounts of protein in the LV (red dye) and RV (green dye) are nearly identical in both species. Major proteins were identified by MS as previously described (2) and indicate that the relative concentration of structural proteins (i.e., actin and myosins) as well as proteins associated with energy metabolism (i.e., oxidative phosphorylation enzymes, aconitase, Mn-SOD, myoglobin, etc.) are essentially the same for both ventricles. It is also notable that no significant isoelectric shifts are detectable for the proteins observed in these DIGE studies. These data suggest that the RV and LV protein detected are in a similar state of posttranslational modifications, at least for those modifications that generate significant isoelectric shifts. For examples of physiological conditions where significant posttranslational modifications are observed, see Refs. 2, 22. To expand on the number of proteins detected and provide a more quantitative analysis of the relative concentration of proteins in the RV and LV, an iTRAQ analysis was conducted.

Fig. 2.

Two-dimension differential in-gel electrophoresis (2D-DIGE) images of the RV and LV proteins from the pig (A) and rabbit (B). RV protein was labeled with the green probe (Cy3), while the LV was labeled with red label (Cy5). A yellow color indicates a near identical protein level. These images represent LV and RV protein taken from a single pig and a single rabbit.


iTRAQ studies focused on porcine hearts, where the RV and LV from four animals were examined. To confirm these results, two rabbit hearts were also studied. Supplemental Table S1 provides a complete list of the proteins detected in the porcine study, including proteins that did not make the standard deviation criteria discussed below. The four RV and LV pairs were labeled and run via MS simultaneously, using different molecular weight markers. Inclusion criteria for analysis in this study included proteins for which at least 1 unique peptide was detected across all four experiments. Additionally, proteins for which the standard deviation of the four composite RV to LV ratios exceeded 0.4 were also excluded.

The iTRAQ studies quantified the relative concentration of more than 600 RV and LV proteins from both the porcine and rabbit heart. In general, the iTRAQ data were in excellent agreement with the 2D-DIGE studies and the optical spectroscopy data (specific for complex IV). As illustrated by the histogram in Fig. 3A for porcine heart and Fig. 3B for rabbit heart, few proteins were detected with RV to LV ratios that exceeded ±20%. These iTRAQ studies reveal that a vast majority of cardiac proteins have nearly equal expression levels in both ventricles. This applies to proteins across various cellular compartments, including proteins in the cytosol, nucleus, sarcolemma, sarcoplasmic reticulum, and mitochondria.

Fig. 3.

Histogram of the RV/LV ratio of proteins detected in iTRAQ analysis. These histograms are composed of the means porcine heart proteins (n = 4, A) and rabbit heart proteins (n = 2, B) that met our mass spectrometry protein inclusion criteria.

Table 1 presents the relative expression level of RV and LV proteins from several categories, including contraction and cell structure, oxidative phosphorylation and oxygen metabolism, intermediary metabolism, and calcium signaling and regulation. Across these categories, there was again excellent correlation between the iTRAQ data and 2D-DIGE studies (see Supplemental Table S4 for quantitative comparison). This was specifically noted for subunits of myosin, tropomyosin, subunits of complexes I, III, IV and V; SOD; myoglobin; aconitase; and creatine kinase. The iTRAQ studies and optical spectroscopy data also revealed nearly identical relative contents of complex IV between RV and LV.

View this table:
Table 1.

Selected proteins from porcine LV and RV free walls via iTRAQ analysis (n = 4)

Consistent with the lack of isoelectric shifts and thus posttranslational modifications observed between the RV and LV in the 2D-DIGE studies, no expression differences were noted for the protein kinases and phosphatases detected via iTRAQ (Supplemental Table S1). Examples of these regulatory enzymes include: cAMP-dependent protein kinase types I and II, calcium/calmodulin-dependent protein kinase, integrin-linked protein kinase, glycogen synthetase kinase, and serine/threonine-protein phosphatase 2A.

Although the vast majority of RV and LV proteins were expressed within 20%, a few proteins exceeded this threshold. These proteins are presented in Table 2. We were unable to ascribe particular functional outcomes to these protein differences, largely since the expression differences were mild and because no common biochemical pathway was differentially expressed. These proteins are presented to illustrate the variability in expression levels of RV and LV proteins, even though the physiological significance of these alterations is uncertain.

View this table:
Table 2.

Under- and overexpressed proteins in the porcine RV relative to the LV (>20% threshold)

The more limited rabbit study revealed a similar pattern of protein expression as observed in the pig, as shown in Fig. 3A. As observed in the porcine heart, nearly all of the proteins identified in the 2D-DIGE as being similar in the rabbit RV and LV were found to be within 20% via iTRAQ analysis, including [RV/LV means (n = 2)]: cytochrome oxidase subunit 5A (0.92), Mn-SOD (0.92), aconitase hydratase (0.94), myoglobin (0.94), creatine kinase M-type (1.00), and myosin light chain 2 (1.1). While the majority of rabbit RV and LV proteins were within ± 20%, a few proteins exceeded this threshold (see Supplemental Table S2). Interestingly, no correlation was found between the RV proteins under- or overexpressed in the pig and rabbit. However, it was noted that in the rabbit heart the major overexpressed proteins in the RV include proteoglycans [RV/LV means (n = 2)]: mimecan (1.60), lumican (1.59), and decorin (1.47). These data support our finding that dissection of connective tissue and fat from the RV is more difficult in the rabbit heart than the porcine heart. Ultimately, this more limited rabbit iTRAQ study confirms our observation in porcine heart that a high homogeneity of protein expression and posttranslational modification state exists in the mammalian RV and LV. The complete rabbit iTRAQ data set can be found in Supplemental Table S3.

Enzymatic activity.

In vitro activity assays were conducted for porcine heart aconitase [RV/LV 0.94 ± 0.16 (n = 3)] and Mn-SOD [1.0 ± 0.1 (n = 3)] as addition markers of mitochondrial activity not related to oxidative phosphorylation. Neither assay revealed a significant difference in the activity of these enzymes between porcine RV and LV. As the assays agree with the protein content measurements described above, these results further suggest that no posttranslational modifications exist that differentially alter the enzymatic activity of aconitase (intermediary metabolism) or Mn-SOD (oxidative metabolism) between the RV and LV.

It is important to emphasize that these activity measurements are normalized as a function of cardiomyocyte protein content, not gram wet weight. Previous studies have compared activities between enzymes in the RV and LV. In studies where activity is normalized per gram wet weight, enzyme activities in the LV are generally found to exceed those in the RV (18, 20), presumably due to the inclusion of connective tissue and fat from the RV. However, consistent with our results, when activity was normalized to cardiomyocyte protein concentration (i.e., mg protein/g tissue were equal in the RV and LV), no difference in metabolic rates was noted between the ventricles (6, 7, 26).


Many differences exist between the RV and LV with regards to embryological origin (31), afterload pressure development, overall workload, gross contractile measurement (11), coronary blood flow (8), oxygen consumption (i.e., metabolic rates/gram wet weight) (15), and systemic signaling (29). To add insight into the underlying mechanism of these differences, the current study used a variety of techniques to determine the relative protein content and assess protein modifications in the RV and LV of porcine and rabbit hearts. Surprisingly, these data revealed that the translational programming of cellular proteins by nuclear and mitochondrial DNA is essentially the same between the ventricles. A recently published quantitative proteomic study on the mouse heart (22a) supports this observation. These results imply that the metabolic and mechanical potential of RV and LV cardiomyocyte cytosol is nearly identical. In other words, the differences in net workload and metabolic needs between the ventricles appear to be matched by the amount of cytosol, or cardiac muscle, implying that the cell uses a fixed, and possibly optimized, ratio of contractile proteins (actin, myosin, troponin, etc.) and elements of energy conversion (i.e., mitochondrial and cytosolic metabolism).

The five complexes of oxidative phosphorylation were essentially identical in the RV and LV, implying that the cellular aerobic capacity is matched between ventricles. These data are consistent with several morphology studies on the density of cardiac mitochondria in the rat (1), dog (16), sheep (24), cat (10), and pig (22 vol% LV and 21 vol% RV) (23), which show that the relative cytosolic volume of mitochondria are identical across species. The enzymatic activities of aconitase and Mn-SOD were the same in porcine RV and LV, when normalized by cardiomyocyte content, consistent with previous measures of metabolic enzyme activities between the ventricles (6, 7, 26). Taken together these results indicate that the processes of energy metabolism within the right and left heart are essentially identical and that the maximum capacity for metabolic flux is the same per gram of cardiac tissue.

What is the relative metabolic stress of the LV and RN? Assuming the maximum rate of ATP production in the LV and RV is proportional to the mitochondrial content. The observation that the LV and RV have the same mitochondria content/gram suggests that the metabolic stress can simply be determined by comparing the rate of respiration/gram of cardiomyocyte in the two ventricles. The measurement of coronary blood flow and oxygen extraction from the right heart, which is required to determine the respiratory rate, is a difficult task. Primary studies include Kusachi et al. (15) in open-chest dogs and the remarkable measures by Bian et al. (4) and Hart et al. (12) in conscious dogs. Kusachi et al. (15) is the only study to measure oxygen consumption in both ventricles; this study showed large differences in myocardial oxygen consumption (MV̇o2) between the RV and LV, when normalized to wet weight. Over a wide range of conditions, including control, pacing, isoproterenol and afterload challenges, the ratio of LV to RV MV̇o2 was relatively uniform at 2.0 ± 0.3 for seven different conditions. Although Hart et al. (12) measured the RV MV̇o2 in exercising dogs, no parallel data were collected for the LV. Since heart rate is the major determinant of cardiac workload, we combined the RV data from Hart et al. (12) with the LV study from Tune et al. (25), and normalized to heart rate (Fig. 4A). Additionally, we included the 33% RV mass adjustment for cardiomyocyte content, as determined above from our porcine studies, which adds a third data set and brings the RV and LV data closer together. Despite this mass correction, the metabolic rate of the RV remains significantly lower than the LV. Since oxygen extraction of the RV per wet weight is lower than the LV (32), the oxygen delivery is presumably adequate for oxidative phosphorylation and is not rate limiting in the RV. Taking into account the identical metabolic capacity of the RV and LV cytosol, these data demonstrate that the RV has a lower metabolic stress with a larger metabolic reserve capacity than the LV. That is, the right heart is operating further below its metabolic, workload and coronary flow (32) capacities compared with the left heart.

Fig. 4.

Influence of heart rate and rate pressure product (RPP) on RV and LV myocardial oxygen consumption (MV̇o2) in the close-chest dog. A: effect of heart rate on RV and LV MV̇o2. ▲, LV data from Tune et al. (25). ■, RV data from Hart et al. (12). □, RV data correcting for the connective tissue (33%) determined in this study on the pig RV. Errors are SE following the convention of Hart et al. B: effect of RPP on RV and LV MV̇o2. Symbols are the same as in A. RPP was calculated using mean arterial blood pressure for the LV and pulmonary peak pressure in the RV based on available data. Numbers next to the lines represent the slope these lines. Complete linear fits: RV (not corrected for connective tissue) MV̇o2 = −1.7 + (0.002144 × RPP), RV (corrected for connective tissue) MV̇o2 = −1.3 + (0.002859 × RPP), LV MV̇o2 = −4.0 + (0.001203 × RPP).

On first consideration this notion seems to oppose the general features of symmorphosis (30), where the metabolic capacities are matched to demands. However, heart rate is only one of the terms used to calculate the total workload of a ventricle. Since the stroke volume of the RV and LV is identical, the rate pressure product (RPP) (heart rate·systolic blood pressure) is a reasonable measure of relative external work of the ventricles. A graph of the RV and LV MV̇o2 as a function of their respective RPPs (where pressure is peak pulmonary and mean arterial) is presented in Fig. 4B. This graph reveals that the slope of the RV MV̇o2 vs. RV RPP exceeds the corresponding LV plot by a factor of ∼2.4, which is effectively identical to the LV/RV mass difference(21). The relatively diminutive mass of the RV likely causes the steeper RPP-MV̇o2 relationship than the massive LV. This difference in RV and LV slope of this plot suggests that near maximum workloads, the RV and LV metabolic rates/gram tissue will approach each other. Furthermore, the higher dependence of metabolic rate on RPP in the RV also requires that the RV has a larger contractile and energy conversion reserve capacity at low workloads as calculated above. Regrettably, no RV RPP or MV̇o2 data are available at LV RPP approaching 35,000 (mean arterial pressure), which is believed to be near the maximum performance of the dog (27, 28). The large difference in RV and LV MV̇o2 detected by Kusachi et al. (15) in open-chest animals is likely due to the fact that the LV RPP values generated in that study were well below 16,000 (or <50% of the maximum values), likely resulting in a large RV reserve capacity compared with the LV.

Taking into account the different response function of the RV and LV metabolic rate to workload (Fig. 4B), it is likely that the peak work requirements/gram tissue are similar between ventricles, which may account for their near identical protein programming (Fig. 2 and Table 1). Thus, the concept of symmorphosis, discussed above, may still be obeyed when considering the requirements of the ventricles at peak workloads. The large differences in total work between the LV and RV (∼5-fold) are not due to changes in cellular composition, but rather are dependent on the amount of muscle, or cytosol, present. The steep relationship between MV̇o2 and RPP of the RV may in part explain the high sensitivity of RV hypertrophy to pulmonary hypertension. Small increases in pulmonary resistance drives up the right heart RPP, thereby requiring the RV to generate more cellular mass to compensate for the increased workload, since modifications to improve the performance of the cytosol is not feasible.

These data are consistent with the notion that there is a near optimal quantitative relationship between the contractile elements and metabolic machinery of a heart cell and its unique constant work requirements. This results in a near fixed volume of mitochondria within mammalian heart tissue. Indeed, only a small increase in mitochondrial content was observed in this study for rabbit and porcine hearts, based on cytochrome a,a3 measurements, despite the fact that the metabolic and resting rate of the rabbit heart far exceed the pig. This notion is supported by the observation that the cardiac mitochondrial volume remains nearly constant for animals ranging in size from ∼2 kg to ∼900 kg (13), even with enormous differences in metabolic and work rates. Small animals, such as mice and shrews, show more significant increases in cardiac mitochondrial content (13). Collectively, the current study on the mammalian RV and LV implies that the mechanism for modifying the workload capacity in the heart is to vary the amount of cytosol, as opposed to modifying the molecular composition or activity status of the cellular machinery. These data cannot distinguish a difference in cellular cytosol due to the size of individual cells or the number of cells.


These studies were funded by the National Institutes of Health Division of Intramural Research.


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: D. P. and R. S. B. conception and design of research; D. P., A. M. A., R. C., E. N., Z. -X. Y., and R. S. B. performed experiments; D. P., A. M. A., R. C., and R. S. B. analyzed data; D. P., A. M. A., R. C., E. N., and R. S. B. interpreted results of experiments; D. P., A. M. A., R. C., E. N., Z. -X. Y., and R. S. B. prepared figures; D. P., A. M. A., R. C., E. N., and R. S. B. drafted manuscript; D. P., A. M. A., R. C., and R. S. B. edited and revised manuscript; D. P., A. M. A., R. C., E. N., Z. -X. Y., and R. S. B. approved final version of manuscript.


  • 1 The online version of this article contains supplemental material.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 22a.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
View Abstract