Introduction

A canonical view of biochemistry confers upon myoglobin (Mb) a role as an oxygen reservoir or as a facilitated transporter of O2 (Wittenberg and Wittenberg 1989; Wittenberg 1970). Even though in vitro experimental evidence supports such a view, in vivo experiments have produced mixed results. Some experiments support a prominent cellular O2 role (Wittenberg and Wittenberg 2003). Other studies, however, have raised questions. In spontaneously beating rat heart, the Mb O2 store and can only prolong normal heart function for a few seconds (Chung and Jue 1996). Cellular Mb also diffuses too slowly to play a prominent O2 transport role under steady state normoxic condition (Lin et al. 2007a, b). Even without myoglobin, a mouse model shows no respiration impairment (Garry et al. 1998; Godecke et al. 1999).

Many ideas on Mb function have emerged from experiments following Mb concentration change during physiological adaptation. At high altitude, Mb expression increases (Gimenez et al. 1977; Terrados et al. 1990). With exercise training, some studies show an increase in tissue Mb (Beyer and Fattore 1984; Harms and Hickson 1983). Others, however, do not (Masuda et al. 2001; Svedenhag et al. 1983). The discordant observations may arise from a dependence on a standard optical technique to determine the Mb contribution in blood-perfused tissue. As proposed by Reynafarje, the standard algorithm removes the interfering hemoglobin (Hb) signals by assuming identical intensities for Hb α (568 nm) and β (538 nm) bands (Reynafarje 1963). Even though both Mb and Hb exhibit similar spectral profiles, the signal intensity ratios at 538 and 568 nm differ. Hb displays a ratio of 1, while Mb shows a ratio of about 0.8. As a consequence, the signal intensity difference at 568 and 538 nm divided by the Mb extinction coefficient should yield the Mb concentration in all blood-perfused tissue. Such spectral differencing strategy underpins the recent determination of tissue neuroglobin concentration (Williams et al. 2008).

The general validity of such optical differencing technique does raise some concerns. Yet, its simplicity, in contrast to the alternative, semiquantitative immunohistochemistry or antibody approach, maintains its popularity (Kunishige et al. 1996; Nemeth and Lowry 1984). Unfortunately, an unmindful application of the Reynafarje method leads to an erroneous Mb concentration determination and a concomitant misunderstanding of Mb function and O2 transport. Because the 1H NMR spectra of HbCO and MbCO show distinct CH3 Val-E11 signals, an opportunity now arises to validate the optical differencing algorithm, present modifications to improve the optical differencing technique, and establish the basis for an alternative methodological approach (Ho and Russu 1981; Kreutzer et al. 1992). Given the results, the analysis suggests that any increase in Mb concentration with exercise training does not change significantly the equipoise PO2 the PO2 where Mb and O2 contribute equally to the O2 flux, and militate against a simplistic interpretation of how Mb increase can influence O2 metabolism.

Materials and methods

Experimental animals

Male C57/BL6 mice (25–35 g) were used for the present experiment. All mice were housed in a temperature-controlled room at 23 ± 2°C with a light–dark cycle of 12 h and maintained on mice chow and water ad libitum. All procedures performed in this study conformed to the Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences in the University of California, Davis. All surgical procedures were performed under pentobarbital anesthesia.

Tissue preparation

Animals were anesthetized by an intraperitoneal injection of sodium pentobarbital (65 mg/kg) and heparinized (1,000 U kg−1). The heart was quickly isolated and either extracted immediately or perfused with buffer. Hearts for perfusion were placed in ice-cold buffer solution until aortic cannulation. The heart was then perfused in Langendorff mode, with Krebs–Henseleit buffer containing (in mM) 118 NaCl, 4.7 KCl, 1.2 KH2PO4, 1.8 CaCl2, 20 NaHCO3, 1.2 MgSO4, 15 glucose. The buffer was equilibrated with 95% O2, and 5% CO2, and passed through 5 and 0.45 μm Millipore filters. A circulating water bath (Lauda MT-3) and temperature jacketed reservoir and tubings maintained the temperature at 35°C. A peristaltic pump (Rainin Rabbit) maintained a constant, non-recirculating perfusate flow of 2.0–2.5 ml min−1.

After isolation or perfusion, the heart was immediately weighed, thoroughly minced using stainless scissors, and homogenized in an ice bath with phosphate buffer (0.04 M, pH 6.6) bubbled with carbon monoxide (CO). The homogenate was then centrifuged at 14,000g for 30 min at 4°C (Centra-MP4R, IEC, USA). The clear supernatant was transferred to a small glass tube and again equilibrated with carbon monoxide to ensure myoglobin binding with carbon monoxide (MbCO). Tubes were capped tightly and kept at 4°C until optical spectroscopy and NMR measurements.

Blood sample

Mouse blood from the abdominal veins of the anesthetized mouse was taken with a heparinized syringe. The blood was lysed with H2O, and the resultant sample was centrifuged at 9,000g for 5 min. The supernatant was transferred to a small tube and was equilibrated with carbon monoxide (CO) to form HbCO. The HbCO solution was stored at 4°C until measurement by optical spectroscopy and NMR.

Optical measurement

Optical measurement (380–900 nm) of Mb and Hb used either a UVIKON 941 (Kontron Instruments) or an HP8452 spectrophotometer (Hewlett Packard). For MbCO, the concentration determination used the extinction coefficients 14.7 × 103 cm−1 M−1 and 12.3 × 103 cm−1 M−1 for the respective maxima at 540 and 577 nm. Similarly, hemoglobin concentration was determined based on the optical density at 540 and 568 nm (extinction coefficient for HbCO maxima at 540 and 568 nm = 13.4 × 103 cm−1 M−1) (Antonini and Brunori 1971). The optical density at 538 nm (β band) and 568 nm (α band) was used for calculation of both myoglobin and hemoglobin concentration in the heart tissue as described by Reynafarje (Reynafarje 1963).

Reynafarje method

Assuming only two components, HbCO and MbCO, the following equations describe the signal intensity difference at 538 nm (β) and 568 nm (α):

$$ {\text{OD}}_{538} - {\text{OD}}_{568} = \left( {\varepsilon _{538,{\text{HbCO}}} - \varepsilon _{568,{\text{HbCO}}} } \right)C_{{\text{HbCO}}} + \left( {\varepsilon _{538,{\text{MbCO}}} - \varepsilon _{568,{\text{MbCO}}} } \right)C_{{\text{MbCO}}} $$

where OD is the optical density or absorbance or signal intensity, ε = extinction coefficient, and C MbCO and C HbCO = concentration of MbCO and HbCO, respectively.

The Reynafarje method assumes an identical Hb extinction coefficient at 538 and 568 nm, \( \varepsilon _{538,{\text{HbCO}}} = \varepsilon _{568,{\text{HbCO}}} , \) so the first term on the right cancels. In addition, it assumes for Mb an extinction coefficient of ε538,MbCO = 14.7 × 103 cm−1 M−1 and ε568,MbCO = 11.8 × 103 cm−1 M−1 at 538 and 568 nm, respectively. As a consequence, the equations for C MbCO and C HbCO become

$$ \begin{aligned} C_{{\text{MbCO}}} & = \frac{{{\text{OD}}_{538} - {\text{OD}}_{568} }}{{\varepsilon _{538,{\text{MbCO}}} - \varepsilon _{568,{\text{MbCO}}} }} = \frac{{{\text{OD}}_{538} - {\text{OD}}_{568} }}{{14.7 \times 10^3 - 11.8 \times 10^3 }} = \left( {{\text{OD}}_{538} - {\text{OD}}_{568} } \right) \times 3.45 \times 10^{ - 4} {\text{M}}, \\ C_{{\text{HbCO}}} & = \frac{{{\text{OD}}_{538} - \varepsilon _{538,{\text{MbCO}}} C_{{\text{MbCO}}} }}{{\varepsilon _{538,{\text{HbCO}}} }} \\ \end{aligned} $$

NMR measurement

1H-NMR experiments used an Avance 500-MHz Bruker spectrometer equipped with a 5 mm tri-axial gradient proton/broadband probe. The 1H 90° pulse was 9.0 μs, calibrated against the H2O signal from a 0.1 mM HbCO solution. A modified 1-τ-3-τ-3-τ-1-pulse sequence suppressed the water signal and excited the Val-E11 γ-CH3 signal of HbCO and MbCO at −1.8 and −2.4 ppm, respectively (Ho and Russu 1981). A typical spectrum required 16 k scans, 10,000-Hz spectral width, 2,048 data points, and 110-ms repetition time. Zero-filling the free induction decay (FID) and apodizing with a Gaussian window smoothed the spectra. A spline fit interpolation improved the baseline for the analysis. Sodium-3-(trimethylsilyl) propionate-2,2,3,3-d4 (TSP) served as the chemical shift (0 ppm) reference and the concentration standard, based on a difference spectrum before and after the addition of a known amount of TSP in the sample.

Spectral simulation

All analysis and simulation methods used MATLAB, version 6.5/Release 13 (The Mathworks, Natick, MA). The reference spectra comprised of the optical density versus wavelength data matrix of the MbCO and HbCO spectra from 500–600 nm. All analysis employed a spline interpolation to correct the baseline. A least squares minimization algorithm then compared the heart homogenate with the reference spectra to optimize the weights assigned to the MbCO and HbCO components, which minimized the residual error and, therefore, gave the best estimate of the fractional contribution (Press et al. 2007).

Statistical analysis

All data are expressed as means ± SD. To compare the values among different methods, the repeated one-way analysis of variance was utilized. Scheffé’s post-hoc test was conducted if the analysis of variance indicated a significant difference. Regression and correlation analysis were performed to evaluate the relationship among variables. The level of significance was set at P < 0.05.

Results

Figure 1 shows the optical spectra of HbCO, MbCO, and a mixture of MbCO and HbCO. For solution HbCO from mouse blood, the peak maxima appear at 537 nm (β) and 567 nm (α) (Fig. 1a). For MbCO from buffer-perfused mouse myocardium, the two maxima appear at 540 nm (β) and 577 nm (α) (Fig. 1b). The spectrum of tissue homogenate obtained from the unperfused myocardium containing blood, shows contribution from both Hb and Mb.

Fig. 1
figure 1

Visible spectra of HbCO, MbCO, and MbCO/HbCO mixture: a HbCO extracted from mouse blood; b MbCO extracted from buffer-perfused mouse heart; c 1/1 mixture of mouse MbCO/HbCO

Full size image

Figure 2 shows the MbCO and HbCO peak displacement in the expanded spectral region between 500 and 600 nm. The HbCO 537 nm (β) and MbCO 540 nm (β) peaks do not coincide. Similarly, the HbCO 567 nm (α) and the MbCO 577 nm (α) do not match up. Moreover, the HbCO 537 nm (β) and 567 nm (α) peaks do not exhibit identical maxima intensity.

Fig. 2
figure 2

Visible spectra of HbCO and MbCO: a mouse HbCO extracted from mouse blood; b mouse MbCO extracted from buffer-perfused mouse heart. For HbCO, the peak maxima appear at 537 nm (β) and 568 nm (α). MbCO has maxima at 540 and 577 nm

Full size image

Figure 3 displays the 1H NMR spectra for HbCO solution, for MbCO from buffer-perfused heart, and for MbCO and HbCO from unperfused heart. The solution HbCO spectrum extracted from mouse blood exhibits the ring current shifted γ-CH3 Val-E11 signals from the α and β subunits at −1.72 and −1.92 ppm (Fig. 3a). Because of the non-uniform pulse excitation profile, the α subunit peak at −1.72 ppm has slightly lower signal intensity. Figure 3b displays the 1H NMR spectrum of the buffer-perfused heart homogenate. Only the MbCO γ-CH3 Val-E11 signal appears and resonates at −2.40 ppm. Figure 3c displays the spectrum of unperfused heart homogenate, which contains the well-resolved signals of MbCO and HbCO.

Fig. 3
figure 3

1H NMR spectra of HbCO, MbCO, and MbCO/HbCO mixture: a HbCO extracted from mouse blood; b MbCO extracted from perfused mouse heart; c 1/1 mixture of mouse MbCO/HbCO. For HbCO, the peaks at −1.72 and −1.92 ppm arise from the Val-E11 γ-CH3 resonances of the α and β subunits. For MbCO, the Val-E11 γ-CH3 resonance appears at −2.40 ppm. The MbCO/HbCO mixture shows the distinct peaks from Hb and Mb

Full size image

Neither the MbCO and HbCO maxima wavelength nor the extinction coefficients have the same values for all species. Tables 1 and 2 summarize some reported values in the literature. Because of the variation, the Reynafarje method assumption of a constant extinction coefficient at 538 and 568 nm can lead to significant error in the Mb determination.

Table 1 HbCO absorbance property
Full size table
Table 2 MbCO optical absorbance property
Full size table

The data matrix of absorbance versus wavelength for pure MbCO and HbCO establishes spectra basis sets. A least squares algorithm minimizes the error in matching the heart homogenate spectra with different contribution from the pure HbCO and MbCO spectral components. Using baseline corrected spectra leads to an improved estimate of tissue Mb in blood-perfused heart, as displayed in Fig. 4. Figure 4a shows the actual spectrum from unperfused myocardium homogenate, containing a mixture of Mb and Hb, while Fig. 4b shows the computer generated spectrum, which approximate the Mb and Hb mixture based on weights assigned to the MbCO and HbCO pure spectra. Subtracting the two spectra (Fig. 4b − Fig. 4a) yields the residual spectrum, Fig. 4c.

Fig. 4
figure 4

Deconvolution of Mb and Hb in the optical spectra: a observed optical spectrum from unperfused mouse heart containing both Mb and Hb; b Computer modeling of Mb and Hb contribution; c spectral difference (a − b). The difference spectrum shows no significant residual error

Full size image

The standard Reynafarje method yields an Mb concentration of 0.26 ± 0.06 mM from the perfused heart homogenate and 0.36 ± 0.07 mM from the unperfused heart homogenate. In contrast, NMR analysis of perfused and unperfused heart homogenate reveals a consistent tissue Mb concentration of 0.28 ± 0.04 and 0.26 ± 0.03, respectively. A deconvolution algorithm using baseline corrected and weights of pure HbCO and MbCO spectra also yields an Mb concentration for perfused and unperfused heart homogenate of 0.26 ± 0.06 and 0.26 ± 0.05. Table 3 summarizes the results.

Table 3 Mb concentration in mouse myocardium
Full size table

Discussion

Mb concentration in mouse heart

Studies have reported tissue myoglobin in many terrestrial mammals range from 0.2–0.5 mM. The heart contains about 0.2 mM, while skeletal muscle has about 0.4–0.5 mM. If the cell volume calculation excludes mitochondria (35% of cell volume) and the sarcoplasmic reticulum (4% of the cell volume), the concentration in heart cell would rise to 0.33 mM (Wittenberg and Wittenberg 2003). Muscle Mb concentration in marine mammals can rise to 4.5 g per 100 g tissue or about 3.8 mM in the cytoplasm (Ponganis et al. 1993, 2002).

Determining the tissue Mb concentration has relied heavily on the direct application of visible spectroscopy methods and a spectral differencing to separate the fractional contribution of Mb and Hb. The algorithm relies on equal absorbance intensities for the HbCO at 538 and 568 nm and the corresponding unequal intensities for MbCO at the same wavelengths.

Hearts perfused with saline buffer contain no Hb. Optical measurements of perfused heart homogenate reveal a tissue Mb concentration of 0.26 ± 0.06. From unperfused heart, the homogenate contains both Hb and Mb. The tissue Mb concentration, however, should remain the same. The Reynafarje method no longer yields an accurate assessment. It overestimates the Mb concentration as 0.36 ± 0.07 mM corresponding to an error of 38%.

In contrast, the 1H NMR signals of MbCO and HbCO do not overlap and appear in clear spectral region. Specifically, the HbCO γ-CH3 Val-E11 signals for the α and β subunits resonate at −1.72 and −1.92 ppm, whereas the MbCO γ-CH3 Val-E11 signal appears at −2.40 ppm. The spectra show no MbO2 or metMb signals at −2.8 ppm and −3.7 ppm. Only the MbCO signal appears (Chung et al. 1996; Kreutzer and Jue 2004). Similarly, no HbO2 and/or metHb signals appear (Ho and Russu 1981). Moreover, the spectra show no detectable signals of cytochrome C (Feng et al. 1990). The NMR analysis yields tissue Mb of 0.28 ± 0.04 and 0.26 ± 0.03 from perfused and unperfused heart homogenate, respectively. These values agree with the optical determination of Mb from buffer-perfused heart.

Mb concentration and equipoise PO2

The Mb concentration influences the intracellular oxygen delivery by establishing an equipoise PO2, the PO2 where the contribution from free O2 flux equals the Mb O2 flux as expressed in the equation

$$ \frac{{F_{{\text{O}}_2 }^{{\text{Mb}}} }}{{F_{{\text{O}}_2 }^{{\text{O}}_2 } }} = \frac{{D_{{\text{Mb}}} C_{{\text{Mb}}} }}{{K_0 (P{\text{O}}_2 + P50)}} $$

where \( F_{{\text{O}}_2 }^{{\text{Mb}}} \) is the O2 flux from Mb, \( F_{{\text{O}}_2 }^{{\text{O}}_2 } \)is the O2 flux from free O2, D Mb is Mb diffusion coefficient, C Mb is Mb concentration, K 0 is Krogh’s diffusion constant for free O2, PO2 is the partial pressure of O2 at the cell surface, and P50 is PO2 that will half saturate Mb (Lin et al. 2007a, b). Given the experimentally determined D Mb of 7.85 × 10−7 cm2 s−1 in heart at 35°C and a literature-reported Mb concentration of 0.19 mM, previously reported analysis has determined an equipoise PO2 of 1.7 mm Hg (Lin et al. 2007a, b). With the newly determined value of 0.26 mM in this report, the equipoise PO2 rises only to 1.8 mm Hg. With a resting intracellular PO2 well above 10 mm Hg and a fully saturated Mb signal even at 2× the basal work load in the heart, a 37% increase in Mb concentration from 0.19 to 0.26 mM alters insignificantly the equipoise PO2 (Kreutzer et al. 2001; Zhang et al. 1999). Such a viewpoint raises questions about any simplistic interpretation of O2 delivery enhancement following Mb increase with exercise training (Hickson 1981; Masuda et al. 1998).

Maxima and extinction coefficient error

The standard Reynafarje method deconvolutes the optical spectra of Mb and Hb predicated on the assumption that the HbCO peak at 538 nm exhibits the same intensity as the peak at 568 nm. In contrast, the corresponding Mb peaks show different extinction coefficients at these wavelengths, 14.7 versus 11.8 × 103 cm−1 M−1. With a mixture of MbCO and HbCO found in blood-perfused tissue, the optical density difference at 538 and 568 nm should yield only the Mb contribution. Dividing the intensity difference spectra by MbCO extinction coefficients at 538 and 568 nm leads to the determination of Mb in the presence of Hb (de Duve 1948; Reynafarje 1963). In a similar approach, Nakatani has used the Soret instead of the α and β bands (Nakatani 1988).

Some of the underlying assumptions of the standard method to deconvolute the Mb from Hb do not always hold. The Hb absorbance maxima of different species do not always appear at the same wavelength or with the same intensities. Consequently, the peaks at 538 and 568 nm can exhibit quite different intensity values. In mammalian Hb, the maxima positions can vary by at least ±1 nm, and the two extinction coefficients can vary from 11.4–15.0 × 103 cm−1 M−1. As a consequence, the specified absorbance values at 538 and 568 nm will also vary. In particular, mouse Hb has extinction coefficients of 15.1 and 14.9 × 103 cm−1 M−1.

Similarly, the Mb absorbance maxima and extinction coefficients also vary widely. The literature shows the MbCO β band appearing from 532 to 542 nm and having extinction coefficients ranging from 11.9–14.8 × 103 cm−1 M−1. The corresponding α band can appear from 562 to 580 nm and can have an extinction coefficient between 10.6–12.3 × 103 cm−1 M−1. Specifically, mouse Mb has the β and α bands at 540 and 577 nm, respectively. The Reynafarje method, however, sets MbCO extinction coefficients as 14.7 and 11.8 × 103 cm−1 M−1 and HbCO extinction coefficients as 14.7 and 14.7 × 103 cm−1 M−1 at 538 and 568 nm for all species.

Modified deconvolution algorithm

Simply assuming a constant extinction coefficient of 11.8 × 103 cm−1 M−1 for the MbCO peak intensity at 568 nm already starts the Mb determination with an input error. For example, horse and sperm whale MbCO have estimated extinction coefficients of 12.0 and 10.6 × 103 cm−1 M−1 at 568 nm. Using 11.8 × 103 cm−1 M−1 instead of the actual extinction coefficients introduces an error of 1.6 and 11.3%, respectively.

Using a constant 14.7 × 103 cm−1 M−1 at 538 nm also introduces a similar error. Human and horse MbCO spectra have estimated extinction coefficients of 13.5 and 14.8 × 103 cm−1 M−1 at 538 nm, leading to an error of 8.2 and 0.6%.

These errors can then propagate. In sperm whale, the MbCO exhibits absorbance maxima at 542 and 579 nm with respective extinction coefficients of 14.0 and 12.2 × 103 cm−1 M−1. Assuming the mouse and sperm whale MbCO spectra share similar features in the β and α bands leads to an extrapolation of the extinction coefficients as 13.6 and 10.6 × 103 cm−1 M−1 at the required Reynafarje wavelengths of 538 and 568 nm. For human MbCO, the corresponding extinction coefficients differ, being 13.5 and 10.9 × 103 cm−1 M−1, respectively. Based on the Reynafarje method, the Mb concentration, C MbCO = (OD538 − OD568) × (14.7 − 11.8)−1 = (OD538 − OD568) × 3.45 × 10−1. Using the estimated extinction coefficients at 538 and 568 nm from the sperm whale spectra yields, C MbCO = (OD538 − OD568) × (13.6 − 10.6)−1 (OD538 − OD568) × 3.33 × 10−1. Using the values from human MbCO leads to a C MbCO = (OD538 − OD568) × (13.5 − 10.9)−1 = (OD538 − OD568) × 3.85 × 10−1. Using the Reynafarje extinction coefficients at 538 and 568 leads to a 3.6% error in sperm whale Mb determination but a 10% error in human Mb determination. Different MbCO will exhibit a different degree of error propagation. Given the small propagated error range, however, the variation in extinction coefficients cannot account for all the observed error in the optical difference approach, which overestimates the Mb concentration by 38%. Another significant error source must exist.

One significant contribution arises from baseline errors during spectral subtraction, which reflects individual spectrophotometer’s performance characteristics. Relying simply on constant extinction coefficients at fixed wavelengths, 538 and 568 nm, would overlook these errors. Instead, using a baseline corrected visible spectra of pure 1 mM Mb and Hb as the reference basis set and applying a nonlinear least squares routine to approximate the baseline corrected spectra from blood-perfused tissue, yields a Mb value of 0.22 mM, in excellent agreement with the optical and NMR analysis of the buffer-perfused heart homogenate (Press et al. 2007). Such a deconvolution approach presents a simple modification of the Reynafarje method and appears to dramatically improve the accuracy of the tissue Mb determination. It serves as a simple alternative to the more sophisticated technique employing partial least squares fit of second derivative spectra but requires further study to clarify the underlying chemometrics (Marcinek et al. 2007).

Mb in marine mammals

The error in the Reynafarje method might appear to apply only to terrestrial mammalian tissue, where blood Hb can interfere significantly with the determination of the low tissue Mb concentration. In marine mammals, the high Mb concentration and the very low Hb interference would seem to rise above the error concerns in the present study. An investigator, who overlooks the negligible Hb contribution and uses only the prescribed Mb extinction coefficient of 11.8 × 103 cm−1 M−1 at 568 nm to determine the Mb concentration of sperm whale, would introduce immediately a 10% error. For sperm whale, the Reynafarje spectral differencing method appears to produce only a 3.6% error based on the extinction coefficient variation. For other marine mammals, however, the error may deviate significantly and depends upon the measured extinction coefficients for Mb and Hb. However, a large error contribution does not even relate to the issue of extinction coefficients. It arises from baseline errors in spectral differencing methodology, which contributes a significant fraction of the observed 38% error observed in this study (Truscott et al., unpublished observation). In essence, the study results argue for a reassessment of the tissue Mb concentration in marine mammals to clarify the basis for interpreting the role of Mb in regulating metabolism during a dive.

Conclusion

The study shows that the unmindful use of the standard Reynafarje method to determine Mb concentration in the presence of Hb can lead to significant errors. Variations in extinction coefficients and baseline fluctuation contribute to these errors. Comparative analysis of homogenate from unperfused and perfused heart reveals that a modified algorithm that relies on a basis set of pure Mb and Hb spectra will improve the Mb assay. The newly determined Mb concentration of 0.26 mM reflects a 37% increase in Mb concentration relative to the previously determined value of 0.19 mM. Such a rise in Mb concentration, however, does not significantly alter the equipoise PO2 appreciably and, therefore, the O2 transport function of Mb, at least in the mammalian myocardium. The study suggests that a reassessment of the tissue Mb concentration in different species would help clarify the role of Mb in regulating metabolism.