This article has been Cell and organelle systems can be informative, but interpretation of data on cell fragments needs to be bound by their relevance to what happens in vivo. Mitochondrial lactate oxidation is known to occur in vivo (Chatham et al. 2001), but the labile nature of mLDH gives rise to variability of results obtained among laboratories and the ensuing disagreements on data interpretation.
With regard to their paper, the following technical issues may be considered in general and for the paper of Yoshida et al. in particular:
(a)When muscles are vigorously homogenized the mitochondrial reticulum is shattered. Consequently, homogenization provides opportunity for components of the mitochondrial domain (neighbourhood) to be lost, and non-residents admitted (vide infra).
(b)Isolation of intermyofibrillar (IMF) mitochondrial fragments involves treatments with a protease, Subtilisin A. In our reading of Yoshida et al. it appears that monocarboxylate transporters mMCT1 and mLDH were lost in the process (vide infra).
(c)To ‘remove contaminating subcellular materials’, Yoshida et al. ran both subsarcolemmal (SS) and IMF mitochondrial fragments on a Percoll gradient. Those procedures increased possibility of loss of LDH from the mitochondrial lactate oxidation complex.
(d)Appreciating that they were dealing with LDH-deficient mitochondrial fragments, it is our view that in their Table 2, Yoshida et al. essentially proved the ILS in vivo. According to them, for their preparations to respire lactate at a rate equivalent to pyruvate, the lactate concentration would have to be 1–2 orders of magnitude greater than pyruvate; but such is the case in vivo according to a previous paper from the same group (Stellingwerff et al. 2006).
(e)A more critical analysis of their observation of mLDH is required. In contrasting their results with other reports in the literature, Yoshida et al. stated: ‘LDH content in mitochondria is typically observed to be only 0.5–2.0% relative to cytosolic LDH.’ However, the volume density of mitochondria in non-cardiac tissues is only a few per cent, meaning that LDH is relatively as abundant in the mitochondrial reticulum as in the cytosol. Nonetheless, Yoshida et al. stated: ‘a 1% LDH contamination’[of a muscle mitochondrial fragment preparation]‘would increase the rate of lactate oxidation by 450–1330%.’ To reiterate (d) above, in working muscle [lactate] and mLDH are sufficiently abundant for mitochondria to respire lactate in vivo.
(f)In Fig. 7C and D, the Western blots of Yoshida et al. showed results indicative of both loss and contamination:
(i)Yoshida et al. showed substantial presence of fatty acid translocase (FAT)/CD36 in all (SS and IMF) mitochondrial fragments from both red and white rodent muscles. However, in her review (Kiens, 2006) Kiens was unconvinced that FAT/CD36 is anywhere else than on the muscle cell surface. Based on what is known about fibre type heterogeneity, the bands for FAT/CD36 in Western blots of IMF from white muscle are probably false positive results.
(ii)Yoshida et al. detected no MCT1 in Western blots of IMF preparations, but showed substantial MCT2 in all (SS and IMF) mitochondrial preparations and MCT4 in SS mitochondria. These observations are incompatible with our micrographs which show MCT1 throughout the mitochondrial reticulum in cultured L6 cells (Hashimoto et al. 2006) and adult rat plantaris (Hashimoto et al. 2005), and only weak cell surface MCT2 labelling of type I fibres and no MCT2 labelling of type II fibres, but only cell surface MCT4 labelling of type II fibres (Hashimoto et al. 2005). Hence, we conclude that both LDH and MCT1 were lost in their preparations and that MCT2 and -4 bands were false positive responses.
(g)The use of new technologies was not fully considered in Yoshida et al. Regarding the use of confocal laser scanning microscopy to colocalize components of mitochondrial lactate oxidation complex, Yoshida et al. stated: ‘there is [also] increasing recognition that fluorescent microscopy is fraught with dangers.’ However, readers of The Journal of Physiology will realize that previously we observed mLDH by means of gold particle immunolabelling and electron microscopy (Brooks et al. 1999), and in our most recent report (Hashimoto et al. 2006) we showed colocalization of cytochrome oxidase, LDH, MCT1 and CD147 using combinations of stringent confocal scanning microscopy, immunocoprecipitation and Western blotting after cell fractionation. Therefore, we have shown the presence of mLDH using multiple technological approaches.
To reiterate, new ideas must withstand challenges to be accepted and that challenges to new concepts inevitably make them better. For radically different concepts such as the lactate shuttle, the need for independent evaluation is particularly acute, so we accept the attempts of Yoshida et al. to be part of the process. However, their efforts are unfulfilling because they offered no real tests of our data or ideas. Regrettably, Yoshida et al. disregarded results of studies on the mitochondrial proteome (Mootha et al. 2003; Taylor et al. 2003) that list LDH, basigin, MCT1 and hexokinase, but not FAT/CD36 among hundreds of proteins in the mitochondrial domain. Moreover, Yoshida et al. did not reference a single paper on the burgeoning field of lactate shuttling and oxidation in the brain (Schurr, 2006). Model differences aside, Yoshida et al. have realized centrality of the intracellular lactate shuttle in the regulation of energy substrate flux and its logical imperatives, but they stumbled technically and conceptually over how to make a CCLS work in the absence of an ILS.