**4. Discussion**

Palmitoyl-l-carnitine and other fatty acids of di fferent chain length and degree of saturation di fferently a ffect the various enzymes, carriers and functions of mitochondria [1,2,32]. Therefore, we investigated the role of decanoate, octanoyl-l-carnitine, palmitoyl-l-carnitine and oleoyl-CoA (plus carnitine) in the regulation of mitochondrial respiration. In contrast to oleoyl-CoA, short- and medium-chain fatty acids (up to twelve carbon atoms) can cross mitochondrial membranes bypassing the carnitine dependent transport system [33]. Decanoate is activated to acyl-CoA inside mitochondria before being directed to oxidation. This reaction is catalyzed by the medium-chain acyl-CoA synthetase and requires both CoA and ATP. Meanwhile, oleoyl-CoA does not need activation.

Despite of di fferences in transport pathways, some reactions preceding β-oxidation and the enzyme systems responsible for β-oxidation, decanoic acid and oleoyl-CoA, similarly decreased the apparent K m ADP (Figure 2). Based on these findings and the related e ffects obtained previously with saturated fatty acids [9], it could be concluded that this property is general for all fatty acids. Furthermore, the decrease in the apparent K m ADP was similar during palmitoyl-l-carnitine oxidation at concentrations resulting in the maximal State 3 respiration rate (9 μM) and the suboptimal one (at 2.2 μM and even at 0.5 μM; Figure 3). In addition, when pyruvate and malate were added immediately after the completion of the oxidation of a limited quantity of palmitoyl-l-carnitine (Figure 6), the apparent K m ADP of the mitochondria in cardiac fibers remained at a high level, showing that the e ffect of fatty acids on the apparent K m ADP of mitochondria in permeabilized rat cardiac fibers was reversible.

Levitsky and Skulachev [34] demonstrated that palmitoyl-l-carnitine, when transported into isolated rat liver mitochondria, induced a swelling of mitochondrial matrix. Neely and Feuvray [35] observed morphological changes of isolated heart mitochondria after incubation with palmitoyl-l-carnitine, and Piper et al. [36], with oleoyl-l-carnitine, and in addition, the surfactant properties of palmitoyl-l-carnitine were revealed (for review, see [37]). We examined in this study the possible influence of palmitoyl-l-carnitine on the swelling of respiring, isolated rat heart mitochondria and the intactness of their outer membrane. Our investigations showed clearly that the decrease of the apparent K m ADP due to fatty acids oxidation described above cannot be attributed, neither to the increase in the volume of mitochondria (no mitochondrial swelling at 9 μM of palmitoyl-l-carnitine was observed; Figure 7), nor to the injury of mitochondrial outer membrane (no injury of mitochondrial outer membrane was demonstrated neither at 9 μM nor at much higher 80 μM of palmitoyl-l-carnitine, where large swelling was obvious; no injury of the mitochondrial outer membrane was also confirmed electron microscopically using dextran T-70; see Figure 12c and the discussion below).

Under some conditions (starvation, fasting) fatty acids play a major role in the heart energy metabolism. In the other cases, mostly both glucose and fatty acids are used as energy fuel, and as proposed [38], both are required for the optimal function of the failing heart; simultaneous oxidation of these two substrates is the best for heart function [39]. In this regard, our study shows the important effects of fatty acids on the regulation of the kinetics of oxidative phosphorylation (expressed as the apparent K m ADP value) when they were used alone or in the combination with pyruvate and malate.

We compared the kinetics of the regulation of mitochondrial respiration stimulated both by the exogenous ADP and exogenous ATP. In the latter case, the endogenous ADP is produced by ATPases in the myofibrils and in the sarcoplasmic reticulum, and is delivered directly to the mitochondria [8]. In accordance with the low apparent K m ADP, also the low apparent K m ATP values for saponin-permeabilized fibers respiring on palmitoyl-l-carnitine were demonstrated (for the first time with this substrate). In addition, our data are in accordance with the data obtained by Seppet et al. [8], where they demonstrated very similar high (300 μM) apparent K m values for both exogenous ADP and exogenous ATP for cardiac fibers respiring on non-fatty substrates. Thus, the apparent K m of oxidative phosphorylation for ADP in saponin-permeabilized rat cardiac fibers did not depend on the external or internal source of ADP.

When endogenous ADP is produced by ATPases in the myofibrils and in the sarcoplasmic reticulum, it is directly channeled to mitochondria without significant release into the medium if the mitochondrial oxidative phosphorylation is active [8]. This evidence was obtained by using an exogenous ADP-trapping system consisting of PK + PEP, which e ffectively competes with mitochondria for the extramitochondrial ADP, and therefore, decreases the respiration rate in the State 3. In our experiments, the PK e ffect is exceptionally from the outside of the mitochondria, since the exogenous cytochrome c test (no stimulation of respiration in State 3 by cytochrome c) shows the intactness of the mitochondrial outer membrane and the inaccessibility of PK to ADP localized in the intermembrane space of the mitochondria. Thus, neither the increase in temperature (from 25 to 37 ◦C), nor fatty acid oxidation, which both largely increase the a ffinity of mitochondrial oxidative phosphorylation for exogenous and endogenous ADP (decrease in the apparent K m ADP and apparent K m ATP), a ffect the inhibitory e ffects of the exogenous ADP-trapping system (Figure 9), and thus did not enhance the permeability of the mitochondrial outer membrane for ADP (Figure 9). The inhibitory e ffects of PK + PEP on the respiration of saponin-permeabilized rat cardiac fibers oxidizing glutamate+malate (25 ◦C) obtained by us are in good agreemen<sup>t</sup> with the data of other investigators [8,10].

The stimulating e ffects of creatine on fibers respiration with palmitoyl-l-carnitine assessed at low concentration (60 μM) of ADP reflects maintenance of functional coupling in the intermembrane space between the ADP/ATP carrier and mi-CK. Our data (Figure 10) show that the functional coupling between the ADP/ATP carrier and mi-CK is preserved in mitochondria despite the significant decrease in the apparent K m ADP induced by fatty acid oxidation. Noteworthy, the latter phenomenon is observed not only at 37 ◦C [9], but also at lower, i.e., 20 ◦C temperature. The finding that creatine significantly decreases the apparent K m ADP in the mitochondria oxidizing octanoyl-dl-carnitine (Figure 5) is in a good agreemen<sup>t</sup> with the data of other investigators [40,41], when pyruvate+malate, i.e., non-fatty respiratory substrate, was used.

There are two mechanisms suggested to explain the e ffective interaction between mi-CK and the ADP/ATP carrier: the dynamic compartmentation of ATP and ADP [42], and the direct transfer of ATP and ADP between the proteins [43,44]. According to the first mechanism, the functional coupling between mi-CK and the ADP/ATP carrier can be explained by di fferences between the concentrations of ATP and ADP in the intermembrane space and those in the surrounding solution due to some limitations of their di ffusion across the outer mitochondrial membrane [42]. Beside dynamic compartmentation, facilitated di ffusion was also suggested as the potential mechanism of the action of the phosphocreatine shuttle [45]. According to the second mechanism of coupling, ATP and ADP

are directly transferred between mi-CK and the ADP/ATP carrier without leaving the complex of proteins [44]. Kuznetsov et al. [41] investigated the apparent Km for ADP in cardiac, slow-twitch and fast-twitch skeletal muscle fibers, and they provide a strong argumen<sup>t</sup> against the possibility that the diffusion problems may be related to the increased apparent Km for ADP in cardiac and slow-twitch skeletal muscle fibers, as in fast-twitch skeletal muscle fibers, it remains low and comparable to the apparent Km for ADP in isolated mitochondria [41]. Also, they found that in the ghost fibers where the myosin and 10–20% of cellular proteins were removed from the cells by KC1 treatment, the apparent Km for ADP remained unchanged [41]. Under their conditions the sarcolemma was almost completely removed, and could not act as a diffusion barrier even for the big protein molecules [41]. Their data together with our results that creatine did not increase Vmax in skinned cardiac fibers oxidizing octanoyl-dl-carnitine (Figure 10), and that the addition of dextran did not affect the apparent Km for ADP in the skinned cardiac fibers oxidizing pyruvate and malate (Figure 11a), sugges<sup>t</sup> that the changes in the apparent Km for ADP are unlikely related to the diffusion problems.

In vivo mitochondria are embedded in the cytoplasm in which the protein content may be as high as 20–30% (*m*/*v*) [46]. Macromolecules, like bovine serum albumin, dextran, ficol, polyvinylpyrrolidone, added to the medium, can restore the morphological changes in the outer mitochondrial compartment that occur during the isolation of mitochondria [47–49].

It has been also shown that, in isolated rat heart mitochondria, 10% of bovine serum albumin and 5–25% dextran strongly increase the apparent KmADP of oxidative phosphorylation [15,50] and of mitochondrial creatine kinase [50]. Dextrans or other macromolecules decrease the conductivity of porin pores in artificial membranes [48,51], the volume of the intermembrane space in isolated mitochondria, and increase the number of contact sites between both mitochondrial membranes [52]. Morphological changes of isolated mitochondria are accompanied by a reduced permeability of the mitochondrial outer membrane for adenine nucleotides [50]. In this study, respirometric investigation of saponin-permeabilized rat cardiac fibers demonstrated that an addition of 5% dextran into the incubation medium (to mimic the oncotic pressure of the cellular cytoplasm) markedly increased the low apparent KmADP value of mitochondria respiring on palmitoyl-l-carnitine, but did not affect the high apparent KmADP of mitochondria respiring on pyruvate and malate (Figure 11). Interestingly, the apparent KmADP values during the oxidation of palmitoyl-l-carnitine in the medium containing 5% dextran (110.6 ± 20.7 μM ADP), despite some increase (compared with medium devoid of dextran), remained far below high values characteristic for the oxidation of pyruvate + malate (apparent KmADP: 255.7 ± 42 μM) (Figure 11).

Electron microscopy was used to evaluate the effects of fatty acid oxidation and dextran T-70 on the morphology of rat heart mitochondria in situ (Figure 12). Palmitoyl-l-carnitine oxidation induced marked alterations in the mitochondrial ultrastructure, which is different from that observed with pyruvate and malate, and similar to the changes induced by the incubation of fibers with the ADP/ATP carrier inhibitor carboxyatractyloside. When mitochondria were incubated with compounds, fixing the ADP/ATP carrier in M (i.e., matrix-oriented) conformation, the cristae were compressed, the matrix space enlarged, and the separation of the inner membrane space forming vacuoles was noticed [53]. We obtained similar results after the incubation of mitochondria with bongkrekic acid. In the contrary, when mitochondria were incubated with compounds, fixing the ADP/ATP carrier in C (i.e., cytosol-oriented) conformation, e.g., with carboxyatractyloside, such that the cristae volume was enlarged at the expense of matrix space, the electron densities increased [53]. Furthermore, these two distinct mitochondrial conformations corresponded to condensed and orthodox mitochondrial conformation, accordingly [25,26]. Ultrastructural changes of mitochondria in cardiac fibers respiring on palmitoyl-l-carnitine were completely prevented by dextran addition into the incubation medium. Furthermore, most mitochondria in the fibers incubated with palmitoyl-l-carnitine in the presence of dextran appeared dark, indicating that their outer membranes were intact (Figure 12c), in accordance with the results of the cytochrome c test, presented in the Results section. Partial "normalization", i.e., the increase by dextran of the apparent KmADP and complete "normalization" of the ultrastructure

of mitochondria in saponin-permeabilized cardiac fibers respiring on palmitoyl-l-carnitine could be due to fatty acid oxidation-induced alterations of the mitochondrial ultrastructure.

The high apparent Km for exogenous ADP is determined by low mitochondrial outer membrane permeability for ADP and the concerted action of tight complexes (called as intracellular energetic units) of mitochondria and other cellular ADP-producing systems (ATPases) in myofibrils and sarcoplasmic reticulum [8,10]. A highly organized intracellular structure and the arrangemen<sup>t</sup> of mitochondria are also crucial for the increase in the affinity of mitochondrial oxidative phosphorylation for ADP and the decrease in the apparent KmADP value [16,18,22].

The calculations by Lizana et al. [54] using a simplistic model suggested that mitochondria (and small biological compartments in general) may regulate the dynamics of interior reaction pathways (e.g., the Krebs cycle) by volume changes. It was hypothesized (for review, see [55]) that the cristae shape can be modulated by different pathways/signaling molecules (ROS, the NADH/NAD ratio, the ATP/ADP ratio, etc.). Even low concentration of long-chain fatty acid palmitoyl-l-carnitine oxidation in skeletal muscle and heart mitochondria has been associated with significantly higher ROS production and increased mitochondrial proton leak (2 times lower respiratory control ratio, i.e., coupling), compared with the oxidation of NADH-linked substrates (pyruvate + malate or glutamate + malate) [56–58].

The greater ROS formation, similar at higher and lower mitochondrial membrane potential, was maintained despite of the activation of the uncoupling mechanisms of the ADP/ATP carrier during palmitoyl-l-carnitine (18 μM) oxidation [56]. Furthermore, functional (and possibly structural) interaction of ADP/ATP carrier and VDAC and its modulation/interruption by the ADP/ATP carrier inhibitor atractyloside-induced change of ADP/ATP carrier conformation has been also demonstrated [59]. Structural changes of mitochondria induced by carboxyatractyloside, similar to those induced by palmitoyl-l-carnitine, were observed by us in this study. The functional interaction of ADP/ATP carrier and VDAC facilitated the channeling of nucleotides and other metabolites between cytosol and mitochondrial matrix [59].

The low apparent KmADP in mitochondria from cancerous cells [22,23] and fetal or neonatal mitochondria [24] could be related to mitochondrial ultrastructural changes. Adult mitochondria are usually in condensed conformation, with enlarged mitochondrial matrix volume and the decreased intermembrane space, compared to neonatal mitochondria which are mostly in orthodox conformation [25,26]. Our results showed that palmitoyl-l-carnitine oxidation induced ultrastructural changes of mitochondria, with a marked increase in mitochondrial intermembrane space and decrease of mitochondrial matrix (Figure 12b), and similar changes were obtained in the presence of the ADP/ATP carrier inhibitor carboxyatractyloside (Figure 12d).

Recent studies have revealed that fatty acids could be the cofactors [60] of the newly discovered action mode of the ADP/ATP carrier when protons are transported to mitochondrial matrix [61]. Cytosolic fatty acids could interfere with ADP/ATP transport [60] by activating the proton current to mitochondrial matrix [61]. They are able to bind to the ADP/ATP carrier from the cytosolic side in the C- or M-state, but could not induce C–M conformational change or be transported by the ADP/ATP carrier [61]. It was suggested that the ADP/ATP carrier could have two transport modes: C–M conformational change-related electrogenic ADP/ATP exchange and cytosolic fatty acid-activated conformation-independent proton channel [61]. We have also observed that palmitate, palmitoyl-CoA and palmitoyl-l-carnitine could not induce the decrease of apparent KmADP when their oxidation was prevented by the absence of necessary cofactors or blocked with rotenone [9]. The lower apparent KmADP was related to the oxidation, but not to the transport of fatty acids into mitochondria [9]. In the study of Divakaruni et al. [19], fatty acids have been demonstrated to change the conformation of the uncoupling protein, that is structurally, and to some extent, functionally, similar with the ADP/ATP carrier [20,21]. ADP/ATP exchange is driven by the mitochondrial membrane potential, which has been shown to affect the distribution of the binding sites of the ADP/ATP carrier between inside and outside, as well as the distribution of ADP/ATP carrier molecules between the M- and the

C-state, respectively [62]. Thus, Krämer and Klingenberg argued that the membrane potential could indirectly influence the C- and M-conformational transition of the ADP/ATP carrier [62]. The tumor and fetal/neonatal mitochondria are not only characterized by the low apparent K m ADP, but also have higher mitochondrial membrane potential, i.e., are hyperpolarized [63,64]. According to the data of our colleagues, the mitochondrial membrane potential of isolated rat heart mitochondria respiring on pyruvate+malate or palmitoyl-l-carnitine as substrates was accordingly 146 ± 2 mV and 124 ± 1 mV (*n* = 5, assessed with TPP<sup>+</sup> electrode; R. Baniene, unpublished data, 2020); i.e., the membrane potential did not increase, rather it slightly decreased during palmitoyl-l-carnitine oxidation. These values are in line with the data from the study of Seifert et al. where lower energization and lower membrane potential of the skeletal muscle mitochondria were reported in the case of palmitoyl-l-carnitine oxidation compared to pyruvate and malate oxidation [56]; however, it could be related to the higher uncoupling protein content in skeletal muscle mitochondria [65]. It is technically complicated to quantify correctly the mitochondrial membrane potential in skinned cardiac fibers due to the distribution of the cationic potential probes within the cellular structures present in this object. However, it could not be excluded that the mitochondrial membrane potential in the skinned fibers might di ffer from the mitochondrial potential in the isolated mitochondria.

Thus, the hypothesis that the fatty acid oxidation-induced conformational change of the ADP/ATP carrier (M-state to C-state, condensed to orthodox mitochondria) a ffecting the oxidative phosphorylation affinity for ADP could be driven by the higher membrane potential generated during fatty acid oxidation might be an interesting question to address in the future research.

Overall, our results imply that the fatty acid oxidation could regulate cellular energy metabolism by increasing oxidative phosphorylation a ffinity for ADP due to fatty acid oxidation-induced ADP/ATP carrier switch from M- to C-state, and corresponding mitochondrial transition from condensed to orthodox conformation. Furthermore, this mechanism could be responsible for the altered mitochondrial metabolism [66] when fatty acid oxidation is increased during the development of chronic [67,68] and age-associated disorders, such as cardiovascular diseases, diabetes, neurodegenerative diseases and cancer [69,70], and the ADP/ATP carrier modulation could be a promising target in the search of novel therapies to restore the normal mitochondrial function.

**Author Contributions:** Conceptualization, A.T. and D.M.K.; methodology, A.T., S.T., J.L., N.P.; validation, A.T., J.L., S.T., D.M.K.; formal analysis, J.L., S.T., A.T.; investigation, A.T., S.T., N.P., J.L., L.K.; resources, A.T., D.M.K.; data curation, L.K., D.M.K.; writing—original draft preparation, A.T. and D.M.K.; writing—review and editing, J.L., S.T., A.T., D.M.K.; visualization, N.P., L.K. and D.M.K.; supervision, A.T.; project administration, A.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors wish to thank Vilmante Borutaite and Rasa Baniene for valuable comments and suggestions.

**Conflicts of Interest:** The authors declare no conflict of interest.
