Transport pathways--proton motive force interrelationship in durum wheat mitochondria

In durum wheat mitochondria (DWM) the ATP-inhibited plant mitochondrial potassium channel (PmitoKATP) and the plant uncoupling protein (PUCP) are able to strongly reduce the proton motive force (pmf) to control mitochondrial production of reactive oxygen species; under these conditions, mitochondrial carriers lack the driving force for transport and should be inactive. However, unexpectedly, DWM uncoupling by PmitoKATP neither impairs the exchange of ADP for ATP nor blocks the inward transport of Pi and succinate. This uptake may occur via the plant inner membrane anion channel (PIMAC), which is physiologically inhibited by membrane potential, but unlocks its activity in de-energized mitochondria. Probably, cooperation between PIMAC and carriers may accomplish metabolite movement across the inner membrane under both energized and de-energized conditions. PIMAC may also cooperate with PmitoKATP to transport ammonium salts in DWM. Interestingly, this finding may trouble classical interpretation of in vitro mitochondrial swelling; instead of free passage of ammonia through the inner membrane and proton symport with Pi, that trigger metabolite movements via carriers, transport of ammonium via PmitoKATP and that of the counteranion via PIMAC may occur. Here, we review properties, modulation and function of the above reported DWM channels and carriers to shed new light on the control that they exert on pmf and vice-versa.

Mitochondria play a key role in all aerobic eukaryotic cells, as the site of respiration and ATP synthesis via oxidative phosphorylation. In addition to cellular energy supply, mitochondria provides, through the tricarboxylic acid cycle, reducing equivalents for other compartments and carbon skeletons for biosynthesis of nucleotides, amino acids, fatty acids and vitamin cofactors. In plants, mitochondria are also involved in plant-specific metabolic pathways, including nitrogen assimilation, photorespiration, and photosynthesis in C4 and CAM plants, as well as utilization of carbon and nitrogen storage compounds during seed germination [1–4]. So, for the maintenance of basic energy function and the complex metabolic network involving mitochondria and other cell compartments, a rapid and controlled movement of solutes in and out of mitochondrion is required. The flux of hydrophilic solutes across the inner mitochondrial membrane, that represents a diffusion barrier generally impermeable to charged and polar molecules, is facilitated and regulated by a large number of specific hydrophobic transport proteins, including carriers and ion-conducting channels.

Currently, with respect to plant mitochondrial transporters in plants, several metabolite transporters of inner mitochondrial membrane have been identified at the molecular level and recognized as members of the mitochondrial carrier family (MCF) in light of their structural features similar to their counterparts from animals and yeasts [4]. However, some of these plant MCF genes have not yet been unambiguously associated with a function and with a defined physiological role in plant metabolism [2,3]. In sequenced plant genomes the number of genes encoding putative MCF proteins varies from 37 to 125, thus being larger than that of Saccharomyces cerevisiae and comparable with that of Homo sapiens. In particular, in Arabidopsis thaliana 58 MCF genes have been reported. On the basis of substrate specificity, four functional subfamilies of the proteins of the MCF have been recently proposed [4,5].

The first subfamily comprises nucleotide and nucleotide derivate transporters [4]. In Arabidopsis thaliana three ADP/ATP carriers (AAC1–3) have been identified, which are responsible for the electrogenic export of ATP4− generated in the mitochondrial matrix via oxidative phosphorylation in counter-exchange with cytosolic ADP3− [6–8]. Recently, in the same plant species, the ADNT1 carrier has been identified as responsible for the counter-exchange of mitochondrial ATP with cytosolic AMP, and only to a lesser extent ADP [9]; moreover, three putative ATP/Pi carriers (APC1-3) are supposed to catalyze ATP provision by facilitating the electroneutral exchange between MgATP2− and Pi2− [5]. Two NAD+ transporters (NDT1-2) mediate the fast import of the adenylated cofactor into mitochondria in counter-exchange with various nucleotides (preferentially ADP and AMP), as well as a slow unidirectional NAD+ transport [10]. Moreover, evidence has been reported of the existence of two mitochondrially located MCF proteins related to the mitochondrial thiamine pyrophosphate carriers (TPC) of animals and Saccharomyces cerevisiae [11], as well as of two mitochondrial CoA transporters [12].

The second subfamily comprises the proteins of the MCF that mediate the passage through the inner mitochondrial membrane of di- and tricarboxylates or ketoacids [4]. In Arabidopsis thaliana these proteins include three dicarboxylate carriers (DIC1–3) [13], a dicarboxylate/tricarboxylate carrier (DTC1) [14] and a specific succinate/fumarate carrier (SFC1) [15,16].

A third functional MCF subfamily includes transporters for amino acids, carnitine and S-adenosylmethionine [4]. In Arabidopsis thaliana two basic amino acid carriers (BAC1–2) have been identified: BAC1, involved in arginine/ornithine exchange during mobilization of storage proteins in germinating seeds [17,18], and BAC2, responsible for the transport of arginine, ornithine and citrulline and, therefore, involved in proline, polyamine and alkaloid synthesis or in the shuttling of ammonia between mitochondria and plastids [17–20]. As far as the carnitine transport, the BOU1 carrier identified in Arabidopsis thaliana has been hypothesized to import acetyl-carnitine (generated in peroxisomes from acetyl-CoA) in exchange for carnitine to directly feed the Krebs cycle with acetyl units [21]. Arabidopsis thaliana also possess two transporters (SAMC1 and SAMC2), which mediate the specific import of S-adenosylmethionine exclusively synthesized in the plant cytosol and act as a donor of methyl-groups, in counter-exchange with S-adenosylhomocysteine [22,23].

Carriers responsible for the transport of other substrates, like uncoupling proteins (UCP) and Pi carriers (PiC), constitute the fourth MCF subfamily [4]. In Arabidopsis thaliana six MCF sequences encode plant UCPs (PUCP1–6), which catalyze a nucleotide-sensitive and free fatty acid (FFA)-mediated re-entry of protons into matrix, resulting in an energy dissipation [24–27]. Two proven and one putative Pi carriers (PiC1–3) facilitates an electroneutral ΔpH-driven Pi/OH− antiport or Pi/H+ symport, so allowing Pi supply required for ATP synthesis and for metabolite uptake by means of other MCF proteins [28].

Functional, structural and evolutionary aspects of plant MCF carriers are dealt in detail in some recent reviews [3–5,7].

Along a large number of carriers, the inner membrane of plant mitochondria also contains ion-conducting channels that endow it with controlled permeability to small ions [29]. Plant inner membrane ion channels have been functionally characterized using both classical bioenergetic measurements and electrophysiological techniques, but very little information is available about their molecular identity. The electrogenic transport of K+ ions into the mitochondrial matrix has been described in different plant species and is mediated by selective potassium channels exhibiting different biochemical and electrophysiological properties: ATP-regulated and ATP-insensitive K+ channels; large conductance Ca2+-activated and Ca2+-insensitive K+ channels [30,31]. It should be considered that the electrophoretic K+ influx through potassium channels is coupled to K+ extrusion through an electroneutral K+/H+ exchanger [32,33]. As for other cations, Ca2+-transporting systems have been proposed in plant mitochondria [34,35] although they are elusive to date. Recently, a major mitochondrial iron transporter, essential for plant growth and development, has been identified and characterized in rice [36,37]. With respect to anion conductance pathways, the occurrence of a plant inner membrane anion channel (PIMAC), catalyzing the electrophoretic uniport of different anions, has been functionally demonstrated in mitochondria from a few species [38–40]; a chloride channel activity has been described in potato [41], as well as chloride channel protein performing PIMAC activity has been identified in maize [42].

As to the driving-force dependence of mitochondrial transporters, it should be remembered that substrate oxidation by the mitochondrial respiratory chain is accompanied by protons ejection in the intermembrane space, thus generating the proton motive force (pmf). Pmf represents the driving force not only for ATP synthesis but also for many protein-mediated transport activities across the inner mitochondrial membrane. Pmf consists of two components: the electrical membrane potential (ΔΨ), which is the main component in plant mitochondria ([1], and references therein), and the proton gradient (ΔpH), which is lower than in mammalian organelles.

In recent years, our research group has studied different transport pathways in durum wheat mitochondria (DWM), thus showing the occurrence of an electrogenic K+ uptake via PmitoKATP channel acting together with a K+/H+ exchanger; a fatty acid-mediated H+ transport via Plant Uncoupling Protein (PUCP); a PIMAC-like anion-conducting pathway; an ATP/ADP exchange and a Pi import via ATP/ADP and Pi carriers, respectively; a succinate/malate exchange, probably involving the DICs carriers and/or SFC; and, finally, a malate/oxaloacetate exchange, expected to involve DICs. Moreover, a proline uptake has been described in DWM by Di Martino et al. [43]. In most of these papers, attention has been made on functioning of these transport systems in relation to driving force. Here, we summarize current knowledge on DWM channels and carriers in comparison with other plant mitochondria, focusing attention on the interrelationship between transport activities and pmf in a physiological perspective. In particular, two aspects are addressed: (i) pmf decrease due to the operation of dissipative transport systems (PmitoKATP channel and PUCP); (ii) functioning of ΔΨ- and ΔpH-driven transport pathways in conditions of decreased pmf. This second aspect is of particular interest since it is poorly understood.

DWM represent the source in which a protein-mediated K+ uptake has been described for the first time in plant mitochondria [44]. Its existence was assessed firstly by using swelling technique and by following channel-dependent ΔΨ changes [44], and it has been successively confirmed by means of the first patch clamp study on plant mitochondria [45]. In these mitochondria, K+ uptake is mediated by an ATP-inhibited channel named plant mitoKATP (PmitoKATP) by analogy to the possible animal counterpart, the mitoKATP [46]. Several other mitochondrial potassium channels have been described since then; a cyclosporine A-activated and ATP-inhibited mitochondrial potassium channel was described in pea [47], in soybean suspension cell cultures [48], in three coniferous species [49,50] and in Arum spadix and tubers [51]. A quinine-inhibited but ATP-insensitive potassium channel was instead reported in potato, maize and tomato mitochondria [52]. Other less characterized K+ pathways were reported by our research group in mitochondria from bread wheat, spelt, rye, barley and spinach [44], potato [53], topinambur, triticale, lentil and Arabidopsis thaliana. By using an electrophysiological approach based on planar lipid bilayer technique, three potassium channels were described in potato tuber mitochondria: a large conductance Ca2+-activated and iberiotoxin-sensitive potassium channel, namely the mitoBKCa [54], an ATP-sensitive mitoKATP and a large-conductance Ca2+-insensitive and iberiotoxin-sensitive channel [41].

DWM also possess a very active K+/H+ antiporter [44]. In DWM, PmitoKATP catalyses the electrophoretic uniport of K+ through the inner mitochondrial membrane; the co-operation between PmitoKATP and the K+/H+ antiporter allows the operation of a K+ cycle (Figure 1). In the course of substrate oxidation, protons are ejected in the intermembrane space by complexes I, III and IV of the respiratory chain, thus generating the pmf; the K+ cycle causes re-entry of protons in the mitochondrial matrix, that strongly reduces both ΔΨ and ΔpH ([44], and references therein). As regards the operation of the K+ cycle in DWM, the PmitoKATP is very different from other mitochondrial potassium channels.

Recently, an investigation carried out on isolated potato tuber mitochondria demonstrated that potato mitoKATP activity caused a ΔΨ decrease of only few mV (up to 2.5 mV) [41]. This different behavior between DWM and potato tuber mitochondria may depend on the different experimental systems used. In potato tuber mitochondria experiments, NADH was used as respiratory substrate and ΔΨ decrease was measured using TPP+ in an indirect manner, i.e., as difference between measurements carried out in the presence of diazoxide (a mitoKATP activator) plus ATP and that carried in the presence of diazoxide alone; on the contrary, in DWM experiments, succinate was the respiratory substrate and the probe safranin O was used to directly measure KCl-induced ΔΨ decrease (see [55] for a discussion about these differences). However, it should be underlined that potato mitoKATP activity, evaluated by the same protocol employed for DWM, was found to be lower with respect to DWM-PmitoKATP [44]. This difference may also depend on the phylogenetic distance between potato (a dicotyledonous) and durum wheat (a monocotyledonous).

Similarly to potato tuber mitochondria, in heart mitochondria the increased K+ influx associated to potassium channel opening was small and it was able to depolarize by only 1–2 mV [56]; this is probably due to the fact that, in mammalian mitochondria, the maximal rate of the K+ cycle, given by the Vmax of the K+/H+ antiporter, is only about 20% of the maximal rate of proton ejection [57].

In DWM, PmitoKATP presents a conductance of 150 pS in 150 mM KCl [45]; how this low conductance may be related to ability of the channel to strongly affect ΔΨ remains to be established. As for potato tuber mitochondria, the mitoBKCa shows a conductance ranging from 502 to 615 pS [54]; for the ATP-sensitive mitoKATP and the large-conductance Ca2+-insensitive and iberiotoxin-sensitive channel, a conductance of 164 and 312 pS was measured, respectively [41].

Moreover, DWM-PmitoKATP shows a strong voltage dependence, with activity that remains constant between 95 and 140 mV and rapidly increases from 140 to 175 mV [44]. Although under physiological conditions, the channel is expected to mainly transport K+; it may work as an ATP-sensitive cation channel showing low selectivity for K+ toward Na+ in patch clamp experiments [45]. On the other hand, swelling and ΔΨ experiments show some selectivity: Cs+ > K+ = Rb+ > Na+ = Li+ [44]. Interestingly, in DWM the channel has been reported to actively transport NH4 + (see Section 3.1), thus suggesting that it has other potentially novel physiological roles. As for potato tuber mitochondria, by using ΔΨ experiments, a high selectivity for K+ toward Na+, Li+, Cs+ and Rb+ was measured for all three K+ channels [41,54].

Like mammalian mitoKATP, DWM-PmitoKATP is inhibited by ATP and ADP, and ATP inhibition is prevented or reversed by GTP and diazoxide; the ATP inhibition of DWM-PmitKATP is noncompetitive with a Ki value equal to 290 μM as evaluated by ΔΨ measurements [44]. Consistently, patch-clamp technique shows half inhibition at 500 μM ATP [45]. It should be noted that in DWM PmitoKATP affinity for ATP is significantly lower (about 10–15-fold) than the mitoKATP found in rat liver and beef heart mitochondria [46]. In DWM, ATP inhibition of PmitoKATP is independent on the presence of Mg2+ ions [44], while ATP inhibition of mammalian mitoKATP exhibits an absolute requirement for Mg2+ [58]. Interestingly, DWM-PmitoKATP is insensitive to glyburide, that is known to inhibit K+ ATP channel activity in mammalian mitochondria [59], as well as in plant mitochondria obtained from pea stems [60] and potato tubers [41]. DWM-PmitoKATP is also stimulated by the sulfydryl group reagents mersalyl and N-ethylmaleimide (NEM). As for physiological modulators other than ATP, DWM-PmitoKATP activity is decreased by NADH and mainly increased by superoxide anion, which induces an approximately 100% increase in the PmitoKATP activity [44]. FFAs also stimulate PmitoKATP activity by 2–4-fold, and an even higher increase (5–12-fold) occurs in the presence of their acyl-CoA ester derivatives [61], in contrast to mammalian mitoKATP that is inhibited by palmitoylCoA [58].

For the K+ channel activities detected in plant mitochondria, the information available on the molecular identity is still very scarce. Although a series of hypotheses have been reported [45], to the best of our knowledge, only a congress report is available about the partial purification of a potato mitoKATP, that has been proposed to contain Kir and SUR type subunits [62], and the recognition, by a proteomic approach, of a β regulatory subunit of a non-identified voltage-gated K+ channel in rice mitochondria [63]. More information is available about the animal channel counterpart with the interesting recent identification of the renal outer medullary potassium channel (ROMK, Kir1.1) as a subunit of the mitoKATP channel in bovine heart mitochondria ([64], and references therein).

The uncoupling protein (UCP) is a specialized protein that catalyses the FFA-mediated, purine nucleotide-inhibited re-entry of protons across the inner mitochondrial membrane, thus uncoupling the electron transport along the respiratory chain and the ATP synthesis by the ATP synthase (Figure 1).

In mammals, five distinct isoforms have been identified as members of the UCP family: the UCP1, found exclusively in brown adipose tissue and responsible for the adaptive thermogenesis, the ubiquitous UCP2, the skeletal muscle-specific UCP3, and the brain-specific UCP4 and brain mitochondrial carrier protein (BMCP1) (for review see [24]).

For plants, before the discovery of any mammalian UCP other than UCP1, Vercesi and coworkers demonstrated in 1995 the existence of a UCP in potato mitochondria, and named it PUMP (plant uncoupling mitochondrial protein) [65]. To avoid a possible confusion with the term “pump” used for transporters responsible for active transport, the acronym PUMP has been successively replaced with PUCP (plant uncoupling protein). Since the discovery of the first UCP homologue in plants, the existence of this mitochondrial protein has been demonstrated in a variety of organs and tissues of higher plant species including monocots and dicots, and C3, C4, and CAM plants [27,66]. As said above, upon the completion of the genome sequencing, six genes that encode PUCPs have been identified in Arabidopsis thaliana [25]; beside Arabidopsis thaliana, UCP homologue cDNAs have been identified in the dicot species mango, tomato and the thermogenic Aracea ([66], and references therein). As far as monocots, five PUCP expressed sequences have been identified in the sugarcane EST database [25]; PUCP cDNAs have been also isolated from rice [67], wheat [68] and maize [69]. Sequence analysis has revealed that PUCP are phylogenetically closer to UCP2 and UCP3 than to UCP1.

As far as DWM, the existence of an active PUCP has been demonstrated more than a dozen years ago [70]. In these mitochondria, calculation of the PUCP proton conductance values were carried out as proton leak rate/ΔΨ considering only the ATP-sensitive portion [71]. In fact, a part of uncoupling may be due to alternative pathways insensitive to ATP, such as non-protein membrane pores or protein/lipid interfaces (“proton leak”), or transporters other than the PUCP, i.e., aspartate/glutamate antiporter, PiC and AAC, which are known to mediate the FFA cycling; so, PUCP is responsible for partial ΔΨ depolarization. However, in DWM, linoleate, one of the most abundant unsaturated fatty acids in plants, is able to cause a fast and complete dissipation of ΔΨ when added to DWM at a concentration ranging between 5 and 12 μM and the linoleate-induced ΔΨ dissipation in DWM is largely recovered by the addition of 0.5 mM ATP, a powerful PUCP inhibitor [70], thus demonstrating high activity of PUCP.

So, the above reported findings demonstrate that, with the notable exception of UCP1 in brown adipose tissue [72], the DWM-PUCP is probably the most active uncoupling protein so far described among animal and plant mitochondria. The rate of linoleate-induced ΔΨ decrease shows the typical features of a protein-mediated transport, with a sigmoidal saturation dependence as a function of linoleate concentration and a K0.5 of 16 μM; a substrate specificity, with a preference for unsaturated fatty acids; a dependence on ΔΨ, with a largely enhanced rate at ΔΨ values between 120 and 130 mV [70]. In addition, DWM-PUCP shows features typical of other known mammalian and plant UCPs [24], such as inhibition by ATP, GTP, GDP and Mg2+ [70].

Interestingly, DWM-PUCP has been the first uncoupling protein for which evidence of activation by reactive oxygen species (ROS) has been reported; indeed, our findings demonstrated an increase (about 40%) in the rate of linoleate-induced depolarization in the presence of either hydrogen peroxide or superoxide anion, this latter produced in situ by the xanthine/xanthine oxidase system [70]. Successively, evidence has been reported that the activation by ROS represents a general feature of UCPs from both mammalian [73–76] and plant sources [77,78], and that it occurs indirectly, through lipid peroxidation products: superoxide releases iron from iron-sulfur centered proteins, which then generates carbon-centered radicals that initiate lipid peroxidation, yielding breakdown products, such as 4-hydroxy-2-trans-nonenal, hydroxyhexenal malondialdehyde, most or all of which can activate the proton conductance of UCPs [75].

As far as the physiological role of the PUCP, the widespread presence of these proteins in climacteric or non-climacteric, as well as in thermogenic or non-thermogenic plants suggests that they may have functions other than heat production. In particular, ROS are one of the major components of a wide array of biotic and abiotic stresses, and mitochondria are a major intracellular source of ROS [79–81]; therefore, the ROS-induced activation of DWM-PUCP may play a role in plant defence against adverse environmental stimuli (see Sections 2.3 and 4).

Skulachev’s theory postulates an important function for ROS-activated dissipative pathways, that is, to cause mild uncoupling in response to any ROS over-production, thus leading to a decrease in ΔΨ and a consequent reduction in further mitochondrial ROS generation according to a feedback mechanism [82,83]. In DWM, ROS-activated PmitoKATP and PUCP may cause a deep reduction of both ΔΨ and ΔpH (Figure 1), so exerting a positive strong control of dangerous mitochondrial ROS generation. In particular, it is well known that cellular ROS production increases as a result of plant exposure to various environmental stresses, thus inducing oxidative stress [79–81], and that mitochondria, in particular, increase their ROS generation under drought and salt stress [84]. Consistent with this, we observed an increase in the rate of superoxide anion production in mitochondria purified from osmotic- and salt-stressed durum wheat seedlings by about 40% under moderate stress conditions (germination in either 0.25 M mannitol or 0.125 M NaCl, respectively) and 120% under severe stress conditions (germination in either 0.42 M mannitol or 0.21 M NaCl, respectively) without significant differences between osmotic and salt stress [71]. Under the same conditions, we also observed an increase of FFAs and possibly of Acyl-CoA esters [61], these latter two deriving from the activation (up to about two times) of a mitochondrial phospholipase A2 (PLA2) [85], and a concurrent decrease of ATP [86,87], able to reduce the channel brake [87]. This change in the balance between modulators under stress favors the activation of the PmitoKATP (Figure 2); indeed, channel activity significantly increases under all stress conditions, and the activation is more evident under osmotic stress, with a more than 4-fold increase under severe conditions and a considerable increase even under moderate stress [71].

Once activated, the PmitoKATP may lower ΔΨ and ΔpH up to a complete dissipation, thus damping further large scale ROS production according to a feedback mechanism [88] (see also Figure 2). For example, the opening/closure of the PmitoKATP in a 100 mM KCl medium may cause a variation in the superoxide anion production up to about 35-fold [55]. This result was observed also in other plant mitochondria; indeed, in pea stem, the succinate-dependent H2O2 formation was progressively inhibited when mitochondria were resuspended in media containing increasing concentrations (from 50 to 150 mM) of KCl [89].

DWM-PUCP may cooperate with PmitoKATP to control ROS under stress condition. PUCP activation has been shown under osmotic and, in particular, under salt stress; the activation is stress-intensity dependent, and it is evident even under moderate stress [71]. We have also reported the capability of DWM-PUCP to reduce mitochondrial ROS generation in stressed DWM; in particular, a 70% reduction of the mitochondrial superoxide anion generation has been observed in the presence of externally added linoleate [71]. Interestingly, the increased PUCP functioning is not due to an increased de novo synthesis of the protein, as the expression levels of two PUCP-related genes are not affected by the stress imposition [68], but rather to an activation of the PUCP, able to quickly switch on the decrease on ΔΨ and ΔpH and, consequently, of ROS generation. Indeed, as already observed for PmitoKATP, under stress conditions PUCP is strongly activated by endogenous ROS [71]; moreover, under hyperosmotic stress conditions, the above reported increase in the endogenous FFAs is able to increase PUCP functioning according to a second activation pathway [61].

Moreover, in contrast with unstressed DWM, DWM from hyperosmotically-stressed seedlings present a PUCP insensitive to ATP. A lower UCP sensitivity to nucleotides as a result of stress has been already observed in mitochondria from cold stressed rat skeletal muscle [90] and dehydrated slices of Jerusalem artichoke tubers [78]. In phosphorylating potato tuber mitochondria, the redox state of ubiquinone (Q), which can antagonistically be varied with antimycin A and n-butyl malonate, modulates inhibition of the palmitic acid-activated UCP-sustained H+ leak by GTP. This regulatory parameter enlightens on the most probable physiological role of UCP homologues, namely the control of energy metabolism balance of the cell. The UCP-sustained H+ leak can only be inhibited when Q reduction level is low, corresponding to a low availability in oxidizable substrates and a high ATP demand, in order to preserve ATP synthesis efficiency. Contrarily, the highly reduced state of Q, occurring at high substrate supply or low ATP demand, disables purine nucleotide inhibition, leading to an activation of UCP that consequently decreases the production of harmful ROS [91]. Interestingly, all trans-retinal-induced proton leak and 4-hydroxy-2-nonenal-induced proton leak in potato tuber mitochondria are also regulated by changes in the membranous Q reduction level. So, the inhibition of plant UCPs by purine nucleotides is under the control of mitochondrial Q reduction level, regardless of the type of activator involved (FFA or aldehyde).