Control of the removal of reducing equivalents from the cytosol in perfused rat liver.

Abstract The rat liver perfusion previously employed was adapted to a flow-through system in which samples of effluent perfusion medium were collected for measurement of metabolic products at 2-min intervals by means of a fraction collector. Overnight fasted rats were used to avoid complications of glucose formation from endogenous glycogen. Since xylitol conversion to glucose and lactate by the liver is associated with the production of NADH in the cytosol, the rates of production of these metabolic end products can be used to provide a measure of the rate of NADH production and utilization. Xylitol metabolism was predominantly cyanide sensitive showing an obligatory involvement of mitochondrial respiration for removal of reducing equivalents from the cytosol. The participation of various shuttles for the transport of reducing equivalents into the mitochondria was investigated by infusion of the oxidized or reduced partners of cytosolic or mitochondrial NAD-linked dehydrogenases (e.g. ethanol, β-hydroxybutyrate), or specific inhibitors such as rotenone, amobarbital, cyanide, and the transaminase inhibitor aminooxyacetate. The influence of flux in the mitochondrial electron transport chain on the rate of removal of NADH from the cytosol was studied by infusion of 2,4-dinitrophenol to produce an uncoupling of oxidative phosphorylation, ornithine plus ammonia to provide an energy drain on mitochondrial ATP, and artificial electron acceptors to bypass rate-limiting electron transport steps. The studies show that xylitol metabolism was not controlled by the activity of NAD-xylitol dehydrogenase but by the rate of reoxidation of NADH. This process was regulated by the rate of transfer of reducing equivalents into the mitochondria and by the rate of electron flux to oxygen. Evidence was obtained for the participation of both NAD- and flavin-linked shuttles in the over-all transport of reducing equivalents from cytosol to mitochondria. The flavin-linked shuttle may be identified with the α-glycerophosphate shuttle which operates by virtue of the abnormally high cytosolic α-glycerophosphate concentration produced during xylitol metabolism. Involvement of the NAD-linked malate-aspartate shuttle was shown by the sensitivity of glucose formation from xylitol to aminooxyacetate and β-hydroxybutyrate. The relevance of these findings to the control of ethanol utilization by the liver is discussed.

linked shuttle may be identified with the ar-glycerophosphate shuttle which operates by virtue of the abnormally high cytosolic a-glycerophosphate concentration produced during xylitol metabolism. Involvement of the NAD-linked malateaspartate shuttle was shown by the sensitivity of glucose formation from xylitol to aminooxyacetate and fl-hydroxybutyrate. The relevance of these findings to the control of ethanol utilization by the liver is discussed.
In the previous paper it was shown that xylitol is rapidly metabolized in the perfused rat liver by a cytosolic NAD-linked xylitol dehydrogenase, and that glucose is the main carbon end product (1). Competition between dehydrogenases was illustrated by the use of [14C]lactate added to the perfusion fluid separately or in combination with xylitol. The results showed that glucose formation from lactate was inhibited much more than glucose formation from xylitol when both substrates were present together.
Glucose formation from n-xylulose, the product of NAD-xylitol dehydrogenase, involves phosphorylation of the pentose by ATP to D-xylulose-5-l' which is converted to fructose-6-P and triose phosphates by the transaldolase and transketolase reactions of the pentose phosphate cycle (2). Thus, unlike lact,ate gluconeogencsis, xylitol gluconeogenesis is associated with the net generation of NADH in the cytosol, which is equal to 1.2 times the rate of glucose production.
Reoxidation of cytosolic NADH, therefore, is an obligatory step for xylitol metabolism, and its rate can be conveniently monitored by measuring the rate of glucose production by livers perfused with xylitol as substrate (3).
Data described in this paper illustrate the USC of a flopthrough liver perfusion system for quantitat.ivc investigations of the control of cytosolic-mitochondrial interactions mcdiatcd by NAD-and flavin-linked carbon shuttles. The results obtained by the use of xylitol as a donor of reducing equivalents to NAD in the cytosol provide basic information on the control of intracellular hydrogen transport processes which is of relevance to the problem of ethanol utilization by the liver (4).
Perfusion Techniyue-In general, the experimental methods rrcre similar to those used in the previous paper (I).
However, the liver perfusion was adapted to a flow-through system in order to follow variations in t'hc rate of glucose formation more conveniently than in the normal recirculating perfusion system. Because of the large volumes of fluid required, albumin was omitted from the perfusion medium. Krebs bicarbonate buffer, containing half-calcium concentration to compensate for omission of albumin, was used as the basic perfusion fluid. The lowering of oncotic pressure by the absence of protein or other high molecular weight substitute appeared to have little deleterious effect on liver funct,ion .for perfusion periods up to 90 min. Thus, liver swelling was no greater than in the presence of albumin, oxygen consumpt,ion was not affected, and intracellular K+ was well retained.
The perfusion fluid was contained in a Z-liter volumetric flask immersed in a water bath maintained at 37", and was saturated with a gas mixture containing 95oj, 0, and 57; COZ. Fluid was pumped from this reservoir at a rate of 25 to 30 ml per min into a rotating disc oxygenator (5), which ensured saturation of the fluid with the oxygen gas mixture, and was pumped out at the same rate through the liver via a cannula placed in the portal vein. Fluid left the liver via a cannula placed in the superior vena cava, and passed immediately through a 0.2.ml chamber containing a micro Clark oxygen electrode to a streamsplitt,ing device which allowed part of the fluid to run to waste while the remainder (approximat,ely 2 ml per min) was collected at 2-min intervals with a time-programmed fraction collector. Substrates and inhibitors were continuously added by infusion pumps to the fluid immediately prior to the liver to give the calculated art,erial concentrations indicated in the figure legends.
Analytical Techniques-Glucose in the collected effluent fractions was assayed enzymatically by the glucose-6-P dehydrogenase-hexokinase method (6). Automatic pipetting and sampling devices attached to a spectrophotometer fitted with a flow cell (Gilford) permitted the assays to be completed with a lag of only a few minutes. Lactate was assayed similarly by means of lactate dehydrogenase, with the reaction complete within 5 min.
Fluorescence measurements from the surface of the liver and absorption measurements through a liver lobe were monitored as previously described (1,7,8).
Expression of Results-Data obtained with individual livers are given in the figures in order to increase their clarity.
However, each experiment was repeated four or five times with variations, and representative examples were chosen for presentation.

RESULTS
Egect of Substrate Oxidants and Reductants on Glucose Pormation from Xylitol and Lactate-The rate of glucose production by livers from fasted rats perfused with the flow-through method in the absence of added substrate was 5 to 10 pmoles/lOO g rat weight per hour. Fig. 1 shows that infusion of 5 InM xylitol produced a gradual increase in the rate of glucose production over a 15-to 20.min time interval to a maximum of about 90 ~moles/lOO g rat weight per hour.
In general, the rate of glucose production by livers perfused with the flow-through method was somewhat lower than those obtained with the recirculation system (l), but control experiments showed t'hat it 1. Effects of acetoacetate (5 mMj, nn-fi-hydroxybutyrate (10 DIM) and tetramethyl-p-phenylenediamine (TMPD) (25 PM) infusion on glucose production from xylitol in perfused rat liver. Xylitol (5 mM) was infused throughout the experiment.  Fig. 1 caused an immediate slight stimulation of glucose production, which was not sustained. On the other hand, infusion of 10 mM fi-hydroxybutyrate produced a prompt 25 to 300/, inhibition.
This inhibition was relieved, and glucose formation stimulated above control levels by infusion of 25 PM artificial electron acceptor tetramethyl-p-phenylenediamine.
The reproducibility of the xylitol stimulation of glucose production and inhibition by /3-hydroxybutyrate is illustrated in Fig. 2, which also shows the rapid stimulation of glucose formation produced by 5 pM phenazine methosulfate. iMethylene blue (20 j.4M) produced similar effects. This stimulation of glucose formation by artificial electron acceptors was cyanide sensitive.
The rate of oxygen consumption by livers respiring on endogenous substrates in the flow-through liver perfusion system was in the range of 500 to 600 patoms/lOO g rat weight per hour.
The effect of xylitol infusion on respiration was small and variable, but &hydroxybutyrate infusion invariably increased respiration by 30 to 40 patoms of oxygen per 100 g rat weight per hour. The artificial electron acceptors, on the other hand, stimulated oxygen consumption to more than 1000 patoms/lOO g rat weight per hour.
The above results show that when reducing equivalents are generated directly in the mitochondria by addition of the reduced partner, namely P-hydroxybutyrate, of a mitochondrial NAT)-linked-dehydrogenasc, the rate of reoxidation of NADH generated from xylitol dehydrogenase is diminished. Fig. 3 shows that infusion of 2 mM ethanol, a reducing agent of the cytosolic NAD pool, also produced a substantial inhibition of glucose formation from xylitol.
The figure also illustrates the rapid reversibility of the P-hydroxybutyrate inhibition once infusion was stopped.
In this experiment, respiration of the liver decreased by 50 patoms/lOO g rat weight per hour when infusion of fl-hydroxybutyrate was stopped, but ethanol infusion had a negligible effect on Dhe oxygen consumption.
Other experiments showed that infusion of 1 mM acetaldehyde in the presence of xylitol produced a stimulation of glucose formation and an inhibition of lactate formation, indicating that acetaldehydc acted more as an NADH acceptor with alcohol dehydrogenase than as an NADH donor via acetaldehyde dehydrogenase . However, in the presence of 1 m& pyrazole to inhibit alcohol dehydrogenase (9), an inhibition of glucose formation from xylitol was observed, suggesting that under these conditions it functioned as a donor of reducing equivalents.
Ethanol infusion also inhibited glucose formation from lactate in the flow-through perfusion system. Fig. 4A shows that when 2 mM ethanol was infused together with 1.5 mlLr L(+)-lactate, the rate of gluconeogenesis was about 55 ~molcs/lOO g rat weight per hour.
Upon stopping ethanol infusion, glucose production increased to 80 to 90 pmoles/lOO g rat weight per hour. In Fig. 4B, it is seen that infusion of 4 m&x /3-hydroxybutyrate stimulated gluconeogenesis from lactate, contrary to its inhibitory effect on glucose formation from xylitol.
Shown also in  3. Effects of DL-@hydroxybutyrate (10 mM) and ethanol (2 mm) infusion on glucose production from xylitol in perfused rat liver. P-hydroxybutyrate dehydrogenase affects gluconeogencsis by a different mechanism than when reducing equivalents arc generated directly in the cytosol via alcohol dehydrogenasc.

Effect of Increased Electron
Transport on Glucose Production from Xylitol- Fig.  5 shows the effect of infusing increasing concentrations of 2,4-dinitrophenol to a liver perfused with 10 mM xylitol.
Dinitrophenol conc~entrations of 5 and 20 pM produced successive increases of oxygen consumption and glucoxr production. With 100 pM dinitrophenol, although respirat,ion was further increased, glucose production became inhibited. Mowever, it was restored to cont,rol levels when the dinitrophenol infusion was terminated.
This experiment shows that when flux through the mitochondrial electron transport chain was increased due to partial uncoupling of oxidative phosphorylation, flux in the pathway from xylitol to glucose increased.
ITowcver, since ATP is required for the phosphorylation of n-xyluloxc, severe impairment of mitochondrial ,4TP production is assoriated with an inhibition of xylitol conversion to glucose. Flux through the phosphorylating electron transport chain can also be increased by increasing the rate of mitochondrial ATP utilization for biosynthetic reactions.
As shown by Hems et al. (lo), this may be achieved in the perfused rat liver by addition of ornithine and NH&l, which induces a stimulation of oxygen consumption and urea formation. Fig. 6 shows that infusion of 2 mM ornithine together with 10 mM N&Cl, after prior infusion of xylitol, produced a 40%) stimulation of glucose production.
Ornithine alone had a negligible effect on ghlcose formation.
Oxygen consumption increased from 635 to 717 patoms/lOO g rat weight per hour. Subsequent infusion of 10 m&f P-hydroxybutyrate decreased glucose production from 120 to 55 l.tmoles/lOO g rat weight per hour although the oxygen consumption of the liver increased by 150 patoms/ LOO g rat weight per hour. Infusion of 25 pM methylene blue caused a prompt restoration of glucose formation. 5. Effects of 2,4-dinitrophenol infusion on oxygen consumption and glucose production from xylitol in perfused rat liver.
Pathways Involved in Transport of Reducing Equivalents from Cytosol to Mitochondria--In order to assess the relative significance of the malate-aspartate and a-glyccrophosphate shuttles for the transport of NADH into the mitochondria, the transaminase inhibitor aminooxyacetatc (11) was used to inhibit flux through the maMe-aspartate cycle (12) in livers perfused with xylitol. Fig. 7 shows that 0.5 mu aminooxyacetate initially produced a 20%. inhibition of glucose production but that after about 10 min the inhibition was completely reversed. Lactate production increased from 5 to 15 pmoles/lOO g rat weight, I:er hour after xylitol infusion and was not appreciably affected by aminoosyacetate.
Simultaneous infusion of 4 JYIM P-hydroxybutyrate together with xylitol and aminooxyacetate producrd a partial inhibition of glucose formation and an almost complete inhibition of lactate formation, while 25 FM mcthylcne blue reversed these changes. Oxygen consumption decreased initially by 40 ~atomsj100 g rat \\-tight per hour after aminoosyacetate addition and increased by about the same amount after P-hydrosybut,yrate infusion.
The inhibitions of lactate and glucose formation from sylitol in the presence of alninoox~-acetate and /%hydroxybutyrate indicate that the mcchanixm responsible for t;he transfer of reducing equivalent's across the mitochondrial membrane is still sensitive to competition with mitochondrial NAD-liuked dehydrogenases. The effectiveness of aminooxyacetate as an inhibitor of the rnalate-aspartate shuttle in livers perfused mit)h sylitol \T-as investigated further with t,he aid of the surface fluorometry tcchnique for inessurcmfnt of flavin and pyridine nUcleotide oxidation-reduction changes (7). These measurements were made on livers perfused with recirculation of the perfusion fluid rather than with the flom-through &em. Fig. 811 shows the responses obtained with normal liTera, while Fig. 8B shows rcsults obtained in the presence of aminooxgacetatc. &sides flavin and pyridine nucleotidc oxidation-reduction changes, the absorption change at 560 nnl (575 nm as reference), attributable to cytochrome b was also rccotdctl.
In E'ig. 8A, an anoxic cycle is shown first for internal calibration of the traces. Addit,ion of xylitol after 60 min of perfu&n resulted in a partial reduction of both flavin and pyridinc nuclcot,ides, and a small but definite increase of cytochromc b abwrption.
Subsequent addition of ethanol caused further, but still incomplete, reduchions of the flavin and pyridine nucleotitlc~ components.
Pig. 812 shows that addition of xylitol and ethanol in the presence of aminooxyacetate caused successive further reduction of pyridine nucleo-  Although increased cytochrome b6 absorption may contribute to the total absorption changes during the anoxic cycle, it is unlikely lo interfere with cytochrome b oxidation-reduction changes in the aerobic state. Consequently, the increased absorption in the 560.nm region observed after xylitol addition may be interpreted as the response of cytochrome b to an increased steady state level of reduction of mitochondrial flavin and pyridine nucleotide pools. This observation lends support to the previous conclusion that flavin fluorescence changes in the perfused liver are mainly of mitochondrial origin and reflect principally oxidation-reduction changes of lipoic dehydrogenase (7). The pyridine nucleotide fluorescence changes observed after xylitol and ethanol additions to control livers indicate an enhanced state of reduction of both cytosolic and mitochondrial NAD pools. Since lipoic dehydrogenase is considered to be in approximate equilibrium with the mitochondrial NAD pool (13,14), the increased level of reduction of the pyridine nucleotides observed after xylitol and ethanol additions in the presence of aminooxyacetate, which occurred in the absence of flavin fluorescence changes, can be ascribed solely to the cytosolic NAD pool. The conclusion is reached, therefore, that in the perfused liver aminooxyacetate is effective in inhibiting the transfer of reducing equivalents from the cytosolic to the mitochondrial space via the malate-aspartate shuttle.
The data also indicate that other possible NAD-linked hydrogen transport shuttles arc probably relatively inactive under the conditions of the experiment.
Transport of reducing equivalents into the mitochondria by means of the malate-aspartate and or-glycerophosphate shuttles can also be distinguished by the use of rotenone or amobarbital which inhibit mitochondrial electron transport between NADH and cytochrome b. a-Glycerophosphate donates electrons to a flavoprotein pool situated on the oxygen side of the site of inhibition, so that t'he cr-glycerophosphate shuttle should be insensitive to inhibition by rotenone or amobarbital. Fig. 9 shows the effect of xylitol addition after rotenone on the fluorescence responses of flavin and pyridine nucleotides, and on the oxygen tension of fluid leaving the perfused rat liver.
The numbers on the effluent oxygen trace refer to oxygen consumption rates in microatoms per 100 g rat weight per hour. The liver was depleted of endogenous substrates by prior perfusion for 1 hour. Addition of rotenone caused a partial reduction of the flavin and pyridine nucleotides, these responses being largely mitochondrial (1). Oxygen uptake by the liver was inhibited from 835 to 546 patoms/lOO g rat weight per hour. The subsequent addition of xylitol produced a small further increase of the flavin reduction state, and a large increase in the state of reduction of the pyridine nucleotides as the cytosolic NAD systems became reduced with addition of substrate.
The oxygen consumption of the liver increased by 50 to 60 patoms/lOO g rat weight per hour after a short delay, indicating increased flux through the cr-glycerophosphate shuttle.
Thus, xylitol provides the carbon substrate required to build up the intracellular a-glycerophosphate pool, and produces increased electron flux from the high potential flavin pool (14, 15) to oxygen. A second cycle of anoxia shows that both the flavin and pyridine nucleotide pools were incompletely reduced by the combination of rotenone and xylitol. Antimycin A and cyanide caused further successive increases in the state of reduction of both flavin and pyridine nucleotides along with successive inhibitions of the oxygen consumption.  6. Effects of rotenone, xylitol, antimycin A, and cyanide on flavin and pyridine nucleotide fluorescence and the elffuel oxygen tension of perfused rat liver. Rates of oxygen uptake in natoms per 100 g of rat weight per hour are given on the efiluen oxygen trace.
The data presented in Fig. 9 indicate that xylitol metabolism was active in the presence of rotenone due to removal of reducing equivalents from the cytosol by means of a flavin-linked transport mechanism, which may be identified as the cr-glycerophosphate shuttle. Fig. 10 shows measurements of the rates of glucose and lactate production from xylitol in a liver perfused with the flow-through system after successive additions of amobarbital (amytal) and cyanide. Other experiments showed that similar results were obtained with rotenone instead of amobarbital, and antimycin A instead of cyanide. Addition of amobarbital produced a rapid fall in the rate of glucose production from 86 to 40 ~moles/lOO g rat weight per hour and a rise of lactate production from 16 to 55 ~moles/lOO g rat weight per hour. Since amobarbital inhibits the reoxidation of mitochondrial NADH, oxidation of pyruvate is prevented, so that glucose and lactate become the only end products of xylitol met,abolism. The rate of NADH production from xylitol metabolism thus becomes equal to the rate of xylitol uptake, which may be calculated from the sum obtained by multiplying the rate of glucose production by 1.2 and the rate of lactate production by 0.6. These data are plotted as solid triangles in Fig. 10. Conversion of xylitol to pyruvate and subsequent oxidation of pyruvate in the citric acid cycle in livers perfused with xylitol alone may initially account for 500/, of the xylitol uptake, as previously shown (1). In this experiment, the calculated xylitol uptake uncorrected for pyruvate oxidation prior to addition of amobarbital was 100 pmoles/lOO g, rat weight, per hour, and decreased t,o 80 pmoles/lOO g rat weight per hour after amobarbital addition. Thus, the inhibitor certainly produced some decrease of xylitol metabolism although the exact amount cannot be calculated from the available data. Diminished glucose and increased lactate production can be accounted for by increased phosphofructokinase activity, since amobarbital is known to increase ADP and AMP levels in the liver (16). After inhibition of the mitochondrial electron transport chain by addition of cyanide, glucose production fell rapidly to about 5 pmoles/lOO g rat weight per hour, while lactate production initially increased (probably due to additional adenine nucleotide activation of phosphofructokinase), and subsequently decreased as the capacity of the liver to reoxidize NADH diminished.
The calculated xylitol uptake fell to 40 pmoles/lOO g rat weight per hour after cyanide addition, and decreased further by 75% upon disoxygenation of the liver. The small residual rate of lactate formation of 20 pmoles/lOO g rat weight per hour obtained with anaerobic perfusion probably represents a small oxygen leak from the surface of the liver or breakdown of residual glycogen. From this data we may conclude that the maximum rate of NADH reoxidation by extramitochondrial pathways is about 30 pmoles/lOO g rat weight per hour, and that flux through the a-glycerophosphate shuttle is about 40 ~moles/lOO g rat weight per hour.
Loss of carbon from xylitol by oxidation of pyruvate in the citric acid cycle can also be prevented by arsenite inhibition of lipoic dehydrogenase of the pyruvate dehydrogenase complex. Fig. 11 shows that the response of the liver to 0.2 mM arsenite infusion with xylitol as substrate was similar to that obtained after amobarbital inhibition, namely a decrease of glucose formation after an initial stimulation, and an increase of lactate as an artificial electron acceptor increased the rate of lactate formation and produced a small stimulation of oxygen consumption (from 270 to 380 patoms/lOO g rat weight per hour) indicating an incrraacd rate of NADH reosidation in the cytosol, presumably via the cytochrome P-450 and cytochrome bj electron transport pathways (17). DISCUSSION ConGderable confusion exists in the literature concerning the nature and control of the physiological pathways responsible for the transfer of reducing equivalents from the cytosol to the mitochondria in the intact cell (18). Largely on the basis of enzyme distribut'ion patterns in the cytosol and mitochondria of various tissues and the known impermeability of the mitochondrial membrane to nTAT)I-I, tliichrr and Klingenberg (19) postulated an indirect trannfrr of reducing equivalents into mitochondria via cr-glycerophosphate and malate. In a critical evaluation of a numbtlr of substrate oxidation-reduction cycles, Borst (12) concluded that the a-glycerophosphate cycle could account for all the glycolytic NADH oxidation in flight mu;;&, for part of it in rat liver, but for relatively little in rat heart.
The role of the cr-glS-cerol,hosphatc shuttle for transport of extramitochondrial NADH to t,hc rrspiratory chain is now well established for insect muscle (20,21), some tumor t,issuea (22,23), and yeast (24). The original postul& of a simple malatc-oxalacetate exchange across the nlitochondrial membrane was amended by Borst (25) on t,he grounds of the poor pc,rmeabilit,y of the mitochondrial membrane to oxnlncetate at physiological concentrations. blare recent observations have confirmed this point, although it is clear that oxalacetate transport can occur at nonphysiological concentrations (26,27). Borst proposed an intramitochondrial transamination of osalacetate with glutamate to form aspart'ate and ol-kctoglutarate, followed by transport of these products to the cytoxol for transamination back to glutamate and oxalacetate.
The malnte-aspartate cycle, therefore, involves an influx of rnalate and glutamate and efflux of oc-ketoglutarate and aspartate from the mitochondria.
Evidence for the existence of specific dicarboxylic anion transport mechanisms across t,hc membrane of most mammalian mitochondria has been obtained from a number of laboratories (28)(29)(30)(31)(32).
A possible difficulty with this type of NAD-linked shuttle can bc raised on thermodynamic grounds (12) since free permeability of each of the anions can result only in an equilibration of the oxidation-reduction potential on both sides of the membrane without appreciable transport.
On the basis of measurernrnts of the substrate partners of NAD-linked dehydrogenases located in the cytosolic and mitochondrial spaces, which may be used to calculate the oxidation-reduction potential of free iY,ZD (19,33), it is generally accept,ed that the mitochondrial NAD oxidation-reduction potential under a wide variety of metabolic conditions is considerably more negative than that of the cytosol.
This has been demonstrated, too, in perfused rat liver under conditions of ethanol (34) or xylitol (1) metaboliwm, when NAT11I transport into mitochondria is high. Transport of reducing equivalents into mitochondria, therefore, appears to be against a concentration gradient when mediated via NADlinked shuttles, and should be energy linked. So far no definitive measurements have been made of the stoichiometry of the energy requirements for the malate-aspartate shuttle, although an energy dependency has bern illustrated in the sense by guest on March 24, 2020 http://www.jbc.org/ Downloaded from that the rate of NADH transport has been shown to be inhibited by uncoupling agents with isolat,ed mitochondria and a reconstituted malate-aspartate shuttle (29,35,36). Studies with rat heart mitochondria showed that aspartate efflux was energy dependent, being inhibited when energy production by oxidative phosphorylation was prevented (37,38). As an alternative or additional mechanism, an energy drpcndcnce for the transport of various other anions a)cross the nlitochondrial membrane has been suggested (39)(40)(41).
Thus, a high energy state of the mitorhondrial membrane, or even R suitable proton or charge gradient (42,43) may be required for anion transport via the malate-aspartate shuttle. Although one or more anion transport steps have been shown to be energy dependent, the energy requirements may be small and Ilot proportional to flux. Furthermore, the fi-hydroxybutyrat,e dehydrogenase subst,rate couple, as normally measured in total tissue extracts or in the mitochondrial incubation medium, may not provide a valid index of the intramitochondrial NXD oxidation-reduction Ftate, even on the ,supposition of a single intrnmitoehondrial XX) pool (44). On the basis of a lack of direct energy utilization for the malateaspartnte shuttle, and an ,U) I'-cont,rollcd fully phosphorylating electron transport chain, KAl)IT transport via the malate-aspartat,e shuttle should be associated with a lower oxygen consumption than N.1DII transport via the a-glycerophosphate shuttle due to the diffcrcnt I':0 rwtios fnr NAD-and flavinlinked ,cubstrates.
Reccnlly, a third potentially important system for transport of XADIT into mitochondria has been proposed (45) and inrestigated with a reconstituted mitochondrial system (46). This mechanism involves chain elongation of fatty acyl-CoA deriva-t&s with the utilization of ?JAl)H in the rxtramitochondrial space, transfer of the fatty ac$CoX into the mitochondrial matrix, followed by fi oxidation to a Fhorter chain fatty acyl-CoA ester and its transport back into the cxtramitochondrial space for a fnrther cycle of chain elongation.
This system would be energetically favorable for inward transport of NADH because of t,he Aavin step of ,B oxidation, but suffers from the disadvantage that /3 oxidation of long chain fatty acids, once initiated, is thought) to proceed to completion (47, 48). As with the malateaspartnte cycle, the fatty acid chain elongation cycle should be rot,cnone sensitive.
Dat#a presented in this lpapcr show that the rate of xylitol metabolism in rat liver is controlled by the rate of reoxidation of NADH in the cytosolic compartment and not by the capacity of xylit 01 dehydrogrnase. This is most clearly seen from the large stimulation of glucose formation from xylitol induced by a varictp :)f artificial electron acceptors.
On the basis of previous work with a number of tissues this conclusion can probably be extended to other reduced substrates of cytosolic NAD-linked dehytlrogenases such as ethanol or lactate (49-51).
The large inhibition of calculated xylitol uptakr observed after addition of 1 1~~x1 cyanide or antimycin .Z indicates that most of the XADH reoxidation in the absence of artificial electron acceptors is linkctl Co mitochondrial respiration. The cyanide insensitive respiration nccountcd for the reoxidation of about 30 pmolcs of NADTT per 100 g rat weight, per hour (Fig. lo), and could not be greatly stimulated by addition of artificial electron acceptors (Fig. 11). Control of the rate of NADH reoxidation in the cytosol by the rate of the mitochondrial electron transport was also revealed by the stimulatory effect of low concentrations of the urlcoupling agent 2,4-dinitrophenol, and addition of orni-thine plus ammonia which induces an intramitochondrial utilization of ATP for citrulline synthesis. 2,4-Dinitrophenol has also been shown to increase ethanol utilization by rat liver slices (52) and rats in viva (53), suggesting that the rate-limiting step of et,hanol metabolism was the reoxidation of NADH to NAD rather than the capa& of the dehydrogenase. These authors also showed that most of t)he ethanol uptake by the rat liver slices was cyanide sensit)irc.
The incomplete inhibition of xylitol conversion to glucose by amobarbital suggests that part of t'hc NADH transport into the mitochondria was mrdiated by the or-glyccrophosphate shuttle. The dat,a in Fig. 10 provide a value of 40 prnoles/lOO g rat weight per hour for the flux, while Fig. 9 shows that xSlito1 addition in the prcscnce of rotenone increased respiration by 57 patoms per oxygen per 100 g rat wright per hour. However, part of the respiratory increase may be accounted for by a stimulation of the cyanide insensitive r~~~pirat~ion induced by t)he increased availability of cytosolic XhT)(l')H f, a +er substrate addition.
Earlier studies with rat liver mitochondria, slices or perfused rat liver (54)(55)(56) have indicated that the a-glycc,rophosphate shuttle has a very low capacity in this tisnlle, compared with the total rate of NADTI transport, c~en nhcn the mitochondrial cu-glycerophosphatc oxidase activity is incrcasetl many-fold, as in the hyperthyroid animal.
Furthermore, parallel studies with isolated liver mitochondria prepared from normal and hyperthyroid rats showed that cr-glycerophosphate osidasr activity at the or-glycerophosphatc concentrations preT-ailing in the liver was inadequate to account for the observed rates of [Z3H]glycerol-3-P detritiation in the perfuPed liver. The increased activity of the a-glycerophosphate oxidase in mitochondria of hyperthyroid rats was largely offset by an increase in the apparent K, for the substrate (from 0.6 to 3.4 mu).
It may be concluded, therefore, that. normally in rat liver the ol-glycerophosphate shuttle accounts for very little of the transport of NADH into rnitochondria unless tissue oc-glycerophosphate concentrations increase suflicicntly to produce Vn,,, flux through the mitochondrial ac-glycerophosphate oxidnse, as is the case apparently with xyli-to1 metabolism.
However, even with xylitol as substrate, when tissue oc-glycerophosphate conceutrations rise to 12 rnM (I), the strong competition observed betwren transport of NADH in mit,ochondria and intramitochondrial NXD-dehydrogenases indicates the involvement of NADlinked shuttles. Operation of the fatty acid chain elongation shutt,le, although unlikely, cannot be excluded from the present data. The initial inhibit,ion of glucose production from xylitol by aminooxyacetate (Fig. 7). clearly shows participation of the malate-aspartate shuttle in the transport of reducing cquivalmts from t'he cytosol to the respiratory chain.
Reversal of tht inhibition after about 10 min is probably caused by an increased contribution from t'hc a-glyccrophosphatc shut,tle as the cr-glgcerophosphate levels rise. A similar substitution of increased flux through the or-glycerophosphate shuttle upon inhibition of the malate-aspartate shuttle, associated with increased tissue a-glycerophosphate levels, has been obscrvcd in rat hearts perfused with glucose (58 take since there was a lack of precursors for ar-glycerophosphate formation. The greater inhibitory effect of P-hydroxybutyrate on glucose production from xylitol observed in Fig. 6, when flux was stimulated by ornithine and ammonia addition (cf. Figs. 1 and 2), suggests a greater contribution of the malate-aspartate shuttle possibly as a result of a change of the intramitochondrial pyridine nucleotides to a more oxidized state as availability of ADP increased due to the utilization of ATP for citrulline synthesis. The small increased rate of respiration observed after /I-hydroxybutyrate addition, together with an increase in the state of reduction of the mitochondrial pyridine nucleotides, suggests the possibility that the electron transport chain may not be fully saturated with substrate during xylitol metabolism.
The small effect of xylitol, and ethanol (56), on oxygen uptake by the liver indicates that NADH generation in the cytosol followed by transport to the mitochondria for oxidation can substitute for NADH generation by mitochondrial dehydrogenases without an appreciable loss of energetic efficiency, indicating that the energy requirements for operation of the hydrogen shuttles must be minimal.
Neither xylitol nor ethanol (1,34) are capable of suppressing completely intramitochondrial generation of reducing equivalents by the citric acid cycle dehydrogenases indicating that transfer of reducing equivalents into the mitochondria as well as flux through the mitochondrial electron transport chain are both involved in the regulation of the utilizat,ion of the reduced substrate.
The observation that P-hydroxybutyrate decreased gluconeogenesis from xylitol but increased it from lactate is of particular interest when the flux of anions across the mitochondrial membrane under the two conditions is taken into account. Transport of reducing equivalents via the malate-aspartate shuttle involves an influx of malate and glutamate and efflux of a-ketoglutarate and aspartate. Gluconeogenesis from lactate involves an influx of pyruvate and glutamate and efflux of o(ketoglutarate and aspartate. Since pyruvate carboxylase is the rate-limiting step for gluconeogenesis from lactate, the more reduced state of the mitochondrial NAD system after addition of @hydroxybutyrate must cause an activation at this site, possibly by a decrease of the inhibitor acetoacetyl-CoA (59). The increased flux of anion transports probably follows from the increased rate of generation of mitochondrial oxalacetate. On the other hand, inhibition of glucose formation from xylitol by P-hydroxybutyrate probably involves an interaction between malate dehydrogenase and &hydroxybutyrate dehydrogenase in the mitochondria and an inhibition principally of the malateaspartate shuttle, possibly by altering the malate gradient across the mitochondrial membrane (see also References 60,61). Flux in the phosphorylating electron transport chain, as controlled by the availability of phosphate acceptor, is a determinant of the intramitochondrial NAD oxidation-reduction state. It would appear, therefore, that transport of reducing equivalents into mitochondria via the malate-aspartate shuttle, being regulated by the mitochondrial NAD oxidation-reduction state, must also be regulated by the electron transport chain. Details of the factors affecting competition between the utilization of reducing equivalents delivered from the cytosol to the mitochondria and internally generated reducing equivalents from p oxidation and the reactions of the citric acid cycle require further evaluation, and are presumably determined by the control characteristics of the individual dehydrogenase and mitochondrial anion transport mechanisms.
Aclcnowledgments-Technical assistance was provided by Gary Gross. We are indebted to Dr. Britton Chance for advice and help with regard to adsorption measurements in the perfused liver.