transporter in the mitochondrial inner membrane pro- vides a potential site .... mM succinate (potassium salt), 1 mM EGTA, 5 p~ rotenone, 1 p~. TPMP (bromide ..... mM o-octanoyl-L- carnitine chloride (carnitine carrier) (351, 500 pM a-cyano-4-.
Vol. 267, No. 21, Issue of July 25, pp. 14637-14646, 1992 Printed in U.S.A.
OF BIOLOGICAL CHEMISTRY THEJOURNAL
0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.
Choline Transportinto Rat Liver Mitochondria CHARACTERIZATION AND KINETICS OFA SPECIFICTRANSPORTER* (Received for publication, February 14,
1992)
Richard K. Porter$., JohnM. Scott, and MartinD. Brands From the Department of Biochemistry, Trinity College Dublin, Dublin 2, Ireland and the §Department of Biochemistry, University of Cambridge, Cambridge CB21 Q W, United Kingdom
Rat liver mitochondriapossessa specific choline of rats (5). Homocysteine can also be remethylatedby methitransporter in the inner membrane. The transporter onine synthase (EC 2.1.1.13), also a cytosolic enzyme, which shows saturable kinetics at high membrane potential is present in all mammalian cells (4).Methionine synthase is with a K, of 220 W M and a Vmaxof 0.4 nmol/mg of a cobalamin-dependent enzyme, using 5-methyltetrahydrofolprotein/min atpH 7.0 and 25 OC. At physiological con- ate asa methyl donor, and is inactivated in vivoby inhalation centrations of choline, the rate of choline uptake by of nitrous oxide, which irreversibly oxidizes the cobalamin the transporter shows a linear dependence on memcofactor during catalysis ( 6 ) . brane potential; uptake is distinct from the nonspecific Inhalation of nitrous oxide has been shown to produce a cation diffusion process. Hemicholinium-3, hemicholi- neurological lesion in humans (7), monkeys (8))pigs (9), and nium-15, quinine, and quinidine,all analogues of cho- fruit bats (10) similar to that found in humans with vitamin line, are high affinity competitive inhibitors of cholineBlz deficiency and in patients with AIDS dementia complex transport with K i values of 17, 55, 15, and 127 PM, respectively. The choline transporter is distinct from (11).However, rats exposed to nitrous oxide do not develop the neurological lesion (12). Weir et al. (9) measured levels of other knownmitochondrialtransporters.Ratheart mitochondria do not appearto possess a choline trans- S-adenosylmethionine (AdoMet)’ and S-adenosylhomoporter. Evidence suggests that the transporter is an cysteine (AdoHcy) in various tissues from pigs and rats exelectrophoretic uniporter. Analogue studies have posed to nitrous oxide and showed markedly reduced AdoMet/ AdoHcy ratios in neural tissue of pigs compared to controls. shown that the hydroxyl and the quaternary ammoHowever, near normal AdoMet/AdoHcy ratios were measured nium groups of choline are necessary for binding to the transporter. A comparison of molecular models of cho-in neural tissue of rats. Many in vivo methylation reactions line and the high affinity inhibitors has provided evi- are controlled by the ratio of AdoMet to AdoHcy (13)) and dence for the preferred conformation of choline for markedly elevated AdoHcylevelswere measured in neural binding to the transporter. The presence of a choline tissue of pigs exposed to nitrous oxide; on the other hand, transporter in the mitochondrial inner membrane proAdoMet levels were comparable to those in controls (9). The vides a potential site for control of choline oxidation products of all AdoMet-dependent methylation reactions are and hence supply of endogenous betaine. AdoHcy and methylated product. AdoHcy is hydrolyzed to adenosine and homocysteine, and tissue levels of AdoHcy are dependent onthe rateof removal of the product homocysteine (13). In mammals, the main route for removal of homocysteine In mammals, choline is a precursor for the biosynthesis of is by remethylation to methionine or its degradation via the neurotransmitter acetylcholine and the membrane phos- cystathionine to cysteine (14). As already mentioned, homopholipids phosphatidylcholine, sphingomyelin, and plasmal- cysteine can be remethylated by two enzymes, methionine ogens (1).In addition, duringits oxidation, the methyl groups synthase and betaine-homocysteine methyltransferase. Conof choline can be salvaged and utilized in carbon-1metabolism sequently, in the presence of nitrous oxide, the latter enzyme (1). provides the only means by which to remethylate homocysCholine oxidation occurs mainly in theliver (2) and kidney teine tomethionine. (3), where it is oxidized by FAD-linked choline dehydrogenase Administered betaine has been efficacious in treatment of (EC 1.1.99.1) to betaine aldehyde, which in turn is oxidized homocystinuria in humans (15) and has delayed onset of the t o betaine by betaine-aldehyde dehydrogenase (EC 1.2.1.8), neurological lesion in fruit bats (16). However, betaine is not an NAD-linked enzyme. Betaine canact asa source of methyl present in significant amounts in normal diets; and thus, the groups by remethylating homocysteine to methionine via the usual source of betaine is by oxidation of choline (17). The enzyme betaine-homocysteine methyltransferase (EC 2.1.1.5) site for choline oxidation to betaine is in the mitochondria. (4).This enzyme is a cytosolic enzyme (5) and is present in Choline dehydrogenase is situated on the inner side of the the liver and kidney of humans andpigs, but only in the liver inner mitochondrial membrane (IS), whereas betaine-aldehyde dehydrogenase activity is in the mitochondrial matrix * This work was supported by Alexander Porter, Grainne Porter, and Bernie Butler; the Luker-Cobbe Bursary, Trinity College (Dub- (19). Consequently, for oxidation to occur, choline must first lin) and Girton College (Cambridge); and the Trinity Trust, Trinity traverse the mitochondrial inner membrane. The knowledge College (Dublin). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact. $ TOwhom correspondence shouldbe addressed Dept. of Biochemistry, University of Cambridge, Tennis Court Rd., Cambridge CB2 lQW, United Kingdom. Tel: 0223-333649;Fax: 0223-333345.
The abbreviations used are: AdoMet, S-adenosylmethionine; AdoHcy, S-adenosylhomocysteine; A*, electrical potential across the mitochondrial inner membrane; EGTA, [ethylenebis(oxyethy1enenitrilo)ltetraacetic acid; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; TPMP, triphenylmethylphosphonium.
14637
14638
Mitochondrial Choline Transporter
of the mechanism of choline transport is essential to our methanol/NH3 (lO:lO:l, v/v). RFvalues were determined from unlaunderstanding of choline oxidation and hence supply of en- beled standards.The approximate RF values for choline, betaine dogenous betaine to betaine-homocysteine methyltransferase. aldehyde, and betaine are 0.25, 0.4, and 0.65, respectively. Once the separation was completed, theplate was dried inair,and 1-cm2 DeRidder (20) was the first to suggest the existenceof such a samples of silica were scraped into scintillation vials to which 4 ml cholinetransporter.Later work by Tyler (21) and,most of scintillant was added. The vials were left for 24 h and mixed before recently, by Brown and Brand (22) seemed to contradict this counting. Molecular Modeling-Modeling was carried using the program view. In this paper, we present comprehensive evidence for the Macromodel (27) onan Evans & Sutherland PS390 graphics terminal existence of a specific choline carrier in the inner membrane with a VAX/VMS computer. The choline and quinine models were derived from crystallographic data (28). The model of hemicholiniumof rat liver mitochondria. 15 was built usingMacromodel as no crystal structure could be found. The model of quinidine was made by inverting the quinine crystal structure about its chiral center. The choline moieties in quinine, Isolation of Mitochondria-Standard methods were used for the quinidine, and hemicholinium-15 were superimposed on the crystal preparation of mitochondria from rat liver (23) and rat heart (24). structure of choline using Macromodel. Mitochondria, from both tissues,were isolated from 200-250-g female Materiak-[methyl-'4C]Choline chloride (55 mCi/mmol), [U-"C] Wistar rats in an ice-cold medium containing 250 mM sucrose, 5 mM sucrose (215 mCi/mmol), and 'H20 were from Amersham InternaTris,and 1 mM EGTAadjusted topH 7.4 with HCI at 25 "C. tional (Amersham, Buckinghamshire, United Kingdom). ['HITPMP Mitochondrial protein was assayed by the biuret method (25) against bromide (52.8 Ci/mmol) was from New England Nuclear Chemicals a bovine serum albumin standard. (Hamburg, Federal Republic of Germany). Amytal was from Eli Lilly Measurement of Mitochondrial Matrix Volume-The mitochonCo. Ltd. (Basingstoke, Great Britain). Carnitine and o-octanoylcardrial matrix volumes were determined using the following procedure. nitine were from Otsuka Pharmaceutical Co. (Tokushima, Japan). Mitochondria (3 mg of mitochondrial protein/ml) were incubated in Nigericin and carboxyatractyloside were from Boehringer Mannheim a medium containing 120 mMKC1, 5 mM HEPES/KOH, pH 7.0, 5 (Mannheim,FRG).FCCP was from Aldrich Chemical Co. (GilmM succinate (potassium salt), 1 mM EGTA, 5 p~ rotenone, 1 p~ lingham,Dorset, UK). KC1 and succinic acid were from Fisons T P M P (bromide salt), [U-'4C]sucrose (0.1 pCi/ml), 3H20 (1.0 pCi/ Scientific Equipment (Leicestershire, UK). KH2P0,, MgC12, N,N'ml), and nigericin (100 pmol/mg of mitochondrial protein) and stirred dicyclohexylcarbodiimide, quinine, and LaC13were from British Drug a t 25 "C. All incubations were performed in a single vessel from which Houses (Poole, Dorset, UK). All other chemicals were from Sigma 1 ml samples were taken at defined time intervals and placed in 1.5- Chemical Co. (Poole, Dorset, UK). Polygram si1 G thin-layer chroml minifuge tubes. Experiments were stopped by brief centrifugation matographic plates (0.25-mm silica gel) were obtained from Machin abenchtop centrifuge (1 min at 10,000 X g) that pelleted the erey-Nagel & Co. (Duren, FRG). mitochondria. 500 p1 of the supernatant was removed and added to 4.0 ml of scintillant in a scintillation vial. Residual supernatant was RESULTS removed with a tissue. The mitochondrial pellet was resuspended in Accumulation of Choline by Rat Liver Mitochondria-Mi40 pl Triton X-100 (20%, v/v), and the base of the minifuge was cut off into a scintillation vial containing 4.0 ml scintillant. The vials tochondria were incubated ina highpotassium medium in the were counted on a Pharmacia LKB Biotechnology 1217 Rackbeta presence of nigericin and succinate as described under "Exscintillation counter set up for double isotope counting with appro- perimental Procedures." Addition of nigericin to mitochonpriate crossover and quenchcorrections. The matrix volume was dria suspended ina high potassium medium abolishes the pH calculated from the difference in the 3H20 space and ['4C]sucrose gradient, and as a result, A* increases so that the protonspaces as described by Brown and Brand (26). Simultaneous Measurement of Radiolabeled Choline Uptake and motive force equals A 9 (29). A time course of [rnethyl-14C] Mitochondrial Membrane Potential-Radiolabeled choline uptake ex- choline uptake by rat liver mitochondria showed that radioperimentsand mitochondrialmembranepotentialmeasurements label was accumulated in the matrix at high membrane potenwere determined using the conditions described for mitochondrial tial (Fig. 1).The uptake began to plateau at 15 min with an volume measurements, except that ['HITPMP, iodide salt, (0.2 pCi/ apparent maximumcholine accumulation ratioof %fold, corml), and [methyl-'4C]choline were present instead of radiolabeled water and sucrose. Experiments were initiated by addition of radio- responding to a matrix concentration of 8 mM. Thin-layer labeled choline chloride, the concentration and specific activity of chromatography was performed on matrix and medium samwhich are described in the figure legends. Choline uptake was meas- ples at defined time points during the time course. Fig. 2A ured from the accumulation ratio of [methyl-"C]choline across the shows that choline, although accumulated, was also oxidized mitochondrial inner membrane. Choline accumulation was measured to betaine within the matrix despite the presence of 5 pM a s the [methyl-"Clcholine space minus the [U-'4C]sucrose space, the result of which was divided by the matrix space. The [U-'4C]sucrose rotenone. Furthermore, radiolabeled betaine, produced as a medium after space was determined from the matrix volume experiments, which result of choline oxidation, was detectable in the 15 min, and its concentration increased linearly with time were performed in parallel. The membrane potential was measured simultaneously with choline uptake from the distribution of the (Fig. 2B). lipophilic cation ["HITPMP. Themembrane potential was calculated Addition of the uncoupler FCCP in the presence of nigericin a s described by Brown and Brand (26). 0 dissipated the membrane potential completely. a t time Mitochondrial Swelling in Choline Acetate-Mitochondria (3 mg of Under these conditions, choline uptake occurred only until protein) were incubated in 3 ml of 120 mM choline acetate, 5 mM HEPES, 1 mM EGTA, 1 p~ T P M P bromide, 5 p M rotenone, and 1 equilibrium was established, i.e. -1-fold accumulation (data pg of oligomycin/ml, pH 7.0, and the initial rate of absorbance was not shown). Addition of FCCP at 20 min, during energized followed at 540 nm afteraddition of 30 pl of 500 mM succinate uptake, caused immediate efflux of choline from the matrix (choline salt). Membrane potential and matrix volume measurements (data not shown). were performed in a parallel experiment as described by Brown and Choline Uptake by Rat Liver Mitochondria: Effect of Choline Brand (22). mitochondria were suspended in a Thin-layer Chromatography-After choline uptake was stopped, a Concentration-Rat liver presence of 5-pl sample of the supernatant was removed for thin-layer chromat- range of choline acetate concentrations in the ographic analysis. The remaining supernatant was removed, and the rotenone. Initial swelling rates were measured as a function pellet was dried as already described. 50pl of trichloroacetic acid of choline concentration after addition of succinate. It was ( 5 % , w/v) was added to the pellet in the minifuge tube, mixed, and shown that the initial rate of swelling of mitochondria was centrifuged in the benchtop centrifuge for 5 min. A 5-yl sample of directly proportional to the concentration of choline in the the resulting supernatant was transferred to a thin-layer chromatographic plate. Thin-layer chromatography was performed on 0.25- medium (Fig. 3). Furthermore, by titration of the membrane potential (A*) with FCCP, it was confirmed that the initial mm silica gel plates. Therunning solvent was NaCl (6%, w/v)/ EXPERIMENTALPROCEDURES
Transporter CholineMitochondrial 220 200 180 160 140 120
- 8 - 7
- 6 - 5 - 4
100
- 3
80 60
2 0
14639
4 7 20
-
0
0
5 1 0 1 5 2 02 53 03 5
4 0 4 55 0
0
5
10
15
tirne(rnin)
2 02 5
30 35
44 05
time(min)
FIG. 1. Time course of choline accumulation by rat liver mitochondria at high membrane potential.Rat liver mitochondria (90 mg of protein) were incubated in 30 ml of 120 mM KCI, 5 mM HEPES/KOH, pH 7.0, 5 mM succinate (potassium salt), 1 mM EGTA, 5 p~ rotenone, 1PM T P M P (bromide salt), [3H]TPMP(iodide salt) (0.2 pCi/ml), and nigericin (100 pmol/mg of protein) at 25 “C. The experiment was initiated by addition of 1 mM choline chloride containing [methyl-14C]cholineto give a final specific activity of 80 pCi/mmol. Samples (1 ml) of incubation medium were subsequently taken at 1-,5 - , and 5-min intervals and processed as described under “Experimental Procedures.” The graph shows the accumulation of choline ( 0 )and membrane potential (0)as a function of time. The experiment was repeated for three separate mitochondrial preparations, and each determination was made in triplicate (mean f S.E., n = 3).
0 0
swelling rate shows a non-ohmic relationship with membrane potential (Fig. 4) (22). Radiolabeled choline uptake was also measured as a function of choline concentration, but using much lower concentrations within the physiological range. The rate of choline uptake was measured a t high membrane potential and showed saturation kinetics (Fig. 5 ) . A K , of 220 p~ and a V,, of 0.4 nmol/mg of protein/min were calculated for choline uptake data fitted to a rectangular hyperbola for Michaelis-Menten kinetics using the Enzfitter program(Elsevier Biosoft, Cambridge, UK). The standard errors associated with the curve fit aref 3 0 pM for the K, and f0.02nmol/mg of protein/min for the V,,,,,. The rate of radiolabeled choline uptake was also measured asa function of membrane potential using a choline concentration of 1 mM. The rate of uptake showed a linear dependence on membrane potential, and a rate of uptake of -0.1 nmol/mg of protein/min could be measured at zero potential (Fig. 6). Choline Uptake across Mitochondrial Inner Membrane: Effect of Specific High Affinity Analogues-A number of high affinity inhibitors of mitochondrial choline transport were identified. These compounds areanalogues of choline. Fig. 7A shows a Lineweaver-Burk plot (30) for the choline uptake processinthe presence andabsence of the high affinity inhibitor hemicholinium-3. Inhibition by hemicholinium-3 is competitive with a K, of 17 pM. The Kivalues for the other high affinityinhibitors of choline uptakewere calculated using Dixon plots (31). The type of inhibition was determined using Cornish-Bowden plots (data not shown) (32). Quinine is a competitive inhibitor of choline uptake with a Kiof 15 p~ (Fig. 7 B ) . Hemicholinium-15 is a competitive inhibitor with a K; of 55 p~ (Fig. 7C), and quinidine, a stereoisomer of quinine, shows competitive inhibition with a I C i of 127 p~ (Fig. 7 0 ) . The inhibitoryeffects of many other compounds on choline uptake were also investigated; the results are summarized in Tables I and I1 and will be referred to under “Discussion.” Effect of Other Mitochondrial TransporterInhibitorson
5
1 0 32150505
40
45
time (min)
FIG. 2. Thin-layerchromatographicanalysis of choline time course. Rat liver mitochondria (30 mgof protein) were incubated in 10 ml of medium as described for Fig. 1. The experiment was initiated by addition of 1mM choline chloride containing [nethyl“Clcholine to give a final specific activity of 0.8 mCi/mmol. Samples were taken at the times indicated, and thin-layer chromatography was performed as described under “Experimental Procedures.” A, total radiolabeled compounds (M) and the amountof the total thatis choline (0), betaine (O),and betaine aldehyde (0)in the matrix as a function of time; B, total radiolabeled choline transported into the matrix (B) and total radiolabeled choline transported into thematrix plus effluxed radiolabeled betaine (A). The experiment was repeated for two separate mitochondrial preparations (mean and range, n = 2).
Choline Uptake-To investigatewhether choline istransported across the mitochondrial inner membrane on other known transport systems, choline uptake was tested in the presence of inhibitors of these transporters. The following inhibitors were investigated and were found to be without effect: 30 pM mersalyl (phosphate carrier) (33), 90 p M N ethylmaleimide (phosphate carrier) (34), 5 mM o-octanoyl-Lcarnitine chloride (carnitine carrier) (351, 500 pM a-cyano-4hydroxycinnamate (pyruvate and carnitine carriers)(36, 37), 2 p~ carboxyatractyloside (adenine nucleotide translocator) (38), 1.5 pM ruthenium red (calcium uniporter) (39), 120 p~ lanthanum chloride (calcium uniporter) (40), 150 p~ p-chloromercuribenzoate (neutral amino acid carrier) (41), 10 mM butyl malonate(dicarboxylate carrier) (34), and 10 mM phenyl succinate (a-ketoglutaratelmalate exchanger) (42), but 10mM carballylic acid (tricarboxylate carrier) (43) lowered A 0 and inhibited choline uptake rate by 30%.Choline uptake was also investigated in the presence of L-ornithine at 10 times the K , value for the ornithine carrier (44); however, choline uptake was not inhibited. Furthermore, in the presence of 10 mM L-carnitine, choline uptake was not inhibited (in contrast with Ref. 19). N,N‘-Dicyclohexylcarbodiimide(6 PM) (45) and oligomycin (1 p M ) (46), which inhibit ATPase activity, likewise had no effect on choline transport. Finally, the rate
Mitochondrial Choline Transporter
14640 l4
0.35-
r c
0.0
0
I
I
I
I
I
25
50
75
100
125
[Choline]mM FIG. 3. Rate of mitochondrial swelling at high membrane potential as function of choline acetate concentration. Rat liver mitochondria (3 mg of protein) were suspended in 3 ml of 25-125 mM choline acetate with osmolarity being maintained by corresponding additions of sucrose, 5 mM HEPES/KOH, pH 7.0, 1 mM EGTA, 5 p~ rotenone, 1 p M TPMP at 25 "C, and swelling was initiated by addition of 30 ~1 of 500 mM succinate (choline salt)to a final concentration of 5 mM. Initial swelling rates were determined from the change in percent transmission in the first minute after succinate addition. The change in percent transmissionwas converted to uptake in nanomoles/milligram of protein/minute by reference to a plot of inverse osmolarity against percent transmission. The line through the points was fitted by eye to pass through the origin. The experiment was repeated for three separate mitochondrial preparations (mean k S.E., n = 3). 0.04
0.2
0 4 0.6 0.8 [Choline] mM
1.0
FIG. 5. Uptake of radiolabeled choline at high membrane potential as function of choline concentration. Mitochondria (18 mg of protein) were incubated in 6 ml of medium as described for Fig. 1. The experiment was initiated by addition of 0.01-1 mM choline chloride containing [methyl-'4C]cholineto give a final specific activity of 8-0.08 mCi/mmol. The curve was fitted to a rectangular hyperbola for Michaelis-Menten kinetics using the Enzfitter program. A K , of 220 p M and a V,, of 0.4 nmol/mg of protein/min were calculated for the transporter. The experiment was repeated for eight separate mitochondrial preparations (mean -C S.E., n = 8).
0.5 0.6
r 0 0 0
0
30
60
90
120 180150
membranepotential (mV)
FIG. 6 . Rate of choline uptake at physiological concentrations of choline as function of membrane potential. Mitochon-
0 0 0.00 L 0 L
0
0 . .
I
50 150
.
1
.
1
100
1
200
membranepotential (mV)
FIG. 4. Rate of mitochondrial swelling incholine acetate as function of membrane potential. Rat liver mitochondria (3 mg of protein) were suspended in 3 mlof 125 mM choline acetate, 5 mM HEPES/KOH, pH 7.0, 1 mM EGTA, 5 p~ rotenone, 1 pgof oligomycin/ml, 1 p~ TPMP,[3H]TPMP (iodide salt) (0.2 pCi/ml) at 25 "C, and swelling was initiated by addition of 30 p1 of 500 mM succinate (choline salt) to a final concentration of 5 mM. The membrane potential was titrated with FCCP (0-100 pmol/mg of mitochondrial protein), and swelling rates and volume determinations were determined as previously described. The graph represents the results from a single experiment, and each determination was made in triplicate.
dria (18mg of protein) were incubated in 6 ml of medium as described for Fig. 1. The experiment was initiated by addition of 1 mM choline chloride containing [methyl-"Clcholine to give a final specific activity of 80 pCi/mmol. The membrane potential was titrated by addition of FCCP (0-100 pmol/mg of protein). The line through the points was fitted by linear regression. The experiment was repeated for three separate mitochondrial preparations (mean S.E., n = 3).
*
porter and Kf uniporter processes using experimental conditions as described (47, 48). In both cases, hemicholinium-3 and hemicholinium-15 had no effect on either the K+/H+ antiporter or the K+ uniporter. Quinine, on the other hand, inhibited both these transporters to the extent reported in the literature (47, 48). It was further shown that hemicholinium-3, hemicholinium-15, and quinine had noeffect on state 3 and 4 respiration rates, indicating that they had no effect on the dicarboxylate carrier or the adenine nucleotide transof choline uptake did not differ when the HEPESbuffer was locator. changed to a phosphate buffer. Therefore, HEpES did not Tissue specificity of ChQline Transporter-In vitro activity inhibit uptake of choline despite the similarity in their stmc- of choline dehydrogenase is negligible in rat heart, whereas high activity has been reported for rat liver (2). Therefore, tures. Quinine, however, isknown to inhibit the K+/H+antiporter mitochondrial choline transport activity by these two tissues in Me-depleted mitochondria (47), and quinine and quini- was compared. Rat heart mitochondria were incubated under dine partially inhibit the K+ uniporter (48). To investigate the same conditions used to investigate [rnethyl-14C]choline the possibility that choline is transported on these transport uptake in liver mitochondria. Rat heart mitochondria did not systems, we compared the inhibitory effects of hemicholi- accumulate choline. After a 10-min incubation, choline accunium-3, hemicholinium-15, and quinine on the K+/H+ anti- mulation in rat heartmitochondria was -4% that of rat liver
:iyy 14641
Mitochondrial Choline Transporter A Competitive Ki for hemicholinium-3=17pM ,. ,
B2 . p 1 5 4
2m sm
O
p
fb
M inhibitor
35
12
-4:,z m
x -
2 0 10
-
._ .s
:s
-
4
-
- 0
2 s
-6 -5 -4 -3 -2 -1 0
1 2
t/ICholine] mM'
30 r
400pM choline
2 .=A .-
OpM inhibitor
I
C
Competitive
.nIB
:
.g x
lOOpM choline
-10
-20
3 4 5
0 10 20 [Quinine] pM
'
/
30
40
100pM choline
- 2 0 0 -100 0 100 2 0 0 300 400 [Hemicholinium-15] pM
. 0
-100
-150
.
8
.
'
1
.
-50 0 [Quinidine) pM
50
100
FIG. 7. Determination of Ki values for high affinity inhibitors of choline transport. Mitochondria (18 mg of protein) were incubated in 6 ml of medium as described for Fig. 1. Inhibitors were present prior to addition of choline, and uptake rates were measured as the difference in uptake between the first and the fifth minute. A, Lineweaver-Burk plot of choline uptake rate in the presence ( 0 )and absence (0)of 50 pM hemicholinium-3; B , Dixon plot of choline uptake rate in the presence of 10, 20, 30, and 40 pM quinine at choline concentrations of 100 pM ( 0 )and 400 p~ (0);C, Dixon plot of choline uptake rate in the presence of 100,200,300, and400 FM hemicholinium15 at choline concentrations of 100 pM (0)and 400 p~ (0);D,Dixon plot of choline uptake rate in the presence of 25, 50, 7 5 , and 100 p~ quinidine at choline concentrations of 100 pM ( 0 )and 400 p~ (0).The lines through the pointswere fitted by linear regression. Experiments were repeated for three separate mitochondrial preparations (mean, n = 3).
from the matrix, demonstrating that accumulation is an electrophoretic process. Conflicting reports have previously appeared in the literaDISCUSSION ture on the inhibitory effects of the respiratory chain inhibitor rotenone on choline oxidation in isolated mitochondria (49A time course of choline uptake was measured at high membrane potential. Fig. 1 shows that at high A*, in the 52). It is clear from our results that choline is being oxidized presence of rotenone. Theinsensitivity presence of succinateandrotenone, choline is maximally to betaine despite the of choline oxidation to respiratory chain inhibitors may be accumulated &fold, equivalent toa matrix concentration of 8 Qiu and mM. The timecourse began to plateau after15 min, although explained by two observations in the literature. First, Lin (53) showed that choline dehydrogenase can reduce atthere is no reduction in membrane potential. Similar time mospheric oxygen in vitro and suggested that the soluble courses were measured when(i)rotenone was substituted with Amytal and the mitochondria were energized with suc- enzyme in vivo can use oxygen as an acceptor. Second,Tsunge et al. (54) have reported that betainealdehyde can be oxidized cinate, (ii) rotenone plus myxothiazol were present and the by choline dehydrogenase, thus providing a means by which mitochondria were energized with ascorbate plus TMPD, and to circumvent the betaine-aldehydedehydrogenase NAD-de(iii) when rotenone plus cyanide were present and the mitopendent oxidation step. An example of such a system does chondria were energized with an ATP-regenerating system. exist in nature. The choline dehydrogenase enzyme of EschAnalysis of the time course energized with succinate in the erichia coli can reduce atmospheric oxygen and oxidize betaine presence of rotenone, using thin-layerchromatography, aldehyde to betaine (55), and we propose that this is also showed that althoughcholine was being transported into the happening in our experiments. matrix, it was also being metabolized to betaine (Fig. 2 A ) . In The concentrationdependence of the choline uptake system addition, betaine produced as a result of choline oxidation was theninvestigatedtoestablishwhetheruptake was a was detectable in the medium after 15 min, and its concentra- saturable or nonsaturable process. Brown and Brand(22) had tion increased linearlywith time(Fig. 2B). Therefore, would it shown, using swelling experiments, that cations can diffuse appear that choline is being taken up continuously over the across the mitochondrial inner membrane athigh A*. Fig. 3 time period and that the plateau, apparently representing shows that at high A*, the swelling rate is a linear function maximum accumulation, is in fact due to efflux of radiolabel of external choline concentration,indicatingthat choline in the formof betaine. diffuses across the mitochondrial inner membrane under these In the absence of A*, after the addition of the uncoupler conditions. Furthermore,by titration of the A* with FCCP, FCCP, choline uptake occurs only until equilibrium is estab- it was confirmed that the initialswelling rate is a non-ohmic lished, i.e. -1-fold accumulation. Addition of FCCP at 20 min, function of A* (Fig. 4) (22). during energized uptake, caused immediate efflux of choline However, radiolabeled choline uptake was also measured as mitochondria at the same time point, a figure that probably represents the contributionof nonspecific uptake.
14642
Mitochondrial Choline Transporter TABLE I
Percent inhibition of choline uptake and structural formulas for a range of inhibitors of choline uptake Mitochondria (18 mg of protein) were incubated in 6 ml of standard incubation medium and under the conditions described under “Experimental Procedures.” Analogues were present prior to initiation of uptake at a concentration of 10 mM. The experiment was initiated by addition of 500 p M choline chloride containing [methyl-14C]cholinea t a specific activity of 0.2 mCi/mmol. Values represent percent reduction in the choline uptake rate in the presence of inhibitor. The results are from two separate mitochondrial preparations; each determination was made in triplicate (average, n = 2). INHIBITOR
RO INHIBITION
STRUCTURE
45%
35%
NN-DIMETHYWMINOPROf’ANPANOL
30%
LcARNlTINE
14%
BETAWE ALDEHYDE
13%
ETHANOLAMINE
8%
N,WDIETHYL€THANOLAMINE
6%
TRIGTHANOLAMINE
0%
BET-
0%
+ H&-
M2-
N-UE~HANOLAMINE
a function of choline concentration, but using much lower concentrations within the physiological range (60-340 PM) (2, 21). The rateof choline uptake was measured as thedifference in choline accumulated within the matrix between the first and the fifth minute. It was assumed that binding occurred
CHz-
P
I”-
09i
0%
within the first minute of incubation. This procedure automatically corrected for the small amount of apparent uptake due to binding of radiolabel. Furthermore, because the time course showed linearity for the first 10 min of uptake (data not shown), it was assumed that initial choline uptake rates
Transporter CholineMitochondrial
were being measured. The rate of choline uptake was measured over a range of choline concentrations at high A\k and showed saturation kinetics (Fig. 5). This indicated the involvement of a carrier-mediated process. A K,,, and VmaXof 220 p~ and 0.4 nmol/mg of protein/min, respectively, were calculated for this saturable choline transport process. The
14643
rate of this choline-saturable uptake process was also measured as a function of A*. By titration of A* with FCCP, it was shown that the rateof carrier-mediated choline uptake is a linear function of A* (Fig. 6). There are thus two modes of choline accumulation: a high capacity nonsaturable diffusion process showing a nonlinear
14644
Transporter CholineMitochondrial
dependence on A s and a low capacity high affinity saturable dria, it has been shown that betaine efflux is not coupled to carrier-mediated process linearly dependent onA s . From the choline uptake (62). Other putative pathways might include slope of the graph showing the cholineswelling rate as a cotransport of choline with other cations present in the mefunction of choline concentration (Fig. 3), it is possible to dium such as K+. However, matrix accumulation of choline make a rough estimate of the choline diffusion rate at high has been reported innonionic low K+ media (20)to equivalent A s . At a K,,, of 220 pM, the estimated diffusion rate from the concentrations seen in a high K' medium (Fig. 1).In short, graph is 0.02 nmol/mg of protein/min, which is -10% of the the sole driving force forcholine uptake a t physiological total choline uptake rate at that concentration. Therefore,at concentrations of choline would appear to be the choline physiological choline concentrations (60-340 p ~ (2,) 21), the electrochemical potential. The results are therefore consistent carrier-mediated process is dominant. with this carrier-mediated process being a cholineuniporter. In addition, a number of high affinity inhibitors of mitoHigh choline dehydrogenase activities have been measured chondrial choline transport were identified. The inhibitors for rat liver and kidney (2, 3). If the rates of choline uptake with the highest affinity were found to be hemicholinium-3 and oxidation obtainedin vitro are applicable to intact cells, ( K , = 17 pM) (Fig. 7 A ) , hemicholinium-15 (Ki = 55 p M ) (Fig. then thecholine transporter may have significant control over 7 B ) ,quinine ( K ,= 13 p M ) (Fig. 7C), and quinidine(Ki = 127 choline oxidation. However, further experimentation will be p M ) (Fig. 70); they are all competitive inhibitors of choline required before the contributionof the choline transporter to uptake. These inhibitors are also known to inhibit choline control of betaine supply to betaine-homocysteine methyltransport across plasma membranes (56, 5 7 ) , and allhave transferase is known. structural similarity tocholine, a topic that will be discussed In an endeavor to map the choline-binding site of this later. transporter, selected choline analogues were investigated for To establish whether the choline transporter was unique or inhibitory effects on choline transport. Table I shows the whether choline was being transported on another liver mi- chemical formulas and the degree of choline uptake activity tochondrial transport system, uptake was investigated in the in the presenceof these analogues. Table I1 shows the strucpresence of known mitochondrial transporter inhibitors. The tures andK, values of choline analogues with high affinity for results show that noneof the inhibitors investigated inhibitedthe choline transporter.Thepercentinhibition values in choline transport. The threeexceptions were carballylic acid, Table I can be compared toKivalues in Table I1 by reference quinine, and quinidine. Carballylic acid apparently inhibited to N,N-dimethylaminopropylchloride, which has a Kiof 660 M reduces choline uptake rate to55%. First, the necesthe choline uptake rate, but further investigation showed that ~ L and sity of the quaternary nitrogen for binding of choline to the carballylic acid decreasedA\k and that the apparent inhibition could be attributed to thisAs decrease. Quinine, on the other binding site was investigated. The degree of inhibition by N, N-dimethylethanolamine(deanol)(35%),N-methylethanohand, is known to inhibit the K+/H+ antiporter (47), and quinine and quinidine partially inhibit theK' uniporter (48). lamine (O%), and ethanolamine (8%)demonstrates that seHowever, because of the insensitivityof these transporters to quential removal of the methyl groups from the quaternary hemocholinium-3 and hemicholinium-15, it was concluded nitrogen reduces affinity for the transporter. Thisconclusion that the choline transport system is different than those for is corroborated further by the observation that triethanolaK' transport and K+/H+ exchange. It was also shown that mine, a tertiary ammoniumanalogue, doesnot inhibit uptake, hemicholinium-3 and quinine had no effect on state 3 and 4 whereas tetramethylammonium (Table II), a competitive inrespiration rates. In isolated mitochondria, respiring on suc- hibitor of choline uptake, shows a greater degree of inhibition = 459 p ~ >55%) ; than does N,N-dimethylethanolamine. cinate, thedicarboxylate and adeninenucleotide carriers have (Ki substantial controlover state 3 respiration (58). An effect on Second, a comparison of inhibitory action of N,N-dimethylthe respiratory rates might have been expected if these com- ethanolamine (deanol) (35%) and N,N-diethylethanolamine pounds had inhibited the dicarboxylate carrier or the adenine (6%) suggests that substitution of the methyl groups on the tertiary nitrogenreduces affinity. Third, the productsof chonucleotide translocator. Uptake of choline by rat heart mitochondria was also in- line oxidation show little or no inhibition of choline uptake. vestigated. Rat heart mitochondria did not accumulate radi- Choline, betaine, and betaine aldehyde all have quaternary of ammonium groups. Substitution of the hydroxyl of the organic olabeled choline at high membrane potential. The absence a choline transporter in heart mitochondria and itspresence cation choline ( K , = 220 PM) with a carboxyl to yield betaine, in liver mitochondria correlatewith the tissue distributionof a neutral zwitterion, completely abolished affinity (0% inhibition).This abolition of affinity could be due to (i) the choline dehydrogenase in the rat (2). The sole driving force for carrier-mediated choline uptake absence of a hydroxyl, (ii) the presence of a carboxyl, or (iii) would appear tobe the choline electrochemical gradient. The the fact that betaineis neutral overall (at pH 7.0). However, rationale for this interpretation is as follows. Choline is ac- the evidence from betaine aldehyde (13% inhibition) shows cumulated to 4 mM in the matrix a t high membrane potential that the charge on the whole molecule plays no significant withinthefirst 5 min of thetime course (Fig. 1). Thus, role in inhibition. The presence of a carbonyl instead of a assuming a mole for mole stoichiometry between the chemical carboxyl still results in dramatically reduced affinity. If the species and choline, any putative chemical coupled to choline carbonyl or carboxyl is absent, as in tetramethylammonium, uptake would have to exist at 4 mM in the matrix. It is clear affinity is increased. If the carboxyl or carbonyl is replaced that uptake of choline is electrophoretic; therefore, a mono- by a hydroxyl, as in choline, affinity is increased even more. valent cation/choline antiporter or a monovalent anion/cho- The evidence suggests ( a ) that the hydroxyl is important for binding, and ( b ) that thereis an unfavorable environment for linesymporteris energetically not possible. Ontheother hand, anexchange of choline with aneutral organic zwitterion carbonyls and carboxyls in the vicinity of that hydroxylseems a more physiological possibility. However, the pool of binding site. The conclusion is corroborated by the observaneutral amino acids (59) and other putative neutral counter- tion that L-carnitine, which possesses a quaternary nitrogen ions within the matrix of isolated mitochondria such as car- and a hydroxyl group on the carbon equivalent to that in nitine (60) and citrulline (61) is
