Binding of Straight-Chain Saturated Dicarboxylic Acids to Albumin
`James H. Tonsgard,* Stephen A. Mendelson,* and Stephen C. Meredith
`Departments of**Pediatrics, *Neurology, and Pathology, and *the Joseph Kennedy Mental Retardation Center, Pritzker Medical
`School, The University ofChicago, Chicago, Illinois 60637
`
`Abstract
`
`Dicarboxylic acids are prominent features of several diseases,
`including Reye's syndrome. Long-chain dicarboxylic acids
`have profound effects on the function and structure of isolated
`mitochondria, suggesting that they could contribute to the mi-
`tochondrial dysfunction in Reye's syndrome. Binding of fatty
`acids to albumin and the intracellular fatty acid-binding pro-
`teins is important in regulating the transport and metabolism
`of fatty acids and protects against the toxic effects of unbound
`fatty acids. We studied the binding of dicarboxylic acids to
`defatted albumin using equilibrium dialysis to assess to what
`extent dicarboxylic acids are likely to be bound in the plasma
`of patients. Dicarboxylic acids bind weakly to albumin in a
`molar ratio of 3.8, 4.2, 1.6, 0.8, and 0.7 to 1 for octadecane-
`dioic, hexadecanedioic, tetradecanedioic, dodecanedioic, and
`decanedioic acid, respectively. The dissociation constants for
`long-chain dicarboxylic acids are 100-1,000-fold larger than
`those of comparable monocarboxylic acids. Oleate competes
`with dicarboxylic acid and reduces the moles of dicarboxylic
`acid bound per mol of albumin to < 1. Octanoate inhibits di-
`carboxylic acid binding. Our observations indicate that in
`Reye's syndrome, substantial concentrations of dicarboxylic
`acids of patients may be free and potentially toxic to mito-
`chondria and other cellular processes.
`
`Introduction
`
`Dicarboxylic acids are the product of omega-oxidation in the
`microsomes. Only 5-10% of FFA are metabolized by this
`pathway in ketotic rats (1). However, dicarboxylic acids have
`recently been shown to be a prominent feature of several dis-
`eases, including Reye's syndrome (2-4), neonatal adrenoleu-
`kodystrophy (5), Zellweger's syndrome (5), and defects of fatty
`acid metabolism (6, 7). In Reye's syndrome, dicarboxylic acids
`make up as much as 55% of the total serum FFA (4).
`The hallmark ofReye's syndrome is a transient generalized
`impairment of mitochondrial enzymes and swelling and dis-
`tortion of mitochondrial ultrastructure (8). Several toxins, in-
`cluding aspirin, have been suggested to predispose children to
`this illness (9). Serum from patients with Reye's syndrome
`impairs ATP formation and induces swelling and distortion of
`isolated mitochondria (10). These effects correlate directly
`
`Address all correspondence to Dr. James H. Tonsgard, Department of
`Pediatrics, Box 228, The University of Chicago, 5841 S. Maryland
`Avenue, Chicago, IL 60637.
`Receivedfor publication 20 January 1988 and in revisedform 13
`June 1988.
`
`J. Clin. Invest.
`© The American Society for Clinical Investigation, Inc.
`$2.00
`0021-9738/88/11/1567/07
`Volume 82, November 1988, 1567-1573
`
`with the concentration ofdicarboxylic acids in the serum sam-
`ples (10). Studies with isolated mitochondria have demon-
`strated that dicarboxylic acids, particularly long-chain dicar-
`boxylic acids, at concentrations comparable to those in plasma
`of patients with Reye's syndrome, profoundly inhibit the en-
`zymes of the terminal respiratory pathway (1 1), inhibit ATP
`formation (10), and induce an irreversible expansion of mito-
`chondria characteristic of an uncoupler of oxidative phos-
`phorylation (12). Dicarboxylic acids thus may contribute to
`the mitochondrial dysfunction that is central to Reye's syn-
`drome (10). The regulation of dicarboxylic acid metabolism
`may therefore be important in Reye's syndrome and other
`diseases in which dicarboxylic acid formation is prominent.
`The binding of monocarboxylic fatty acids to albumin and
`the intracellular fatty acid-binding proteins modulates the
`transport and metabolism ofthese fatty acids (13, 14). Binding
`of potentially toxic ligands to albumin protects against the
`toxic effects of the ligands, as toxicity is proportionate to the
`concentration of unbound ligand (13, 15, 16). We undertook
`this study to determine the affinity and capacity ofalbumin for
`dicarboxylic acids.
`
`Methods
`
`Monocarboxylic and dicarboxylic acids. Unlabeled monocarboxylic
`and dicarboxylic acids were purchased from Sigma Chemical Co. (St.
`Louis, MO), Applied Science (Warrenville, IL), and Foxboro Analabs
`(North Haven, CT). The purity of the unlabeled acids was assessed by
`gas-liquid chromatography and was found to be at least 99.5%. [3H]-
`Dicarboxylic acids, 1-[44C]C8.o,1 and 1-[(4C]C18.1 were purchased from
`Amersham Corp. (Arlington Heights, IL) and 1-['4C]decanedioic acid
`was purchased from Pathfinder Laboratories, Inc. (St. Louis, MO).
`The [3H]dicarboxylic acids prepared by Amersham Corp. were ob-
`tained by reaction with tritiated water (method TR.8; Amersham In-
`ternational, Amersham, England) at a very high specific activity. The
`chemical purity of the labeled compounds was assessed by gas-liquid
`chromatography (3) and found to be at least 97%. The radiopurity of
`the labeled compounds was assessed by thin-layer chromatography
`(17). Silica G plates were spotted with a mixture of labeled and unla-
`beled methyl esters ofthe dicarboxylic acids, separated using pentane/
`ether/acetic acid (92:7:1), and visualized with iodine vapors. The plates
`were scraped in 1-cm bands and the gel fractions were solubilized with
`water and then the radioactivity was measured. The radiopurity ofthe
`dicarboxylic acids was at least 93.4%.
`Potassium salts ofa mixture oflabeled and unlabeled monocarbox-
`ylic or dicarboxylic acid were made by dissolving the acid in a small
`amount of ethanol and adding 10 meq of 1 M methanolic KOH to 1
`meq acid. Phenolphthalein was added as a pH indicator. The flask
`containing the solution was connected to a refluxing apparatus and
`heated at 80°C for 30 min in a water bath. The solution was cooled,
`then dried with N2, and then redissolved in phosphate buffered solu-
`tion containing 0.1 16 M NaCI, 0.0049 M KCI, and 0.016 M sodium
`
`1. Abbreviations used in this paper: C18.1, oleic acid; C18.0, stearic acid;
`C16.0, palmitic acid; C14.0, myristic acid; C12.0, lauric acid; C10.0, de-
`canoic acid; C8.0, octanoic acid.
`
`Dicarboxylic Acid Binding to Albumin
`
`1567
`
`MPI EXHIBIT 1040 PAGE 1
`
`

`

`phosphate, pH 7.4. This salt solution was rinsed with chloroform to
`remove the monocarboxylic or dicarboxylic acids that were not in salt
`form. The concentration of the carboxylate salt was determined by
`gas-liquid chromatography by comparing the detector response to a
`calibration curve of a known amount ofthe unlabeled compound. The
`extraction efficiency was determined by comparing the radioactivity of
`aliquots before and after extraction. The specific activity ofthe carbox-
`ylate salt was determined from the counts per minute of the extracted,
`derivatized sample divided by the concentration of the carboxylate
`salt.
`Albumin. Essentially fatty acid-free (< 0.005%) crystalline BSA
`was from Sigma Chemical Co. and further purified as described by
`Spector, John, and Fletcher ( 18). The monocarboxylic fatty acid con-
`tent of the purified albumin was determined by gas-liquid chromatog-
`raphy (4) and found to have < 0.02 mol fatty acid/mol of albumin.
`Equilibrium dialysis. Binding ofdicarboxylic acids to defatted BSA
`was determined using equilibrium dialysis as described by Ashbrook,
`Spector, and Fletcher (19). Dialysis chambers and membranes were
`from Bel-Art Products (Pequannok, NJ). The chamber contains two
`l-ml compartments separated by a dialysis membrane that is imper-
`meable to compounds with a molecular weight > 6,000; albumin does
`not cross the dialysis membrane (19). The dicarboxylic acids reached
`equilibrium within 18 h at 370C. We chose a 26-h incubation period to
`ensure equilibrium. Binding was assessed by varying concentrations of
`dicarboxylic acid (0.05-1.5 mM) in a near physiologic salt solution
`containing 0.1 16 M NaCl, 0.0049 M KCl, and 0.0 16 M sodium phos-
`phate, pH 7.4. Dicarboxylic acid was added to one side ofthe chamber
`and defatted albumin in the same salt solution was added to the other
`side to a final concentration of0.05 mM. Because octadecanedioic acid
`is less soluble than the other dicarboxylic acids (20), binding of this
`fatty acid was assessed using 0.010 mM albumin and 0.010-0.200 mM
`octadecanedioic acid. Dialysis chambers were incubated in a shaking
`water bath at 370C and at the end of the incubation period, 100 ,l was
`removed from each side, and the radioactivity of the aliquot was mea-
`sured using an Isocap/300 liquid scintillation counter (G. D. Searle &
`Co., Des Plaines, IL). The concentration of free dicarboxylic acid was
`determined from the counts per minute in the aliquot from the side of
`the chamber that did not contain albumin, corrected for the volume of
`the two sides of the chamber into which the free dicarboxylic acid was
`distributed, and multiplied by the specific radioactivity of the dicar-
`boxylic acid solution. The concentration of bound dicarboxylic acid
`was determined by subtracting the counts per minute in the aliquot
`from the side of the equilibrium chamber that did not contain albu-
`min, from the counts per minute in the aliquot from the albumin-
`containing side. This was then corrected for the volume of the albu-
`min-containing chamber and multiplied by the specific radioactivity of
`the dicarboxylic acid solution. Experiments were performed in tripli-
`cate. The recovery ofradioactivity was between 88 and 10 1%. The data
`shown are the results of between four and eight separate experiments.
`Monocarboxylic acid competition. Experiments involving the
`competition of the I-[14C]CI8.I with 3H-labeled dicarboxylic acids were
`performed as described by Ashbrook, Spector, and Fletcher (19). It has
`previously been observed that C18.1 does not readily cross the dialysis
`membrane (15, 19). This observation was confirmed in preliminary
`experiments. The competition experiments were performed as de-
`scribed above, except that either 0.1, 0.3, or 0.5 mM I-[14C]C18.1 was
`added to the albumin-containing side ofthe equilibrium chamber. The
`extent to which CI8.I was bound was determined by gel filtration (21,
`22). At the end of the incubation, 100 ,u was removed from the albu-
`min-containing side of the chamber and added to a l-ml Sephadex
`G-25 column (Pharmacia Fine Chemicals, Piscataway, NJ). The col-
`umn was prepared by placing a slurry of Sephadex G-25 in buffered
`salt solution in a 1-ml plastic syringe. The buffer was removed by
`centrifugation at 150 g for 10 min at 25°C. 100 ul from the albumin
`side of the chamber was added to the column and the column was
`centrifuged again at 150 g for 10 min. Protein content ofthe eluate was
`monitored spectrophotometrically at 280 nm and fatty acid content
`was assessed by scintillation counting. At least 90% of the albumin is
`eluted after a single centrifugation. Albumin was completely eluted
`
`1568
`
`J. H. Tonsgard, S. A. Mendelson, and S. C. Meredith
`
`from the column with two additional rinses of 100 Ml of salt solution.
`Fatty acids not associated with albumin remained bound to the col-
`umn when the column was eluted with additional rinses of aqueous
`buffer. Recovery of the FFA was achieved by rinsing the column three
`times with 0.5 ml of 2:1 chloroform/methanol. The buffer and chloro-
`form rinses were collected separately and transferred to scintillation
`vials. The samples were evaporated to dryness and the amount of
`['4C]C18.1 was determined by liquid scintillation. The mean recovery of
`albumin was determined to be 95.8%. The mean recovery of ['4C]C18.,
`was 98.3%. We demonstrated that reequilibration ofbound fatty acid is
`sufficiently slow that it does not occur to any significant extent, as
`albumin is propelled by centrifugation through the column.
`In contrast to C18.1, C8.0 freely diffuses across the dialysis mem-
`brane (19). Equilibration ofC8.0 occurs within 24 h. Therefore, binding
`of['4C1C8.0 was assessed from the partitioning of'4C counts per minute
`as described previously for the dicarboxylic acids.
`Analysis ofbinding. Binding isotherms were analyzed using a non-
`linear least-squares fit computer program; in general, the damping
`Gauss-Newton method was used (23). The points in the figures are
`experimental data and the lines are the best fit of the data to the
`theoretical equation as determined by the least-squares fit computer
`program. Dissociation constants are expressed as the mean±SD.
`
`Results
`
`The binding of dicarboxylic fatty acids to albumin was ana-
`lyzed using the equation of reversible, saturable binding to
`one, two, or three classes of noninteracting equivalent sites,
`i.e., the equation of a Langmuir isotherm (24). For the case of
`one class of sites, Kd = (FAf)(Sf)/CX = (FAf)(St - FAb)IFAb,
`where Kd = dissociation constant (micromolar), FAf = free
`dicarboxylic acid (micromolar), FAb = bound dicarboxylic
`acid (micromolar), C, = dicarboxylic acid-site complex (mi-
`cromolar), and Sf and St = free and total binding sites, respec-
`tively. The number of sites/albumin molecule can be calcu-
`lated from St and the albumin concentration. We used the
`Akaike information coefficient (23) to ascertain whether the
`best fit of the data could be obtained by assuming one, two, or
`three classes of binding sites.
`Affinity of dicarboxylic acids for albumin. We first exam-
`ined the binding of hexadecanedioic acid (Fig. 1). Analysis of
`the binding shows two different types of binding sites with
`dissociation constants±SD of 1.2±1.4 and 67.8±15.6 ,M, re-
`spectively (Table I). The first site binds 1 mol of dicarboxylic
`acid and the second class of binding sites binds the remaining
`acid. We observed a molar ratio of bound hexadecanedioic
`acid to albumin of 4.2:1. The calculated saturation of albumin
`occurs at a molar ratio of 5:1 (0.9±0.3 plus 4.2±0.2 mol,
`Table I).
`
`0
`
`0
`
`o3,
`
`0
`
`o 0
`
`0
`
`250.00 -
`
`o 200.00 -
`
`2 150.00 -.
`
`w 100.00 -,
`a
`
`50.00
`
`Z
`
`0.00
`0.00
`
`200.00
`400.00
`FREE HEXADECANEDIOIC ACID (itM)
`
`Figure 1. Binding of
`hexadecanedioic acid to
`defatted albumin. Bind-
`ing was assessed at pH
`7.4, using varying con-
`centrations of [3H]-
`dicarboxylic acid in an
`equilibrium dialysis
`chamber in which one
`side contained 0.050
`mM defatted albumin.
`The points are experi-
`mental data and the
`lines are the best fit of
`the data to the theoreti-
`cal equation.
`
`MPI EXHIBIT 1040 PAGE 2
`
`

`

`Table L Equilibrium Constantsfor Saturated Straight-Chain Dicarboxylic Acids
`
`Fatty acid
`
`Bound
`dicarboxylic
`acid
`
`Dissociation constants
`
`Kd,
`
`Kd,
`
`Predicted saturation
`
`Site 1
`
`Site 2
`
`Number ofdata points
`
`mol/mol BSA*
`
`jsM±SD
`
`mol/mol BSA±SD
`
`Octadecanedioic
`pH 7.4
`
`Hexadecanedioic
`pH 6.8
`pH 7.4
`pH 8.0
`
`Tetradecanedioic
`pH 7.4
`
`Dodecanedioic
`pH 6.8
`pH 7.4
`pH 8.0
`
`Decanedioic
`pH 7.4
`
`3.8
`
`4.2
`4.2
`4.2
`
`1.6
`
`0.80
`0.88
`0.80
`
`.70
`
`1.1±1.7
`
`19.4±18.2
`
`1.4±1.4
`
`3.2±1.1
`
`2.5±1.4
`1.2±1.4
`3.9±0.9
`
`27.8±6.1
`67.8±15.6
`216.0±46.0
`
`1.1±0.4
`0.9±0.3
`2.1±0.2
`
`3.3±0.3
`4.2±0.2
`4.3±0.9
`
`23.2±9.8
`
`292.6±26.0
`
`1.0±0.1
`
`1.1±0.1
`
`30.1±6.2
`75.2±4.3
`83.3±13.2
`
`31.5±9.3
`
`0.9±0.1
`1.0±0.02
`1.0±0.1
`
`0.8±0.1
`
`28
`
`35
`51
`41
`
`49
`
`22
`28
`21
`
`17
`
`Equilibrium constants for saturated straight-chain dicarboxylic acids. Binding was assessed in the presence of 0.050 mM albumin for all of the
`dicarboxylic acids except octadecanedioic acid, which was incubated with 0.010 mM albumin. * Indicates the mol ratio of fatty acid to albumin
`at saturation as observed experimentally.
`
`experimentally for dodecanedioic and decanedioic acids is 0.8
`and 0.7 mol/mol of albumin (Fig. 2, Table I). The dissociation
`constants for dodecanedioic and decanedioic acid are
`75.2±4.3 and 31.5±9.3 aM, respectively.
`Effect ofpH on dicarboxylic acid binding. The effect of pH
`of the buffer solution on dicarboxylic acid binding was also
`examined. Incubations were performed at pH 6.8 and 8.0 with
`hexadecanedioic acid and dodecanedioic acid and compared
`with incubations at pH 7.4 (Table I). The pH of the buffer
`solution does not affect the saturation level of the albumin.
`The binding curves for both hexadecanedioic and dodecane-
`dioic acid shift slightly to the left with a more acidic pH (Fig.
`3). The pH ofthe buffer has no significant effect on the K, for
`hexadecanedioic acid but the Kd decreases at the more acidic
`pH. The dissociation constant for dodecanedioic acid also de-
`creases at pH 6.8 (Table I).
`Competition with monocarboxylic fatty acids for binding.
`We examined the effect of the monocarboxylic acid, oleic acid
`(C18.1) on the binding of dicarboxylic acids using 0.1, 0.3, and
`0.5 mM C18.1 (Fig. 4, Table II). C18.1 competes with hexade-
`canedioic acid for binding; as a result, the apparent dissocia-
`tion constant of the dicarboxylic acid increases 15-35-fold
`(from 1.2±1.4 to 41.0±7.3 ,uM, Table II). When the monocar-
`boxylic acid concentration is increased to 0.5 mM (a 10:1
`C18.1/albumin ratio), all but 0.70 mol of hexadecanedioic acid
`is displaced (Fig. 4, Table II). A similar result was obtained
`when the binding ofoctadecanedioic acid was examined in the
`presence of C18.1. All but 0.7 mol of octadecanedioic acid is
`displaced in the presence of near saturating concentrations of
`C18.1. As shown in Fig. 5, when the concentration oflong chain
`dicarboxylic acid is high enough, long-chain dicarboxylic acid
`can bind to albumin and displace some C18.1.
`The effect of monocarboxylic fatty acids on the binding of
`dodecanedioic acid was also examined (Fig. 6, Tables II and
`
`Dicarboxylic Acid Binding to Albumin
`
`1569
`
`The binding of octadecanedioic acid is
`; similar to that of
`lioic acid is bound
`hexadecanedioic acid (Fig. 2). Octadecane
`in a 3.8:1 molar ratio (Table I). As with hex
`cadecanedioic acid,
`inding sites with a
`there appear to be two distinct types of bi
`dicarboxylic acid.
`theoretical saturation of 4.6±1.4 mol of the
`In contrast to the longer chain dicarbox:
`ylic acids, albumin
`only binds 2 mol of tetradecanedioic acid p
`Per mol of albumin
`(Table I, Fig. 2). There are again two dist
`tinct binding sites,
`each binding 1 mol of tetradecanedioic acii
`d per mol of albu-
`min with dissociation constants of 23.2±9.
`.8 and 292.6±26.0
`jM (Table I). There is only a single binding
`site for dodecane-
`dioic and decanedioic acids. The maximal
`binding achieved
`
`100.00
`C18 1
`
`C14
`s
`
`00
`
`age
`
`i
`
`'do.o......
`C10
`
`50.00
`
`I 0.00
`
`50.00
`
`1 00.00
`C12
`
`~~~~0
`
`0
`
`0
`
`40.00
`
`1-10
`
`0.00
`0.
`
`.00
`
`tP
`
`LLI
`Z40.00
`
`D0m
`
`0.00
`
`io~o
`
`aD
`
`61..
`
`00
`
`0
`
`200.00
`
`400.00
`
`0.00
`0.
`
`00
`
`500 0
`.500.00
`
`0.000.00
`1000
`1 000.00
`I)
`FREE FATTY ACID (uM
`d albumin. Binding
`Figure 2. Binding of dicarboxylic acids to defattec
`Lion that the bind-
`was assessed as described in Fig. 1 with the except
`10 mM albumin.
`ing of octadecanedioic acid was assessed with 0.01
`I (C14), dodecane-
`Octadecanedioic acid (C18), tetradecanedioic acid
`dioic acid (C12), and decanedioic acid (CO0).
`
`MPI EXHIBIT 1040 PAGE 3
`
`

`

`- 0.1 mM OLEIC ACID
`
`200.00
`
`150.00
`
`06
`
`C-)
`
`05b
`
`iz 100.00
`
`A
`
`250.00 -J
`
`o 200.00-
`
`o 1 50.00 --
`
`0
`
`0
`
`a
`
`0
`
`0
`
`0
`
`50.00
`
`r
`u.uu I
`)rr
`0.c
`
`0 xI
`
`0z
`
`0m
`
`00
`
`0
`
`100.00 I
`
`50.00
`
`0 pH 6.80
`
`0 pH 8.00
`
`wz w
`
`0 xL
`
`i
`::
`0z
`
`FREE HEXADECANEDIOIC ACID (uM)
`Figure 4. Competition of oleic acid and hexadecanedioic acid for
`binding to albumin. Binding of [3Hldicarboxylic acid was assessed in
`the presence of 0.05 mM albumin incubated with 0.1 (o), 0.3 (o),
`and 0.5 (A) mM ['4C]oleic acid.
`
`.00
`
`and the experimentally observed binding capacity of dodecan-
`edioic acid is reduced by more than half (Table III).
`The medium-chain length monocarboxylic acid, octanoic
`acid (C8.0) has a more profound effect on the binding of dode-
`canedioic acid (Table III). When 0.5 mol of C8.0 per mol of
`albumin are bound, only 0.1-0.2 mol of dodecanedioic acid is
`bound and when 1 mol of C8.0 per mol of albumin is bound,
`the binding of dodecanedioic acid is further inhibited.
`
`Discussion
`
`Our results indicate that dicarboxylic acids bind to albumin
`with lower affinity than monocarboxylic acids of the same
`chain length. There is a single low-affinity site for dodecane-
`dioic and decanedioic acid. There is a single higher affinity site
`for the longer chain dicarboxylic acids (C18-C14) with be-
`tween one and four additional low-affinity sites, depending on
`the chain length of the dicarboxylic acid. The affinity of long-
`chain dicarboxylic acids for albumin is 100-1,000-fold less
`than that oflong-chain monocarboxylic fatty acids and is more
`comparable to the affinity of many drugs for albumin (25).
`The molar ratio of bound dicarboxylic acid to albumin ob-
`served is also significantly less than that reported for monocar-
`
`400.00
`200.00
`FREE HEXADECANEDIOIC ACID (uM)
`
`600.00
`
`o
`
`0
`
`o
`
`pH 6.80
`
`0 pH 8.00
`
`0.00
`0.0
`)O
`
`0m
`
`50.00
`
`3 0
`
`-40.00 -
`
`0
`
`B
`
`0 30.00 -
`20.00
`z 10.00
`w
`
`z u
`
`0 0
`
`1 0.00
`
`0 z
`
`D 0
`
`0.00
`0.(
`
`FREE DODECANEDIOIC ACID (uM)
`Figure 3. The effect of buffer pH on the binding of dicarboxylic
`acids. Binding of hexadecanedioic acid (A) and dodecanedioic acid
`(B) was assessed at pH 6.8 (o) and 8.0 (o) as described in Fig. 1.
`
`III). In the absence of any competition, between 0.8 and 0.9
`mol are bound per mol of albumin. When 2-3 mol of C18.1 are
`bound to albumin (0.10-0.15 mM C18.), the affinity for dode-
`canedioic acid is reduced almost fivefold (Table II), but there is
`no significant effect on the maximal binding capacity (Tables
`II and III). When 4-5 mol of C18.1 are bound to albumin
`(0.3-0.5 mM C18.1) the affinity is reduced ninefold (Table II)
`
`Table II. Competition ofOleic Acid with Dicarboxylic Acidsfor Binding
`
`Fatty acid
`
`Hexadecanedioic
`with
`0.1 mM oleic
`0.3 mM oleic
`0.5 mM oleic
`
`Dodecanedioic
`with
`0.1 mM oleic
`0.5 mM oleic
`
`Bound
`dicarboxylic
`acid
`
`mol/mol BSA*
`
`4.4
`
`3.3
`1.4
`0.7
`
`0.9
`
`0.9
`0.4
`
`Dissociation constants
`
`Predicted saturation
`
`Kd,
`
`Site 1
`
`Site 2
`
`Number of data points
`
`1AM±SD
`
`mol/mol BSA±SD
`
`1.2±1.4
`
`67.8±15.6
`
`0.9±0.3
`
`41.0±7.3
`15.0±1.2
`34.4±9.6
`
`75.2±4.3
`
`339.3±116.0
`707.4±171.7
`
`4.2±0.2
`
`3.7±0.1
`1.4±0.2
`0.8±0.1
`
`1.0±0.02
`
`1.4±0.2
`0.7±0.1
`
`51
`
`29
`14
`14
`
`28
`
`15
`22
`
`Competition of oleic acid with dicarboxylic acids for binding. * Indicates the mol ratio of fatty acid to albumin, as observed experimentally.
`
`1570
`
`J. H. Tonsgard, S. A. Mendelson, and S. C. Meredith
`
`MPI EXHIBIT 1040 PAGE 4
`
`

`

`ACID
`
`50.00
`
`3.
`1-
`
`:2
`3
`A:
`C)
`
`0C
`
`LX
`
`z L
`
`uI
`
`000 0z 0m
`
`Figure 6. Competition of 0.5 mM oleic acid with dodecanedioic acid
`for binding to albumin. Binding was assessed as described in Fig. 1.
`Binding in the presence of 0.5 mM oleic acid (A) is compared with
`binding in the absence of competition (o).
`
`The albumin molecule is composed of three nonidentical
`cylindrical domains (13), each of which has a narrow hydro-
`phobic channel able to accomodate only one or two hydrocar-
`bon chains. The ends of each domain form nonidentical sub-
`domains containing positively charged amino acid side chains.
`Recent studies indicate that medium-chain monocarboxylic
`fatty acids bind almost exclusively via hydrophobic interac-
`tions, whereas the binding oflong-chain monocarboxylic acids
`is dependent on both electrostatic and hydrophobic interac-
`
`Table III. Competition ofDodecanedioic and Monocarboxylic
`Fatty Acids
`
`Concentration of fatty acids
`
`Bound
`dodecanedioic
`
`Bound oleic
`
`Bound
`octanoic
`
`AM dodecanedioicl
`juM oleic or octanoic
`
`mol/mol BSA
`
`10/0
`200/0
`300/0
`
`0/150
`
`10/150
`200/150
`300/150
`
`0/320
`
`10/320
`200/320
`300/320
`
`0/100
`200/100
`300/100
`0/500
`
`200/500
`300/500
`
`0.03
`0.50
`0.60
`
`-
`
`0.02
`0.48
`0.65
`
`0.01
`0.19
`0.27
`
`0.16
`0.25
`
`0.16
`0.19
`
`-
`
`2.60
`
`2.50
`2.70
`2.70
`
`4.88
`
`4.66
`4.67
`4.53
`
`-
`
`0.46
`
`0.49
`0.47
`
`1.25
`
`1.10
`1.10
`
`Competition of dodecanedioic with the monocarboxylic acids octan-
`oic, and oleic. Binding was assessed as described in Fig. 5. The con-
`centration of albumin was 0.050 mM.
`
`Dicarboxylic Acid Binding to Albumin
`
`1571
`
`OLEIC ACID
`
`HEXADECANEDIOIC ACID
`
`250.00
`
`11,
`
`m'200.00
`
`0
`
`150.00-
`
`100.00-
`
`3.
`
`0zv
`
`0m c <
`
`50.00-
`
`I
`
`I
`
`4bb.00
`
`0.0.
`
`0
`
`200.00
`FREE HEXADECANEDIOIC ACID (AM)
`Figure 5. Competition of 0.3 mM oleic acid with hexadecanedioic
`acid for binding to albumin. Binding was assessed as described in
`Fig. 4. The amount of bound oleate was determined after separation
`of free and bound fatty acid by centrifugation through a Sephadex
`G-25 column. o, micromoles of bound oleic acid; o, micromoles of
`bound hexadecanedioic acid. The binding of hexadecanedioic acid is
`analyzed using the equation of a Langmuir isotherm. The points are
`experimental data and the lines are the best fit of the data to the the-
`oretical equations obtained using the nonlinear least-square com-
`puter program as described in the text.2
`
`boxylic fatty acids of the same chain length. Spector and co-
`workers (18, 19) reported binding ratios of 6.5, 7.2, 7.0, 8.4,
`and 13.9 for C18.0, C16.0, C14.0, C12.0, and C1o.o, respectively
`compared with our observations of 3.8, 4.2, 1.6, 0.8, and 0.7
`for dicarboxylic acids of the same chain length.
`The pH of the buffer exerts only a modest effect on the
`binding of dicarboxylic acids. The pK of each of the carboxyl
`groups for octanedioic acid is 4.52 (26). The dissociation con-
`stants for longer chain dicarboxylic acids are probably close to
`this value. Thus, the dicarboxylic acids are > 99% deproton-
`ated in the physiologic pH range we examined. The modest
`shift in binding at pH 6.8 may be due to a change in amino
`acid charge at one or more binding sites which facilitates bind-
`ing of dicarboxylic acids.
`As might be expected from a comparison of the dissocia-
`tion constants of moncarboxylic and dicarboxylic acids,
`monocarboxylic acids competitively inhibit dicarboxylic acid
`binding. However, the competition studies demonstrate that at
`high enough concentrations, dicarboxylic acid can bind and
`displace some C18.I.
`
`2. The binding of hexadecanedioic acid is analyzed using the equation
`of a Langmuir isotherm, Kdaw = (Sm" - Sb) (Df)/Db, where KdaJP
`= apparent dissociation constant, St"PP = total sites on albumin avail-
`able for binding hexadecanedioic acid under these experimental con-
`ditions, Sb = concentration of sites occupied by hexadecanedioic acid,
`= concentration of bound hexadecanedioic acid, and Df = concen-
`tration of free hexadecanedioic acid.
`The desorption of oleic acid at various concentrations of hexade-
`canedioic acid is analyzed using the equation Kd2 = (SfdXDf)/Obd = (Ob
`- Dt,'XDf)/Obd-o- Ob. In this equation, the total sites from which
`oleic acid can be displaced = 0bd-- Dbm", where Obd=O = oleic acid
`bound in the absence of hexadecanedioic acid, and Db" = maximal
`hexadecanedioic acid able to bind to displaceable oleic acid sites on
`albumin. Df = concentration of free hexadecanedioic acid, 0b = con-
`centration ofbound oleic acid, and Kd2 = dissociation constant (M) for
`hexadecanedioic acid under these experimental conditions.
`
`MPI EXHIBIT 1040 PAGE 5
`
`

`

`tions (27). Three ofthe albumin subdomains are primary bind-
`ing sites for long-chain monocarboxylic fatty acids. The other
`three subdomains are primary binding sites for medium- and
`short-chain fatty acids and drugs (13). Drugs bind along the
`rim of the subdomains through electrostatic interactions with
`positively charged amino acid groups.
`The binding of dicarboxylic fatty acids to albumin resem-
`bles that of the monocarboxylic acids. Like medium-chain
`monocarboxylic fatty acids (13), the medium-chain dicarbox-
`ylic acids appear to bind to a single class of binding sites; the
`dissociation constants for these dicarboxylic acids are in the
`range of 10-1 M. The binding of the long-chain dicarboxylic
`acids resemble that of the long-chain monocarboxylic acids in
`that there are at least two classes of binding sites, one of rela-
`tively high affinity and one of low affinity. In the case of hexa-
`decanedioic acid, for example, the Kd for the primary binding
`site is 1.2 AM, and the Kd for the secondary binding sites is
`68 .M.
`There are, however, several important differences between
`the binding of monocarboxylic and dicarboxylic acids to albu-
`min. First, the affinity of monocarboxylic acids for albumin is
`generally at least two orders of magnitude greater than that of
`dicarboxylic acids of the same chain length. Second, whereas
`there are three primary or high-affinity sites on albumin for
`long-chain monocarboxylic acids, there is only one such site
`for long-chain dicarboxylic acids. Third, whereas albumin
`binds up to 13 mol of medium-chain monocarboxylic acid per
`mol of albumin (19), only 1 mol of medium-chain dicarboxy-
`lic acid is bound per mol of albumin.
`Our observations suggest that the single high-affinity site
`for long-chain dicarboxylic acids may be one ofthe high-affin-
`ity sites for monocarboxylic fatty acids. As shown in Table II,
`when the high-affinity monocarboxylic acid binding sites are
`largely occupied (0.1 mM C18.1), there appears to be only a
`single class of low-affinity sites for dicarboxylic acids, having a
`dissociation constant similar to that of the low-affinity sites in
`the absence of any competition. On the other hand, the low-
`affinity sites for long- and medium-chain length dicarboxylic
`acids are distinct from the primary binding sites for long-chain
`monocarboxylic acids, as when these latter sites are occupied
`by monocarboxylic acids, all the low-affinity sites for hexade-
`canedioic and dodecanedioic acids are still available (Tables II
`and III); the dicarboxylic acids are only displaced when the
`bound C18.1 exceeds 3 mol/mol of albumin. Moreover, C8.0 is
`more effective than C18.1 (Table III) in inhibiting the binding at
`the low-affinity site. We infer from these considerations that
`some or all of the low-affinity sites for medium- and long-
`chain dicarboxylic acids are in the subdomains ofalbumin that
`bind medium-chain monocarboxylic fatty acids and drugs.
`These studies confirm our hypothesis that dicarboxylic
`acids are not as tightly bound to albumin as monocarboxylic
`fatty acids (9), and suggest that in Reye's syndrome, in which
`the fatty acid/albumin ratio approaches or exceeds 4:1 (4, 28),
`substantial concentrations of dicarboxylic acids may be free.
`Long-chain dicarboxylic acids are readily displaced from the
`higher affinity binding site. Binding to the lower affinity sites is
`probably also significantly impaired in patients with Reye's
`syndrome. The concentration of dicarboxylic acids in those
`patients is often as high as 0.5 mM in comatose patients, with
`the majority of the dicarboxylic acids being long chain (C14-
`C18) (4). In our experiments, when the binding of 0.5 mM
`hexadecanedioic acid was assessed in the presence of near satu-
`
`1572
`
`J. H. Tonsgard, S. A. Mendelson, and S. C. Meredith
`
`rating concentrations of C18.1, only 16% of hexadecanedioic
`acid was bound (Fig. 4). Moreover, Goodman and others (29,
`25) have shown that when the fatty acid/albumin ratio exceeds
`2:1, drugs, with affinity for albumin comparable to that which
`we observed for dicarboxylic acids, are largely unbound. Our
`studies suggest that the low-affinity sites for dicarboxylic acids
`are the same as the binding sites for medium-chain fatty acids
`and drugs, and thus may compete with these compounds for
`binding. Some investigators have, in fact, shown elevations in
`medium-chain fatty acids in Reye's syndrome (30). In addi-
`tion, aspirin and other protein-bound medications are fre-
`quently found in patients with Reye's syndrome and have
`been suggested to play a role in the pathogenesis of the illness
`(9). Thus, it seems likely that with the competition oflong- and
`medium-chain length monocarboxylic acids and drugs for
`binding to albumin, dicarboxylic acids are largely unbound in
`the plasma of patients with Reye's syndrome.
`
`Acknowledaments
`
`The authors thank E. van Mele, Francis Ko, and John O'Connell for
`their excellent technical assistance and Godfrey S. Getz, M.D., Ph.D.
`for advice and critical review of the manuscript.
`The work was supported in part by the Children's Research Fund
`and the Jennifer Wieser Fund, the University ofChicago, a grant from
`the Markey Trust, and grants HD-04583, NS-23616, and HL-15062
`from the United States Public Health Service.
`
`References
`
`1. Bjorkhem, I. 1978. On the quantitative importance of w-oxida-
`tion of fatty acids. J. Lipid Res. 19:585-590.
`2. Tonsgard, J. H. 1985. Urinary dicarboxylic acids in Reye syn-
`drome. J. Pediatr. 107:79-84.
`3. Ng, K. J., B. D. Andresen, M. D. Hilty, and J. R. Bianchine.
`1983. Identification of long chain dicarboxylic acids in the serum of
`two patients with Reye's syndrome. J. Chromatogr. 276:1-10.
`4. Tonsgard, J. H. 1986. Serum dicarboxylic acids in Reye syn-
`drome. J. Pediatr. 109:440-445.
`5. Rocchiccioli, F., P. Aubourg, and P. F. Bougneres. 1986. Me-
`dium- and long-chain dicarboxylic aciduria in patients with Zellweger
`sy