J Biol Inorg Chem (2006) 11:991–998
`DOI 10.1007/s00775-006-0156-z
`
`O R I G I N A L P A P E R
`
`Ion-exchange and potentiometric characterization
`of Al–cystine and Al–cysteine complexes
`
`Denise Bohrer Æ Vania Gabbi Polli Æ Paulo Cı´cero do Nascimento Æ
`Jean Karlo A. Mendonc¸a Æ Leandro Machado de Carvalho Æ
`Solange Garcia Pomblum
`
`Received: 6 May 2006 / Accepted: 2 August 2006 / Published online: 24 August 2006
`Ó SBIC 2006
`
`Abstract The interaction between aluminium and
`cysteine and cystine was evaluated by means of ion-
`exchange
`experiments
`and
`potentiometry.
`Ion-
`exchange experiments included other ligands with
`affinity for aluminium and two kinds of resins, either a
`Na+-form or an Al3+-form exchanger. The ability of
`the ligands to keep aluminium in solution in the
`presence of the Na+ exchanger or to withdraw it from
`the Al3+-form resin was evaluated. Aluminium quan-
`tification was carried out by either graphite-furnace or
`flame atomic absorption spectrometry. Aluminium
`extraction isotherms were linearised using the Scat-
`chard plot, and stability constants were obtained from
`the curves’ slopes. The experiments showed that the
`ability of the ligands to withdraw aluminium from
`the Al3+-form resin increased following the order
`cysteine < oxalate < citrate = cystine < nitrilotriacetic
`
`acid < ethylenediaminetetraacetic acid. Potentiomet-
`ric titrations, carried out in aqueous solution with
`constant
`ionic strength and temperature,
`showed
`that the predominant species in solution have a metal–
`ligand proportion of 1:1 for both amino acids. The
`main species are Al(OH)3L, with log K of 6.2 for
`cysteine, and AlL and Al(OH)L, with log K of 10.3
`and 1.7, respectively, for cystine. Stability constants
`obtained from the Scatchard plots showed a linear
`correlation with the stability constants obtained by
`potentiometry for cystine and cysteine in this work
`and those collected from the literature for the other
`ligands. These results show that cysteine and cystine
`extract and maintain aluminium in solution, which
`may explain elevated concentrations of aluminium in
`parenteral nutrition solutions containing these amino
`acids.
`
`Electronic Supplementary Material Supplementary material
`is available to authorised users in the online version of this
`article at http://dx.doi.org/10.1007/s00775-006-0156-z.
`
`Keywords Aluminium Æ Cystine Æ Cysteine Æ
`Ion-exchange Æ Scatchard plot
`
`D. Bohrer (&) Æ P. Cı´cero do Nascimento Æ
`J. K. A. Mendonc¸a Æ L. M. de Carvalho
`Departamento de Quı´mica,
`Universidade Federal de Santa Maria,
`97111-970 Santa Maria, RS, Brazil
`e-mail: ndenise@quimica.ufsm.br
`
`V. G. Polli
`Departamento de Fı´sica,
`Universidade Federal de Santa Maria,
`97111-970 Santa Maria, RS, Brazil
`
`S. G. Pomblum
`Departamento de Ana´ lises Clı´nicas e Toxicolo´ gicas,
`Universidade Federal de Santa Maria,
`97111-970 Santa Maria, RS, Brazil
`
`Introduction
`
`Aluminium is a nonessential element to which humans
`are frequently exposed. The toxicity of this element is
`evidenced in chronic renal patients on haemodialysis
`treatment, its association with diseases of bone and
`brain in patients receiving long-term parenteral nutri-
`tion therapy [1, 2], and the aluminium accumulation in
`bones of premature infants receiving total parenteral
`nutrition therapy [3–5]. Aluminium is also related to
`Alzheimer’s disease and, in spite of several studies, it is
`not clear yet how this element reaches the human
`brain. Approximately 80% of the aluminium present in
`
`Nexus Ex. 1011
`Page 1 of 8
`
`123
`

`

`992
`
`J Biol Inorg Chem (2006) 11:991–998
`
`blood is carried by proteins, basically transferrin and
`albumin, and the remaining 20% is believed to be
`transported by small ligand molecules such as citrate
`[6–8]. Amino acids can be included in this category of
`ligands because they are able to bind metals through
`complex formation; therefore, amino acids present in
`blood may also be responsible for metal transportation
`in the human body.
`In a previous work [9], we observed that the alu-
`minium present as a contaminant in amino acid solu-
`tions for nutritional purposes is leached from glass
`containers owing to the interaction of formulation
`constituents with the glass surface. In the experiment,
`the highest interaction occurred in solutions of cysteine
`and cystine . These amino acids individually stored in
`glass containers extracted 10 times more than the
`others and as much as ethylenediaminetetraacetic acid
`(EDTA) and other complexing agents. Aluminium was
`also found in the raw material used to prepare these
`formulations [10]. However, while the aluminium level
`found in cysteine samples (2.7 lg/g) was not much
`higher than the mean level in the other amino acids
`(1.8 lg/g), cystine samples presented a mean of 70 lg
`Al/g. Fifteen different samples of cystine from differ-
`ent suppliers were analysed, and contamination levels
`from 30 to 95 lg/g were observed.
`Stability constants for the aluminium complexes of
`cystine and cysteine were not found in the literature
`[11–13], and the possibility of complex formation with
`these species has not been considered, probably be-
`cause of the low affinity of aluminium (a hard acid) for
`sulphur (a soft base).
`In a study on aluminium speciation in biological
`fluids, Dayde´ et al. included amino acids as potential
`ligands for aluminium metabolism, especially in the
`gastrointestinal tract. Their experiments included gly-
`cine, representing the classic apolar amino acids, ser-
`ine, histidine, and threonine as amino acids with polar
`side chains [14] and glutamic acid [15]. Owing its extra
`carboxyl group, glutamic acid presented an ability to
`maintain Al3+ ions in solution superior to that of gly-
`cine-like amino acids. The authors concluded that
`whereas the first group of amino acids, given the low
`affinity of the Al3+ ion for the amino group, presented
`an interaction of limited extent, glutamic acid was able
`to neutralise Al3+ ions as succinate, tartrate, and ma-
`late do. Since cystine and cysteine would play a similar
`role to glycine in complexing aluminium, owing to the
`low affinity of the Al3+ ion for thiol groups, the authors
`did not include these amino acids in their studies [14].
`However,
`in spite of cystine being a sulphur-
`containing amino acid, the four binding sites of cystine
`are hard nitrogen and oxygen containing groups. If
`
`cysteine itself is not a strong complexing agent for
`aluminium, its oxidation to cystine, by converting the
`monomer into a dimer through a disulphide bridge,
`may be the reason why a high aluminium level was
`found in cysteine solutions.
`
`HS – CH2– CH – COOH
`
`HOOC – CH – CH2 – S - S – CH2 – CH – COOH
`
`NH2
`
`H2N
`
`NH2
`
`Cys
`
`Cys2
`
`Up to now there have been no studies on the
`interaction of aluminium with both amino acids;
`however, more attention should be paid to this pos-
`sibility. In two recent editorials in the Journal of
`Health-System Pharmacy, Discroll and Discroll [17]
`and Canada [18] refer to the aluminium concentration
`in parenteral nutrition additives. Both works quote
`cysteine hydrochloride as one of the three most con-
`taminated formulations. The other two are sodium
`phosphate and calcium gluconate, which are very well
`known as major aluminium sources in parenteral
`nutrition. The aluminium level in cysteine solutions
`(15,000 lg/l) was even higher than that in calcium
`gluconate (12,000 lg/l). Like in the other two prod-
`ucts [19], the presence of such a high aluminium level
`in cysteine solutions might be related to the affinity of
`the amino acid for the metal.
`In this work we investigated not only the interaction
`of cysteine but also of cystine with aluminium through
`ion-exchange-equilibrium and potentiometric mea-
`surements. The interaction of aluminium with cystine
`can be as important as that with cysteine or even more
`so, since cysteine is rapidly oxidised to cystine in human
`blood [20, 21]; therefore, cystine is the predominant
`species in this fluid. Owing to the difficulties inherent in
`the determination of stability constants of aluminium
`complexes by potentiometry, we carried out experi-
`ments with ion exchangers in order to confirm complex
`formation. The experiments also included well-estab-
`lished complexing agents for aluminium. By comparing
`the ability of these ligands to keep aluminium ions in
`solution in the presence of a cation-exchanger, we could
`estimate the interaction of cysteine and cystine with
`aluminium in the same conditions. A second approach
`was carried out using an Al3+-form exchanger. In this
`case the ability of the ligands to withdraw aluminium
`ions from the resin was evaluated and the abilities were
`compared with each other. The extraction profiles were
`linearised by means of a Scatchard plot, and the sta-
`bility constants obtained were compared with those
`obtained by potentiometric titration.
`
`Nexus Ex. 1011
`Page 2 of 8
`
`123
`

`

`J Biol Inorg Chem (2006) 11:991–998
`
`993
`
`Experimental
`
`Apparatus
`
`A Varian SpectrAA 200 atomic absorption spectrom-
`eter equipped with flame accessories and a GTA-100
`graphite furnace with an autosampler (Melbourne,
`Australia), a Trox class 100 clean bench (Curitiba,
`Brazil), a Corning pH meter (UK), and an Edmund
`Bu¨ hler
`(Tu¨ bingen,
`7400 KL1 mechanical
`shaker
`Germany) were used. The conditions for operation are
`described in Table S1 in the supplementary material.
`
`Reagents
`
`All chemicals were of analytical-reagent grade. The
`solutions were prepared with distilled then deionised
`water (Milli-Q high-purity grade, Millipore, Bedford,
`USA). An aluminium standard solution containing
`1,000 mg Al/l
`(Merck, Darmstadt, Germany) was
`used to prepare working standard solutions by suit-
`able dilution with purified water. Cystine and cyste-
`ine were purchased from Sigma (St. Louis, USA) and
`the solutions were prepared by dissolution of the
`amino acid in Milli-Q purified water or in 0.1 M
`HCl. KCl, NaOH, KOH, oxalic acid, citric acid, ni-
`trilotriacetic acid (NTA), and EDTA were all from
`Merck.
`
`Contamination control
`
`To avoid contamination, all laboratoryware (pipette
`tips, volumetric flasks, etc.) was made of plastic
`(polyethylene). It was immersed for at least 48 h in a
`10% HNO3 in ethanol (v/v) mixture and washed with
`Milli-Q purified water shortly before use.
`To avoid contamination from the air, all steps in the
`sample and reagent preparation were carried out in a
`class 100 clean bench.
`
`Preparation of the resins
`
`A strong cationic-exchanger resin (sulphonate–poly-
`styrene resin, Dowex 50 WX 8, 60–170 mesh, H+ form,
`Merck), with an exchange capacity of 4.16 mEq/g (dry
`weight), was used after conditioning by washing with a
`0.1 M NaCl solution and water. The resin was con-
`verted into the Al3+ form by passing 500 ml of 0.5 M
`Al(NO3)3 solution, through a column (1-cm inner
`diameter) containing 10 g resin, with a flow rate of
`1 ml/min. The resin was then washed with water until
`the eluate was relatively free of aluminium (less than
`
`2 lg/l, the limit of detection of graphite-furnace atomic
`absorption spectrometry, GFAAS).
`
`Interaction with the Na exchanger
`
`Interactions of Al with the Na exchanger
`
`Two hundred millilitres of solutions containing from 1
`to 10 mM Al and 0.3 g resin was placed in a closed
`plastic flask and shaken for 135 min. At intervals of
`15 min, aliquots of 1 ml were taken and the Al con-
`centration was measured by flame atomic absorption
`spectrometry (FAAS) or GFAAS.
`
`Interaction of Al with the Na exchanger in the presence
`of amino acids
`
`Individual solutions containing 0.1 mM Al and 5 mM
`ligand (cystine, cysteine, EDTA, NTA, oxalic acid, and
`citric acid) were prepared. Two hundred millilitres of
`these solutions was mixed with 0.3 g Na+-form resin,
`aliquots (500 ll) were taken, and the Al concentration
`measured.
`
`Extraction experiments
`
`Cystine and cysteine solutions in concentrations from 1
`to 10 mM were stored separately in 250-ml polyethyl-
`ene flasks with 0.15 g Al3+-form resin. The flasks were
`stoppered and shaken at room temperature (25 °C).
`Aliquots were taken as follows: each 30 min, for 2 h;
`then daily for 3 days; and after 7, 30, 60, 120, 240, and
`300 days. The Al content was measured by FAAS or
`GFAAS and a blank control was carried out with
`purified water adjusted to the same pH with 0.1 M HCl.
`The solutions were prepared in polyethylene volu-
`metric flasks previously decontaminated as already
`described. The same assay was carried out with solu-
`tions of EDTA, NTA, oxalic acid, and citric acid, with
`aliquots only taken up to 30 days. All test samples
`were assayed in triplicate.
`
`Stability of solutions of cysteine
`
`Since cysteine can be oxidised to cystine, two different
`experiments were carried out to evaluate the stability
`of cysteine solutions. Firstly, the stability of cysteine
`solution was evaluated by measuring its concentration
`with Ellman’s reagent [5,5†-dithiobis(2-nitrobenzoic)
`acid, DTNB]
`[21] at regular periods of
`time for
`30 days. Secondly, the ion-exchange experiments were
`done in an inert atmosphere. The extraction experi-
`ments were carried out as already described but
`
`Nexus Ex. 1011
`Page 3 of 8
`
`123
`

`

`994
`
`J Biol Inorg Chem (2006) 11:991–998
`
`nitrogen was bubbled into all solutions (cystine and
`cysteine). For this, the flasks were covered with latex
`tops and needles were used for gas inlet and outlet.
`Solutions were aired off before contact with the resin
`and also after the flasks were opened to take aliquots
`of the solutions for analysis.
`
`Potentiometric experiments
`
`Potentiometric measurements were performed in a
`jacketed cell thermostated at 25.0 ± 0.1 °C under an
`inert atmosphere of purified nitrogen. To avoid Al
`contamination, the traditional glass cell was replaced
`by a cell made of acrylic. An automatic titration set
`including a Dosimat 665 burette (Metrohm, Herisau,
`Switzerland) and a Corning 350 pH meter fitted with
`glass and calomel reference electrodes was used. In
`order to follow the kinetics conditions established by
`Venturini and Berthon [22], 0.1-ml aliquots of 0.1 M
`KOH were added to the cell at intervals of 2 min. The
`ionic strength was adjusted to 0.01 or 0.1 M with KCl.
`The apparatus was calibrated to read pH, in terms of
`–log [H+], rather than hydrogen activity. Stock solu-
`tions of HCl and KOH were standardised against
`sodium carbonate and potassium hydrogen phthalate,
`respectively. The Gran plot constructed ensured that
`the KOH solution was not contaminated by CO2 [23].
`Amino acid protonation constants, complex forma-
`tion constants, and Al hydrolysis constants were
`determined from three independent titrations (42–109
`points for each titration). The pH range studied was
`4.0–11.5. Solutions with 0:1, 1:1, 1:2, and 1:3 metal-to-
`ligand ratios were investigated with ligand concentra-
`tions of 8.10 mM for cysteine and 0.35 mM for cystine.
`Log Kw for
`the system, defined in terms of
`log [H+][OH–], was found to be –14.09 and –14.23 at
`ionic strengths 0.01 and 0.1, respectively, and it was
`maintained fixed during refinement of the constants.
`Protonation constants were calculated by fitting the
`potentiometric data with the PKAS program. The
`potentiometric data were converted into equilibrium
`constants with the aid of the program BEST [24]. For
`the analysis of the titration curves, the formation of Al
`hydroxo complexes was taken into consideration. For
`the determination of species distribution, the program
`SPE was used. All refinements were carried out in
`order to show rfit < 0.03 [25].
`
`Results and discussion
`
`Experiments with ion-exchanger resins are normally
`done in order to collect either the labile fraction or the
`
`free fraction of an ion that is present in a solution
`containing ligands. The supposition is that, being
`bound to a ligand, the ion is not available to be sorbed
`by the resin [26–28].
`
`Ion-exchange experiments
`
`Preliminary experiments were carried out with both
`metal and ligand in solution in the presence of the
`exchanger. An Na+-form resin was chosen in order to
`avoid or minimise changes in pH, as would occur with
`an H+-form resin. The pH of the solutions was not
`adjusted, since buffers would introduce different spe-
`cies from those derived from Al and amino acids.
`Previous studies on the interaction of Al ions with an
`Na+-form resin (Amberlite IRA-120) showed that even
`in the presence of an excess of Na ions, the trivalent
`metal is always preferably sorbed by the resin [19].
`Moreover, Al ions were totally exchanged after 120 min
`in contact with the Na+-form resin. This means that the
`Na ion, as a ligand counter ion, is not a competitor for
`Al to the resin. The Na/Al exchange was investigated to
`calculate the exchange capacity towards Al ions. The
`exchange isotherm indicated a resin capacity of
`1.35 mmol Al/g (dry weight) and showed that Al ions
`were totally sorbed by the resin in 2 h. The results show
`(Fig. S1 in the supplementary material) that in the
`presence of EDTA, NTA, oxalate, citrate, cystine, and
`cysteine, Al remained in solution. Al ions were there-
`fore not available to be exchanged, inferring the amino
`acids form stable complexes with Al
`in the order:
`EDTA > NTA > cystine > citrate > oxalate > cysteine.
`The results in Fig. 1 show that cystine and cysteine also
`remove Al from the exchanger. The relationship be-
`tween the amount extracted and the stability constants
`of the respective complexes was established. Since the
`conditions were the same (molar concentration, tem-
`perature, and time of contact), the extraction yield
`could be associated to the affinity of the ligand for Al.
`Thus, cysteine forms the least stable Al complex,
`whereas cystine forms a complex as stable as that with
`oxalate and citrate.
`The extraction rates, calculated from the isotherms,
`increased with the ligand concentration (Fig. 2). The
`difference in the extraction rates among the ligands
`confirms the ability to remove Al from the resin and
`the ability to maintain it in solution depends on the
`affinity of the ligand for the metal. Isotherms were
`linearised using the Scatchard approach [29]. Generi-
`cally, the equilibrium in which a metallic cation M and
`a ligand L react to form ML is given by K = [ML]/
`[M][L] (neglecting activity coefficients). In a series of
`solutions in which increments of ligand are added to a
`
`Nexus Ex. 1011
`Page 4 of 8
`
`123
`

`

`J Biol Inorg Chem (2006) 11:991–998
`
`995
`
`constants and those calculated from Scatchard plots
`have good a relationship and present a linear correla-
`tion, with a coefficient of 0.9966, as can be seen in
`Fig. 4.
`
`Potentiometric titrations
`
`Owing to the low solubility of cystine (0.11 g/l at 25 °C,
`approximately 0.46 mM),
`the ligand concentration
`used was 0.35 mM, and the metal-to-ligand ratios were
`0:1, 1:1, 1:2, and 1:3 (0.01 M ionic strength). Under
`these conditions, no Al precipitation was observed.
`The titration of cysteine was carried out with 8.10 mM
`solution, the same metal-to-ligand ratios as for cystine,
`and 0.1 M ionic strength.
`The choice of an adequate ligand concentration to
`carry out potentiometric tirtation seems controversial.
`Kiss et al. [30] recommend the use of a ligand con-
`centration up to 0.5 M as Al complexation should be
`distinguished from hydrolysis occurring at
`ligand
`concentrations greater than 20 mM. In other studies,
`the ligand concentration was low (0.004 or 0.002 M)
`either without explanation [31, 32] or in order to
`avoid stacking between ligand molecules [33]. The
`metal-to-ligand ratio is also important; several studies
`used a ratio in the range 1:1–1:3 [34–36], whereas
`Dayde´ et al. [14] suggest a metal-to-ligand ratio of at
`least 1:10 to avoid Al hydrolysis. We tried to follow
`all the details of the guidelines (different approaches
`such as metal-to-ligand ratio, metal and ligand con-
`centrations,
`inclusion of Al hydrolysis constants,
`waiting time between titrant addition) given by
`experienced researchers [14, 24, 37, 38], but even so
`we did not get rfit lower than 0.03 for the calculations
`carried out with a set of values collected at the con-
`ditions used.
`
`0.2
`
`0.1
`
`[L]bound/[L]free
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`Al solution (%)
`
`EDTA
`
`NTA
`Cys2
`Cit
`Ox
`Cys
`
`0
`
`5
`
`10
`
`20
`15
`time (day)
`
`water pH 4
`
`25
`
`30
`
`35
`
`Fig. 1 Aluminium extracted from an Al3+-form cation exchan-
`ger by the action of ligands as a function of the time. Ligand
`concentration 5 mM. EDTA ethylenediaminetetraacetic acid,
`NTA nitrilotriacetic acid, Cys2 cystine, Cit citrate, Ox oxalate,
`Cys cysteine
`
`constant amount of metal, the total concentration of
`metal is M0, and the following relation may be written:
`[M] = M0 – [ML]. The equilibrium expression can be
`rearranged as
`ML½
`Š= L½
`Š ¼ K M½
`
`Š ¼ K M0 ML½
`ð
`
`Š
`
`Þ:
`
`ð1Þ
`
`The Scatchard plot is the graph of [ML]/[L] versus
`[ML], the slope of which is –K.
`To construct
`this plot, M0 was considered the
`amount of Al in the exchanger and L the ligand con-
`centration in solution. The concentration of ML was
`obtained from the amount of Al measured in solution.
`To calculate the free fraction of the ligand, the amount
`of Al measured in solution (in moles) was subtracted
`from the number of moles of the ligand (initial con-
`centration of the ligand). The plots of [ML]/[L] versus
`[ML] for the ligands investigated are shown in Fig. 3.
`The four ligands, cysteine, and cystine gave good linear
`relationships with correlation coefficients (R2) between
`0.9756 and 0.9968. Table 1 compares the present data
`with published stability constants for the predominant
`species existing in solution. The published stability
`
`Al solution (%)
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`EDTA
`
`NTA
`
`Cys2
`Cit
`Ox
`
`Cys
`
`0
`
`0
`
`2
`
`4
`
`6
`ligand conc. (mM)
`
`8
`
`10
`
`12
`
`Ox
`
`Cys
`
`0
`
`0.0005
`
`Fig. 2 Extraction isotherms of aluminium from an Al3+-form
`cation exchanger by the action of ligands
`
`Fig. 3 Scatchard plots obtained from the extraction isotherms
`
`NTA
`
`EDTA
`
`Cit
`
`Cys2
`
`0.0010
`[L]bound (M)
`
`0.0015
`
`0.0020
`
`Nexus Ex. 1011
`Page 5 of 8
`
`123
`

`

`J Biol Inorg Chem (2006) 11:991–998
`
`r2 = 0.9966
`
`EDTA
`
`NTA
`
`Cys2
`
`Cit
`
`Ox
`
`Cys
`
`0
`
`5
`
`10
`log b
`
`15
`
`20
`
`3
`
`2
`
`1
`
`0
`
`log Scatchard slope
`
`Fig. 4 Correlation of the stability constants from the Sctachard
`plots and from potentiometric titration for cysteine and cystine
`complexes calculated in this work and for EDTA, NTA, citrate,
`and oxalate collected from the literature [11, 25, 42, 43]
`
`The protonation constants obtained by titration for
`the stepwise amino acid dissociation by computer-
`simulated titration curves are presented in Table 2.
`The overall fits were r = ± 0.017 pH units and
`r = ± 0.008 pH units for cysteine and cystine, respec-
`tively. Since the calculated constants are in agreement
`with the proton dissociation constants published for
`both amino acids [11, 39], the experimental approach
`can give reliable data for the calculation of stability
`constants of their Al complexes. The metal depresses
`the titration curve for both amino acids by the release
`of protons according to the ability of Al to bind the
`amino acids. No 1:2 complexes were derived from the
`Al–cysteine system under the experimental conditions
`used in this work. The overlapping of the titration
`curves for 1:1 and 1:2 metal-to-ligand ratios for the
`cystine–Al system provides, according to O¨ hman [40],
`evidence that complexation in this case is characterised
`3–2n species.
`by the formation of a simple AlLn
`A representative distribution diagram of the com-
`plexes as a function of pH is depicted in Fig. 5. For
`cystine, the complexation started at pH 4.5 with the
`formation of the complex ML. At pH 5, MLH–3 started
`being formed and it is the major species above pH 6.
`No turbidity or precipitation was observed during the
`titration. The stability constants for these complexes
`are shown in Table 2. Because Al
`forms several
`hydrolytic species in aqueous solution, Al hydrolysis
`(log b(ML) = 8.5;
`log b(ML2) = 16.7;
`constants
`log b(ML4) = 34), determined in the
`log b(ML3) = 27;
`absence of the ligand [41–43], were introduced in the
`refinement of the constants.
`For cysteine, the complexation started at pH 4.0
`with the formation of the complex MLH–1 up to pH
`
`bObtainedbypotentiometry
`aDatafromspeciesdistributiondiagramsfromthereferencecitedinthelastcolumn
`EDTAethylenediaminetetraaceticacid,NTAnitrilotriaceticacid,Citcitrate,Oxoxalate,Cys2cystine,Cyscysteine
`
`Thisworkb
`Thisworkb
`
`[11]
`
`[25,42]
`
`[42,43]
`
`[25,42]
`
`0.1
`0.01
`
`0.1
`
`0.1
`
`0.1
`
`0.1
`
`6.2
`10.3
`10.93
`5.97
`
`6.23
`7.98
`1.9
`11.1
`
`2.14
`
`14.4
`
`[MH–1L]/[MH–1][L]
`[ML]/[M][L]
`[ML]/[M][L]2
`[ML]/[M][L]
`(OH)][H]
`
`[MH–1L]/[M(H–1L)
`[ML]/[M][L]
`[MHL]/[ML][H]
`[ML]/[M][L]
`[MHL]/[ML][H]
`[ML]/[M][L]
`
`1.0
`1.0
`0.05
`0.95
`
`0.2
`0.8
`0.13
`0.87
`0.25
`0.75
`
`AlH–1Cys
`AlCys2
`Al(Ox2)–
`AlOx+
`
`Al(H–1Cit)–
`AlCit0
`AlHNTA
`AlNTA
`AlHEDTA
`AlEDTA
`
`2.5–3.1
`
`1.9–2.9
`
`2.0–2.5
`
`4–5
`4–5
`
`0.9878
`0.9892
`
`–11.94x+0.018
`–20.39x+0.052
`
`1.6–2.1
`
`0.9968
`
`–13.01x+0.021
`
`0.9756
`
`–18.83x+0.468
`
`0.9895
`
`–30.23x+0.070
`
`Nitrilotriaceticacid
`
`acid
`
`0.9866
`
`–46.60x+0.113
`
`Ethylenediaminetetraacetic
`
`Cysteine
`Cystine
`
`Oxalate
`
`Citrate
`
`Reference
`
`(M)
`strength
`Ionic
`
`logQ
`
`Quotient,Q
`
`species
`ofeach
`Fraction
`
`thispHa
`tobepresentat
`Speciessupposed
`
`pH
`Solution
`
`plot
`R2Scatchard
`
`plot
`Sctachard
`Regression
`
`Ligand
`
`996
`
`thesepHvalues(T=25°C)
`Table1DataoftheScatchardlinearisationplotfortheAl–ligandcomplexes,thepHoftheligandsolutions,andthestabilityconstantsoftheAlcomplexespresentinsolutionat
`
`Nexus Ex. 1011
`Page 6 of 8
`
`123
`

`

`J Biol Inorg Chem (2006) 11:991–998
`
`997
`
`well documented by the group of Berthon. They
`consider that, regarding Al hydrolysis, three situations
`are possible [14]: (1) Al–ligand complexation is strong
`and renders hydrolysis negligible; (2) Al–ligand com-
`plexation is medium, hydrolysis is negligible within the
`acidic pH range and formation constants of minor
`hydroxides may be refined; (3) Al–ligand complexa-
`tion is weak, hydrolysis prevails over the whole pH
`range, true equilibrium is never reached, and hydro-
`lysis constants determined in the absence of ligand
`must be applied to keep the Al–ligand complex for-
`mation constant as realistic as possible. We believe
`that in the case of cystine, the complexation is strong
`enough to reach a true equilibrium without any pre-
`cipitation and to make hydrolysis negligible within the
`pH range investigated. On the other hand, for cyste-
`ine, a weaker complex is formed and Al hydrolysis was
`not negligible.
`If the species present in solution are those displayed
`in the species distribution diagram at pH 4–5, their
`constants are in agreement with the ones calculated by
`the regression of the curve obtained by correlating the
`slope of the Scatchard plots (log) with the stability
`constants found in the literature (Fig. 4).
`
`Stability of cysteine solutions
`
`The evaluation of the stability of cysteine solutions by
`reaction with DTNB showed that, at the conditions
`established for the experiment, the conversion of cys-
`teine into cystine occurred at a rate of 0.3% per day.
`After 100 days, the Al extraction yield of the cysteine
`solution is practically the same as that of the cystine
`solution when the experiment is conducted in the
`presence of oxygen (air) (Fig. S2 in the supplementary
`material). In the absence of oxygen, the Al extraction
`in the cysteine solution stopped increasing after 7 days,
`as it did for solutions of the other ligands. The slow
`increase in the percentage of Al extracted is due to the
`conversion of cysteine into cystine, which has more
`affinity for Al. Since this reaction is very slow, there
`was no interference during the potentiometric titration
`in cysteine systems.
`
`Conclusions
`
`A cation exchanger in Al form was used as a substrate
`to produce Al ions in solution. Solutions containing
`species able to form Al complexes, such as EDTA,
`NTA, and citric acid, extracted Al from the resin
`according to the relative affinity of the ligand for Al.
`Similar behaviour was observed with the amino acids
`
`Table 2 Protonation constants of cysteine and cystine, and
`(log b)
`aluminium stability constants
`for
`their
`respective
`complexes at 25 °C and I = 0.1 M (cysteine system) or
`I = 0.01 M KCl (cystine system)
`
`System
`
`Proton–cysteine
`
`Al–cysteine
`
`Proton–cystine
`
`Al–cystine
`
`Species
`
`HL
`H2L
`H3L
`ML
`MH–1L
`MH–2L
`MH–3L
`MHL
`HL
`H2L
`H3L
`H4L
`ML
`MH–1L
`MH–2L
`MH–3L
`
`log b
`
`11.04 ± 0.06
`19.79 ± 0.06
`22.77 ± 0.13
`6.45 ± 0.10
`6.15 ± 0.09
`–2.80 ± 0.08
`–15.47 ± 0.22
`6.38 ± 0.10
`9.80 ± 0.04
`18.50 ± 0.28
`22.55 ± 0.11
`23.52 ± 0.05
`10.28 ± 0.13
`1.65 ± 0.05
`–5.61 ± 0.01
`–5.22 ± 0.01
`
`bpqs = [MpLqHs]/[M]p [L]q [H]s
`
`9.0, where the species MLH–2 formed. From pH 4.6 to
`10.0, a slight
`turbidity was observed. Because of
`this turbidity, hydrolysis probably occurred and Al
`hydrolysis constants, determined in the absence of the
`ligand, were introduced in the refinement of
`the
`constants.
`The problem of Al hydrolysis and Al precipitation
`during the titration has been discussed [41–43] and
`
`(a)
`
`100
`
`CysH2
`
`CysAl(OH)2
`
`CysAl(OH)
`
`80
`
`60
`
`40
`
`20
`
`0
`
`3
`
`% species
`
`(b)
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`4
`
`% species
`
`CysH3
`
`4
`
`5
`
`6
`
`7
`
`pH
`
`8
`
`9
`
`10
`
`11
`
`Cys2H2
`
`Cys2H3
`
`5
`
`6
`
`Cys2Al(OH)3
`
`Cys2Al
`
`7
`
`pH
`
`8
`
`9
`
`10
`
`Fig. 5 Species distribution diagram for the systems a cysteine
`and b cystine as a function of pH. CAl = Cligand = 1 mM
`
`Nexus Ex. 1011
`Page 7 of 8
`
`123
`

`

`998
`
`J Biol Inorg Chem (2006) 11:991–998
`
`cystine and cysteine. The extraction rates and stability
`constants determined by potentiometric measurements
`follow the order EDTA > NTA > cystine @ citric
`acid > oxalic acid > cysteine.
`This study explains why solutions used clinically for
`nutritional support and containing cystine/cysteine can
`have high concentrations of Al. In particular, if cystine
`is in contact with Al from containers or other compo-
`nents during production, it may cause the metal to
`migrate into the solution owing to complex formation.
`It is advisable therefore to ensure that materials, e.g.
`glass or plastics, used during formulation, preparation,
`and storage of parenteral nutrition solutions do not
`contain easily exchangeable aluminium.
`
`Acknowledgement The authors are grateful to CNPq (Conse-
`lho Nacional de Desenvolvimento Tecnolo´ gico) for financial
`support.
`
`References
`
`1. Koo WW, Kaplan LA (1988) J Am Coll Nutr 7:h199–h214
`2. Klein GL (1998) Nutrition 14:149–152
`3. Sedman AB, Klein GL, Merritt RJ, Miller NL, Weber KO,
`Gill WL, Anand H, Alfrey AC (1985) N Engl J Med
`312:1337–1343
`4. Bishop NJ, Morley R, Day JP, Lucas A (1997) N Engl J Med
`336:1557–1561
`5. Koo WW (1996) Clin Biochem 29:429–438
`6. Venturin M, Berthon G (1989) J Inorg Biochem 37:69–90
`7. Berthon G (2002) Coord Chem Rev 228:319–341
`8. Sanz-Medel A, Soldado Cabezuelo AB, Milacic R, Polak TB
`(2002) Coord Chem Rev 228:373–383
`9. Bohrer D, do Nascimento PC, Binotto R, Becker E (2003)
`J Trace Elem Med Biol 17:107–115
`10. Bohrer D, do Nascimento PC, Binotto R, Becker E, Pomb-
`lum SC (2002) J Parenter Enter Nutr 26(6):383–388
`11. Martell AE, Smith RM, Motekaitis RJ (1998) NIST standard
`reference database 46, NIST critically selected stability
`constants of metal complexes database, version 5.0
`12. Perrin DD (1979) Stability constants of metal-ion complexes.
`Part B: organic ligands. IUPAC chemical data series no 22.
`Pergamon, Oxford
`13. Berthon G (1995) Pure Appl Chem 67:1117–1240
`14. Dayde´ S, Champmartin D, Rubini P, Berthon G (2002) Inorg
`Chim Acta 339:513–524
`
`15. Dayde´ S, Brumas V, Champmartin D, Rubini P, Berthon G
`(2003) J Inorg Biochem 97:104–117
`16. Bohrer D, do Nascimento PC, Binotto R, Carlesso R (2001)
`J Trace Elem Med Biol 15:103–108
`17. Driscoll M, Driscoll D (2005) Am J Health Syst Pharm
`62:312–315
`18. Canada T (2005) Am J Health Syst Pharm 62:315–318
`19. Bohrer D, do Nascimento PC, Martins P, Binotto R (2002)
`Anal Chim Acta 459:26 7–276
`20. Elliott WH, Elliott DC (1997) Biochemistry and molecular
`biology. Oxford University Press, Oxford
`21. Burtis CA, Ashwood ER (eds) (2001) Tietz fundamentals of
`clinical chemistry. Saunders, Philadelphia
`22. Venturini M, Berthon G (1987) J Chem Soc Dalton Trans
`1145–1148
`23. Rossotti FJC, Rossoti H (1965) J Chem Ed 42:375–378
`24. Martell AE, Motekaitis RJ (1992) Determination and use of
`stability constants, 2nd edn. Wiley-VCH, Weinheim
`25. Martell AE, Motekaitis RJ, Smith RM (1990) Polyhedron
`9:171–187
`26. Burba P (1994) Fresenius’ J Anal Chem 348:301–311
`27. Yuchi, Sato T, Morimoto Y, Mizuno H, Wada H (1997) Anal
`Chem 69:2941–2944
`28. Pesavento M, Alberti G, Biesuz R (1998) Anal Chim Acta
`367:215–222
`29. Scatchard G (1949) Ann N Y Acad Sci 51:660–672
`30. Kiss MT, So´ va´ go´ I, To´ th I, Lakatos A, Bertani R, Taparo A,
`Bombi G, Martin RB (1997) J Chem Soc Dalton Trans 1967–
`1972
`31. Motekaitis R, Martell AE (1984) Inorg Chem 23:18–23
`32. Li Y, Martell AE (1995) Inorg Chim Acta 231:159–163
`33. Atka´ ri K, Kiss T, Bertani R, Martin RB (1996) Inorg Chem
`35:7089–7094
`34. Champmartin D, Rubini P, Lakatos A, Kiss T (2001) J Inorg
`Biochem 84:13–21
`35. Rubini P, Lakatos A, Champmartin D, Kiss T (2002) Coord
`Chem Rev 228:137–152
`36. Marklund E, O¨ hman LO (1990) Acta Chem S