a study of the interaction between humic acids and acidic metabolites of phenanthrene by capillary...

Enrica DonatiChiara M. Polcaro

Institute of ChemicalMethodologies, NationalResearch Council, Research Areaof Rome 1, MonterotondoStazione (RM), Italy

Short Communication

A study of the interaction between humic acids andacidic metabolites of phenanthrene by capillaryelectrophoresis

Humic acids (HA) are able to interact with a wide range of pollulants and caninfluence their solubility, transport and bioavailability. In order to study the inter-action between polar aromatic hydrocarbons and these macromolecules, an affinitycapillary electrophoretic method, the Hummel–Dreyer (HD) method in its modifiedversion, has been used. Two acidic metabolites of phenanthrene: 1-hydroxy-2-naph-thoic acid (1-HNA) and 3,4-dihydroxybenzoic acid (3,4-DBA) were studied. The anal-ysis for the binding studies was carried out by injecting a solution of HA in 25 mMammonium hydrogen carbonate buffer (pH 9) into an uncoated fused-silica capil-lary, filled with buffer solutions of 1-HNA or 3,4-DBA in varying and increasingamounts. The results obtained indicate that both compounds bind to HA, as hadbeen confirmed by dialysis experiments and literature data. CE proved to be a usefultechnique for investigating the link between xenobiotics and environmental macro-molecules.

Keywords: CE / Complexes / Humic acids / PAH /

Received: May 12, 2006; revised: July 19, 2006; accepted: July 21, 2006

DOI 10.1002/jssc.200600197

1 Introduction

In many biological systems, binding with macromole-cules is frequently employed for the storage or mobilisa-tion of low molecular weight organic compounds andmetals. Many different techniques have been used tostudy this kind of interaction, among which are equili-brium dialysis, chromatography, quenching of fluores-cence and, more recently, CE. This technique is a veryimportant tool for the study of electrically charged mole-cules and consequently has been utilised in the researchconcerning the association or reaction equilibriabetween proteins and pharmaceuticals [1, 2].

Humic acids (HA) are an important class of macromole-cules found in soil and water. They are polymeric com-pounds of different molecular weights formed by themicrobial degradation of vegetal matter. Their chemicalstructure has not yet been defined precisely, but, in anycase, the HA molecules contain a great number of aro-

matic rings, aliphatic chains and functional groups (car-boxylic, phenolic, hydroxylic, carbonylic, amine andamides). Consequently, they are able to interact with awide range of pollulants and can influence the solubility,transport, bioavailability and lower the toxicity of watersolutions of dangerous organic compounds, such as poly-cyclic aromatic hydrocarbons (PAH), for aquatic organ-isms [3].

However, the influence of HA on the microbial degrada-tion of organic pollulants is not completely clear [4]: Infact, an enhanced water solubility has a positive effect onbiodegradation, but a lack of bioavailability can protectit against microbial degradation. Furthermore, the bind-ing between HA and organic compounds is related to thechemical structure of their molecules: The inclusioncomplexes between HA and PAH are formed as a result ofhydrophobic interaction between aromatic rings,whereas more polar bacterial oxidation products canalso interact in different ways with functional groups inthe HA molecules. Radiochemical studies on PAH biode-gradation in the presence of HA have demonstrated thebiogenic formation of chemically bonded compounds[5]. Chemical degradation techniques made evident thepresence of reaction products (mainly esters) formedbetween PAH acidic metabolites and HA [3, 4]. Besides, itwas observed that a fraction of the acidic PAH was notchemically bonded but formed inclusion compounds [6].

Correspondence: Dr. Chiara M. Polcaro, Institute of ChemicalMethodologies (IMC), National Research Council (CNR), ResearchArea of Rome 1, 00016 Monterotondo Stazione (RM), ItalyE-mail: [email protected]: +39-06-90672269

Abbreviations: 3,4-DBA, 3,4-dihydroxybenzoic acid; HD, Hum-mel–Dreyer; 1-HNA, 1-hydroxy-2-naphthoic acid; PAH, polycyclicaromatic hydrocarbons; HA, humic acids

i 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

J. Sep. Sci. 2006, 29, 2853 –2857 E. Donati et al. 2853

2854 E. Donati et al. J. Sep. Sci. 2006, 29, 2853 –2857

In the present work the capillary electrophoreticmethod, namely the Hummel–Dreyer (HD) method in itsmodified version [7, 8] usually utilised for binding stud-ies, was applied as an experiment to evidence the interac-tions between HA and two acidic metabolites of Phenan-threne, a wide-spread PAH. These compounds, 1-hydroxy-2-naphthoic acid (1-HNA) and 3,4-dihydroxybenzoic acid(3,4-DBA) (Fig. 1), have been demonstrated to link withHA both covalently and by means of the formation ofinclusion complexes [5].

2 Experimental

2.1 Chemicals and reagents

3,4-DBA was purchased from Fluka (Buchs, Switzerland)and 1-HNA was obtained from Aldrich (Deisenhofen, Ger-many). All water used was purified with a Milli-Q waterpurification system (Millipore, Billerica, MA, USA). ACNwas HPLC grade, all other reagents were of analytical-grade purity and were purchased from Carlo Erba(Milano, Italy).

The HA was supplied by Fluka: As described in the techni-cal information report provided by the manufacturersfor the sample used in the present work (lot no. 53680),the origin is brown coal, which comes from the northernpart of Germany. The HA Na salt was prepared by alkalinedigestion, then a hydrochloric acid solution (pH l1) wasadded, left to stand for a 2–3-month period and decantedoff. The brown residue was dried and the final producthas the characteristics described in Table 1.

2.2 HPLC apparatus and procedures

The HPLC apparatus was composed of a LabFlow 3000Labservice Analytica (Anzola Emilia, Italy) pumping sys-

tem, a 7125 Rheodyne Valve (USA) as an injector and anSPD-M6A diode array detector (Shimadzu, Kyoto, Japan).The HPLC analyses were performed with an RP18 column(Synergi Fusion-RP 80, 150 mm64.60 mm, Phenomenex,USA). Chromatographic analyses were conducted in iso-cratic conditions with solutions of ACN in water, towhich 0.7% acetic acid was added, as mobile phases. AnACN content of 60% was used for 1-HNA analysis and of8% for 3,4-DBA determination; both metabolites weredetected in UV at 254 nm.

2.3 CE apparatus and procedures

A Hewlett Packard 3DCE instrument (Agilent Technolo-gies, USA) equipped with a diode array detector was used.Data were acquired and stored by a ChemStation soft-ware. Separations were carried out in an uncoated fused-silica capillary (total length 33 cm, effective length24.5 cm, 75 lm id) purchased from Composite Metal Ser-vices (Hallow, Worcestershire, UK).

2.3.1 CE experiments

Before being used for the first time a new capillary wasconditioned as follows: 10 min 1 M NaOH at 408C,10 min 0.1 M NaOH at 408C, 10 min water at 308C andfinally 15 min buffer at 208C. The capillary was washedevery morning by flushing with water for 5 min, 0.1 MNaOH for 20 min, water for 10 min, buffer for 10 minand between analyses by rinsing with water, 0.1 MNaOH, water and buffer, each for 5 min. In the bindingstudies the capillary was filled with the buffer contain-ing a known amount of metabolite for 3 min before eachrun. In order to obtain a stable and reproducible EOF, theelectrolyte in inlet and outlet vials was changed at everythree runs.

Hydrodynamic injection at a pressure of 50 mbar per 5 sfrom the positive end was used for sample introduction.Electrophoretic runs were carried out at a temperatureof 208C and a constant voltage of 15 kV, in a 25 mMammonium hydrogen carbonate buffer, pH 9, passedthrough a 0.45 lm disposable membrane filter (LIDAKenosha, WI, USA) before use. Wavelength detection wasset at 214 nm.


Figure 1. Chemical structure of phenanthrene acidic meta-bolites.

Table 1. Characteristic parameters of Fluka HA – lot no.53680

Molecular weight 2000–500 000(mainly 20 000–50 000)

Solubility 0.1 g in 10 mL H2OpH 6.5Residue on ignition 15.8%Carbon content 48.95%Hydrogen content 4.34%Nitrogen content 0.72%

J. Sep. Sci. 2006, 29, 2853 –2857 CE study of the interaction between humic acids and acidic PAH 2855

2.3.2 Sample preparation

2.3.2.1 HA

A specific amount of HA was dissolved in a 25 mM ammo-nium hydrogen carbonate buffer (pH 9) to obtain a work-ing solution of 0.5 mg/mL. An ultrasound bath wasapplied for 10 min to enhance the dissolution. The sam-ple was filtered through a 0.45 lm microporous mem-brane before use.

2.3.2.2 Metabolites

Stock solutions of 0.5 mg/mL of 1-HNA and 3,4-DBA wereprepared by dissolving the corresponding weight in a25 mM ammonium hydrogen carbonate buffer, pH 9. Inthe case of CE analysis, the stock solutions were dilutedwith the buffer in the range 0.01–0.15 mg/mL for 1-HNAand 0.03–0.18 mg/mL in the case of 3,4-DBA. Before useall metabolite working solutions were filtered through a0.45 lm microporous membrane.

2.3.3 Calibration and binding studies

Preliminarly, the linear correlation between the injectedmetabolite concentration and the peak area value waschecked by injecting single metabolite solutions ofincreasing concentrations in the capillary, filled withthe buffer. A linear response was obtained in the concen-tration range: 0.01–0.15 mg/mL in the case of 1-HNA;0.03–0.18 mg/mL in the case of 3,4-DBA.

The calibration plot for each metabolite was obtained byfilling the capillary with the buffer containing differentand increasing amounts of metabolite in the calibrationrange previously tested and injecting an ammoniumhydrogen carbonate buffer.

Binding was determined by injecting the HA solution inthe capillary filled with the metabolite (1-HNA or 3,4-DBA) containing buffer, under the same conditions as inthe calibration plot. Each point of the calibration graphswas the mean of values obtained from three distinctpeak area measurements (Figs. 2 and 3).

2.4 Dialysis experiments

Dialysis experiments were performed using Visking dia-lysis membranes (1350 MWCO) (Medicell International,London, UK): Prior to use the membranes were boiled inHPLC grade water and soaked overnight.

Two solutions of HA and 1-HNA or 3,4-DBA in a 25 mMammonium hydrogen carbonate buffer (pH 9) were pre-pared, the one at a concentration of HA 0.5 mg/mL and 1-HNA 0.15 mg/mL and the other at a concentration of HA0.5 mg/mL and 3,4-DBA 0.18 mg/mL. The dialysis tubeswere filled with 10 mL of each solution and placed in abeaker containing 100 mL of the same buffer. Solutionswere agitated in the dark at room temperature (208C)

with a magnetic stirrer. All experiments were repeatedthree times and the results are expressed as meanvalues € SD. The concentration of the free compound inthe external solution was checked every hour by HPLC.Quantitation of the metabolites in the dialyzed solutionswas obtained from the corresponding calibration curvesat 254 nm.

When an equilibrium state was reached, the percentagefraction of HA-bound PAH (Fb %) was calculated accordingto the equation:

Fb % = [(Qtot – Qfree)/Qtot]6100

where Qtot is the total amount of metabolite at zero timeand Qfree is the amount of the free metabolite at equili-brium time.


Figure 2. Calibration and working curves obtained by inject-ing the buffer, or its HA solution, into a capillary filled with thebuffer containing different amounts of 3,4-DBA.

Figure 3. Calibration and working curves obtained by inject-ing buffer, or its HA solution, into a capillary filled with thebuffer containing different amounts of 1-HNA.

2856 E. Donati et al. J. Sep. Sci. 2006, 29, 2853 –2857

3 Results and discussion

Owing to their particular nature, HA are known to inter-act with zwitterions and to complex metal cations, parti-cularly sodium salts [9]; furthermore, phenolic or car-boxylic groups present in their molecules react withboric acid or borates [10–12]. Therefore special attentionmust be paid to the choice of buffer used for the CE bind-ing studies, because a possible interaction between HAand buffer ions is highly probable. With this aim inmind, the more suitable electrolytes are ammoniumacetate and ammonium hydrogen carbonate buffers,because they present little or no interaction with humicsamples [8]. Both HA and the studied PAH metabolites aredissociated at neutral pH, but the shape of HA electro-phoretic peaks was very broad, with acetate as BGE: Thisbehaviour, together with a lower solubility than underalkaline conditions, created serious detection problems.Consequently, in the present work an ammonium hydro-gen carbonate buffer (pH 9) was used. In order to avoidsystem peaks, due to different ionic strength distribu-tions within the capillary, the ionic strength of BGE waskept at 25 mM and the applied voltage at 15 kV [10].

Unfortunately under these conditions, also due to theslight differences in molecular weight and electriccharge, the HA and HA–PAH complex mobilities werevery similar. For this reason the modified HD methodwas used, because it is known to ensure accurate resultsin binding parameters when the mobilities of the macro-molecules and their complexes are equal [13] and, inaddition, it involves a limited number of HA–metabolitecomplex equilibria, as the equilibrium is reached duringthe CE run [2].

The HA–metabolite interaction was studied by injectingHA solutions in the capillary that had already been filledwith the buffer containing 1-HNA or 3,4-DBA. First of all,a calibration curve was calculated for each compound byfilling the capillary with the buffer containing varyingand increasing amounts of metabolites and injecting theammonium carbonate buffer, as described in Section2.3.3. Owing to the presence of an absorbing compoundin the BGE, the injection of a neat buffer caused avacancy, which resulted as a negative peak at the samemigration time of the metabolite in the electrophero-gram (Figs. 4a and b; 5a and b). The more the compoundconcentration in the BGE increased the more the area ofthe negative peak grew linearly. The calibration curvewas identical to the one obtained by the injection of themetabolite (Section 2.3.3).

When the buffer solution of HA is injected, the electro-pherograms show a positive peak, corresponding to thefree HA comigrating with HA–PAH complexes, and anegative peak at the same migration time of the acidicPAH (Figs. 4c and 5c). If a complex is formed, the negative

peak area is bigger than the area obtained by injectingthe neat buffer, as it is the sum of this value and thevacancy corresponding to the amount of HA-bound com-pound. Figures 2 and 3 represent, respectively, the curvescorresponding to 3,4-DBA and 1-HNA obtained by plot-ting on the same graph the negative peak areas, relatedto the buffer injection or HA injection versus the metabo-lite concentration in the running buffer. Both the curvesare linear and the difference between them correspondsto the extent of the HA bound – acidic PAH.

As the HA commercial samples are a mixture of a greatvariety of compounds with an undetermined number ofactive sites (Section 2.1), the observed interactionsbetween the acidic PAH and HA is the sum of the interac-tions with each single component of the HA mixture.The extent of this binding is expressed as the percentageof the bound substance.

In the case of 3,4-DBA, the extent of binding depends onthe PAH concentration in the BGE, ranging from 10% €0.4 SD at 0.03 mg/mL to 19% € 0.6 SD at 0.18 mg/mL,while in the case of 1-HNA it is quite constant in therange 0.03–0.15 mg/mL in the BGE (between 5% € 0.3 SDand 7% € 0.4 SD).


Figure 4. Electropherograms obtained by injecting into anuncoated capillary: (a) 1-HNA 0.03 mg/mL in a 25 mMammonium hydrogen carbonate buffer (pH 9); BGE: a25 mM ammonium hydrogen carbonate buffer (pH 9). (b) A25 mM ammonium hydrogen carbonate buffer (pH 9); BGE:1-HNA 0.03 mg/mL in a 25 mM ammonium hydrogen carbo-nate buffer (pH 9). (c) HA 0.5 mg/mL in a 25 mM ammoniumhydrogen carbonate buffer (pH 9); BGE: 1-HNA 0.03 mg/mLin a 25 mM ammonium hydrogen carbonate buffer (pH 9).

J. Sep. Sci. 2006, 29, 2853 –2857 CE study of the interaction between humic acids and acidic PAH 2857

The equilibrium dialysis was used to confirm the bindingcapacity of HA towards the two acidic PAH. As describedin Section 2, two separate series of experiments were per-formed: 3,4-DBA or 1-HNA was dissolved with HA in thesame buffer used for electrophoretic runs. The equili-brium state was obtained in 7 h in the case of 3,4-DBAand in 27 h in the case of 1-HNA, and the bound fractionwas 22% € 0.8 SD and 8.9% € 0.5 SD, respectively.

The values of the percentage of binding obtained by dia-lysis are shown in Table 2.

4 Concluding remarksThe binding studies performed by both CE and dialysisunder the above-mentioned conditions demonstrate thepresence of an interaction between HA and PAH polarmetabolites. A thorough investigation over the nature ofthis interaction was not possible owing to the use of acomplex mixture of macromolecular compounds such asthe commercial HA sample: More interesting resultscould be obtained by the performance of CE-bindingexperiments using chromatographically purified frac-tions of HA. Furthermore, different conditions could betested in order to study a possible pH dependent varia-tion of the binding: This investigation would require theindividuation of interference free electrolytes, whichcan operate in a different range of pH.

In conclusion CE with the HD modified method can beconsidered a fast and useful tool in the study of ionisableenvironmental pollulants–HA interactions and resultsshould be promising in the future.

Our sincere thanks to Professor M. G. Quaglia for having intro-duced the authors to CE-binding studies.

5 References[1] Busch, M. H. A., Carels, L. B., Boelens, H. F. M., Kraak, J.

C., Poppe, H., J. Chromatogr. A 1997, 777, 329 –353.

[2] Quaglia, M. G., Boss�, E., Dell'Aquila, C., Guidotti, M., J.Pharm. Biomed. Anal. 1997, 15, 1033 –1039.

[3] Perminova, I. V., Grechshcheva, N. Y., Kovalevskii, D. V.,Kudryavstev, A. V. et al., Environ. Sci. Technol. 2001, 35,3841–3848.

[4] Lesage, S., Li, W., Millar, K., Brown, S., Liu, D., Proceedingsof 6th Symposium and Exhibition on Groundwater and SoilRemediation, Montral, Quebec, Canada, 18 –21 March1997, 325 –338.

[5] Richnow, H. H., Seifert, R., Hefter, J., Kastner, M. et al.,Org. Geochem. 1994, 22, 671 –681.

[6] Kacker, T., Haupt, E. T. K., Garms, C., Francke, W., Stein-hart, H., Chemosphere 2002, 48, 117 –131.

[7] Busch, M. H. A., Carels, L. B., Boelens, H. F. M., Kraak, J.C., Poppe, H., J. Chromatogr. A 1997, 777, 311 –328.

[8] Oravcov�, J., Bohs, B., Lindner, W., J. Chromatogr. B 1996,677, 1–28.

[9] Pacheco, M., Havel, J., Electrophoresis 2002, 23, 268 –277.

[10] Pacheco, M. L., Pena-M�ndez, E. M., Havel, J., Chemosphere2003, 51, 95–108.

[11] Schmitt-Kopplin, P., Junkers, J., J. Chromatogr. A 2003,998, 1–20.

[12] Fetsch, D., Havel, J., J. Chromatogr. A 1998, 802, 189 –202.

[13] Rudnev, A. V., Alensko, S. S., Semenova, O., Hartinger, C.G. et al., J. Sep. Sci. 2005, 28, 121 –127.


Figure 5. Electropherograms obtained by injecting into anuncoated capillary: (a) 3,4-DBA 0.03 mg/mL in a 25 mMammonium hydrogen carbonate buffer (pH 9); BGE: a25 mM ammonium hydrogen carbonate buffer (pH 9). (b) A25 mM ammonium hydrogen carbonate buffer (pH 9); BGE:3,4-DBA 0.03 mg/mL in a 25 mM ammonium hydrogen car-bonate buffer (pH 9). (c) HA 0.5 mg/mL in a 25 mM ammo-nium hydrogen carbonate buffer (pH 9); BGE: 3,4-DBA0.03 mg/mL in a 25 mM ammonium hydrogen carbonate buf-fer (pH 9).

Table 2. Percentage of the bound fraction of each metaboliteto HA (0.5 mg/mL) determined by equilibrium dialysis andCE methods

Metabolite Concentration(mg/mL)

% Boundfraction (CE)

% Boundfraction (ED)

3,4-DBA 0.18 19 € 0.6 22 € 0.81-HNA 0.15 7 € 0.4 8.9 € 0.5

The results are expressed as mean values from three differ-ent determinations € SD.

a study of the interaction between humic acids and acidic metabolites of phenanthrene by capillary...

Documents