Sie sind auf Seite 1von 15

J Biol Inorg Chem (2009) 14:61–74 DOI 10.1007/s00775-008-0424-1



Simultaneous Cu-, Fe-, and Zn-specific detection of metalloproteins contained in rabbit plasma by size-exclusion chromatography– inductively coupled plasma atomic emission spectroscopy

Shawn A. Manley Æ Simon Byrns Æ Andrew W. Lyon Æ Peter Brown Æ Ju¨ rgen Gailer

Received: 16 April 2008 / Accepted: 23 August 2008 / Published online: 10 September 2008 SBIC 2008

Abstract Analytical methods which are capable of deter- mining the plasma or serum metalloproteome have inherent diagnostic value for human diseases associated with increased or decreased concentrations of specific plasma metalloproteins. We have therefore systematically devel- oped a method to rapidly determine the major Cu-, Fe-, and Zn-containing metalloproteins in rabbit plasma (0.5 mL) based on size-exclusion chromatography (SEC; stationary phase Superdex 200, mobile phase phosphate-buffered sal- ine pH 7.4) and the simultaneous online detection of Cu, Fe, and Zn in the column effluent by an inductively coupled plasma atomic emission spectrometer (ICP-AES). Whereas most previous studies reported on the analysis of serum, our investigations clearly demonstrated that the analysis of

Parts of the work described in this paper were presented at HPLC

2007 in Ghent, Belgium.

Electronic supplementary material

article (doi:10.1007/s00775-008-0424-1) contains supplementary material, which is available to authorized users.

The online version of this

S. A. Manley S. Byrns J. Gailer ( & )

Department of Chemistry, University of Calgary,

2500 University Drive NW,

Calgary, AB T2N 1N4, Canada e-mail:

A. W. Lyon

Department of Pathology and Laboratory Medicine, University of Calgary and Calgary Laboratory Services, 9, 3535 Research Rd NW, Calgary, AB T2L 2K8, Canada

P. Brown

Teledyne Leeman Labs, 6 Wentworth Drive, Hudson, NH 03051, USA

plasma within 30 min of collection results in the detection of one more Cu peak (blood coagulation factor V) than has been previously reported (transcuprein, ceruloplasmin, albumin-bound Cu, and small molecular weight Cu). The average amount of Cu associated with these five proteins corresponded to 21, 18, 21, 30 and 10% of total plasma Cu, respectively. In contrast, only two Fe metalloproteins (fer- ritin and transferrin, corresponding to an average of 9 and 91% of total plasma Fe) and approximately five Zn metal- loproteins (a 2 -macroglobulin and albumin-bound Zn, which corresponded to an average of plasma Zn) were detected. Metalloproteins were assigned on the basis of the coelution of the corresponding metal and protein identified by immunoassays or activity-based enzyme assays. The SEC- ICP-AES approach developed allowed the determination of approximately 12 Cu, Fe, and Zn metalloproteins in rabbit plasma within approximately 24 min and can be applied to analyze human plasma, which is potentially useful for diagnosing Cu-, Fe-, and Zn-related diseases.


chromatography Inductively coupled plasma atomic

emission spectrometry Metalloproteins

Blood plasma Size-exclusion


All organisms must regularly ingest sufficient quantities of essential trace elements, such as Cu, Fe, and Zn, to main- tain the continuous in vivo assembly of biologically active metalloproteins, which are inherently associated with health [1, 2]. In humans, for instance, about 1% of the total body Zn content is replenished daily by the diet [3]. Fol- lowing the absorption of essential trace elements from the gastrointestinal tract into the systemic blood circulation,



J Biol Inorg Chem (2009) 14:61–74

Table 1 Molecular properties and relative abundances of the major metalloproteins and metallopeptides in human plasma or serum


Metalloprotein or entity which contains bound metal


Number of metal atoms bound per protein

Plasma or serum protein concentration


mass (kDa)





10–250 lg/L





1.8–3.7 g/L



Blood coagulation factor V



* 10 mg/L





* 180 lg/L a

[4, 32]




0.2–0.6 g/L





36.1–53.6 g/L

[4, 9]




[40, 49]




Peptides and amino acids



a 2 macroglobulin



1.1–3.7 g/L

[4, 9]




36.1–53.6 g/L









SOD superoxide dismutase, EC-SOD extracellular Cu,Zn superoxide dismutase

a Rat plasma

specific plasma proteins subsequently distribute these ele- ments to internal organs [4], where absorption occurs by highly specific uptake mechanisms, such as endocytosis in the case of Fe [5, 6]. Among the most abundant transition metals present in human plasma are Cu, Fe, and Zn, which are present at total concentrations of 0.84–1.45 lg Cu/mL, 0.46–1.67 lg Zn/mL, and 2.47–13.1 lg Fe/mL [7, 8]. Therefore, the analysis of plasma for the contained major Cu-, Fe-, and Zn-containing metalloproteins (Table 1) will provide insight into the roles of trace elements in the bio- chemistry and pathophysiology of both healthy and diseased states. It is well known that numerous genetic human diseases are associated with increased or decreased plasma concentrations of specific metalloproteins [9]. Wilson’s disease, for instance, is a Cu-overload disease associated with low plasma Cu [10] and hereditary hemochromatosis is an Fe-overload disease associated with elevated plasma Fe [6]. Conversely, it is chemically feasible that the exposure of humans to certain environmental pollutants or that the physiological response to infection could also result in increased or decreased plasma concentrations of specific Cu, Fe, and Zn metalloproteins [9]. Therefore, an instrumental analytical method which can rapidly deter- mine the major Cu-, Fe-, and Zn-containing plasma metalloproteins would represent an innovative tool to assist in screening or diagnosing human diseases associated with altered essential trace elements, regardless of whether the latter have a genetic origin or are the result of exposure to environmental pollutants (chemical or bacterial). From an analytical point of view, the direct liquid chromatography (LC) analysis of plasma in conjunction


with an element-specific detector, such as a flame atomic absorption spectrometer, a graphite furnace atomic absorption spectrometer, an inductively coupled plasma atomic emission spectrometer (ICP-AES), or an induc- tively coupled plasma mass spectrometer (ICP-MS), should allow the detection of individual plasma metalloproteins assuming that the metal–protein bond(s) in the latter remain intact during the LC separation process [1114]. Despite this attractive proposition, and even though numerous investigations have been reported on the speci- ation of metals and metalloid compounds in other biological fluids and tissues, comparatively few studies have been carried out to attempt this goal in undiluted mammalian plasma or serum [15]. All studies which reported on the direct LC analysis of mammalian plasma or serum for metalloproteins are lis- ted in Table 2 (only those which detected at least two of the elements of interest are listed). In particular, size- exclusion chromatography (SEC), anion-exchange chro- matography (AEX), and reversed-phase chromatography have been employed in conjunction with various element- specific detectors. The pioneering work of Dawson et al. [16], published in 1981, represents the first study of its kind to detect metalloproteins in human plasma and involved SEC analysis (Sephacryl S-300) followed by the detection of Cu and Zn in the collected fractions by flame atomic absorption spectrometry. The reconstruction of a Cu- and Zn-specific chromatogram revealed one Cu and three Zn peaks [16]. The same approach (Sephadex G-100) was applied for the analysis of human serum and graphite furnace atomic absorption spectrometry of the fractions showed one Cu, one Fe, and three Zn peaks

J Biol Inorg Chem (2009) 14:61–74


Table 2 Applications of liquid chromatography coupled with element-specific detectors for the identification of metalloproteins in mammalian plasma/serum reported in the literature (in chronological order)

Biological fluid




Stationary phase d

Mobile-phase composition


analyzed b

detected c





Human plasma (0.5)

Cu (1) Zn (3)


57 9 0.9

Sephacryl S-300 (25–75)

0.1 M Tris/HCl pH 8.0 ? 0.5 M NaCl


Human serum (1.0)

Cu (1) Fe (1) Zn (3)


100 9 2.6

Sephadex G-100 (40–120)

0.05 M Tris/HCl pH 7.4, 4 C


Human serum (5.0)

Cu (2) Fe (1) Zn (3)


100 9 2.6

Sephacryl S-300 (25–75)

0.1 M Tris/HCl pH 7.4, 22 C


Human serum (0.25)

Cu (1) Fe (2) Zn (4)


TSK G 3,000 SW (10)

0.1 M HEPES ? 0.1 M NaCl pH




Rat serum (0.1)

Cu (4) Zn (3)


30 9 0.7

TSK G 3,000 SW (10)

0.1 M Tris/HCl pH 7.2


Human serum

Cu (4) Fe (1) Zn (6)


25 9 0.2

SynChropak GPC 300 (5)

0.1 M Tris/HCl pH 6.9


SRM (0.002)


Human plasma? (0.25)

Cu (3) Fe (2)


60 9 1.6

Fractogel EMD BioSEC 650 (20–40)

0.02 M NaH 2 PO 4 ? 0.3 M NaCl pH 6.8, 30 C


Human serum (0.1)

Cu (1) a Fe (2) Zn (4) a


5 9 0.5

Mono Q HR (10)

15 min linear gradient A (0.05 M Tris/HCl pH 7.4) ? B (0.25 M NH4Ac in A)


Human serum (0.2)

Cu (2) Zn (2)


Superose 12 HR (8–12)

0.1 M Tris/HCl ? 2.5 mM CaCl2 pH 7.4


Human serum (0.02)

Cu (1) Fe (1) Zn (1)


25 9 0.46

CHAPS-coated ODS

0.2 mM Tris/HCl ? 0.2 mM CHAPS pH 7.4


Human serum (0.1)

Cu (1) a Fe (2) Zn (4) a


5 9 0.5

Mono Q HR (10)

15 min linear gradient A (0.05 M Tris/HCl pH 7.4) ? B (0.25 M NH 4 Ac in A)


SEC size-exclusion chromatography, AEX anion-exchange chromatography, RP reversed-phase chromatography, Tris tris(hydroxy- methyl)aminomethane, HEPES N-(2-hydroxyethyl)piperazine-N 0 -ethanesulfonic acid, CHAPS 3-[(3-cholamidopropyl)dimethylammonio]-1- propanesulfonate, ODS octadecyl silica, SRM standard reference material

a Artifact

b Volume in mL

c No. of peaks

d Bead diameter in lm

[17]. The separation of human serum on another SEC stationary phase (Sephacryl S-300) and the utilization of direct current plasma atomic emission spectrometry resulted in the detection of two Cu, one Fe, and three Zn peaks [18]. With use of a sequential ICP-AES as the online multielement-specific detector, the SEC analysis (TSK G 300 SW) of human serum brought to light one Cu, two Fe, and four Zn peaks [19]. Applying the same stationary phase for the analysis of rat serum in con- junction with an ICP-MS revealed four Cu and three Zn peaks [20]. Reversed-phase chromatography (octadecyl silica stationary phase) was also employed to analyze human serum using an ICP-MS as the Cu-, Fe-, and Zn-specific detector, but revealed that all metalloproteins containing these elements were essentially coeluted [21]. Yet another SEC material (SynChropak GPC 300) was used for the analysis of a reconstituted human serum standard reference material by an ICP-MS and uncovered

four Cu (one major and three minor), one Fe, and six non-baseline-separated Zn peaks [22]. Human plasma analysis by SEC (Fractogel EMD BioSEC 650) with offline analysis of the fractions by an ICP-AES identified three Cu and two Fe peaks (all baseline-separated) [8]. A double focusing ICP-MS was also employed as an online Cu-, Fe-, and Zn-specific detector for human serum analysis using AEX (MonoQ HR) [23, 24] and employ- ing a mobile-phase gradient. AEX, however, altered the speciation of Cu and Zn and must therefore be avoided when plasma/serum metalloproteins are to be determined for potential diagnostic applications. More recently, human serum analysis by SEC (Superose 12HR) followed by the online detection of Cu and Zn by an ICP-MS revealed two Zn and two poorly separated Cu peaks [25]. In view of the reported variability in the number of Cu, Fe, and Zn metalloprotein peaks that were detected by LC analysis of mammalian plasma or serum (Table 2) and to



J Biol Inorg Chem (2009) 14:61–74

develop this analytical approach into a clinically useful

Materials and methods

diagnostic tool, two key questions must be investigated. First, it should be examined if plasma and serum generate consistent analytical results (whichever contains more

Chemicals and solutions

individual metal peaks in the corresponding Cu-, Fe-, and


dextran, PBS (10 mM phosphate, 2.7 mM KCl, 137

Zn-specific chromatogram inherently contains more infor-


NaCl) tablets, sodium chloride (more than 99.5%

mation). Second, the stability of metalloproteins in plasma/ serum over time must be investigated to establish the maximum time that plasma samples can be kept (e.g., at room temperature) before ex vivo degradation of certain metalloproteins will occur. We have systematically developed a novel LC method for the determination of Cu, Fe, and Zn metalloproteins in rabbit plasma which can be potentially applied to any mammalian plasma. In view of the fact that the binding of transition metals in plasma metalloproteins can be weak [4, 26] and to maintain the integrity of the Cu, Fe, and Zn metalloproteins throughout the entire LC separa- tion process [19], we chose SEC as the separation mechanism. This is because SEC minimizes direct interactions between the analyte molecules and the sta-

pure), sodium acetate trihydrate (more than 99% pure), N,N-dimethyl-p-phenylenediamine monohydrochloride (highly toxic), lysozyme (from chicken egg white), heparin (sodium salt), and the bicinchoninic acid protein determi- nation kit were purchased from Sigma-Aldrich (St Louis, MO, USA), bovine serum albumin (BSA) was from Amersham Pharmacia Biosciences (Little Chalfont, UK), glacial acetic acid (more than 97.7% pure, to make 1 M acetic acid) was from Fisher Scientific (Nepean, ON, Canada), and Plasma PURE HCl (34–37%) was from SCP Science (Baie D’Urfe´, QC, Canada). A mixture of protein standards for SEC column calibration was obtained from Bio-Rad Laboratories (Hercules, CA, USA) which con- tained thyroglobulin (670 kDa), c-globulin (158 kDa), ovalbumin (44 kDa), myoglobin (17 kDa), and vitamin B 12

tionary phase and therefore represents a comparatively

(1.35 kDa). All solutions were prepared with water from a

gentle separation mechanism for the separation of bio-

Simplicity water purification system (Millipore, Billerica,

molecules. In addition, the utilization of physiologically


USA). PBS of pH 7.4 (10 mM phosphate, 2.7 mM

relevant buffers will provide an environment where


and 137 mM NaCl) was prepared by dissolving PBS

conformational changes of the plasma metalloproteins— which could potentially lead to the loss of the metal—are least likely to occur. Therefore, dissociation of a metal from its parent plasma protein is minimized. Furthermore, SEC can be employed in an isocratic separation mode

tablets in the appropriate volume of water (followed by pH adjustment if necessary) and filtration through 0.45-lm nylon filter membranes (Mandel Scientific, Guelph, ON, Canada). Ceruloplasmin oxidase activity was measured in collected fractions according to a published procedure [28].

(which inherently increases sample throughput), whereas AEX often requires salt gradients to elute proteins which can—in turn—sever weak transition metal–protein link-

Development of an isocratic SEC separation method of the major plasma proteins using UV detection

ages. Finally, the availability of SEC stationary phases with much smaller particle sizes (approximately 13 lm)

Screening of prospective stationary phases

and particle size distributions today compared with those that were used in some previous studies (Table 2) will inherently allow better separations owing to the increased chromatographic resolution. Following these consider-

All separations were carried out at room temperature (22 C) and a mobile phase flow rate of 1.0 mL/min (peristaltic pump). Since irreversible binding of plasma

ations to least disrupt weak metal–protein binding

proteins to the stationary phase represents a major obstacle

equilibria during the SEC separation process, we


must be overcome when plasma is to be directly

employed phosphate-buffered saline (PBS, pH 7.4) as the isocratic mobile phase. With regard to the detection of the separated Cu-, Fe-, and Zn-containing metalloproteins in the SEC column effluent, we utilized a state-of-the-art charge injection device based ICP-AES because this multielement-specific detector could be directly hyphen- ated to the separation column for the simultaneous online

analyzed by SEC, we systematically screened three com- mercially available SEC stationary phases (Sephacryl S-500, Superose 6 prep grade, and Superdex 200 prep grade, 25 cm 9 1.0 cm column) for their ability to analyze rabbit plasma (0.5 mL) using various PBS concentrations (39, 29 and 19) and a UV detector (280 nm). Each column was equilibrated with at least 50 mL of the

detection of Cu, Fe, and Zn. In addition, this detector is compatible with LC separations involving mobile phases containing more than 0.5% salt [27], which is required to preclude irreversible adsorption of plasma proteins to the stationary phase.

mobile phase before plasma was injected. To detect irre- versible binding of plasma proteins to the stationary phases, a standard protein mixture (BSA and lysozyme) was chromatographed before and after the analysis of six consecutive plasma samples. A shift of the retention times


J Biol Inorg Chem (2009) 14:61–74


of the standard proteins after the six plasma injections (compared with those obtained before the six plasma injections) would indicate irreversible binding of plasma proteins to the stationary phase. In addition, the column effluent of each plasma injection was analyzed for total protein (bicinchoninic acid assay) and compared with the total protein contained in 0.5 mL of the plasma to establish the percentage protein recovery. After six consecutive plasma injections, no irreversible protein binding was detected with any stationary phase–mobile phase combi- nation for the seventh injection (protein recovery 99 ± 1%, see the supplementary material). On the basis of the chromatograms obtained and our objective to separate plasma proteins into as many chromatographic protein bands as possible (the term ‘‘band’’ is used rather than the term ‘‘peak’’ since hundreds of proteins constitute a single chromatographic ‘‘band’’), Superdex 200 prep grade and PBS (19) were identified as the ideal stationary phase– mobile phase combination. In particular, Superdex 200 prep grade resulted in four protein bands, corresponding to 15% (band 1), 14% (band 2), 69% (band 3), and 2% (band 4) of total protein. In contrast, Superose 6 prep grade produced only three protein bands, with 10, 89, and 1% of total protein, and Sephacryl S-500 resulted in only two protein bands, with 7 and 93% of total protein, respectively.

Increase of column efficiency by using a higher-resolution column

The commercially available stationary phase Superdex 200 prep grade is composed of 34-lm particles. In view of the fact that stationary phases with a smaller particle size generally result in a better chromatographic resolution, we improved the separation of plasma bands by using a pre- packed Superdex 200 column (30 cm 9 1.0-cm inner diameter) which contained 13-lm particles. Again a mix- ture of BSA and lysozyme (1.2 and 0.62 mg in 5.0 mL of water) was used to check the column integrity before and after the injection of six plasma samples. In addition, the number of theoretical plates was calculated (using the lysozyme peak) and increased from approximately 1,000 for Superdex 200 prep grade (25.4 cm 9 1.0-cm inner diameter, 34-lm particles) to approximately 23,000 for the prepacked Superdex 200 column (30 cm 9 1.0-cm inner diameter, 13-lm particles). At least 30 consecutive plasma analyses could be carried out per column without loss of chromatographic resolution of the metal peaks.

Animal experiments

The Animal Care Committee of the University of Calgary approved the procedure to collect blood from New Zealand

white rabbits (Protocol Approval #BI 2005-27). Male New Zealand white rabbits were purchased from Casey Van- dermeer (Edmonton, AB, Canada) and fed ad libitum on a ‘‘high-fiber’’ diet (Lab Diet 5321, Canadian Lab Diets, Leduc, AB, Canada). Blood (5.0 mL) was collected from the marginal ear vein with 20-gauge stainless steel blood collection needles (211 monoject, Sherwood Medical, St Louis, MO, USA) into BD Vacutainer blood collection tubes (no additive, BD Vacutainer, Franklin Lakes, NJ, USA) to which 0.5 mg heparin had been added for the preparation of plasma. For the preparation of serum, the blood clot was removed by centrifugation (described below). The injection of a heparin blank onto the SEC-ICP- AES system (control experiment) revealed no detectable Fe, Cu, or Zn (data not shown). Blood was collected from 4.5-h-fasted rabbits at approximately 13:30 and was cen- trifuged at 1,100g (22 C) for 10 min and the plasma (or serum) obtained was analyzed using the SEC-ICP-AES system within 30 min after blood collection, or at the time points indicated later. Only straw-yellow plasma (free of the characteristic red color of hemoglobin from ruptured erythrocytes) was used throughout the study. To establish the interanimal variation, 18 rabbit plasma samples were consecutively analyzed using the SEC-ICP-AES system. To qualitatively identify the detected Cu, Fe and Zn metalloproteins in collected fractions after the analysis of rabbit plasma, one human plasma sample was chromato- graphed; this was obtained from a healthy male volunteer. The collection of human blood was approved by the Uni- versity of Calgary Conjoint Health Research Ethics Board (approval no. E-21198).

Experimental setup of the optimized SEC-ICP-AES system

The SEC-ICP-AES system (Fig. 1) consisted of a Waters (Milford, MA, USA) model 510 high-performance LC pump, a Rheodyne 9010 PEEK injection valve (Rheodyne, Rhonert Park, CA, USA) equipped with a 0.5-mL PEEK injection loop, and a prepacked Superdex TM 200 10/300 GL Tricorn TM high-performance column (30.0 cm 9 1.0-cm inner diameter, separates globular proteins between approximately 600 and 10 kDa; GE Healthcare, Piscataway, NJ, USA). The exit of the SEC column was connected to the Meinhard concentric glass tube nebulizer of the ICP-AES with fluorinated ethylene–propylene Teflon tubing (30 cm, inner diameter 0.5 mm). Simultaneous multielement- specific detection of C (193.091 nm), S (180.731 nm), Zn (213.856 nm), Fe (259.940 nm), Cu (324.754 nm), and P (213.618 nm) in the column effluent was achieved with a Prodigy, high-dispersion, radial-view ICP-AES (Teledyne Leeman Labs, Hudson, NH, USA) at an Ar gas-flow rate of 19 L/min, an RF power of 1.3 kW, and a nebulizer gas



J Biol Inorg Chem (2009) 14:61–74

Fig. 1 The instrumental analytical size-exclusion chromatography (SEC)– inductively coupled plasma atomic emission spectrometer (ICP-AES) setup. HPLC high- performance liquid chromatography

) setup. HPLC high- performance liquid chromatography pressure of 35 psi. The detector technology utilized in

pressure of 35 psi. The detector technology utilized in the Prodigy allows the simultaneous measurement of the peak and the background emissions to generate the net emission intensity. This capability is critical in experiments where the background emission intensity changes (e.g., when a major protein peak reaches the ICP-AES) so the operator is not mislead into believing that an analytically significant event has occurred when in fact it has not. This advantage, together with the ability of the ICP-AES to handle salt- containing solutions, makes the Prodigy ideally suited for the LC analysis of solutions containing metalloproteins. Time scans were performed using the time-resolved-anal- ysis mode (Salsa version 3.0) and a data acquisition rate of one data point per 2 s. The raw data were imported into SigmaPlot 10 and smoothed using the bisquare algorithm. According to the void volume of the Superdex 200 column (blue dextran), a 7.0-min delay was implemented between injection and the beginning of data acquisition (1,080-s acquisition window). Figure 2 depicts the C-specific SEC- ICP-AES chromatograms of rabbit plasma obtained on a Superdex 200 prep grade (34 lm) and a prepacked (13 lm) column, which clearly displays the increased resolution of

80000 Albumin (~66 kDa) 60000 40000 20000 V 0 0 600 800 1000 1200 1400
Albumin (~66 kDa)
Intensity (counts/s)

Time (s)

Fig. 2 C-specific chromatograms of rabbit plasma on Superdex 200 prep grade (1.0 cm 9 25 cm; 34-lm particle size) (dashed line), and Superdex 200 10/300 GL SEC column (1.0 cm 9 30 cm; 13-lm particle size) (solid line). Phosphate-buffered saline mobile phase (pH 7.4, 22 C), flow rate 1.0 mL/min, injection volume 500 lL, ICP- AES detector (C emission at 193.091 nm). Void volume 600 kDa, inclusion volume 10 kDa

the latter stationary phase. The marginal increase in the retention time for the small molecular weight C peak in the 13-lm column compared with the 34-lm column is caused by the difference in column length of approximately 5 cm. We size-calibrated the analytical Superdex 200 column with known molecular weight protein standards. In addi- tion, the analysis of rabbit blood plasma provided internal molecular weight standards as several proteolytically stable plasma metalloproteins, such as ferritin and transferrin (Fe metalloproteins), ceruloplasmin (Cu metalloprotein), and the most abundant plasma protein albumin (which can be easily identified on the basis of the most intense C and S peaks in the chromatogram), are naturally present in plasma. These can therefore serve as a rough proxy to estimate the size of a detected metalloprotein. Furthermore, it is likely that owing to the sheer complexity of plasma (more than 3,700 proteins), a detected unknown plasma metalloprotein is unlikely to have the same retention time as is suggested from a calibration curve because of its unavoidable interactions with other plasma proteins. Therefore, deductions of the molecular mass of an unknown metalloprotein from its retention time alone should be interpreted with caution.

Identification and quantification of metalloproteins in SEC column effluent

In terms of qualitatively identifying the separated plasma metalloproteins in the column effluent, we did not utilize electrospray ionization mass spectrometry because of the salt content (approximately 1% or approximately 164 mM) of the mobile phase. In general, the salt concentration of aqueous samples that can be analyzed by electrospray ionization mass spectrometry must be below 10 mM. We therefore qualitatively identified the Cu metalloprotein ceruloplasmin in collected fractions using an established ceruloplasmin oxidase activity assay (based on the oxida- tion of N,N-dimethyl-p-phenylenediamine) [28]. Since an antibody-based approach (e.g., enzyme immunoassay) for the identification of rabbit ferritin, transferrin, a 2 macro- globulin, and factor V was not readily available, an alternative way of identifying these metalloproteins had to be pursued and we therefore analyzed human plasma using


J Biol Inorg Chem (2009) 14:61–74


the SEC-ICP-AES system and collected fractions for pro- tein identification purposes. The two major Fe-containing proteins and all Zn-con- taining species had essentially the same retention times as those obtained for rabbit plasma. Fractions were collected for each Fe peak at time points corresponding to the max- imum, shoulders on either side of the maximum, and at the baseline before and after each peak. Similarly, altogether eight fractions were collected of all Zn-containing entities. Human ferritin was quantified in the collected fractions by microparticle enzyme immunoassay technology with an Axsym analyzer (Abbott Diagnostics, Mississauga, ON, Canada) using the manufacturer’s method and calibrators. Human transferrin was measured by immunoturbidometric assay with a Cobas Integra 700 analyzer (Roche Diagnos- tics Canada, Laval, QC, Canada) using the manufacturer’s method and calibrators, and human a 2 -macroglobulin was measured with a Dade Behring BN2 Prospect rate nephe- lometric immunoassay using the manufacturer’s reagents (Dade Behring Canada, Mississauga, OC, Canada). With regard to the identification of factor V, fractions were collected corresponding to the baseline before and after Cu peak 1 as well as the peak itself. Factor Va coagulation activity was determined by performing a modified prothrombin time assay. In this assay, correction of the clotting time of factor V-deficient plasma is pro- portional to the concentration (activity percentage) of factor Va in test plasma, interpolated from a calibration curve [29]. Factor Va coagulation activity was determined using an ACL TOP analyzer (Beckman Coulter, Palo Alto, CA, USA) using factor V-deficient plasma, reagents, and a HemosIL factor V assay protocol supplied by Instrumen- tation Laboratory USA (Lexington, MA, USA). To quantify the metal that corresponded to a detected chromatographic metal peak, we injected increasing doses of each metal onto the chromatographic system without a column and measured the area under each ‘‘peak’’ using SigmaPlot. This allowed us to establish a calibration curve, which was used to calculate the total amount of metal (in micrograms per 0.5 mL plasma) that was associated with a detected metal peak (based on its peak area) in the sub- sequent analysis. Finally, these data were used to calculate the number of micrograms of metal that was present in form of a certain metalloprotein per 1.0 mL of plasma for comparison with literature data [7, 8].

Results and discussion

Although mammalian blood plasma can be easily obtained and contains critical information about the essential trace element status of the organism from which it was obtained, the sheer complexity of the plasma proteome makes it

extremely difficult to extract relevant information about the organism’s health status. In fact, the human serum prote- ome comprises at least 3,700 proteins [30], which poses an almost insurmountable problem from an analytical sepa- ration viewpoint. The very complexity of analyzing plasma for the proteins contained within it, however, can be reduced dramatically if one is able to selectively analyze for a subproteome, such as the ‘‘metalloproteome’’ (in the context of this paper this term refers to all major plasma proteins with bound Cu, Fe, and Zn). This would require the separation of these metalloproteins from each other, which—since the molecular masses of all major metallo- proteins in plasma are known—can be achieved by choosing a SEC stationary phase with the appropriate fractionation range. Following this basic approach, we have developed a rapid SEC-ICP-AES method to directly analyze plasma for the major Cu-, Fe-, and Zn-containing metalloproteins (by essentially determining the retention time of the metals corresponding to these metalloproteins). Conceptually, the detection of metal peaks within the chromatographic window (between the exclusion volume and the inclusion volume) together with the established stability of the major plasma metalloproteins of Cu, Fe, and Zn [1, 3133] would imply that each detected metal peak corresponds to a plasma metalloprotein. Furthermore, the absence of tailing in the detected Cu, Fe, and Zn peaks would further substantiate that each metalloprotein remained intact during the entire LC separation process. With regard to the analysis of rabbit plasma, interanimal variation was expected and experimentally quantified (Table 3). Excluding the standard deviation of the diag- nostically inadequate Cu peak 1 (factor V, which disappears after 0.5 h), the average relative standard deviation for all

Table 3 Average concentration of Cu, Fe, and Zn associated with metalloproteins derived from size-exclusion chromatography–induc- tively coupled plasma atomic emission spectrometry analysis of rabbit plasma samples (N = 18)



Average metal concentration (lg/mL plasma) ± SD


Factor V and transcuprein

0.85 ± 0.73


0.46 ± 0.14


0.65 ± 0.29

Small molecular weight

0.21 ± 0.12



0.27 ± 0.16


2.66 ± 0.99


a 2 -Macroglobulin and unidentified peaks 2–4

0.84 ± 0.19


1.10 ± 0.24

The peak areas of metalloproteins that were not distinct were com- bined for integration

SD standard deviation



J Biol Inorg Chem (2009) 14:61–74

detected Cu, Fe, and Zn metalloproteins was 39%. In con- trast to this, the method reproducibility is excellent (see ‘‘ Stability of plasma metalloproteins ’’). The task of qualitatively identifying the detected plasma metalloproteins is simplified as only approximately ten major Cu, Fe, and Zn metalloproteins have so far been reported in mammalian plasma (Table 1) [1]. In principle, two strategies can be employed to qualitatively identify an individual metalloprotein. First, it can be definitively identified on the basis of either its enzymatic activity or a specific antibody target site on its surface (e.g., using an enzyme immunoassay). However, if neither of these tech- niques is applicable because the metalloprotein has no inherent enzymatic activity (e.g., if it functions exclusively as a transport protein) or because no enzyme immunoassay is readily available (for the organism of interest; in our case rabbits), the second strategy to tentatively identify a metalloprotein must be used. The latter involves the utili- zation of information that is derived from the Cu-, Fe-, and Zn-specific chromatogram in conjunction with literature data. For instance, the retention time of an unknown

metalloprotein relative to a known and abundant protein, such as albumin (66 kDa), can be indicative of its hydro- dynamic radius and thus its approximate molecular mass (assuming minimal protein–protein interactions). In addi- tion, the intensity of a metal peak (corresponding to a metalloprotein) relative to another metal peak (of a dif- ferent metalloprotein containing the same metal) contains information about the relative abundance of metal atoms that are associated with these two proteins in plasma. In this instance, the experimentally determined relative retention time and abundance of both metalloproteins can be compared with their known molecular mass and abun- dance from handbooks on human metalloproteins [9] to tentatively identify both metalloproteins.

Plasma versus serum metalloprotein analysis

To establish whether plasma or serum contains a larger number of individual Cu-, Fe-, and Zn-containing entities and therefore more information with regard to the health status of an organism, we applied the SEC-ICP-AES

Fig. 3 Simultaneous multielement-specific

chromatograms of rabbit plasma on a Superdex 200 10/300 GL (13 lm particle size) SEC column with a phosphate- buffered saline mobile phase (pH 7.4, 22 C); flow rate 1.0 mL/min, injection volume 500 lL, ICP-AES detector. Emission lines for a C at


nm (black), S at


nm (orange), and P at


nm (pink) and for b Cu

at 324.754 nm (green), Fe at


nm (blue), and Zn at


nm (red). Both a and b

were obtained from the same rabbit plasma sample. The SEC

column was size-calibrated with

a mixture of standards

(thyroglobulin 670 kDa, c-globulin 158 kDa, ovalbumin 44 kDa, myoglobin 17 kDa, and vitamin B 12 1.35 kDa). The qualitative identification of the metalloproteins factor V,

a 2 -macroglobulin,

ceruloplasmin, ferritin, and transferrin in collected fractions

by various enzyme-based assays

(see ‘‘ Materials and methods ’’)

is indicated by horizontal bars

fractions by various enzyme-based assays (see ‘‘ Materials and methods ’’) is indicated by horizontal bars


J Biol Inorg Chem (2009) 14:61–74


method developed to analyze plasma (n = 3) and serum (n = 3) of 4.5-h-fasted rabbits. Cu peak 1 in the plasma chromatogram at the bottom of Fig. 3 was absent in the serum chromatogram (data not shown). Because the num- ber and intensity of all other detected peaks remained unchanged (data not shown), the Cu metalloprotein corre- sponding to Cu peak 1 must be either directly or indirectly (e.g., by specific adsorption) involved in the blood clotting process. This Cu peak could possibly represent blood coagulation factor V, which is a single-chain glycoprotein that contains one Cu per molecule [34] and is known to be sensitive to proteolysis [35]. Although factor V has a molecular mass of 330 kDa, it is known to self-associate to form higher multimers [36, 37], which could explain its elution in essentially the void volume. On the basis of these results, plasma contains more information than serum for the desired application of the instrumental analytical method developed for diagnostic purposes.

Stability of plasma metalloproteins

To address a possible degradation of metalloproteins at room temperature (22 C) over time, plasma was analyzed using the SEC-ICP-AES system at 0.5, 1, 1.5, and 2 h after blood collection. This experiment was carried out twice and the results essentially showed the same overall trend. Typical Cu-, Fe-, and Zn-specific time-course chromato- grams are shown in Fig. 4. As depicted in this chromatogram, the Cu that was eluted prior to 800 s disappeared from plasma after the 0.5-h time point. In addition, the peak corresponding to Cu that was eluted at approximately 900 s (it likely corresponds to albumin-bound Cu; see the discussion below) and the one corresponding to the Cu that was eluted at approximately 1,200 s (small molecular weight Cu; see the discussion below) decreased, whereas the most intense Cu peak (ceruloplasmin; see the discussion below) increased to some extent. The discrepancy between the reduction in intensity of some Cu peaks over time versus the increase of the most intense Cu peak must be attributed to the loss of Cu (net Cu loss of approximately 30%) either to the container wall (that the plasma was kept in prior to analysis) or to the stationary phase of the SEC column. These, however, are only the two most likely explanations and at present the exact cause is unknown. Therefore, if one aims to detect all Cu metallo- proteins (including the labile ones) in plasma, the latter must be analyzed within 0.5 h after blood collection. The intensity of the Fe and Zn peaks remained virtually unchanged over the 2-h time period (Fig. 4). These results strongly suggest that the corresponding metalloproteins are stable and—more importantly—that the analytical method itself produces results that are sufficiently reproducible for diagnostic applications.

250 Cu 200 150 100 Vo 50 0 0.5 h Fe 1.0 h 1.5 h

Retention Time (s)

Fig. 4 Simultaneous Cu-, Fe-, and Zn-specific chromatograms of rabbit plasma over a 2-h time period (after collection) on a Superdex 200 10/300 GL (13-lm particle size) SEC column with a phosphate- buffered saline mobile phase (pH 7.4, 22 C); flow rate 1.0 mL/min, injection volume 500 lL, ICP-AES detector. Cu-, Fe-, and Zn- specific chromatograms were obtained in 0.5-h intervals at room temperature and the emission lines of each element (Cu at 324.754 nm, Fe at 259.940 nm, and Zn at 213.856 nm) were plotted on top of each other

C-, S-, and P-specific chromatogram of plasma

The analysis of plasma with the method developed and the simultaneous online detection of C, S, and P using the ICP- AES resulted in the three-element-specific chromatogram shown at the top of Fig. 3 and revealed four major C-containing protein bands. The S-specific chromatogram



J Biol Inorg Chem (2009) 14:61–74

closely resembled that of the first three C-containing pro- tein bands, which is expected since most mammalian proteins contain the S-containing amino acids L-cysteine and/or L-methionine. Interestingly, the fourth S-containing entity was eluted before the fourth C-containing entity (both of these correspond to small molecular weight pep- tides and amino acids). Since albumin is by far the most abundant mammalian plasma protein (approximately 50 g/L) and comprises more than half of the total protein in plasma [9], the most prominent C band—band 3—must be pre- dominantly composed of albumin (a BSA standard had the same retention time). Owing to the utilization of a P-con- taining mobile phase (PBS), the P-specific chromatogram displayed an elevated P baseline throughout the entire chromatographic window (Fig. 3, top). The P-emission intensity also provided an effective measure of the mass transfer of droplets from the nebulizer chamber to the plasma. The detection of the albumin peak did not affect the intensity of the P-emission line, which demonstrates that the injected total protein did not adversely affect the mass transfer from the nebulizer to the plasma (e.g., by a surfactant effect), which is an important prerequisite to accurately measure the total metal that is associated with an eluting plasma metalloprotein. In addition, the P-spe- cific chromatogram also revealed a characteristic dip at a retention time of about 1,210 s, which coincides with the elution of the small molecular weight C band 4, and therefore corresponds to the injected plasma ‘‘plug’’ (which contains less P than the mobile phase) reaching the detector. A similar phenomenon has previously been observed [38].

Cu, Fe, and Zn metalloproteins in plasma

The analysis of plasma (eight different animals) by SEC and the simultaneous online detection of Cu, Fe, and Zn using the ICP-AES resulted in a three-element-specific chromatogram, a representative of which is shown at the bottom of Fig. 3. At first glance, five Cu-containing, two Fe-containing, and approximately five poorly separated Zn- containing entities were detected. The majority of the detected metal peaks displayed an ideal peak shape, which suggests that the metals did not dissociate from their parent protein during the chromatographic separation process. However, the second Fe peak, the last Zn peak, and Cu peaks 2 and 4 displayed a hump on the long retention end, which can be rationalized by the elution of a slightly smaller metalloprotein containing the same metal. Given the inherent limitations of SEC with regard to the chro- matographic resolution of proteins of almost similar size from each other, this is not unexpected. Nevertheless, the detection of approximately 12 metalloproteins demon- strates that the optimized mobile phase–stationary phase


combination is well suited to separate the major Cu-, Fe-, and Zn-containing entities that are contained in rabbit plasma. Importantly, and in contrast to the peaks in the chromatograms at the top Fig. 3, the individual Cu, Fe, and Zn peaks in the chromatograms at the bottom of Fig. 3 correspond to individual metalloproteins and—since Cu peak 5 was in the small molecular mass region— metallopeptides. The Cu-specific chromatogram revealed five peaks, which is one more than has been reported in other studies (Table 2) (peak 1: approximately 515 s, approximately 28% of total Cu, approximately 0.9 lg Cu/mL; peak 2:

approximately 605 s, approximately 13% of total Cu, approximately 0.3 lg Cu/mL; peak 3: approximately 775 s, approximately 27% of total Cu, approximately 0.9 lg Cu/mL; peak 4: approximately 890 s, approximately 19% of total Cu, approximately 0.6 lg Cu/mL; peak 5:

approximately 1,230 s, approximately 13% of total Cu, approximately 0.3 lg Cu/mL; Fig. 3, bottom). The sum of all Cu peaks in this particular plasma sample amounted to 3.0 lg Cu/mL plasma, which is higher than the average total Cu concentration (2.14 lg/mL) in the 18 plasma samples that were analyzed (Table 3). Thus, the total rabbit plasma Cu concentration was higher than what has been reported for other mammalian species (range 0.2–2.0 lg/mL) [7]. Cu peak 1 appeared close to the void volume and was not observed in previous studies as evidenced by comparing the relative intensities of the observed Cu peaks with the Cu- specific chromatograms of previous studies [20, 22, 32]. This discrepancy, however, can be easily explained by the fact that previous studies analyzed either aged plasma or serum, which according to our investigations both lack Cu peak 1 (Fig. 4). On the basis of the observation that Cu peak 1 was absent from serum and disappeared from plasma after 30 min (Fig. 4), it was tentatively identified as factor V, which is known to be labile, contains Cu [34], and is present in human plasma at approximately 10 mg/L [35]. This was corroborated by analyzing collected fractions for factor V coagulation activity (bar in Fig. 3, bottom), which was highest at the maximum intensity of Cu peak 1. The activity, however, extended beyond this peak, which can be ratio- nalized by the fact that under the circumstances of blood sampling factor V is expected to be activated into factor Va, which has a smaller molecular mass (approximately 221 kDa) and could explain the observed tailing of the activity. Factor V is known to be very labile and prone to proteolysis, which may explain the rapid disappearance of Cu peak 1 in the time-course experiments (Fig. 4) [39]. According to our analytical data and assuming an identical stoichiometry in rabbit and human factor V (Table 1), we calculated a plasma concentration of 4.7 g/L, which is sev- eral-fold higher than its concentration in human plasma and is therefore in apparent disagreement. We note, however,

J Biol Inorg Chem (2009) 14:61–74


that evidence in favor of significant interanimal species differences of certain plasma metalloprotein concentrations have been reported [40]. Furthermore, a different metal- to-protein stoichiometry between human and rabbit factor V could significantly affect the calculated plasma con- centration. With regard to the retention time of this metalloprotein (Fig. 3, bottom) it is noteworthy that dimerization of this protein has been observed by others [36, 37]. This would result in a 660-kDa entity and could explain the elution of factor V close to the void volume. On the basis of the previously reported order of elution of Cu plasma metalloproteins from a SEC column [32], Cu peak 2 was tentatively identified as the 270-kDa protein transcuprein, which is also known to dimerize (540 kDa) [4]. This Cu peak had a small shoulder on the long retention end, which is in accord with the Cu-specific chromatogram in Fig. 4. Cu peak 3 had ceruloplasmin oxidase activity (bar in Fig. 3, bottom) and was therefore identified as the glycoprotein ceruloplasmin. On the basis of the experimentally deter- mined total Cu associated with ceruloplasmin in 1.0 mL of plasma and the known stoichiometry of Cu in this protein (Table 1), the plasma ceruloplasmin concentration was calculated at 0.31 g/L, which is within the concentration range reported for human serum. Cu peak 4 was compara- tively broad and appeared 11 s after albumin (dotted line in Fig. 3, bottom). The misalignment of this Cu peak with the albumin peak can be rationalized either by a rather weak binding of Cu to albumin (which has been reported by others [41]) or by the presence of a smaller Cu-containing entity in addition to the expected albumin-bound Cu [4, 32]. On the basis of the simultaneous appearance of Cu peak 5 (Fig. 3, bottom) with the last C band (Fig. 3, top), this Cu peak represents Cu bound to small non-S-containing peptides and amino acids, such as L-histidine [42] (the retention time for S band 4 in Fig. 3, top was different from that for Cu peak 5 in Fig. 3, bottom) and is in general agreement with other studies [32]. We note that the observed order of elution for all major Cu entities is as expected on a SEC column and follows decreasing molecular masses from 660 kDa (putative factor V dimer), 540 kDa (transcuprein dimer), 132 kDa (ceruloplasmin), 66 kDa (albumin), and small molecular weight Cu. The Fe-specific chromatogram revealed two baseline- separated peaks, which is identical to the maximum num- ber of Fe peaks that was previously reported (Table 2) [8, 19, 23, 24, 43, 44] (peak 1: approximately 670 s, 11% of total Fe, approximately 0.3 lg Fe/mL; peak 2: approx- imately 870 s, 89% of total Fe, approximately 2.6 lg Fe/mL; sum of all Fe peaks: 2.9 lg Fe/mL plasma) (Fig. 3, bottom). On the basis of the known molecular size and plasma abundance of the two major Fe-containing metal- loproteins ferritin and transferrin (Table 1), Fe peak 1 is identified as ferritin and Fe peak 2 as transferrin. This peak

assignment was confirmed by enzyme immunoassay and immunoturbidometric assay (bars in Fig. 3, bottom). On the basis of the experimentally derived total Fe associated with ferritin and transferrin in 1.0 mL of plasma and assuming that both Fe metalloproteins were fully loaded with Fe (Table 1), the rabbit plasma concentration of fer-

ritin was calculated as 535 lg/L and that of transferrin as

1.8 g/L. Even though these results are in overall accord

with the established concentrations of these metallopro- teins in mammalian plasma/serum (Table 1), we point out that the method developed cannot inherently determine the metal loading of a metalloprotein in which the metal loading may vary. It is therefore impossible to distinguish if the Fe that is associated with, e.g., ferritin is attributable

to (1) 50% apoferritin and 50% fully loaded holoferritin or (2) the case where all ferritin is 50% loaded with Fe. Nevertheless, our method allows us to determine the dis- tribution of a metal among various metalloproteins, which has inherent diagnostic value that cannot be obtained by conventional antibody-based enzyme assays. The distinct shoulder on the long retention end of Fe peak 2 indicates the presence of another Fe-containing entity. On the basis of previous studies which demonstrated that Fe is bound to human serum albumin in serum [23], this additional Fe- containing entity could be albumin-bound Fe especially since albumin is approximately 14 kDa smaller than transferrin, which would therefore explain its elution after transferrin. In contrast to Cu, however, no detectable Fe was eluted in the small molecular weight range, which indicates that Fe is not bound to peptides and amino acids in rabbit plasma. With respect to Zn, approximately five non-baseline- separated peaks were detected (peak 1: approximately 613 s, 10% of total Zn, approximately 0.1 lg Zn/mL; peaks 2–4: approximately 655 s, approximately 700 s, approximately 770 s, 34% of total Zn, approximately

0.6 lg Zn/mL; peak 5: approximately 880 s, 56% of total

Zn, approximately 1.1 lg Zn/mL; sum of all Zn peaks:

1.8 lg Zn/mL plasma) (Fig. 3, bottom), which is more than

the average number of Zn peaks that has previously been reported (Table 2) [31, 44, 45]. This finding can be ratio- nalized with the higher-resolution SEC column that was used in the present study (13-lm particles) compared with earlier studies. Zn peak 1 likely represents the 725-kDa a 2 -

macroglobulin [46] since between 12 and 31% of human plasma Zn has previously been reported to be tightly incorporated in this protein, which is in accord with our results [25, 31]. This peak assignment was confirmed by enzyme immunoassay (bars in Fig. 3, bottom) and is in accord with another study, which also identified the first Zn compound that was eluted from a SEC column as a 2 - macroglobulin [31]. Even though the putative a 2 -macro- globulin was eluted before ferritin (450 kDa), it was eluted



J Biol Inorg Chem (2009) 14:61–74

after the void volume, which is in discord with what would be predicted if its retention were solely based on its molecular mass. We note, however, that nonideal interac- tions between a 2 -macroglobulin and a SEC stationary phase (TSK-G4000SW) have been observed using PBS when the native protein was treated with chemicals which exposed hydrophobic amino acids to the surface [47] and resulted in the adsorption of this protein to the stationary phase. Similar behavior could also occur between the components of plasma and a 2 -macroglobulin and subse- quently the stationary phase in our experiments. Using the experimentally determined total Zn that is contained in 1.0 mL of plasma in form of this metalloprotein (approx- imately 0.1 lg Zn/mL), we calculated the rabbit plasma a 2 - macroglobulin concentration (using the stoichiometry delineated in Table 1) at 222 mg/L, which is approxi- mately one tenth of its concentration reported for human plasma (Table 1). We note, however, that significant in- teranimal species differences regarding certain plasma metalloprotein concentrations have been reported [40, 48]. On the basis of the identical retention times for Zn peak 5 and albumin (dotted lines in Fig. 3) and since albumin- bound Zn represents the major Zn entity in plasma (56% of total plasma Zn in this study, which is in good accord with previous studies on humans [26, 31] and pigs [44]), this Zn entity likely represents albumin-bound Zn (transferrin does not bind Zn 2? [44]). Zn peaks 2–4 and the Zn shoulder on

Practical applications

metalloproteins of one element, let alone those of more than one element simultaneously. Therefore, this method has the obvious advantage of extracting more information (namely, the relative abundance of the metalloproteins of the three major essential trace metals in plasma as well as the concentration of those metalloproteins in which the metal-to-protein ratio is fixed given that no other metal- loprotein containing the metal of interest is coeluted) from a single analysis in a given amount of time than is possible with other methods that are currently in use. This, in turn, can be helpful to more accurately diagnose the severity of a disease since Wilson’s disease, for instance, is not only associated with decreased plasma concentrations of the Cu metalloprotein ceruloplasmin [58], but can also result in an increased plasma concentration of the Fe metalloprotein hemoglobin during episodes of acute hemolysis [52]. The second application is the utilization of the method developed to probe the nonenzymatic bioinorganic chem- istry of environmentally abundant pollutants, such as toxic metals and metalloid compounds, in the mammalian bloodstream to better understand their chronic toxicity, individually and—more importantly—cumulatively. This latter application appears particularly relevant since bio- inorganic processes in the mammalian bloodstream are likely to be fundamentally involved in the origin of numerous human diseases that are associated with chronic exposure to toxic metals and metalloid compounds [59,

the long retention end of Zn peak 5 could not be qualita-


tively identified. However, the existence of a 165-kDa extracellular secretory glycoprotein Cu,Zn superoxide dismutase (EC-SOD) and that of a 31-kDa Cu,Zn super-


oxide dismutase (Cu,Zn-SOD) have been reported in guinea pig and human plasma [40, 49]. It is therefore likely that EC-SOD represents one of the unidentified Zn peaks 2–4 and that the shoulder on the long retention end of Zn peak 5 could possibly be Cu,Zn-SOD. Similar to the results obtained for Fe, no Zn was detected bound to small molecular weight peptides and amino acids in rabbit plasma.

The daunting analytical task of extracting health-relevant information from plasma can be considerably simplified by determining a ‘‘subproteome’’, such as the Cu, Fe, and Zn metalloproteome. To this end, we have developed a rapid SEC-based separation of the metalloproteins con- tained in rabbit plasma followed by the online analysis of the column effluent by an ICP-AES, which served as the simultaneous Cu-, Fe-, and Zn-specific detector. This novel SEC-ICP-AES method has allowed us to directly

Owing to the fact that the SEC-ICP-AES method devel- oped allows one to determine the plasma Cu, Fe, and Zn metalloproteome within approximately 24 min, two major practical applications of this method can be envisioned. The first application is its utilization as a clinical tool to screen for early- or advanced-stage human diseases by the direct analysis of human plasma or serum [5057]. Even though assays exist to quantify individual plasma metalloproteins, such as ceruloplasmin (e.g., by a spectrophotometric activity assay), few methods have been reported that can simultaneously determine all

analyze rabbit plasma in order to generate the Cu, Fe, and Zn metalloproteome, which is composed of approximately 12 metalloproteins and metallopeptides, within approxi- mately 24 min. From a clinical perspective, this simple and rapid technique to establish the Cu, Fe, and Zn me- talloproteome offers important advantages over individual metalloprotein assays since much more information can be extracted with this method from a single plasma sample. Thus, the detection of the majority of the expected Cu-, Fe-, and Zn-containing entities in rabbit plasma by the SEC-ICP-AES system constitutes an important first step in the development of an instrumental


J Biol Inorg Chem (2009) 14:61–74


analytical technique for the efficient detection of the plasma metalloproteome for potential diagnostic applica- tions in humans. The method developed can also be used to directly probe the bioinorganic chemistry of toxic metals in whole blood and thus has considerable potential to provide exciting new insights into the origin of toxic- metal-related human diseases.

Acknowledgments This research was funded by the Natural Sci- ences and Engineering Research Council (NSERC) of Canada. Teledyne Leeman Labs is gratefully acknowledged for funding the attendance of S.A.M. and J.G. at HPLC 2007 in Ghent, Belgium. Katie L. Pei is gratefully acknowledged for help regarding the col- lection of fractions. We would also like to extend thanks to Raymond J. Turner and especially Arvi Rauk for constructive feedback on the final draft of the manuscript. The staff of the Animal Health Unit (LESARC) at the University of Calgary is gratefully acknowledged for the maintenance of and the drawing of blood from the rabbits. We would also like to extend sincere thanks to one anonymous reviewer who provided valuable comments to significantly improve the final manuscript.



Frausto da Silva JJR, Williams RJP (2001) The biological

chemistry of the elements. Oxford University Press, Oxford


Karlin KD (1993) Science 261:701–708


Cousins RJ, Liuzzi JP, Lichten LA (2006) J Biol Chem




N, Lo LS, Askary H, Jones L, Kidane TZ, Trang T, Nguyen

M, Goforth J, Chu Y-H, Vivas E, Tsai M, Westbrook T, Linder


(2007) J Nutr Biochem 18:597–608



Eijk HG, de Jong G (1992) Biol Trace Elem Res 35:13–24


de Silva DM, Askwith CC, Kaplan J (1996) Physiol Rev 76:31–



Prohaska C, Pomazal K, Steffan I (2000) Fresenius J Anal Chem



Pomazal K, Prohaska C, Steffan I, Reich G, Huber JFK (1999) Analyst 124:657–663


Craig WY, Ledue TB, Ritchie RF (2000) Plasma proteins, clin-


utility and interpretation. Dade Behring, Newark


Winzerling JJ, Law JH (1997) Annu Rev Nutr 17:501–526


Szpunar J (2005) Analyst 130:442–465


Szpunar J, Lobinski R (1999) Pure Appl Chem 71:899–918


Garcia JS, de Magalhaes CS, Arruda MAZ (2006) Talanta




Szpunar J (2000) Analyst 125:963–988


Sanz-Medel A, Montes-Bayon M, Sanchez MLF (2003) Anal Bioanal Chem 377:236–247


Dawson JB, Bahreynitoosi MH, Ellis DJ, Hodgkinson A (1981) Analyst 106:153–159


Gardiner PE, Ottaway JM, Fell GS, Burns RR (1981) Anal Chim Acta 124:281–294


Gardiner PE, Braetter P, Negretti VE, Schulze G (1983) Spec- trochim Acta 38B:427–436


Gardiner PE, Braetter P, Gercken B, Tomiak A (1987) J Anal At Spectrom 2:375–378


Gercken B, Barnes RM (1991) Anal Chem 63:283–287


Inagaki K, Umemura T, Matsuura H, Haraguchi H (2000) Anal




Shum SCK, Houk RS (1993) Anal Chem 65:2972–2976

23. Muniz CS, Gayon JMM, Alonso JIG, Sanz-Medel A (2001) J Anal At Spectrom 16:587–592

24. Bayon MM, Cabezuelo ABS, Gonzalez EB, Alonso JIG, Sanz-Medel A (1999) J Anal At Spectrom 14:947–951

25. Inagaki K, Mikuriya N, Morita S, Haraguchi H, Nakahara Y, Hattori M, Kinosita T, Saito H (2000) Analyst 125:197–203

26. Zalewski P, Truong-Tran A, Lincoln S, Ward D, Shankar A, Coyle P, Jayaram L, Copley A, Grosser D, Muriga C, Lang C, Ruffin R (2006) Biotechniques 40:509–520

27. Pomazal K, Prohaska C, Steffan I (2002) J Chromatogr A


28. Curzon G, Vallet L (1960) Biochem J 74:279–287

29. Kalafatis M, Krishnaswamy S, Rand MD, Mann KG (1993) Methods Enzymol 222:224–236

30. Pieper R, Gatlin CL, Makusky AJ, Russo PS, Schatz CR, Miller SS, Su Q, McGrath AM, Estock MA, Parmar PP, Zhao M, Huang ST, Zhou J, Wang F, Esquer-Blasco R, Anderson NL, Taylor J, Steiner S (2003) Proteomics 3:1345–1364

31. Falchuk KH (1977) N Engl J Med 296:1129–1134

32. Weiss KC, Linder MC (1985) Am J Physiol 249:E77–E88

33. Harris WR, Wang Z, Brook C, Yang B, Islam A (2003) Inorg Chem 42:5880–5889

34. Mann KG, Lawler CM, Vehar GA, Church WR (1984) J Biol Chem 259:12949–12951

35. Church WR, Jernigan RL, Toole J, Hewick RM, Knopf J, Knutson GJ, Nesheim ME, Mann KG, Fass DN (1984) Proc Natl Acad Sci USA 81:6934–6937

36. Laue TM, Johnson AE, Esmon CT, Yphantis DA (1984) Bio- chemistry 23:1339–1348

37. Saenko EL, Yaropolov AI, Harris ED (1994) J Trace Elem Exp Med 7:69–88

38. Gailer J, Madden S, Burke MF, Denton MB, Aposhian HV (2000) Appl Organomet Chem 14:355–363

39. Mann KG, Kalafatis M (2003) Blood 101:20–30

40. Karlsson K, Marklund SL (1988) Biochem J 255:223–228

41. Bal W, Christodoulou J, Sadler PJ, Tucker A (1998) J Inorg Biochem 70:33–39

42. May PM, Linder PW, Williams DR (1976) Experientia 32:1492–


43. Wang J, Houk RS, Dreessen D, Wiederin DR (1999) JBIC 4:546–


44. Chesters JK, Will M (1981) Br J Nutr 46:111–118

45. Raab A, Braetter P (1998) J Chromatogr B 707:17–24

46. Adham NF, Song MK, Rinderknecht H (1977) Biochim Biophys

Acta 495:212–219

47. Gonias SL, Roche PA, Pizzo SV (1986) Biochem J 235:559–567

48. Marklund SL, Holme E, Hellner L (1982) Clin Chim Acta


49. Marklund SL (1984) J Clin Invest 74:1398–1403

50. Kessler H, Pajonk F-G, Meisser P, Schneider-Axmann T, Hoffmann KH, Supprian T, Herrmann W, Obeid R, Multhaup G, Falkai P, Bayer TA (2006) J Neural Transm 113:1763–


51. Akiba S, Neriishi K, Blot WJ, Kabuto M, Stevens RG, Kato H,

Land CE (1991) Cancer 67:1707–1712

52. Attri S, Sharma N, Jahagirdar S, Thapa BR, Prasad R (2006) Pediatr Res 59:593–597

53. Wolf TL, Kotun J, Meador-Woodruff JH (2006) Schizophr Res


54. Konofal E, Lecendreux M, Arnulf I, Mouren MC (2004) Arch Pediatr Adolesc Med 158:1113–1115

55. del Castillo Busto ME, Montes-Bayon M, Blanco-Gonzalez E, Meija J, Sanz-Medel A (2005) Anal Chem 77:5615–5621

56. Arizaga Rodriguez S, Blanco Gonzalez E, Alvarez Llamas G, Montes-Bayon M, Sanz-Medel A (2005) Anal Bioanal Chem




J Biol Inorg Chem (2009) 14:61–74

57. Saito Y, Saito K, Hirano Y, Ikeya K, Suzuki H, Shishikura K, Manno S, Takakuwa Y, Nakagawa K, Iwasa A, Fujikawa S, Moriya M, Mizoguchi K, Golden BE, Osawa M (2002) J Pediatr



58. Scheinberg IH, Gitlin D (1952) Science 116:484–489

59. Gailer J (2002) Appl Organomet Chem 16:701–707

60. Gailer J (2007) Coord Chem Rev 251:234–254