Katie L.
Pei
and
Jürgen
Gailer
*
Department of Chemistry, University of Calgary, 2500 University Drive NW, Calgary, AB, T2N 1 N4, Canada. E-mail: jgailer@ucalgary.ca; Fax: +1-403-289-9488; Tel: +1-403-210-8899
First published on 23rd July 2009
Arsenobetaine, which is frequently ingested by humans via the consumption of seafood, is rapidly excreted unchanged in urine, but not much is known about its transport in the mammalian bloodstream. To assess whether this transport involves binding to plasma proteins, rabbit and human plasma were spiked with arsenobetaine and the mixture was analyzed (after 5 min and again after 6 h) by size-exclusion chromatography (SEC) coupled on-line to an inductively coupled plasma atomic emission spectrometer (ICP-AES). Simultaneous monitoring of the emission lines of As, Cu, Fe and Zn in the column effluent allowed us to determine the elution of arsenobetaine relative to that of the major Cu, Fe and Zn-containing metalloproteins. Over the investigated time period, a single As peak eluted near the inclusion volume on two different SEC columns with fractionation ranges of 600–10 KDa and 7000–100 Da. These results indicate that arsenobetaine did not bind to plasma proteins and that SEC-ICP-AES is a useful tool to rapidly probe toxicologically and pharmacologically-relevant interactions between organometalloid compounds and mammalian blood plasma constituents in vitro.
Although arsenobetaine has proven useful for studies into the mammalian metabolism of arsenite,19 not much is known about the mechanism by which arsenobetaine—after it has been absorbed from the gastrointestinal tract—is transported in the mammalian bloodstream for its subsequent excretionvia the kidneys. Since numerous medicinal drugs which are approved for administration to humans, (e.g.aspirin, warfarin and diazepam) bind to human serum albumin in blood plasma,20arsenobetaine could similarly bind to this or another plasma protein. Even though the latter question has not been investigated experimentally, it would require a liquid chromatographic separation method that is capable of separating the major plasma proteins using a physiologically relevant buffer and a detector that can simultaneously detect arsenobetaine and the separated plasma proteins. If plasma were then to be spiked with arsenobetainein vitro and analyzed, the elution of arsenobetaine relative to the separated plasma proteins would indicate if binding to plasma proteins had occurred. Importantly, this approach would reveal if binding occurred under near physiological conditions.
To this end, we have recently developed an instrumental analytical method for the separation of the major Cu, Fe, and Zn-containing metalloproteins that are contained in rabbit plasma using size-exclusion chromatography (SEC) and the simultaneous on-line detection of the associated metals by a state-of-the-art charge injection device (CID)-based inductively coupled plasma atomic emission spectrometer (ICP-AES).21 Since the latter instrument can simultaneously detect several metals and/or metalloids, this SEC-ICP-AES method should be suited to probe the in vitro binding of pharmacologically interesting metalloid compounds, such as arsenobetaine,16 to plasma proteins in blood plasma. Even though the related instrumental analytical technique, SEC-ICP-MS, has also been successfully used to study the binding of Pt-based anti-cancer drugs to plasma proteins, the majority of these studies employed Tris–HCl buffers for the separation of the proteins and monitored only the element of interest (e.g. Pt) by ICP-MS.22,23 Conversely, the SEC-ICP-AES method that was employed in the present study employs a phosphate buffered saline (PBS) mobile phase for the separation of plasma proteins and simultaneously detects Cu, Fe, and Zn in addition to the element of interest. This allows one to measure the elution of the element of interest relative to the major Cu, Fe, and Zn-containing metalloproteins (the latter are collectively referred to as the plasma metalloproteome in the context of this paper).
Our main focus was to evaluate our recently developed SEC-ICP-AES method for its capability to probe the in vitro binding of the organometalloid compound arsenobetaine to plasma constituents in blood plasma. We therefore added arsenobetaine to fresh rabbit and human plasma and analyzed the resulting mixture after 5 min and again after 6 h by SEC-ICP-AES. Simultaneous monitoring of the emission lines of As, Cu, Fe, and Zn in the SEC column effluent was intended to provide insight into the relative elution of arsenobetaine compared to the major plasma proteins and therefore its overall binding behavior. Our deliberate decision to study the binding of arsenobetaine to plasma proteins in vitro represents the first practical application of our systematically developed SEC-ICP-AES method and will simplify the interpretation of the chromatographic results since this organoarsenic compound is not metabolized by mammals.17 Accordingly, only a single As-peak is expected to elute in the SEC column effluent, whereas choosing an organometalloid compound that is biotransformed in plasma would likely result in the detection of multiple peaks and therefore unnecessarily complicate the interpretation of the results.
After the addition of arsenobetaine to fresh rabbit plasma, the obtained mixture (67 μg As/0.5 mL plasma) was analyzed after 5 min and again after 6 h. The corresponding As, Cu, Fe, and Zn-specific SEC-ICP-AES chromatograms are depicted in Fig. 1A and B. With regard to the result for the 5 min time-point (Fig. 1A), the total number of Cu, Fe and Zn-peaks that were detected as well as their respective retention times and their relative intensity (to each other) were comparable to our earlier study.21 Notably, however, the Cu-peaks that were previously shown to correspond to coagulation factor V and small molecular weight Cu compounds (see single headed arrows in Fig. 1A) were less intense in the present study, which can be explained by the relatively large standard deviation of the peak intensity/peak area that we reported for these peaks (we noticed during our previous study that the analysis of rabbit plasma often results in very small peaks for these two Cu-metalloproteins).21 Importantly, however, only a single As peak was detected, which—owing to the lack of metabolism of arsenobetaine by mammalian organisms17—likely corresponds to arsenobetaine. The fact that the latter peak essentially eluted in the inclusion volume of the employed SEC column (indicated by the molecular weight marker for vitamin B12 on top of Fig. 1), implies that arsenobetaine did not bind to plasma proteins >10 kDa. The simultaneous Cu, Fe, and Zn-specific chromatogram that was obtained for the 6 h time-point was slightly different from that of the 5 min time-point (Fig. 1A and B), which is expected since some plasma proteins will inevitably undergo proteolytic degradation. Nevertheless, the peak areas that were obtained for all Fe and Zn-peaks detected at the 6 h time-point were very similar to those for the 5 min time-point, whereas ∼20% of Cu was lost over the investigated time period (Table 1). The latter result is in accord with our previous findings and may involve the binding of Cu (which was liberated from labile Cu-metalloproteins, such as coagulation factor V) to the container walls.21 The retention time and the peak area of the peak corresponding to arsenobetaine, however, remained essentially unchanged over the investigated time period (see dotted line in Fig. 1A and B and Table 1).
![]() | ||
Fig. 1 Simultaneous As, Cu, Fe and Zn-specific chromatogram of arsenobetaine-spiked rabbit plasma (A: after 5 min; B: after 6 h) and human plasma (C: after 5 min; D: after 6 h) on a Superdex 200 10/300 GL (30 × 1.0 cm I.D., 13 μm particle size) SEC column at 22 °C using a phosphate buffered saline buffer (PBS, pH 7.4) as the mobile phase (rabbit and human plasma were injected onto the same column). Abbreviations: Cp = ceruloplasmin, Alb = albumin , AsB=arsenobetaine. Flow-rate 1.0 mL/min, injection volume 500 μL, detectorICP-AES at 189.042 nm (As), 324.754 nm (Cu), 259.837 nm (Fe) and 213.856 nm (Zn). The retention times of the molecular weight markers are depicted on top of the figure. For explanation of other denotations (lines and arrows) see text. |
Emission line | Total area counts | |||
---|---|---|---|---|
Rabbit (t0) | Rabbit (t6Hr) | Human (t0) | Human (t6Hr) | |
As 189.042 nm | 19![]() |
20![]() |
19![]() |
19![]() |
Cu 324.754 nm | 68![]() |
53![]() |
113![]() |
61![]() |
Fe 259.837 nm | 45![]() |
43![]() |
24![]() |
24![]() |
Zn 213.856 nm | 34![]() |
34![]() |
20![]() |
20![]() |
To corroborate these findings with fresh human plasma, the latter was spiked with arsenobetaine and the obtained mixture (67 μg As/0.5 mL plasma) was analyzed by SEC-ICP-AES as described above. The resulting As, Cu, Fe, and Zn-specific SEC-ICP-AES chromatograms are depicted in Fig. 1C and D. To a first approximation, the same number of Cu, Fe and Zn-peaks were detected in human plasma as in rabbit plasma. In addition, the retention times of the plasma Cu-metalloprotein ceruloplasmin and albumin -bound Zn were identical in both mammalian species (see vertical dashed lines in Fig. 1A–D). In human plasma, however, the retention times of ferritin, transferrin, and α2-macroglobulin were significantly shorter than those obtained for rabbit plasma (see double headed arrow in Fig. 1C corresponding to plasma ferritin; ferritin Δretention time = 104 s; transferrin Δretention time = 26 s; α2-macroglobulin Δretention time=109 s). These results imply different hydrodynamic radii of these metalloproteins in the respective mammalian organisms which could be caused by either a different hydrodynamic radius of the metalloproteins itself, or the aggregation of these metalloproteins with other plasma proteins during the chromatographic separation process. Whereas the comparatively small difference in the retention times for transferrin may be rationalized based on subtle differences in the hydrodynamic radius of the protein itself,28 the marked differences in the retention times for ferritin and α2-macroglobulin (in rabbit and human plasma) are more difficult to explain, especially since the specific cellular origin of mammalian plasma ferritins is currently unknown.28 Importantly, and in accord with the results obtained for rabbit plasma (Fig. 1A and B), however, a single As peak which co-eluted with the small molecular weight Cu-peak was detected at both time points (Fig. 1C and D). Similar to the results obtained for the analysis of rabbit plasma, small changes in the Cu, Fe and Zn-specific chromatogram were detected over the investigated time period (Fig. 1C and D). In addition, the peak areas that were obtained for all Fe and Zn-peaks at the 6 h time-point were very similar to those obtained after the 5 min time-point, whereas a considerably larger amount of Cu, namely ∼45%, was lost over the investigated time period (Table 1). Like the results obtained for rabbit plasma, the retention time and peak area of the As peak corresponding to arsenobetaine remained essentially unchanged over the investigated time period (see dotted line in Fig. 1C and D, Table 1).
These results indicate that arsenobetaine does not bind to plasma proteins >10 kDa in rabbit and human plasma. In order to investigate whether arsenobetaine may bind to plasma proteins <10 kDa, arsenobetaine-spiked human plasma (0.5 mL) was analyzed by SEC-ICP-AES on a Superdex Peptide SEC column, which offers the appropriate fractionation range (7 kDa–100 kDa), after 5 min and again after 6 h. The corresponding As, Cu, Fe, and Zn-specific chromatograms are depicted in Fig. 2. As expected, all Cu, Fe and Zn-containing plasma metalloproteins eluted close to the void volume and essentially in one broad peak. In addition, the area counts obtained for the peaks corresponding to these elements were comparable to the values reported in Table 1. Most importantly, arsenobetaine eluted close to the molecular weight standard GSH over the entire time period, which indicates that it did not bind to plasma constituents >300 Da (Fig. 2A and B).
![]() | ||
Fig. 2 Simultaneous As, Cu, Fe and Zn-specific chromatogram of arsenobetaine-spiked human plasma (A: after 5 min; B: after 6 h) on a Superdex Peptide 10/300 GL (30 × 1.0 cm I.D., 13 μm particle size) SEC column at 22 °C using a phosphate buffered saline buffer (PBS, pH 7.4) as the mobile phase. AsB = arsenobetaine. Flow-rate 1.0 mL min−1, injection volume 500 μL, detectorICP-AES at 189.042 nm (As), 324.754 nm (Cu), 259.837 nm (Fe) and 213.856 nm (Zn). The retention times of the molecular weight markers are depicted in the upper right corner of A. |
Taken together, these findings indicate the absence of an arsenobetaine transport protein in human plasma. In the context of studies which demonstrated a rapid clearance of non-transferrin bound iron from human plasma,29 our results help to explain why ingested arsenobetaine is rapidly excreted from humans in urine.17 From a bioinorganic perspective, the absence of an arsenobetaine transport protein is somewhat unexpected in view of the fact that the most abundant mammalian plasma protein, serum albumin , displays prolific ligand-binding properties for a large variety of metals [e.g. Au(I), Hg(II), Cu(II), Ni(II)], metal-based drugs (e.g. auranofin, cisplatin),30 medicinal drugs (e.g.aspirin, warfarin, bilirubin, diazepam, ibuprofen) and amino acids (e.g.L-tryptophan).20 In order to put the results into a wider context, however, it is instructive to look at the structurally analogous zwitterionglycinebetaine[(CH3)3N+CH2COO−](the nitrogen analog of arsenobetaine), which is naturally present in human blood plasma at concentrations of 20–60 μmol L−1.31 With regard to the function of glycinebetaine in mammalian organisms, this compound has been demonstrated to serve as an osmolyte (to maintain normal cell volume),32 a methyl-donor and as a chaperone to protect proteins against denaturation.33 Since glycinebetaine must presumably be freely available in plasma (i.e. not bind to plasma proteins) in order to serve its role as an osmolyte and as a chemical chaperone, it is only reasonable to expect that the structurally analogous arsenobetaine should behave similarly, which—according to our results—it does (Fig. 1, Fig. 2). Furthermore, the fact that glycinebetaine has been identified as one of the major osmolytes in the renal medulla of kidneys in mammals,32 may explain the unusually high urinary excretion of arsenobetaine from human volunteers that had consumed an As-free diet for 3 days prior to the collection and analysis of urine.18 Based on our data, these results cannot be caused by the slow clearance of arsenobetaine from blood owing its binding to a plasma protein, but could be explained by a likely co-localization of glycinebetaine and arsenobetaine in the kidney and an experimental diet-induced release of the latter into the urine. Even though this explanation is somewhat speculative, it appears scientifically feasible.
According to our investigations into the binding of the biologically active “model metallodrug” arsenobetaine to plasma proteins in native plasma in vitro (Fig. 1 and 2), SEC-ICP-AES is identified as a useful tool to probe metallodrug–plasma protein interactions in a similar fashion. These interactions are of considerable pharmacological importance because the efficacy of new metal or metalloid-based drugs critically depends on their binding to plasma proteins. In fact, many promising new drugs have been rendered ineffective because of their unusually high affinity for albumin 20 and the binding of a promising new drug to albumin has therefore been referred to as the “second step in rational drug design”.20 In addition, interactions between metal-based drugs and plasma proteins are also important because the antitumor activity of a drug is usually modified when the drug is complexed by a protein.23 Finally, interactions between metal-based drugs and protein targets within a given proteome are much less studied than metallodrug–DNA interactions, yet they hold the potential to identify new targets for drug therapy.34 Considering the increased recent interest in such metallodrug–plasma protein interactions,22,34,35 appropriate analytical techniques are likely to become increasingly important.22,35 Even though a related analytical technique, namely SEC-ICP-MS, has also been successfully applied to study the binding of Pt-containing anti-cancer drugs to plasma proteins using 30 mM Tris-HCl buffer of pH 7.2 (to separate the proteins),23 the SEC-ICP-AES method that was employed in the present study uses a PBS-buffer to separate the plasma proteins. This should, in principle, avoid interactions that could possibly occur between plasma metalloproteins and the major Tris-buffer constituent tris(hydroxymethyl)aminomethane during the separation process itself when SEC-ICP-MS is used. In fact, tris(hydroxymethyl)aminomethane is known to form complexes with free metal ions36 and this mobile phase constituent could therefore possibly abstract metal ions from a metalloprotein which could result in ambiguous results when the binding of a metallodrug to plasma proteins is studied. Furthermore, the SEC-ICP-AES approach is based on the simultaneous monitoring of the emission lines of Cu, Fe and Zn in addition to the element of interest which offers the additional capability to visualize a potential dose-dependent effect of the compound of interest on the plasma Cu, Fe and Zn-metalloproteome (which could, for example, result in the disappearance of a specific metalloprotein peak and the corresponding elution of a new metal peak in the inclusion volume).21 Studies are currently underway to explore this potential capability of SEC-ICP-AES.
This journal is © The Royal Society of Chemistry 2009 |