Quantification of Ferritin-Bound Iron in Plant Samples by Isotope

Aug 4, 2009 - Ferritin is nature's predominant iron storage protein. ... to combat dietary iron deficiency, a major public health problem in developin...
1 downloads 0 Views 198KB Size
Download PDF
Anal. Chem. 2009, 81, 7368–7372

Quantification of Ferritin-Bound Iron in Plant Samples by Isotope Tagging and Species-Specific Isotope Dilution Mass Spectrometry Matthias Hoppler,† Christophe Zeder,† and Thomas Walczyk*,‡ Laboratory for Human Nutrition, Institute of Food Science and Nutrition, ETH Zurich, 8092 Zurich, Switzerland, and Departments of Chemistry (Science) and Biochemistry (Medicine), National University of Singapore (NUS), 117543 Singapore Ferritin is nature’s predominant iron storage protein. The molecule consists of a hollow protein shell composed of 24 subunits which is capable of storing up to 4500 iron atoms per molecule. Recently, this protein has been identified as a target molecule for increasing iron content in plant staple foods in order to combat dietary iron deficiency, a major public health problem in developing countries. Here, we present a novel technique for quantification of ferritin-bound iron in edible plant seeds using species-specific isotope dilution mass spectrometry (IDMS) by means of a biosynthetically produced 57Fe-labeled ferritin spike and negative thermal ionization mass spectrometry (NTIMS). Native plant ferritin and added spike ferritin were extracted in 20 mM Tris buffer (pH 7.4) and separated by anion exchange chromatography (DEAE Sepharose), followed by isotopic analysis by thermal ionization mass spectrometry. The chosen IDMS approach was critically evaluated by assessing the (i) efficiency of analyte extraction, (ii) identical behavior of spike and analyte, and (iii) potential iron isotope exchange with natural iron. Repeatabilities that can be achieved are on the order of 95% after isolation and purification.19 In the following, we describe the use of this spike for the (7) Harrison, P. M.; Arosio, P. Biochim. Biophys. Acta, Bioenerg. 1996, 1275, 161–203. (8) Crichton, R. R.; Ponceortiz, Y.; Koch, M. H. J.; Parfait, R.; Stuhrmann, H. B. Biochem. J. 1978, 171, 349–356. (9) Laulhere, J. P.; Lescure, A. M.; Briat, J. F. J. Biol. Chem. 1988, 263, 10289– 10294. (10) Sczekan, S. R.; Joshi, J. G. J. Biol. Chem. 1987, 262, 13780–13788. (11) Vandermark, F.; Vandenbriel, W. Plant Sci. 1985, 39, 55–60. (12) Busto, M. E. D.; Montes-Bayon, M.; Sanz-Medel, A. Anal. Chem. 2006, 78, 8218–8226. (13) Schaumloffel, D.; Prange, A.; Marx, G.; Heumann, K. G.; Bratter, P. Anal. Bioanal. Chem. 2002, 372, 155–163. (14) Busto, M.; Montes-Bayon, M.; Anon, E.; Sanz-Medel, A. J. Anal. At. Spectrom. 2008, 23, 758–764. (15) Rodriguez-Gonzalez, P.; Marchante-Gayon, J. M.; Alonso, J. I. G.; SanzMedel, A. Spectrochim. Acta Part B, Atomic Spectroscopy 2005, 60, 151– 207. (16) Harrington, C. F.; Vidler, D. S.; Watts, M. J.; Hall, J. F. Anal. Chem. 2005, 77, 4034–4041. (17) Reyes, L. H.; Sanz, F. M.; Espilez, P. H.; Marchante-Gayon, J. M.; Alonso, J. I. G.; Sanz-Medel, A. J. Anal. At. Spectrom. 2004, 19, 1230–1235. (18) Deitrich, C. L.; Raab, A.; Pioselli, B.; Thomas-Oates, J. E.; Feldmann, J. Anal. Chem. 2007, 79, 8381–8390. (19) Hoppler, M.; Meile, L.; Walczyk, T. Anal. Bioanal. Chem. 2008, 390, 53– 59.

development and evaluation of a novel method for quantification of ferritin-bound Fe by species-specific IDMS and its application to ferritin analysis in different plant seeds. EXPERIMENTAL SECTION Materials. Dry green peas, yellow peas, red kidney beans, pinto beans, soybeans, and lentils were purchased from local grocery stores in Zurich, Switzerland. Reagents (Sigma-Aldrich, Buchs, Switzerland) were of analytical grade, and high-purity (18 MΩ) deionized water (NANOpure system, Barnstead/Thermolyne, Dubuque, IA) was used in all experiments. Analytical grade, concentrated acids (HNO3 and HCl) were further purified by subboiling distillation. All materials were acid washed with HNO3 prior to use. 57Fe-enriched iron was purchased from Chemgas, Boulogne, France and isotopically enriched [57Fe]ferritin was produced and characterized as described earlier.19 For spiking with natural iron, isotopic reference material certified for iron isotopic composition was used (IRM-014, EU Institute of Reference Materials and Measurements, Geel, Belgium). Total Iron in Plant Seeds. Total iron content of legumes was analyzed by graphite furnace atomic absorption spectrometry (GFAAS; Atom Absorption Spectrometer AA240Z, Varian, Mulgrave, Australia). Plant samples were ground under liquid nitrogen with a rotar mill (ZM1, Retsch, Haan, Germany) using a titanium sieve (0.25 mm mesh) and successively mineralized by microwave digestion using an HNO3/H2O2 mixture. A commercial iron standard (Titrisol; Merck, Darmstadt, Germany) was used for external calibration. IDMS Protocol for Ferritin Analysis. Ferritin-bound Fe was determined in plant seeds by spiking samples with recombinant [57Fe]-ferritin and by iron isotopic analysis of spiked samples after ferritin separation from protein extracts. For extraction of proteins, the plant samples were ground as described above and suspended in a 10-fold volume of ice-cold Tris buffer (20 mM tris(hydroxymethyl)aminoethane-hydrochloride (Tris), 1 mM EDTA, 1 mM phenylmethylsulfonylfluorid (PMSF), 1% polyvinylpolypyrrolidone (PVPP), pH 7.4). Proteins were extracted by homogenizing the slurry for 60 s with a Polytron (PT1200 E, Kinematica, Luzern, Switzerland) set at the highest speed. After centrifugation at 5000g for 20 min at 4 °C, the supernatant was separated from the sediment. To assess the efficiency of plant cell disruption, i.e., to evaluate whether free iron and Fe-binding components (including ferritin-bound Fe) were completely released from the plant cells, iron concentration was measured in the extracts. Phosphate buffer (50 mM phosphate, 1 mM EDTA, 1 mM PMSF, 1% PVPP, pH 2) was used instead of Tris buffer in these extractions as free iron is poorly soluble at a higher pH as opposed to ferritin which is also soluble at pH 7. Sediments were washed three times with phosphate buffer in these experiments to increase iron recovery from the sediment. Supernatants were combined, and iron content was measured by GF-AAS. Plant ferritins in Tris buffer extracts were initially purified by ion exchange chromatography using diethylaminoethyl (DEAE) Sepharose.8-10,20 Chromatography columns (0.7 by 10 cm; Biorad, (20) Oh, S. H.; Cho, S. W.; Kwon, T. H.; Yang, M. S. J. Biochem. Mol. Biol. 1996, 29, 540–544.

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

7369

Reinach, Switzerland) were filled with 2 mL of DEAE Sepharose (GE Healthcare, Munich, Germany) and equilibrated with buffer A (20 mM Tris, 1 mM EDTA, pH 7.4). Plant extracts were applied to the columns and eluted by gravity flow. Columns were washed with buffer A and buffer B (20 mM Tris, 1 mM EDTA, 100 mM NaCl, pH 7.4) to elute Fe not bound to ferritin. Ferritin was eluted using buffer C (20 mM Tris, 300 mM NaCl, 1 mM EDTA, pH 7.4). Eluted ferritin fractions were further purified using ultrafiltration spin columns (100 kDa mw cutoff, Sartorius AG, Go¨ttingen, Germany) and centrifugation at 2000g. Isolates were washed by passing water 3 times through the columns before purification by gel filtration following previously described protocols.8-10,20 Gel filtration was performed using a Superdex 200 column with a fractionation range of 10-600 kDa (1 by 30 cm, GE Healthcare, Munich, Germany) and phosphate buffer for elution (0.15 mM sodium phosphate, 0.15 mM NaCl, pH 7) at a flow rate of 0.5 mL/ min. Fractions of 1 mL were collected and iron concentrations of all collected fractions were measured by GF-AAS. Ferritin was eluted approximately 20 min after injection, as previously determined under the same conditions.19 Mass Spectrometry. Collected ferritin fractions after ion exchange and ultrafiltration steps were mineralized using a HNO3/ H2O2 mixture and microwave digestion. Iron was isolated from potentially interfering elements by anion exchange chromatography, as described earlier, and further purified by Fe extraction into diethylether. Combined organic phases were evaporated to dryness.21 Iron isotopic analysis was performed by negative thermal ionization mass spectrometry (NTIMS) using FeF4- molecular ions and a rhenium double-filament ion source.22 In brief, evaporation as well as the ionization filament were coated with BaF2 to promote the formation of negatively charged ions. Sample iron was loaded as FeF3 in HF (40%) on top of the BaF2 layer on the evaporation filament and coated with a solution of AgNO3 in HF (20%). All mass-spectrometric measurements were carried out with a magnetic sector field mass spectrometer (MAT 262; Finnigan MAT, Bremen, Germany) equipped with a multicollector system for simultaneous ion beam detection. Measured iron isotope ratios were translated into ferritin-bound Fe content of the sample following isotope dilution principles.21 Blank measurements were performed by applying the IDMS protocol to the isotopically enriched [57Fe]-ferritin alone. Detection limits for the methods were assessed as three times the standard deviation of the blank measurements. RESULTS AND DISCUSSION IDMS in Elemental Speciation Analysis. Quantification of ferritin-bound Fe in plant samples may serve as a good example to demonstrate the relevance of element speciation in proteomics. While the amount of ferritin protein permits conclusions about expression efficiency of the protein itself, it does not allow one to study its primary function as an iron storage molecule, since iron saturation of the protein shell may vary widely.8-11 Only speciesspecific elemental analysis reveals the amount of iron that is bound to ferritin, i.e., the amount of storage iron in a sample. (21) Walczyk, T.; Davidsson, L.; Zavaleta, N.; Hurrell, R. F. Fresenius’ J. Anal. Chem. 1997, 359, 445–449. (22) Walczyk, T. Int. J. Mass Spectrom. Ion Processes 1997, 161, 217–227.

7370

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

Quantitative elemental speciation in proteomics usually involves the isolation of the metallo-protein of interest followed by elemental analysis. The most direct, nonisotopic approach includes one or more separation steps followed by element analysis of collected isolates. Accuracy of such measurements, however, can be subject to matrix effects as matrix composition usually varies significantly in chromatographic eluates. This limitation can be overcome by employing isotope dilution mass spectrometry (IDMS). By mixing a known amount of isotopically enriched element (spike) to the isolated element species of interest, the amount of element bound to this species can be calculated from induced changes in isotopic composition. This can be done either off-line after isolation of the metallo-protein or by mixing the spike continuously after column chromatography to the eluate before isotopic analysis.15 This approach, known as species-unspecific IDMS, is less sensitive to matrix effects as compared to nonisotopic techniques. Measurable isotope ratios of the isotopic spike and the natural element in the sample are affected by sample matrix to the same extent. In IDMS, isotopic spiking translates element amount/concentration in the sample into a specific isotope ratio. Provided that mixing is complete and the spiked sample is isotopically homogeneous, element loss has no effect on analytical accuracy after spiking. In species-unspecific IDMS, this controls for bias caused by fluctuations in the efficiency of sample transfer to the measuring device, e.g., when the chromatographic system is directly coupled to ICPMS and postcolumn spiking is employed for IDMS analysis. In species-specific IDMS, the basis of the presented technique, the isotopically tagged element species is added to the sample prior to species separation. Thus, analyte loss during species separation has, ideally, no effect on analytical accuracy. In nonisotopic approaches, nonquantitative analyte recovery during separation is usually the dominant source of analytical error. This refers in particular to metallo-proteins which often require multiple separation and purification steps for isolation. Protocol for Ferritin Purification from Plant Samples. Crude plant protein extracts contain different iron species of different molecular weight. This becomes apparent from the gel filtration chromatogram of a red kidney bean extract shown in Figure 1a. Initially, we followed earlier protocols for plant ferritin separation which were based on DEAE ion exchange chromatography followed by gel filtration chromatography as an additional purification step.8-10,20 For IDMS analysis, we have found that the gel filtration step can be omitted. Fractions containing ferritinbound Fe eluted at the same time as pure recombinant bean ferritin during gel filtration (see Figure 1b). Gel chromatograms shown in Figure 1c-h demonstrate for a variety of plant seeds that all iron other than ferritin iron has already been removed from the extracts by DEAE chromatography. We conclude that DEAE chromatography alone is sufficient for ferritin separation from plant samples as (a) iron species can be safely expected to be the same across different plant species, (b) no other iron species exist which are of similar molecular weight as ferritin, and (c) any remaining protein or other impurities in the isolate do not alter the iron isotopic composition of separated ferritin as long as they do not contain iron. Processing of [57Fe]-ferritin samples, following the described protocol, yielded a total chemical blank of 74 ± 20 ng of iron in isolated ferritin samples

Figure 1. Chromatographic separation of ferritin containing solutions and plant extracts by gel filtration (Superdex 200) using GF-AAS for iron detection: (a) red kidney bean extract, (b) pure recombinant bean ferritin, (c) red kidney bean extract purified by DEAE chromatography, (d) pinto bean extract purified by DEAE chromatography, (e) soybean extract purified by DEAE chromatography, (f) lentil extract purified by DEAE chromatography, (g) green pea extract purified by DEAE chromatography, and (h) yellow pea extract purified by DEAE chromatography. Iron contents of the collected 1 mL fractions are presented in the percentage of total injected iron. Table 1. Total Iron (n ) 3 per sample), Extractable Iron (n ) 3 per sample), and Ferritin-Bound Iron in Legume Seeds as Determined by Species-Specific Isotope Dilution Mass Spectrometry (n ) 5 per sample)

green peas yellow peas red kidney beans pinto beans lentils soybeans a

total Fe content (µg/g)

extractable Fe (µg/g)

ratio (total Fe/ extractable Fe)

ferritin-Fea (µg/g)

45.8 ± 1.7 45.2 ± 0.8 64.4 ± 1.0 55.6 ± 0.6 59.4 ± 1.2 65.8 ± 0.9

41.9 ± 2.7 47.8 ± 3.6 65.3 ± 1.2 51.6 ± 1.4 59.0 ± 2.4 58.6 ± 1.7

1.09 0.94 0.99 1.08 1.01 1.12

24.0 ± 0.6 (52) 28.2 ± 1.3 (62) 9.7 ± 0.3 (15) 15.9 ± 0.8 (29) 41.0 ± 1.7 (69) 25.2 ± 1.3 (38)

Ferritin-Fe content is stated in parentheses in percent of total Fe.

(n ) 5) for the entire analytical procedure which translates into an analytical detection limit of 60 ng of ferritin-bound Fe or 0.3 µg/g ferritin-bound Fe for a typical sample size of 0.2 g of plant tissue. Critical Evaluation of the Chosen IDMS Approach. The primary reason for choosing species-specific IDMS for quantification of ferritin-bound Fe were well-known difficulties in extracting ferritin quantitatively from biological samples and complexities of previously developed separation schemes. Both factors may result in significant analytical errors due to nonquantitative ferritin recovery. While analyte loss is of less concern in IDMS, other sources of inaccuracy, such as species conversion during storage and analysis, have to be considered instead. Furthermore, complete mixing of spike and sample is needed for the isotopic equilibration of the elemental species in the sample with the added spike. After equilibration, the ratio of metallo-protein coming from spike and sample and, thus, the isotope ratio of the element of interest has to remain constant during extraction, separation, and isotopic analysis.23 In this study, we made use of 57Fe enriched recombinant bean ferritin that we have produced earlier by biosynthesis in E. coli. We have previously demonstrated spike stability and that all detectable iron in the spike solution is ferritin-bound.19 Isotopic equilibration of added [57Fe]-ferritin spike with sample ferritin requires release of ferritin from the plant cell matrix (23) Meija, J.; Mester, Z. Anal. Chim. Acta 2008, 607, 115–125.

into the aqueous phase for extraction. In this study, we could show that virtually all plant iron becomes water extractable by Polytron homogenization (see Table 1). This ensures complete mixing of sample ferritin with added ferritin spike, i.e., isotopic equilibration of ferritin in the sample and the spike, before extraction. For accurate quantification of ferritin-bound Fe by speciesspecific IDMS, it is essential that both native plant ferritin and added [57Fe]-ferritin behave identically during ferritin separation. This requires that the amount ratio of sample ferritin and spike ferritin and, thus, the iron isotopic composition of ferritin are not altered during DEAE chromatography and ultrafiltration. To investigate this, pea extract was spiked with [57Fe]ferritin and the mixture was separated by DEAE chromatography and ultrafiltration. Isolated ferritin was processed for a second time in an identical fashion and iron isotopic compositions were compared before and after the second separation. No change in iron isotopic composition and amount ratios of ferritin-bound Fe coming from sample and spike was detected before and after separation (Table 2). Ferritin separation for species-specific IDMS is potentially vulnerable to isotopic exchange between ferritin-bound Fe and iron of other sources in the sample during processing. Such species conversions would inevitably result in analytical error. To examine this critical issue, 5 µg of natural Fe (IRM 014) was spiked into a solution containing 25 µg of [57Fe]-ferritin and the Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

7371

Table 2. Iron Isotopic Ratios after Two Sequential DEAE Separations of a Green Pea Extract Spiked with [57Fe]-Ferritina ratio (56Fe/57Fe) ferritin-Fe (µg/g) after first DEAE separation after second DEAE separation

0.164 ± 0.001 0.163 ± 0.004

3.45 ± 0.03 3.44 ± 0.11

a Three independent separation experiments using a single green pea extract.

Table 3. Ferritin-Bound Iron Analyzed by Species-Specific Isotope Dilution Mass Spectrometry in an Aqueous Pea Extract at Different Spike/Sample Ratios (n ) 3 per ratio)

mean RSD (%)

ratio [Fe(spike)/Fe(sample)]

ferritin-Fe (µg/g)

4.0 7.0 10.3

2.58 ± 0.03 2.59 ± 0.02 2.55 ± 0.02 2.57 0.75

iron isotopic composition of separated ferritin was compared with and without spiking. Measured 57Fe/56Fe isotope ratios of the separated [57Fe]-ferritin sample with and without spiking with natural iron were found to be very similar. Mean 57Fe/ 56 Fe iron isotope ratios of separated ferritin for five independent spiking experiments using NTIMS were 0.047 ± 0.003 and 0.046 ± 0.001 with and without spiking, respectively. Differences in iron isotope ratios translate into a recovery of (120 ± 35) ng of natural (spiked) Fe in isolated [57Fe]-ferritin samples. After blank correction, this represents less than 0.2% of iron in the [57Fe]-ferritin sample which can be considered negligible. For final assessment of the accuracy of the method, we analyzed an aqueous pea extract at variable amount ratios of added [57Fe]-ferritin spike to native sample ferritin of 4, 7, and 10 (see Table 3). Significant differences between analytical runs can be expected in case of a preferential recovery of ferritin from the spike solution; differences in stability of ferritin from spike and sample during separation or a highly variable chemical blank, to name only a few potential confounders, can be identified by such testing. As can be seen from the data presented in Table 3, analyzed concentrations of ferritin-bound Fe were identical for the different spiking ratios. Repeatability for independent runs was on the order of 1% RSD for the aqueous pea extract.

7372

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

Ferritin-Fe in Plant Samples Determined by SpeciesSpecific IDMS. Little is known about iron speciation and ferritin concentration in legume and other plant seeds. Reported data on ferritin-bound Fe content varied from 15% of total Fe as determined in soybeans by ferritin-specific immunoprecipitation24 to over 90% as determined in pea embryos by densitometric analysis of ferritinFe bands on native PAGE gels.25 Here, we present the first set of data on ferritin-bound Fe in legumes as determined by speciesspecific IDMS. As expected, concentrations of total iron and ferritin-bound Fe varied significantly between studied legumes. Percentages of ferritin-bound Fe in relative total iron were as low as 15% for red kidney beans and as high as 69% for lentils. Repeatability was in general e5% RSD as compared to ca. 1% RSD for the aqueous pea extract. This can be explained by small variations in extraction efficiency of ferritin from the plant matrix. CONCLUSIONS We have developed and successfully applied a novel technique to assess the content of ferritin-bound Fe in plant samples using IDMS. The developed technique may serve as a reference technique for such measurements due to the advantages of species-specific IDMS. The presented technique is independent of the efficiency of ferritin recovery from the sample and sources of uncertainty in the analytical result are fully traceable and quantifiable. In combination with immunochemical techniques for protein detection, quantification of ferritin-bound Fe now permits one to assess iron saturation of ferritin in plant tissues. Thus, the efficiency at which iron stores are filled up or depleted in plants can be studied not only for the evaluation of biofortification strategies but also for a better understanding of iron metabolism in plant biology, in general. ACKNOWLEDGMENT The authors wish to express their gratitude to Prof. R. Hurrell for providing laboratory infrastructure for this research. We also thank Adam Krzystek for conducting parts of the TIMS measurements and Karin Hotz for her scientific advice. Received for review April 24, 2009. Accepted July 16, 2009. AC900885J (24) Beard, J. L.; Burton, J. W.; Theil, E. C. J. Nutr. 1996, 126, 154–160. (25) Marentes, E.; Grusak, M. A. Seed Sci. Res. 1998, 8, 367–375.