A comparative study of element concentrations and binding in transgenic and non-transgenic soybean seeds

Abstract

Transgenic and non-transgenic soybean seeds were compared in terms of total element concentrations, behavior of elements during sequential extraction fractionation and element bioaccessibility using an in vitro simulated gastrointestinal digestion. The analysis were carried out by ICP-sector field-MS or size-exclusion ICP-MS (25 elements in concentrations varying from ng g⁻¹ to the % level). It seems that transgenic and non-transgenic soybean seeds exhibit statistically significant differences in concentrations of Cu, Fe and Sr, which are also reflected by element contents in water extracts and residues. Additionally, contributions of bioaccessible fractions of Cu, Fe and other elements (Mn, S, Zn) for transgenic soybean seeds appear to be larger than those found in non-transgenic soybean seeds.

---

Introduction

Soybean (Glycine max (L) Merrill) is a cultivar of great interest in the world economy, the total crop value in 2009 exceeding in the USA 31.7 billion dollars.¹ It represents a significant source of fatty acids and proteins for human and animal nutrition, and is also applied for non-edible uses, including industrial feed-stocks and production of bio-diesel.² The major source of this commodities is the seed, rich in proteins (∼40%), oil (∼20%) and carbohydrates (∼35%).^3,4

Soybean seeds are also known to be rich in elements, such as Ca, Cu, Fe, Mg, Mn, P and Zn. Calcium has several important functions in human body, such as bone and teeth formations. An adequate Ca intake for healthy adults is 1000–1300 mg day⁻¹. Copper is a component of enzymes required for Fe metabolism. The recommended dietary allowance (RDA) of Cu for adults is about 0.9 mg day⁻¹.⁵ Iron is a constituent of hemoglobin and various enzymes and its RDA for men and women are 8 and 18 mg day⁻¹, respectively. The bioaccessibility of this element is affected mainly by the presence of phytate and fibers.⁶ Magnesium is a co-factor for enzymatic systems and its RDA varies from 240 to 420 mg day⁻¹. Manganese is involved in bone formation and enzymes responsible for amino acids, cholesterol and carbohydrate metabolism, its RDA value varies from 1.6 to 2.3 mg day⁻¹. Phosphorus is responsible for maintenance of pH, storage and transfer of energy and nucleotide synthesis. The RDA for this element is about 700 mg day⁻¹. Finally, Zn is a component of multiple enzymes and proteins. The RDA for this element is 8–11 mg day⁻¹. An accurate determination of these elements in soybean seeds can provide access to their nutritional quality.

Infestation of soybean cultures with weeds can cause prejudice in productivity index, product quality and crop yield. As consumption of herbicides has a high impact on productivity costs,⁷ areas cultivated with transgenic soybeans tolerant to herbicides are growing rapidly. Roundup Ready^® soybean seeds represent more than a half of cultivated areas of genetically modified organisms in the world.⁸ Roundup Ready^® crop lines contain a gene derived from Agrobacterium sp. strain CP4, encoding glyphosate-tolerant enzyme so-called CP4 EPSP synthase. Expression of CP4 EPSP synthase results in glyphosate-tolerant crops, enabling a more effective weed control by allowing post-emergent herbicide application.⁷

Effects of genetic modification are known to greatly change the proteome in general,⁸ but its effects on the elemental composition (metallome) are largely unknown. The goal of this research was the investigation of the effects of the genetic modification on elemental composition in non-transgenic and Roundup Ready^® transgenic soybean seeds. The differences in multielement concentrations were addressed on the total element level, sequential extraction fractionation and during simulated gastrointestinal digestion. The molecular weight distribution of elements in soybean seed water extracts was examined as well.

Experimental

Instrumentation

An Element (Thermo Fisher Scientific, Bremen, Germany) XR ICP-SF-MS, equipped with a demountable Fassel-type torch shielded with a Pt ground electrode and a quartz bonnet, was used throughout. Two different sample introduction systems were used. Aqueous solutions were introduced with a MicroMist glass concentric nebulizer (Glass Expansion, Australia) fitted to a double-pass quartz spray chamber. A standard 1.75 mm ID quartz injector and Pt sample (1.10 mm orifice diameter) and skimmer (0.8 mm orifice diameter) cones were used. Sample and standard solutions were delivered using a CETAC (Omaha, NE) ASX-520 autosampler.

For the analysis of organic solutions, a CETAC modified total consumption micro-flow nebulizer DS-5 attached to a laboratory-made low-dead volume (8 mL) single-pass drain-free glass spray chamber was used. The spray chamber was jacketed and heated to 80 °C with a water–glycol mixture circulating through a NesLab RTE-111 thermostat (Thermo Fisher Scientific, Waltham, MA). A 1.0-mm ID quartz injector for organic matrices and Pt sampler (1.1 mm orifice diameter) and skimmer (0.8 mm orifice diameter) cones were used. O₂ was added to the sample Ar flow through an additional mass flow controller of the spectrometer. Sample and standard solutions were introduced in a micro-flow injection mode (μFI) with tetrahydrofuran (THF) as a carrier. μFI was carried out using a Dionex (Amsterdam, Holland) HPLC system, comprising of a UltiMate 3000 pump, a UltiMate 3000 autosampler and a low port-to-port dead volume μFI valve.

Reagents, solutions and materials

Deionized water from a Millipore ELIX 3 water purification system (Molsheim, France) was used throughout. ACS grade THF and n-hexane, CHROMASOLV LC-MS iso-propanol and methanol, NaOH pellets and reagents used for gastrointestinal digestion were purchased from Sigma-Aldrich. CHROMASOLV chloroform was supplied by Riedel-de-Häen (Seelze, Germany). INSTRA-ANALYZED HNO₃ (69–70%) and H₂O₂ (30%) were obtained from J. T. Baker (Deventer, Holland). A SPEX CertiPrep (Matuchen, NJ, USA) Claritas PPT multi-element standard solution (10 μg g⁻¹ of Au, Hf, Ir, Pd, Pt, Rh, Ru, Sb and Te), a Sigma-Aldrich TraceCERT multi-element standard solution (10 μg g⁻¹ of Be, Cd, Co and Mn, 20 μg g⁻¹ of Cr, Cu, and Ni, 40 μg g⁻¹ of Al, As, Ba, Pb and V, and 100 μg g⁻¹ of B, Fe, Se, Tl and Zn) and SCP Science (Clark Graham, QC, Canada) PlasmaCAL single-element standards (1000 μg g⁻¹ of Ag, Bi, Ca, Hg, K, Li, Mg, Mo, P, S, Si, Sr and Ti) were used to prepare multi-element standard solutions in 5.0% (w/v) HNO₃ for external calibration (0, 1, 2, 5, 10, 20, 50, 100, 200, 500 and 1000 ng g⁻¹) and for standard additions. A 1 μg L⁻¹ SPEX CertiPrep Claritas PPT standard solution (Ba, B, Co, Fe, Ga, In, Li, Lu, Na, Rh, Sc, Tl, U, Y and K) was used to tune the ICP-SF-MS instrument combined with aqueous solutions introduction.

A SCP Science Conostan S-21 multi-element oil-based standard containing 100 μg g⁻¹ of Ag, Al, B, Ba, Ca, Cd, Cr, Cu, Fe, Mg, Mn, Mo, Na, Ni, P, Pb, Si, Sn, Ti, V and Zn, and Conostan 1000 μg g⁻¹ single-element oil-based standards of As, Bi, Li, S, Co, Hg, Sb and Sr were used to prepare multi-element working standard solutions at 10, 20, 50, 100, 200 and 500 ng g⁻¹ in hexane, hexane/iso-propanol (1 : 1) and chloroform–methanol (1 : 1). A 1 ng g⁻¹ tuning solution was prepared by diluting SCP Science PlasmaCAL 1000 μg g⁻¹ mono-element standards of Ba, B, Co, Fe, Ga, In, Li, Lu, Na, Rh, Sc, Tl, U, Y and K in THF.

A Standard reference material of non-fat milk powder (SRM 1549) was purchased from National Institute of Standards and Technology (NIST, Gaithersburg, MD).

Measurements

Isotopes of 25 elements (²⁷Al, ⁷⁵As, ¹³⁷Ba, ²⁰⁹Bi, ⁴⁴Ca, ¹¹¹Cd, ⁵⁹Co, ⁵²Cr, ⁶³Cu, ⁵⁶Fe, ²⁰²Hg, ⁷Li, ²⁴Mg, ⁵⁵Mn, ⁹⁵Mo, ⁶⁰Ni, ³¹P, ²⁰⁸Pb, ³²S, ¹²¹Sb, ⁷⁸Se, ¹¹⁸Sn, ⁸⁸Sr, ¹²⁵Te, ⁶⁶Zn) were measured by ICP-SF-MS in medium resolution (R = 4000) in order to resolve them from plasma-based polyatomic interfering ions, especially in the case of Al, As, Ca, Cr, Cu, Fe, Mg, Mn, Ni, S, Se, P and Zn isotopes. ICP-SF-MS instrumental settings were optimized daily using adequate tuning solutions (Table 1). The highest stable and reproducible (relative standard deviations <3%) signals obtained for ⁷Li, ¹¹⁵In and ²³⁸U isotopes, and the lowest ¹³⁸Ba¹⁶O/¹³⁸Ba intensity ratios were considered for best analytical performance. Mass calibration was carried out at resolutions of 300 and 4000. Mass off-set values were evaluated each time and implemented into data acquisition methods in order to compensate mass drifts related to magnet hysteresis. A quantification mode was used for determining concentrations of elements in aqueous solutions. A time-resolved mode was used for analysis of organic solvents with μFI-ICP-SF-MS. Built-in Element XR ICP-MS software was used for acquisition and integration of measured signal.

Table 1 Relevant operating conditions of ICP-SF-MS with introduction of aqueous (A) and organic (B) solutions

	A	B
Inductively coupled plasma mass spectrometry
Torch position/mm	+2.20 (X), +1.10 (Y), −3.50 (Z)
Forward power, W	1200	1500
Plasma Ar flow rate, L min^−1	16.00	16.00
Auxiliary Ar flow rate, L min^−1	1.00	0.90
Sample Ar flow rate, L min^−1	1.10	0.65
O2 flow rate, L min^−1	—	0.12
Sample flow rate, μL min^−1	300	15
Injection volume, μL	—	5
Drain flow rate, mL min^−1	450	—
Ion transmission
Extraction lens, V	−2000	−2000
Focus lens, V	−1150	−1100
X-deflection lens, V	+0.30	+1.15
Y-deflection lens, V	−2.40	−2.80
Shape lens, V	+130	+125
SEM	+2800
SEM deflection	+570
Data acquisition
Isotopes measured	^7Li, ^24Mg, ^27Al, ^31P, ^32S, ^44Ca, ^52Cr, ^55Mn, ^56Fe, ^59Co, ^60Ni, ^63Cu, ^66Zn, ^75As, ^78Se, ^88Sr, ^95Mo, ^111Cd, ^115In, ^118Sn, ^121Sn, ^125Te, ^137Ba, ^202Hg, ^208Pb, ^209Bi
Mass window (%)	125
Integration window (%)	60
Settling time, ms	1 (Al, Bi, Co, Cu, Fe, In, Mn, Mo, Pb, S, Se, Sn, Te, Zn), 34 (Ni), 35 (Cr), 37 (P), 39 (As), 42 (Ba, Cd), 43 (Ca), 60 (Hg), 71 (Mg), 300 (Li, Sr)
Sampling time, ms	20 (As, Ba, Bi, Cd, Hg, In, Li, Mo, Pb, Sb, Se, Sn, Sr, Te), 50 (Al, Ca, Co, Cr, Cu, Fe, Mg, Mn, Ni, P, S, Zn)
Number of samples per peak	20
Detection mode	Triple
Scan type	EScan
Integration type	Average
Resolution	4000 (medium)

---

Samples and procedures

Soybean seeds. Laboratory grown transgenic (variety MSOY 7575 RR) and non-transgenic soybean seeds (variety MSOY 7501) were donated by Prof. Siu Mui Tsai (CENA/USP, Piracicaba, SP).⁹ These seeds were originally obtained from Monsanto, Brazil.

Total element content analysis. Samples were dried in an oven at 60 °C to constant mass. Then, 0.2 g of dried samples were digested in EasyPrep vessels using 6.3 mL of concentrated HNO₃ and 0.75 mL of 30% (w/v) H₂O₂. Vessels were closed and subjected to digestion at 160 °C for 25 min in a CEM (Matthews, NC, USA) MARS microwave digestion system. In the case of NIST SRMs, 0.2 g samples were digested in 50 mL polypropylene (PP) tubes with 1.2 mL of concentrated HNO₃ and 0.3 mL of 30% (w/v) H₂O₂ at 80 °C for 3 h using a SPC Science DigiPrep MS heating block with time and temperature controller. Resulting digests were cooled, diluted to 25 mL (SRMs) or 50 mL (soybean seeds) with water and analyzed on the total content of elements by ICP-SF-MS with 3 standard additions. Three independent replicates were made; respective blanks were considered in final results.

Sequential extraction. In the first step, performed to separate the water-insoluble fraction, 0.1 g of soybean samples were sonicated in a Branson (Danbury, CT, USA) ultrasonic bath with 2 mL of a hexane/iso-propanol mixture (1 : 1 v/v) for 30 min. This was repeated 3 times; then, supernatants were pooled and analyzed by ICP-SF-MS using external standard solutions prepared in the hexane/iso-porpanol mixture for calibration. For the second step, residues from the previous step were dried in a sample concentrator DB-3 system (Bibby Scientific, Staffordshire, UK) to remove organic solvents and next sonicated with 2.5 mL of water for 2 min using a Bioblock Scientific (Illkirch, France) Vibracell 75115 ultrasonic device at 20% of its maximum power and in a pulse mode (1-s pulses interrupted by 1-s breaks). Supernatants were recovered using a Hettich (Tuttligen, Germany) Universal 16 centrifuge at 3000 rpm for 10 min. This procedure was repeated two times, and the supernatants were pooled and analyzed by ICP-SF-MS with 3 standard additions. Finally, sample residues were digested with 1.2 mL of concentrated HNO₃ and 0.3 mL of 30% (w/v) H₂O₂ at 80 °C for 3 h using 50 mL PP tubes in a SCP ScienceDigiPrep MS heating block. Final solutions were diluted with water to 15 mL and analyzed by ICP-SF-MS with 3 standard additions.

Three independent replicates were made for each digestion and extraction. Respective blanks were considered in the final results.

SEC-ICP-SF-MS analysis. Chromatographic separations were carried out using an Agilent 1100 (Wilmington, DE) high performance liquid chromatography (HPLC) system, comprising an HPLC pump, a degasser, an autosampler and an UV detector, and a Superdex 75 HR 10/30 (Amersham Pharmacia Biotech, Uppsala, Sweden) size exclusion column (10 × 300 mm × 5 μm), with optimum fractionation range of 3–70 kDa. 100 μL aliquots of water extracts were injected into the SEC column and chromatographic run was isocratically made using 100 mmol L⁻¹ ammonium acetate (pH 7.4) buffer as eluent at 0.7 mL min⁻¹. The eluent from the column was directly fed into ICP-SF-MS.

Bioaccessibility test. The protocol used was described elsewhere.¹⁰ Briefly, for gastric extraction, 5 mL of a gastric solution (50 mg of pepsin with 5 mL of 150 mmol L⁻¹ NaCl, pH 2.5) was added to 0.3 g of soybean samples and incubated at 37 °C for 4 h. For simulating intestinal extraction, pH of sample solutions was adjusted to 7.4 with concentrated NaOH solution and then, 10 mL of an intestinal solution, containing solutions of 3.0% (w/v) pancreatin, 1.0% (w/v) amylase and 1.5 g L⁻¹ bile salts, were added and incubated at 37 °C for 4 h. Resulting supernatants were centrifuged in an Universal 16 centrifuge at 3000 rpm for 15 min and measured to determine concentrations of elements using ICP-SF-MS with 3 standard additions. Additionally, sample residues were digested and analyzed to provide the mass balance. Three independent replicates were made, and respective blanks were considered in final results.

Results and discussions

Multielement total analysis

Instrumental detection limits (DL) of studied elements obtained with ICP-SF-MS were assessed according to 3σ criterion (3 × SD of 10 repeated measures of blanks divided by slopes). These DLs can be divided into 3 groups: from 0.0003 to 0.001 μg L⁻¹ (As, Ba, Bi, Cd, Co, Hg, Li, Mo, Pb, Sb, Sr and Te); from 0.02 to 1 μg L⁻¹ (Al, Cr, Cu, Mn, Ni, S, Se, Sn and Zn) and higher than 1 μg L⁻¹ (Ca, Fe, Mg and P). Precision, expressed as the relative standard deviation (RSD) of 3 repeated measurements of multi-element standard solutions was within 1.1 (P, Cr) to 21% (Mo).

Accuracy of ICP-SF-MS was verified by analyzing NIST SRM 1549 (non-fat milk powder). Results of the analysis are given in Table 2. Recoveries of measured elements were within 84–113%, indicating a good agreement between certified and determined concentrations. Precision as RSD for analysis of 3 digested SRM samples varied from 0.4 to 10%.

Table 2 Results of NIST SRM 1549 analysis and total analysis from transgenic (T) and non-transgenic (NT) soybean seeds by ICP-SF-MS after oxidative digestion (average values, n = 6 ± standard deviations)

Element	SRM, μg g^−1		Soybean seeds, μg g^−1
	Found	Certified	T	NT
a Detection limit. b Quantitation limit. c Concentration in %.
Al	1.90 ± 0.18	2	<0.24^a	<0.24^a
As	<0.0009^a	<0.0029	<0.0009^a	<0.0009^a
Ba	—	—	0.104 ± 0.004	<0.002^b
Bi	—	—	<0.001^a	<0.001^a
Ca	1.44 ± 0.07^c	1.3^c	998 ± 92	924 ± 96
Cd	<0.0006^b	<0.0007	<0.0006^b	<0.0006^b
Co	—	—	0.046 ± 0.002	0.028 ± 0.001
Cr	<0.15^a	0.0023 ± 0.0004	<0.15^a	<0.15^a
Cu	0.61 ± 0.02	0.7	9.8 ± 1.0	5.9 ± 0.9
Fe	1.75 ± 0.14	1.78	81.1 ± 1.7	65.9 ± 2.5
Hg	<0.0015^a	<0.0008	<0.0015^a	<0.0015^a
Li	—	—	<0.006^a	<0.006^a
Mg	0.13 ± 0.01^c	0.12^c	3979 ± 407	3693 ± 474
Mn	0.24 ± 0.01	0.26	27.9 ± 0.5	24.8 ± 1.7
Mo	0.34 ± 0.02	0.34	0.32 ± 0.02	0.35 ± 0.05
Ni	—	—	0.50 ± 0.06	0.34 ± 0.07
P	1.13 ± 0.05^c	1.06^c	6589 ± 308	6366 ± 595
Pb	0.019 ± 0.003	0.0181 ± 0.0010	0.006 ± 0.004	0.009 ± 0.003
S	0.33 ± 0.02^c	0.35^c	3083 ± 135	2801 ± 200
Sb	<0.0003^a	<0.006	0.002 ± 0.001	0.002 ± 0.001
Se	0.011 ± 0.01	0.119 ± 0.007	0.11 ± 0.04	<0.02^b
Sn	<0.031^a	<0.0016	<0.031^a	<0.031^a
Sr	—	—	19.9 ± 3.0	30.1 ± 3.5
Te	—	—	<0.0003^a	<0.0003^a
Zn	45.0 ± 2.4	46.1	38.8 ± 5.2	37.5 ± 4.7

---

Total concentrations of elements in transgenic (T) and non-transgenic (NT) soybean seeds obtained after microwave-assisted digestion are given in Table 2. According to t-test (p = 0.005, n = 4), statistically significant differences between T and NT soybean seeds were found for Co, Cu, Fe and Sr. Concentrations of Co, Cu and Fe in T soybean seeds are higher by 39, 40 and 20%, respectively, than corresponding concentrations of these elements in NT soybean seeds. For Sr, a higher concentration was found in NT soybean seeds; and the difference between concentrations was of 34%. Differences in Cu and Fe concentrations between T and NT soybean seeds were lately established by Sussulini et al.⁸ when protein spots cut from a 2D electrophoretic gel were analyzed by ETAAS. It is important to note that elemental concentrations in soybean seeds are dependent on various factors, including soil characteristics, water source composition, that can affect plant development.¹¹ Because the majority of these factors were controlled during the growth of T and NT soybean seeds investigated here, it could be expected that differences in concentrations that were found for Co, Cu, Fe and Sr should be related to genetic modification.

Fractionation analysis

A two-step solvent extraction procedure was carried out to get to know how Co, Cu, Fe, Sr and other elements (Ca, Mg, Mn, Ni, S, Zn) measured in both soybean seeds would be distributed between hydrophobic (by ultrasound assisted extraction with a mixture of hexane/iso-propanol) and water soluble species (by ultrasound assisted extraction with water). Residues left were digested to get mass balance information. Results of this analysis are shown in Table 3 for T and NT soybean seeds.

Table 3 Ca, Co, Cu, Fe, Mg, Mn, Ni, S, Sr and Zn concentrations in each extracted fraction, in residues, and respective mass balances for transgenic (T) and non-transgenic (NT) soybean seeds samples (average values, n = 6 ± standard deviations)

Element	Concentration, μg g^−1						Mass balance (%)
	Organic fraction		Water fraction		Residue
	T	NT	T	NT	T	NT	T	NT
a Quantitation limit.
Ca	<23^a	<23^a	358 ± 32	325 ± 31	500 ± 67	485 ± 45	84.7 ± 1.67	88.2 ± 10.5
Co	<0.1^a	<0.1^a	0.038 ± 0.003	0.023 ± 0.002	0.007 ± 0.001	0.006 ± 0.001	97.6 ± 7.4	103.9 ± 10.5
Cu	<3.4^a	<3.4^a	6.20 ± 0.74	3.11 ± 0.39	2.37 ± 0.40	1.84 ± 0.23	88.1 ± 14.1	82.7 ± 8.6
Fe	<46^a	<46^a	41.2 ± 4.3	29.3 ± 5.6	32.7 ± 7.2	26.9 ± 7.7	92.3 ± 8.0	80.0 ± 6.8
Mg	<0.4^a	<0.4^a	2873 ± 220	2564 ± 232	1058 ± 192	1144 ± 306	105.1 ± 0.3	98.6 ± 17.1
Mn	<0.4^a	<0.4^a	15.2 ± 1.3	11.8 ± 0.7	8.35 ± 1.58	9.80 ± 0.35	87.3 ± 11.7	86.6 ± 1.4
Ni	<2.3^a	<2.3^a	0.40 ± 0.03	0.27 ± 0.03	0.036 ± 0.009	0.037 ± 0.008	90.8 ± 5.9	94.1 ± 7.4
S	<2.4^a	<2.4^a	2742 ± 352	2429 ± 204	778 ± 245	855 ± 115	121.3 ± 7.7	114.1 ± 1.7
Sr	<0.9^a	<0.9^a	2.41 ± 0.35	1.47 ± 0.28	17.1 ± 1.8	24.1 ± 0.9	90.4 ± 10.4	89.4 ± 1.9
Zn	<4.0^a	<4.0^a	40.5 ± 2.5	37.7 ± 3.5	6.48 ± 1.34	9.15 ± 0.34	122.4 ± 8.6	121.6 ± 9.8

---

It can be seen that sums of element concentrations in separated fractions and in digested residues are in good agreement with total concentrations determined. Accordingly, recoveries of elements varied from 80% in the case of Fe in NT soybean seeds to 122% in the case of Zn in T soybean seeds. Only P could be quantified in the organic fraction (∼6% in reference to total concentrations in both samples), probably due to the presence of corresponding phospholipids.

The water fraction, mostly related to the presence of proteins,¹² contributed from 36 (Ca) to 104% (Zn) to total concentrations of elements in T soybean seeds. For NT soybean seeds, water soluble element species accounted for 35 (Ca) to 100% (Zn) in reference to total concentration of these elements. In the case of Sr, it was found that only 5% of this element can be found in water extracts of both soybean seeds. An overwhelming part of Sr was established to be associated with residues. These results are consistent with data reported by Koplík et al.,¹³ who also found high extractabilities of different elements, i.e., from 37% for Ca to 82% for Mo, from soybean flour leached with a Tris-HCl buffer.

Considering concentrations of Co, Cu and Fe determined in water extracts and Sr in residues, a statistically significant difference between T and NT soybean seeds was assessed, as in the case of total concentrations of these elements. Concentrations of Co, Cu and Fe in T soybean seeds water extracts were higher by 39, 50 and 29% as compared to respective concentrations found in water extracts of NT soybean seeds. In the case of Sr, its concentration in the NT soybean seed residue was higher by 34% as compared to the respective one for T soybean seeds.

For further experiments, only Ca, Cu, Fe, Mg, Mn, S and Zn, for which concentrations higher than 2 μg g⁻¹ were determined in water extracts, were considered.

SEC-ICP-SF-MS of water extracts

Molecular weight (MW) distribution of Cu and Fe and other elements (Ca, Mg, Mn, S and Zn) in water extracts was examined by SEC with UV and element specific (ICP-SF-MS) detections. The summary of SEC-ICP-SF-MS analysis on the Superdex 75 column is given in Table 4.

Table 4 Summary of SEC-ICP-SF-MS analysis on superdex 75 column

Element	Transgenic soybean seeds		Non-transgenic soybean seeds
	MW, kDa	Peak (%)	MW, kDa	Peak (%)
a Estimation below lower limit (3 kDa) of the calibration range given by the manufacturer. b Value expressed in Da. c Estimation above upper limit (70 kDa) of the calibration range given by the manufacturer.
Ca	1.9	0.17	2.1^a	0.19
	0.01	100	8.7^a^,^b	100
Cu	148	0.89	155^c	1.1
	54	3.2	55	6.7
	29	1.9	29	1.1
	13	0.89	13	0.28
	2	93	2.1^a	91
Fe	162	87	158^c	81
	62	7.9	62	10
	4.1	5	4.1	8.5
Mg	155	0.004	162^c	0.004
	68	0.06	68	0.05
	0.62	100	0.62^a	100
Mn	74	14	74^c	14
	2.2	86	2.2^a	86
S	57	18	62	18
	19	13	19	13
	6.6	13	6.6	9
	2.1	52	2.1^a	55
	0.08	4.1	0.07^a	4.3
Zn	151	0.11	155^c	0.04
	56	2.8	56	2.5
	26	0.12	26	0.09
	2.2	97	2.2^a	97

---

Apexes of major peaks of Cu, Mn, S and Zn were found at 22.5–23.0 min which corresponds to low molecular weight (LMW) species (∼2 kDa, below the column lower calibration limit). Previous studies devoted to soybean flours^13,14 showed a similar elution behavior of Cu and Zn in water extracts. As reported before, both elements were established to be bound by corresponding LMW compounds. Traces of Cu and Zn were also found to be present in medium and high molecular weight (MMW, HMW) ranges. No HMW species of Mn were detected in water extracts of T and NT soybean seeds as previously reported in the literature,¹⁵ however, it was already assessed that MW distribution of Mn among LMW, MMW, and HMW fractions alters during aging of extracts, and that HMW Mn related species are relatively unstable.¹³ In the case of Fe, major peaks found in water extracts of soybean seeds were eluted around 13.0 min, which corresponds to 160 kDa (above upper column calibration limit) species. These could be some specific metalloproteins or protein non-specific Fe chelators of molecular weight >100 kDa, that are able to form very stable Fe complexes in soybeans.^13,16

Although morphology of chromatograms obtained with ICP-SF-MS detection (Fig. 1) were similar for analyzed water extracts of both soybean seeds, areas of the most abundant peaks for Cu and Fe in T soybean seeds were 3- and 2-fold higher, respectively, than those found in NT samples.

Fig. 1 Chromatograms of Cu and Fe compounds in water extracts of transgenic (T) and non-transgenic (NT) soybean seeds samples.

This could lead to the conclusion that Cu and Fe are associated with compounds more expressed in T soybean seeds. These results corroborate as well to the ones described by Brandão et al.,¹⁷ who studied soybean proteome and provided identification of 10 protein spots with different expressions between T and NT soybean seeds.

Bioaccessibility of elements in soybean seeds

Soybeans are considered to be a great source of proteins (∼40% of their dried mass), fat (∼20%) and carbohydrates (∼35%) for human and animal feeds. They are also recognized to contain relatively high amounts of minerals that can cover element deficiency in the diet.⁴ For that reason, it is important to quantify and compare bioaccessible fraction of elements in T and NT soybean seeds.

A gastrointestinal digestion approach was applied for this purpose. Results obtained for Ca, Cu, Fe, Mg, Mn, S and Zn are shown in Table 5 for T and NT soybean seeds. In reference to total concentrations, contributions of elements in the bioaccessible fraction were found to vary from 95 (Ca) to 9.2% (Fe) in NT soybean seeds; and from 88 (Ca) to 35% (Cu) for T soybean seeds. Except for Ca and Mg, extractabilities of elements achieved for gastrointestinal digestion are lower if compared to respective ones obtained for water extracts. One of the reasons for this could be the presence of fibers and phytates in soybean seeds,¹⁸ typically in amounts of 7 and 2% (w/v), respectively, that are well know to form insoluble complexes with positively charged proteins and cations of Ca, Cu, Fe, Mg, Zn in intestinal pH. In a consequence, absorption of these elements after gastrointestinal digestion can be decreased. As already reported,¹⁹ Fe absorption in soybeans is relatively low. Even after removal of phytates, which slightly enhance absorption of this element, soy proteins themselves are still inhibitory to Fe absorption, likely due to the presence of some Fe-binding peptides.²⁰

Table 5 Ca, Cu, Fe, Mg, Mn, S and Zn concentrations in bioaccessible fractions, residues, and recoveries related to total concentrations of transgenic (T) and non-transgenic (NT) soybean seeds samples (average values, n = 3 ± standard deviations)

Element	Concentration, μg g^−1				Recovery (%)
	Bioaccessible fraction		Residue
	T	NT	T	NT	T	NT
Ca	824 ± 114	817 ± 72	160 ± 17	193 ± 8	105 ± 13.2	117 ± 12.7
Cu	3.52 ± 0.13	1.85 ± 0.02	4.77 ± 0.57	3.39 ± 0.20	80 ± 8.1	85 ± 8.6
Fe	29.1 ± 2.4	5.99 ± 0.26	52.6 ± 3.8	59.0 ± 9.1	101 ± 6.3	100 ± 6.8
Mg	1690 ± 248	1318 ± 78	492 ± 53	511 ± 20	91 ± 11.9	75 ± 7.6
Mn	12.2 ± 0.1	10.5 ± 0.3	7.67 ± 0.90	7.77 ± 0.33	71 ± 9.0	73 ± 10.0
S	1322 ± 56	1961 ± 76	958 ± 64	1476 ± 68	73 ± 8.8	107 ± 9.6
Zn	22.1 ± 0.1	19.6 ± 0.4	15.7 ± 0.6	22.6 ± 0.8	92 ± 2.3	107 ± 7.6

---

Concentrations of bioaccessible fractions for Cu, Fe, Mn, S and Zn were established to be different for both soybean seeds. In the case of Cu, Mg, Mn and Zn, bioaccessible fractions of these elements could be related to LMW species, such as some organic acids (ascorbic, succinic, mallic). These species have already been recognized to increase bioavailability of metals in soybean and soy products.¹⁸

Conclusions

Transgenic and non-transgenic soybean seeds show differences in concentrations of Co, Cu, Fe and Sr which are also reflected by element contents in water extracts and residues. Although element distribution profiles in water extracts of T and NT soybean seeds are similar, Cu and Fe in transgenic soybean seeds exhibit higher areas of major size exclusion chromatographic peaks. Bioaccessible fractions of such elements as Cu, Fe, Mn, S and Zn are larger in transgenic soybean seeds.

Acknowledgements

Lidiane R. V. Mataveli and Marco A. Z. Arruda are grateful to Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for fellowship and financial support. The authors are thankful to Marcelo A. O. da Silva for helping with the cover artwork.

Notes and references

1. SoyStats, The American Soybean Association, available at http://www.soystats.com.
2. J. M. Saz and M. L. Marina, J. Sep. Sci., 2007, 30, 431–451.
3. M. Hajduch, A. Ganapathy, J. W. Stein and J. J. Thelen, Plant Physiol., 2005, 137, 1397–1419.
4. Y. Yip, K. Chan, P. Cheung, K. Poon and W. Sham, Food Chem., 2009, 112, 1065–1071.
5. Food and Nutrition Board of the Institute of Medicine, Standing Committee on the Scientific Evaluation of Dietary Reference Intakes, Subcommittee on Interpretation and Uses of Dietary Reference Intakes, DRI Dietary Reference Intakes. Application in Dietary Assessment, National Academy Press, Washington, DC, 2000.
6. A. G. Gharib, S. G. Mohseni, M. Mohajer and M. Gharib, Radiochem. Radioanal. Letters, 2006, 207, 209–215.
7. T. Funke, H. Han, M. L. H. Fried, M. Fischer and E. Schonbrünn, Proc. Natl. Acad. Sci. U. S. A., 2006, 35, 13010–13015.
8. A. Sussulini, G. H. M. F. Souza, M. N. Eberlin and M. A. Z. Arruda, J. Anal. At. Spectrom., 2007, 22, 1501–1506.
9. C. A. Moldes, L. O. Médici, O. S. Abrahão, S. M. Tsai and R. A. Azevedo, Acta Physiol. Plant., 2008, 30, 469–479.
10. S. Mounicou, J. Szpunar, R. Lobinski, D. Andrey and C. J. Blake, J. Anal. At. Spectrom., 2002, 17, 880–886.
11. J. Naosuka and P. V. Oliveira, J. Braz. Chem. Soc., 2007, 18, 1547–1553.
12. A. Sussulini, J. S. Garcia, M. F. Mesko, D. P. Moraes, E. M. M. Flores, C. A. Perez and M. A. Z. Arruda, Microchim. Acta, 2007, 158, 173–180.
13. R. Koplík, H. Pavelková, O. Mestek, F. Kvasnička and M. Suchánek, J. Chromatogr., B: Anal. Technol. Biomed. Life Sci., 2002, 770, 261–273.
14. J. Schöppenthau, J. Nölte and L. Dunemann, Analyst, 1996, 121, 845–852.
15. S. Yoshida, Agric. Biol. Chem., 1988, 52, 2149–2153.
16. S. Yoshida, Agric. Biol. Chem., 1989, 53, 1071–1075.
17. A. Brandão, H. S. Barbosa and M. A. Z. Arruda, J. Proteomics, 2010, 73, 1433–1440.
18. M. A. Garcia, M. Torre, M. L. Marina and F. Laborda, Crit. Rev. Food Sci. Nutr., 1997, 37, 361–391.
19. A. S. Sandberg, Br. J. Nutr., 2002, 88, S281–S285.
20. S. Ambe, J. Agric. Food Chem., 1994, 42, 262–273.

---