- Email: [email protected]
MR microscopy for noninvasive detection of water distribution during soaking and cooking in the common bean
Magnetic Resonance Imaging 33 (2015) 336–345
Contents lists available at ScienceDirect
Magnetic Resonance Imaging journal homepage: www.mrijournal.com
MR microscopy for noninvasive detection of water distribution during soaking and cooking in the common bean Urša Mikac a,⁎, Ana Sepe a, Igor Serša a, b a b
Jožef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia Department of Biomedical Engineering, Kyung Hee University, 1 Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446–701, Korea
a r t i c l e
i n f o
Article history: Received 10 April 2013 Revised 17 September 2014 Accepted 8 December 2014 Keywords: Common bean (Phaseolus vulgaris) Soaking Cooking Magnetic resonance microscopy
a b s t r a c t Magnetic resonance microscopy (MRM) was used to study water distribution and mobility in common bean (Phaseolus vulgaris) seed during soaking at room temperature (20 °C) and during the cooking of presoaked and dry bean seed in near-boiling water (98 °C). Two complementary MRI methods were used to determine the total water uptake into the seed: the T2-weighted 3D RARE method, which yielded an increased signal from regions of highly mobile (bulk) water and a suppressed signal from regions of poorly mobile (bound) water; and the 3D SPI method, which yielded an increased signal from regions of water restricted in motion and a suppressed signal from the bulk water regions owing to the short repetition time of the method. Based on these results, it can be concluded that during soaking water enters the bean through the micropyle, migrating below the seed coat. The raphe and hypocotyl are hydrated first, while the cotyledon tissue is hydrated next. It was also observed that the imbibition rate increases with an increasing soaking temperature. © 2015 Elsevier Inc. All rights reserved.
1. Introduction A variety of legumes play an important role in the human diet. Of legumes, common beans are widely grown and consumed in various regions of the world. They are high in starch, protein and dietary fiber and are an excellent source of minerals and vitamins [1]. Beans can be dried and stored for a very long time if kept in a cool and dry environment. Most of the legumes used for human food require hydration before cooking to cut down on the amount of cooking time and to reduce the amount of oligosaccharides, which cause flatulence [2]. The hydration of seeds before or during cooking is essential for protein denaturation, starch gelatinization and seed softening. A common procedure for assessing the processing quality of dry beans is to measure their total water uptake, which can be efficiently followed by measuring their weight increase at different hydration times. Different models of hydration introducing various hydration parameters have been proposed based on weight increase [3,4]. The microstructural characteristics of beans during soaking and cooking have been studied by way of a scanning electron microscope [5]. Soaking under different conditions, such as water temperature,
⁎ Corresponding author at: Jožef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia. Tel.: +386 1 477 3314; fax: +386 1 477 31 91. E-mail address: [email protected] (U. Mikac). http://dx.doi.org/10.1016/j.mri.2014.12.001 0730-725X/© 2015 Elsevier Inc. All rights reserved.
relative humidity, and salt concentration, has been studied for a number of bean varieties [6–10]. The effects of atmospheric pressure cooking and high-pressure cooking on the physiochemical and nutritional properties of different bean varieties have been investigated as well [11–13]. In addition, the impact of different storage conditions on bean quality has been examined [14,15]. The various parts of seeds have different water absorption properties. Measuring water distribution in seeds during imbibition and cooking in situ is a demanding task. Several methods have been employed to trace water imbibition in seeds, such as tracer measurement using dyes [16], scanning electron microscopy [17,18], neutron beam [19] and X-ray CT [20]. Magnetic resonance imaging (MRI) provides a very efficient and noninvasive way to follow water distribution in seeds. The method can detect both molecular mobility and localization. In addition, it is fast enough to follow changes during imbibition. MRI has been employed to study water uptake in seeds [21–25] and to follow changes during cooking and baking [26–30]. In our study, MRI was used to measure water distribution inside beans during soaking (hydration) and cooking. Two sets of experiments were performed. In the first set, beans were soaked before cooking, while in the second set they were cooked without presoaking. The water distribution in the beans was followed by MRI using two complementary pulse sequences: the T2-weighted 3D RARE (rapid acquisition with relaxation enhancement) sequence and the 3D SPI (single point imaging) sequence. The T2-weighted 3D RARE sequence enabled the detection of highly mobile water molecules, i.e., those in
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
regions with unrestricted water mobility and therefore relatively long spin–spin (T2) relaxation time. The 3D SPI sequence enabled the detection of bound water regions, i.e., the regions with restricted water mobility and therefore relatively short T2 relaxation time. The combination of both imaging methods enabled an efficient monitoring of water distribution and mobility in the bean seeds during soaking and cooking. The SPI method enabled the detection of bound water at low concentrations at the beginning of imbibition, while the RARE method enabled a more precise determination of water distribution in the seed at later stages, when the amount of water was higher and the T2 relaxation time was longer. 2. Materials and methods 2.1. Beans Fig. 1 shows a schematic presentation of the bean seed structure and a representative T2-weighted MR image in an identical slice. The relevant seed structures are: seed coat, plumule, cotyledon, hilum, micropyle and hypocotyl-radicle axis. A seed coat envelops and protects the embryo, which consists of two cotyledons (food storing structures) that are connected with each other along the radicle (primary root), hypocotyl (shoot system) and plumule (first leaves). The point where the bean is attached to its pod is called the hilum (a scar visible at the bean's exterior). At one end of the hilum is a small pore called the micropyle. Dry common bean seeds (Phaseolus vulgaris), a Slovenian autochthon variety “Savinjski sivček” with weight of 0.35 ± 0.03 g and water contents of 8.5 ± 0.5 %, were used in the study. The water content in the seeds was determined by the gravimetric method. The seeds were weighed first, then dried in an oven at 70 °C until completely dry and finally weighed again. The water content was then determined as the ratio of the water mass loss (the difference between the initial and the final mass of the seed) divided by the mass of the completely dry seed. 2.2. MR imaging of beans during soaking and cooking MRI experiments were performed with an Apollo (TecMag, Houston TX, USA) MRI spectrometer with a superconducting 2.35 T horizontal bore magnet (Oxford Instruments, Abingdon, UK) equipped with micro-imaging accessories (Bruker, Ettlingen, Germany) including radiofrequency (RF) coil inserts of various sizes and micro-imaging gradients with a maximum strength of 250 mT/m. In the MRI experiments, a bean seed was embedded in cotton wool and inserted into a glass tube filled with water. Thus, any eventual bean motion during soaking and cooking was prevented. The tube with the bean seed was then inserted into a 20 mm diameter RF coil, which was in turn inserted into an MR microscopy probe inside the magnet. The probe enabled sample thermoregulation by an air stream of controlled temperature that was blowing on the sample, thus maintaining the sample temperature constant throughout the imaging experiment. Air
PLUMULE
337
temperature was controlled by directing an air stream at a constant flow rate into in a glass Dewar tube containing a heater. The heater power was adjusted dynamically by a temperature controller that received temperature readings from a copper-constantan thermocouple sensor inside the RF coil. Such an experimental scheme allowed for a dynamic following of bean soaking and/or cooking and cooling without removing the bean from the magnet. Two sets of experiments were performed. In the first set, bean seeds were first soaked for 15 h at room temperature, then cooked for 1.5 h and finally cooled in the RF probe for another 3 h. In the second, the seeds were cooked for 3 h (without presoaking) and then cooled for another 3 h. Each type of experiment was repeated three times. The imbibition of water in the bean seed during soaking and cooking was followed by way of sequential MR imaging using the 3D SPI sequence [31,32] and the 3D RARE sequence [33]. The images were acquired every 60 min during the soaking in cold water (approximately 20 °C) and every 20 min during cooking in near-boiling water at 98 ± 1 °C. The 3D RARE images were acquired with an isotropic resolution of 266 μm, imaging matrix 64 × 64 × 64, inter-echo time TE = 1.64 ms, repetition time TR = 2 s, turbo factor 64 and scan time 140 s. The ordering of k-space lines in imaging using the RARE sequence was sequential, thus the corresponding images were T2-weighted. The signal intensity S of the RARE sequence is intricate [33], however, it can be estimated by the expression TEeff TR 1− exp − : SRARE ¼ ρ exp − T2 T1
ð1Þ
Here ρ is the spin density, TR the repetition time, T1 and T2 are the spin–lattice and spin–spin relaxation times, respectively, and TEeff is the effective echo time of the RARE sequence, i.e., the time between the excitation pulse and the echo tagged with a zero phase encode value. The TEeff was 54.12 ms, yielding to T2-weighting for the sample's regions with short T2 values. The signal intensity of the SPI sequence is on the other hand given by [32]
SSPI
3 2 TR 1− exp − 7 tp 6 T 6 1 7 sinφ; ¼ ρ exp − TR 5 T2 4 1−cosφ exp − T1
ð2Þ
where tp is the phase encoding time and φ is the RF excitation pulse flip angle. The spatial resolution of the 3D SPI images was equal to 500 μm isotropic. This is almost half of the resolution obtained with the 3D RARE method. The lower resolution of the SPI method was due to the hardware limitations associated with a limited gradient strength and a need for a short encoding time tp, which was only 0.15 ms. Other parameters of the 3D SPI sequence were: imaging matrix 64 × 64 × 16, RF excitation pulse flip angle φ = 20°, repetition time TR = 5 ms and scan time 330 s. To save imaging
HYPOCOTYLRADICLE AXIS
SEED COAT
PLUMULE SEED COAT
MICROPHYL COTYLEDON
COTYLEDON HILUM
Fig. 1. Bean seed anatomical structures: seed coat, plumule, cotyledon, hilum, micropyle, and hypocotyl-radicle axis. The structures are pointed out by arrows in a drawing (left) and in a representative T2-weighted image in a transverse slice across the bean seed (right).
338
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
time, a short TR was used, yielding to T1-weighting for the sample's regions with long T1 values. In addition, 2D T1 and T2 maps of the bean seeds were measured every 120 min during the soaking in water at room temperature and after the beans were cooked and cooled to room temperature for previously soaked and for un-soaked beans. The T1 map was calculated from images obtained by the spin-echo inversion-recovery (SE-IR) imaging sequence with parameters TE = 4.5 ms, TR = 7 s, and at 6 different inversion times TI varying in the range between 3 ms and 7 s. function for The T1 values were obtained from the h best fit ofthe model i to the experithe SE-IR image signal SðTIÞ ¼ a1 1−b1 exp − TTI1 mental data, where a1 is the proportionality constant and b1 ≅ 2. Similarly, the T2 map was calculated from images obtained by the standard multi-spin-echo (MSE) imaging sequence with the inter-echo time TE = 12 ms, number of echoes N = 60 and TR = 1 s. The T2 values of the map were obtainedfromthe best fit of the model function for the MSE signal SðiÞ ¼ a2 exp − iTE T 2 þ b2 to the experimental data, where i = 1 to N is the echo index and a2 and b2 are two model constants. For T1 and T2 maps, the slice thickness was 2 mm, field of view 17 mm, imaging matrix 64 by 64 and the in-plane resolution 266 μm. 2.3. Water uptake The study of water uptake was performed on ten bean seeds. Five seeds were soaked in water at room temperature (20 °C ± 2 °C) for 15 h. During soaking their weight was measured every hour. For weighing, the seed was taken out of water, wiped of excessive moisture with tissue paper, weighed and then immersed in water again. After the 15-hour soaking, the seeds were cooked for 1.5-hour in near-boiling water (98 °C ± 2 °C). During cooking the weights of the seeds were measured every 15 min. The final measurement of seed weight was performed after the seeds were cooled in water to room temperature. The remaining five seeds were cooked in near-boiling water for 3 h without presoaking. During cooking their weights were measured every 15 min. The final weight measurement was done after the beans were cooled in water to room temperature. The Mitscherlich model [4] was used to analyze water uptake measurements and to determine differences in water uptake at two different temperatures. The model is given by a formula mðt Þ t ¼ 1 þ α 1−β ; m0
ð3Þ
where m0 is a sample mass before soaking, m(t) is its mass after soaking for time t, α is a weight gain at an infinite soaking time, and β is a curve parameter. The value (1 - β) corresponds to the hydration rate, i.e., the larger the value of (1 - β), the faster the bean hydration. 3. Results The T1 and T2 maps as well as the RARE and SPI images of an identical transversal slice across each bean seed after an initial soaking for 2.4 h and after cooking for 1.5 h with 15 h presoaking are shown in Fig. 2a and b, respectively. The corresponding signal intensity profiles along the central line across the seed for three different times of soaking or cooking are provided in Fig. 2c. After 2.4 h of soaking (Fig. 2a), both T1 and T2 were the longest in the void between the cotyledons with the values of T1 ≅ 3000 ms and T2 ≅ 500 ms, indicating that the water in this region was relatively free and highly mobile. As expected, from Eqs. (1) and (2) and known relaxation times, the void region had the highest RARE and the lowest SPI image signal. The relaxation times T1 and T2 decreased significantly from the void region towards the outer edge of the bean causing a disappearance of the RARE signal and an increase in the SPI
signal. In the region below the hilum, where the relaxation times were T1 ≅ 1300 ms and T2 ≅ 32 ms, signals in both RARE and SPI images were high. The situation was quite different after the bean was cooked (Fig. 2b). Here, the void between the cotyledons was much smaller owing to cotyledon swelling. The relaxation times T1 and T2 were still the longest in the void between the cotyledons, but the values of T1 ≅ 1100 ms and T2 ≅ 120 ms were lower than after 2.4 h of soaking, indicating that molecules such as proteins, sugars, polysaccharides etc. were dissolved in the water. In the cotyledon region, the RARE and SPI signals increased after cooking with presoaking owing to an increased amount of water in the region and with it associated increased relaxation times T1 and T2. In the region next to the hilum, where the relaxation times were T1 ≅ 800 ms and T2 ≅ 40 ms, the RARE and SPI signals were still high. If a dry bean seed was cooked for 3 h without presoaking, the distribution of relaxation times was similar to the one obtained in the presoaked cooked bean, and consequently the signal intensities of the RARE and SPI images were also similar. However, the major difference in relaxation times was in the void between the cotyledons, where both T1 and T2 relaxation times of the water in this region were somewhat longer than that of the cooked presoaked bean. This finding indicates that fewer molecules were dissolved in the water of the void region (Fig. 2c). Fig. 2d depicts differences in the time courses of the T2 relaxation times in the void and in the cotyledon region during soaking and after the cooking of dry and presoaked beans. In the void region, the T2 of unbound water decreased during the first 10 h of soaking, while it remained constant during additional 5 h of soaking. The initial T2 decrease can be explained by an increased amount of dissolved substances from the bean tissue. In the cotyledon region, T2 increased during the first 10 h of soaking and remained constant during additional 5 h of soaking. After cooking an additional decrease of T2 was observed in the void region and a slight increase of T2 in the cotyledon region. The corresponding values of T2 for the cooked bean seed without presoaking in comparison to the cooked presoaked bean seed were higher in the void region and practically identical in the cotyledon region (Table 1). Here it should be noted that the T2 values measured by the MR imaging pulse sequence would be underestimated due to a diffusional loss of the signal during the read gradient [34]. 3.1. Soaking and cooking of a common bean seed Fig. 3 shows SPI and RARE images of a bean seed during 15-hour soaking and 1.5-hour cooking in three representative slices in the longitudinal direction across the seed. Two slices have an orientation perpendicular to the cotyledons; one is positioned across the hilum and the other across the center of the seed. The third slice is also positioned across the seed center but in the direction along the cotyledons. At 0.4 h of soaking, water had already entered the seed. At that time, the signal in SPI images was the highest in the micropyle region and was also relatively high in the radicle and hypocotyl regions as well as around the seed. The SPI signal in these regions increased with time, so that at 2.4 h the radicle and hypocotyl regions were the brightest. In addition, the SPI signal intensity increased in the superficial layer, which became thicker due to water penetration through the seed coat in the cotyledon tissue. At 2.4 h, water started to migrate into the cotyledons also from the gap between the cotyledons (Table 1). The void surrounding the hilum appears bright in the RARE images, as it is filled by highly mobile water. From that region, water migrated into the void between the bean coat and the cotyledons, causing changes in the seed coat shape during first 3.4 h of soaking. The thin superficial layer of the seed also gave a bulk water signal in the RARE images; however, it is difficult to distinguish it from the signal of the surrounding water. At 1.4 h of soaking, water filled the gap between the radicle and the cotyledon so that the radicle edge is visible; while
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
(a)
339
(b) T2 map
T1 map
T1 (ms) 2000
T2 (ms) 60
0
T1 map
T2 map
RARE
SPI
0
SPI
RARE
(c)
(d)
RARE (a.u.)
BEAN 60000 40000
T 2 void (soaking at 20
20000
300
o
C)
T2 void cooked un-soaked bean T2 cotyledon cooked pre-soaked bean
2000 0 3000 2000
T2 cotyledon cooked un-soaked bean
200
100
1
T (ms)
C)
T2 void cooked pre-soaked bean
4000
T2 (ms)
SPI (a.u.)
0 6000
o
T 2 cotyledon (soaking at 20
1000
T2 (ms)
0 200
0 5
150 100
10
15
t (h)
50 0
0
2
4
6
8
10
12
profile (mm)
Fig. 2. T1 and T2 maps and RARE and SPI images of an identical transverse slice across a bean seed obtained after the initial soaking for 2.4 h (a) and after cooking for 1.5 h with 15 h presoaking (b). Panel (c) depicts the signal intensity profiles along the central line (red line on the T1 map) for all four methods and for 2.4 h of soaking (red symbols) and cooking with presoaking (1.5 h cooking with 15 h presoaking, black symbols) and without presoaking (3 h cooking, open green symbols). Panel (d) depicts the time courses of the T2 relaxation times in the bean seed void and in the cotyledon during soaking at room temperature and after the cooking of dry and presoaked beans. Note the soaking made a difference in T2 only in the void region but not in the cotyledon region.
at 2.4 h water also started to fill the void between the cotyledons, the void then beginning to shrink after 7.3 h of soaking due to the swelling of the cotyledons.
After the soaking of the bean seed, it was cooked in near-boiling water at 98 °C. At higher temperatures the bulk water signal in the RARE images (Fig. 3) decreased due to an increase in the diffusional
Table 1 Summary of the various bean treatments, used imaging methods, and obtained results. Treatment methods Soaking Follow up
During soaking
During cooking
Results
Hydration
Indication
Released substances Starch gelatinization
Soaking + cooking
Cooking
- SPI, RARE: every 60 min - T1 and T2 maps: every 120 min - Weighing: every 60 min - SPI, RARE: every 20 min - T1 and T2 maps: after cooking and cooling to room T - weighing: every 15 min Slower hydration Faster hydration Same total weight and volume gain Same total weight and volume gain Cotyledon hydration from the outer and inner Cotyledon hydration only from the outer surface surface Intermediate Highest Lowest No Yes Yes
340
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
SPI
RARE
t = 0.4 h t = 1.4 h t = 2.4 h
t = 4.4 h t = 5.3 h
SOAKING
t = 3.4 h
t = 6.3 h
t = 8.3 h t = 10.3 h
t = 16.2 h t = 17.0 h
t = 18.8 h
COOLED
t = 15.5 h
COOKING
t = 14.3 h
Fig. 3. Time series of 3D SPI and 3D RARE images of a bean seed during the 15-hour soaking at room temperature followed by 1.5-hour cooking in near-boiling water; the last set of images was acquired when the seed was cooled back to room temperature. The images are shown in identical three representative slices across the bean seed in a longitudinal orientation; first two are in an orientation perpendicular to the cotyledons (one across the hilum and the other across the center of the seed), and the third is in orientation parallel to the cotyledons. A dark band on the right side of RARE images is an artifact due to the RF field inhomogeneity, while a bright region on the left side of SPI images is a Teflon plug signal.
and convectional motion of the water molecules as the temperature increased. An exception to this was the signal from the thin layer of water next to the seed surface where water molecules were less mobile, resulting in a higher signal. The images also showed a high water signal in the radicle and hypocotyl regions as well as in the gap between the cotyledons. No additional increase in seed volume during cooking was observed. After cooling the seed to room temperature, the free water signal increased, as the mobility of the molecules in the bulk water reduced again. Fig. 4 depicts the time evolution of the signal intensity profiles in the SPI and RARE images acquired during soaking and cooking. The profiles correspond to signals along the central line through the seed oriented perpendicularly to the cotyledons. During soaking, the SPI signal increased due to water migration into the seed, starting in the outer seed regions. At later points in the procedure, the SPI signal also increased in the cotyledon region, and the gap between the
cotyledons started shrinking. Simultaneously with the increase of the water signal in the void between the cotyledons, the signal from the inner side of the cotyledon tissue also increased. The RARE signal was high in the superficial layer of the bean seed. The layer could not be distinguished well from the free water surrounding the seed until the temperature increased, and at which time the RARE signal from the bulk water decreased. The RARE signal in the cotyledon tissue was low at all times and only slightly increased at later soaking and cooking times. After 4.4 h of soaking the RARE signal in the gap between the cotyledons increased owing to the migration of bulk water into that region. 3.2. Cooking of a common bean seed without presoaking Fig. 5 provides the SPI and RARE images of a bean seed during 3-hour cooking without presoaking in three representative slices across the
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
341
(a)
SOAKING
COOKING
signal intensity (a.u)
5x104 4x104 3x104 2x104 1x104
t = 0.5 h t = 1.7 h t = 3.2 h
6x104
signal intensity (a.u)
t = 0.4 h t = 2.4 h t = 4.4 h t = 5.3 h t = 7.3 h t = 9.3 h
6x104
0
5x104 4x104 3x104 2x104 1x104 0
0
5
10
0
5
10
profile (mm)
profile (mm)
SOAKING
COOKING
(b)
signal intensity (a.u)
4x104 3x104 2x104 1x104 0
t = 0.5 h t = 1.7 h t = 3.2 h
4x104
signal intensity (a.u)
t = 0.4 h t = 2.4 h t = 4.4 h t = 5.3 h t = 7.3 h t = 9.3 h
5x104
3x104
2x104
1x104
0 0
5
10
profile (mm)
0
5
10
profile (mm)
Fig. 4. Time series of signal intensity profiles during bean seed soaking and cooking. The profiles correspond to SPI (a) or RARE (b) signals along the line through the center of the bean seed in the direction perpendicular to the cotyledons (red line).
seed identical to those previously presented in Fig. 3. Initially, a high water signal is observed in the SPI images for the micropyle region and for the superficial layer of the seed. Similarly, in the RARE images, a high water signal is initially observed in the thin superficial layer as well as in the area around the hilum. In the initial RARE images, it can also be seen that the seed coat became wrinkled; this effect cannot be observed in SPI images due to the spatial resolution being too low. The SPI images acquired at 0.8 h of cooking have a signal increase to a quite noticeable degree in the radicle and hypocotyl regions; however, in the RARE images the signal increase in these regions is noticeable 0.4 h later (at 1.2 h). At cooking times of 1.2 h or more, the water signal in the RARE images increased in the superficial layer of the seed, which also
became thicker. In addition, a weak RARE signal is observed in the radicle and hypocotyl regions, while the RARE signal is considerably lower in the cotyledon region. The RARE images also show a high signal in some parts of the void between the radicle and the cotyledons. During cooling to room temperature, some water entered the gap between cotyledons and the void between the radicle and the cotyledons. 3.3. Volume and mass increase of a common bean seed during soaking and cooking Fig. 6a depicts the relative mass changes of a bean seed during soaking in water at room temperature (20 °C) and during cooking in
342
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
SPI
RARE
t = 0.5 h
COOKING
t = 0.8 h t = 1.2 h t = 1.4 h
COOLED
t = 3.0 h
t = 4.0 h
Fig. 5. Time series of 3D SPI and 3D RARE images of a bean seed during 3-hour cooking in near-boiling water and when the seed was cooled back to room temperature. The images are of three representative slices across the seed in longitudinal orientation, identical in orientation and position for each method; the first two are in an orientation perpendicular to the cotyledons (one across the hilum and the other across the center of the seed), and the third is in an orientation parallel to the cotyledons.
(a)
m / m0
2.0
1.5
o
T = 20 C o T = 98 C
1.0 0
5
10
15
t (h)
(b)
T = 20°C
T = 98°C 1.6
V(t)/V(0)
1.4 1.2
1.4 1.2
1.0
1.0
3.0
3.0
2.5
2.5
S(t)/S(0)
S(t)/S(0)
V(t)/V(0)
1.6
2.0 1.5 SPI RARE
1.0 0
5
10
t (h)
15
2.0 1.5 SPI RARE
1.0 0
1
2
3
t (h)
Fig. 6. Relative mass m(t)/m0 time dependence (a) and relative volume V(t)/V0 along with relative signal S(t)/S0 time dependences (b) during bean seed soaking at room temperature (20 °C) and during cooking in near-boiling water (98 °C). Relative masses were obtained by sequentially weighing the seed, while relative volumes and signals were obtained from the corresponding SPI and RARE images of the seed during soaking and cooking.
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
near-boiling water (98 °C) without presoaking. Seed soaking resulted in an exponential relative mass m/m0 increase that reached a plateau after 9 h. Seed soaking during cooking in near-boiling water was much faster and reached the plateau just after 2.1 h (Table 1). The data were compared to the Mitscherlich model (Eq.(3)), and the model parameters α and β were determined for soaking at room temperature (20 °C) and at near-boiling temperature (98 °C). The total weight gain is practically identical at both temperatures and yields α = 1.1 ± 0.1; however, the soaking rate is much slower at room temperature (1- β = 0.2 ± 0.05) than at the near-boiling temperature (1- β = 0.7 ± 0.05). Slightly different behavior is observed for the increase of the bean volume during its hydration (Fig. 6b). The bean volume stopped increasing after 7 h of soaking at room temperature and after 1.5 h of cooking. This is in agreement with the SPI signal time dependence: the SPI signal reached a plateau after 7 h of soaking and after 1.5 h of cooking. However, there is a significant delay between the SPI and RARE signal; the RARE signal reaches a plateau after 10.3 h of soaking and after 2.1 h of cooking. The differences among the various bean treatments observed by MRI methods and mass increase are summarized in Table 1.
4. Discussion In the study, changes in a bean seed during soaking and cooking were followed by two complementary MRI methods, RARE and SPI. The RARE method enabled fast image acquisition and was therefore suitable for tracking fast changes during soaking and particularly during cooking. The RARE images were T2-weighted so that they enhanced areas of highly mobile water, while areas of less mobile water gave practically no signal detectable by the RARE method. In contrast, the SPI method enhanced areas of water molecules with restricted mobility. A drawback of the SPI method was slow signal acquisition owing to phase encoding in all three spatial directions. Consequently, the SPI image resolution was reduced, and a fast repetition time was used in order to save imaging time and therefore still enable the tracking of changes in the seed during soaking and cooking. Owing to use of the short repetition time, the bulk water SPI signal was suppressed. SPI imaging is more sensitive than is RARE for the detection of water migrating into a seed, but unfortunately its spatial resolution is insufficient for distinguishing between different anatomical seed structures such as seed coat, the void between the seed coat and cotyledons, and the cotyledon edge. Therefore, it is not possible to determine accurately the migration path of water into the cotyledons; the question remains: did water migrate into the void between the seed coat and cotyledons, through the seed coat, or through the micropyle. The spatial resolution of RARE images is much better and allows a reliable discrimination among the listed seed anatomical structures. The RARE images reveal a high signal of bulk water that entered through the micropyle below the seed coat, suggesting that this water penetrated into the cotyledons. No signal was observed in the cotyledons in the initial RARE images, which implies that the water that penetrated the cotyledons was bound in the cotyledon tissue. This is in accordance with previous observations showing that at the beginning of cotyledon hydration water cannot be stored as free water in it but is absorbed only by the solid matrix of the cotyledon tissue [35]. In the RARE images a water signal was first observed in the gap between the cotyledons at approximately 2.4 h of imbibition. Simultaneously, a strong signal in the inner side of the cotyledons appeared in the SPI images. This supports our hypothesis that water enters the cotyledons immediately after it comes in contact with cotyledon tissue. Identical behavior was observed in kidney and adzuki beans [23], but not in soybeans, where no water penetration in the cotyledons from the void between the cotyledons was observed [24].
343
Our experiments confirmed that water enters the bean seed immediately after its immersion in water. As the initial water migration is very fast it is not possible to determine accurately its entry point and path. Nevertheless, our measurements indicate that most likely water enters the seed through the micropyle and that water penetration through the seed coat is delayed. Fast water uptake was also observed for pinto beans, regarding which it has been proposed that water entered through the hilum/micropyle region and that absorbed water softens the seed coat from inside making it permeable to water [21]. A different imbibition process was observed in kidney and adzuki beans. The water entry into these bean seeds was significantly delayed, and the lens was found to be the water entry point. In the case of kidney beans, the lens was first hydrated for 95 minutes, and water entered the bean after 3 hours of soaking. Regarding the adzuki beans, the lag time before the entry of water was even longer [23]. The MRI measurements of Pietrzak et al. indicated that water enters the soybean seed through the micropyle and the hilum [22]. For the soybean of the Japanese cultivar “Kuromame”, which has a thick seed coat, water uptake was also found to be slow, and it was shown that water entered the seed near the raphe [24]. Our measurements also showed that the seed coat delayed water penetration into the bean. The role of the seed coat in water uptake has been shown to be very important in legume seeds [18,21,24]. At the beginning of soaking, the seed coat delays water uptake in the embryonic axis and governs the direction of water penetration to the embryo. The slow and controlled hydration of a dry seed is essential to preventing soaking injuries of the seed. Eventually, the seed coat serves as a reservoir of water for the hydrating axis. The swelling at higher temperatures follows the same hydration behavior except that the cotyledons are hydrated only from the outer surface (Fig. 5). The reason is the much smaller amount of water that enters into the gap between the cotyledons at 98 °C. At 98 °C, water uptake is much faster than at room temperature, but no significant changes were found in the total weight and volume gain. With regard to water uptake, the same temperature dependence was observed in legume seeds [7,8,36], where the faster water uptake at an increased soaking temperature was attributed to higher water diffusivity resulting from lower fluid viscosity, larger seed pores and the greater water permeability of the seed coat. Our results are in good agreement with the theory that predicts higher water diffusivity at higher temperatures. Namely, the SPI images of the bean soaked at 98 °C showed that 0.8 h of soaking was enough for water to penetrate through the entire seed except the void between the cotyledons; whereas at room temperature, no water was observed in the cotyledons even after 1.4 h of soaking (Figs. 3 and 5). On the other hand, our measurements could not confirm increased water migration through the seed coat at higher temperatures, since the spatial resolution of the SPI images was too low to enable distinguishing between water that migrated through the seed coat into the void between the seed coat and cotyledons and water that migrated into the region through the micropyle. The spatial resolution of the RARE images was twice as good as that of the SPI images, however, their signal was low due to increased diffusional and convectional motions of water with increasing temperature and with it a decrease in the associated RARE image signal. The relaxation times of water measured at different points in the soaking and cooking procedures can provide information on water mobility. A decrease of the T1 and T2 of unbound water in the void between cotyledons during soaking and cooking is associated with an increased amount of dissolved substances in the water. On the other hand, the relaxation time of water in bean tissue is correlated with the amount of water in the tissue, as for example the longer relaxation time of the cotyledon tissue being associated with a higher amount of water in the tissue.
344
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
The time dependence of water T2 in the void between the cotyledons (unbound water) during soaking at room temperature is an indication of bean tissue substances being dissolved. The results in Fig. 2d showed that the T2 of water in the void between the cotyledons decreased during the first 10 h of soaking and remained constant at longer soaking times. An additional decrease was observed after the cooking of the presoaked bean, while the T2 value after the cooking of the dry bean was considerably longer. The T2 changes of unbound water in the void between the cotyledons thus indicate that the total amount of dissolved substances is higher in the case of presoaked than in the case of dry bean cooking. Unfortunately, it is not possible to distinguish among different dissolved substances only on the basis of relaxation time measurements. Further analysis would be needed to confirm differences among the released amounts of dissolved substances and to be able to determine the nature of the dissolved substances. The findings are in agreement with previous studies that have demonstrated that presoaking removes more antinutritient and flatulent substances from bean tissues than solely cooking [37–39]. Therefore, for the most efficient removal of unwanted substances from beans, an optimal combination of soaking and cooking is essential. Soaking reduces certain unwanted heat-stable antinutritient compounds (tannins, raffinose, stachyose, phytates) and the flatulencecausing oligosaccharides, the so-called α-galactosides. Cooking, on the other hand, reduces certain heat-sensitive antinutritients (lectins, protease inhibitors, saponins). Therefore cooking without presoaking removes most of the heat-sensitive antinutritional factors, but the heat-stable and soluble compounds are removed to a limited extent. In addition, soaking shortens cooking time, which is an important factor affecting the protein quality of cooked beans. The correlation between the cotyledon tissue T2 time dependence and time dependence of the total bean mass during soaking in cold water can be observed in Figs. 2d and 6. Both values increased during the first 10 h of soaking and then reached a plateau. Therefore, the increase of T2 during the soaking in cold water can be attributed to the increase in the amount of water in the cotyledons. After cooking, an additional increase of T2 without changes in total bean weight was observed for both presoaked and dry beans (Fig. 2d). The additional increase at higher temperatures can be attributed to starch gelatinization and/or protein denaturation at higher temperatures [40]. As shown previously, T2 increases after starch gelatinization [41] and protein denaturation [40], i.e., the processes associated with heat treatment. A comparison of the graphs in Fig. 6 shows that the seed volume and the SPI signal have approximately identical time dependence and reach a plateau earlier than the seed mass and the RARE signal, which again have approximately identical time dependence. The delay can be explained as water still entering the seed (increasing RARE signal) after the seed stopped expanding (SPI signal plateau). The RARE signal increase that occurred after the SPI signal reached the plateau is not uniform. It is more intense in the cotyledons, which implies that the water that enters the bean after it stopped expanding is mobile, and that it accumulates mainly in the cotyledon region. MRI measurements on kidney beans have also showed that signal intensity does not stop rising when the seed stops expanding [23]. The authors propose two possible explanations for such an effect: one is water replacing gases from small pores in the cells and the second is a molecular structure rearrangement allowing for the retention of more water, which would cause a different physiological stage of water absorption than the previous stages of seed hydration. Our measurements are in a better agreement with the second option. 5. Conclusions Two complementary 3D MRI methods, RARE and SPI, were used for tracking water hydration during the soaking and cooking of
common bean seed. The RARE method enabled the detection of free water and excelled in imaging resolution and signal acquisition speed; while the SPI method enabled the detection of bound water and had relatively low resolution and slow signal acquisition. By combining information from the two imaging methods it was possible to track the mobile and bound water that penetrated the bean seed and to determine the hydration dynamics of different anatomical parts of the seed. The study also demonstrated the great potential of MRI to study water imbibition processes in various seeds as also in many other water susceptible materials.
References [1] Molina MR, Bressani R, Elias LG. Nonconventional legume grains as proteinsources. Food Technol 1977;31:188–90. [2] Jood S, Mehta U, Singh R, Bhat CM. Effect of processing on flatus-producing factors in legumes. J Agric Food Chem 1985;33:268–71. [3] Deshpande SS, Cheryan M. Water-uptake during cooking of dry beans (Phaseolus- vulgaris L.). Plant Food Hum Nutr 1986;36:157–65. [4] Wood JA, Harden S. A method to estimate the hydration and swelling properties of chickpeas (Cicer arietinum L.). J Food Sci 2006;71:E190–5. [5] Berg T, Singh J, Hardacre A, Boland MJ. The role of cotyledon cell structure during in vitro digestion of starch in navy beans. Carbohydr Polym 2012;87:1678–88. [6] Kon S. Effect of soaking temperature on cooking and nutritional quality of beans. J Food Sci 1979;44:1329–35. [7] Hsu KH, Kim CJ, Wilson LA. Factors affecting water-uptake of soybeans during soaking. Cereal Chem 1983;60:208–11. [8] Joshi M, Adhikari B, Panozzo J, Aldred P. Water uptake and its impact on the texture of lentils (Lens culinaris). J Food Eng 2010;100:61–9. [9] Tang JM, Sokhansanj S, Sosulski FW. Moisture-absorption characteristics of laird lentils and hardshell seeds. Cereal Chem 1994;71:423–9. [10] Mnkeni AP, Gierschner K, Maeda EE. Effect of blanching time and salt concentration on pectolytic enzymes, texture and acceptability of fermented green beans. Plant Food Hum Nutr 1999;53:285–96. [11] Landa-Habana L, Pina-Hernandez A, Agama-Acevedo E, Tovar J, Bello-Perez LA. Effect of cooking procedures and storage on starch bioavailability in common beans (Phaseolus vulgaris L.). Plant Food Hum Nutr 2004;59:133–6. [12] Güzel D, Sayar S. Effect of cooking methods on selected physicochemical and nutritional properties of barlotto bean, chickpea, faba bean, and white kidney bean. J Food Sci Technol 2012;49:89–95. [13] Iyer V, Salunkhe DK, Sathe SK, Rockland LB. Quick-cooking beans (Phaseolus vulgaris L.): II. Phytates, oligosaccharides, and antienzymes. Plant Food Hum Nutr 1980;30:45–52. [14] Yousif AM, Deeth HC, Caffin NA, Lisle AT. Effect of storage time and conditions on the hardness and cooking quality of adzuki (Vigna angularis). Lebensm Wiss U Technol 2002;35:338–43. [15] Yousif AM, Batey IL, Larroque OR, Curtin B, Bekes F, Deeth HC. Effect of storage of adzuki bean (Vigna angularis) on starch and protein properties. Lebensm Wiss U Technol 2003;36:601–7. [16] Collins EJ. The structure of the integumentary system of the barley grain in relation to localized water absorption and semi-permeability. Ann Bot 1918;32: 381–414. [17] McDonald MB, Vertucci CW, Roos EE. Soybean seed imbibition: water absorption by seed parts. Crop Sci 1988;28:993–7. [18] McDonald MB, Vertucci CW, Roos EE. Seed coat regulation of soybean seed imbibition. Crop Sci 1988;28:987–92. [19] Nakanishi TM, Matsubayashi M. Nondestructive water imaging by neutron beam analysis in living plants. J Plant Physiol 1997;151:442–5. [20] Lammertyn J, Dresselaers T, Van Hecke P, Jancsok P, Wevers M, Nicolai BM. MRI and X-ray CT study of spatial distribution of core breakdown in ‘Conference’ pears. Magn Reson Imaging 2003;21:805–15. [21] Heil JR, McCarthy MJ, Özilgen M. Magnetic resonance imaging and modeling of water-uptake into dry beans. Lebensm Wiss U Technol 1992;25:280–5. [22] Pietrzak LN, Frégeau-Reid J, Chatson B, Blackwell B. Observations on water distribution in soybean seed during hydration processes using nuclear magnetic resonance imaging. Can J Plant Sci 2002;82:513–9. [23] Kikuchi K, Koizumi M, Ishida N, Kano H. Water uptake by dry beans observed by micro-magnetic resonance imaging. Ann Bot 2006;98:545–53. [24] Koizumi M, Kikuchi K, Isobe S, Ishida N, Naito S, Kano H. Role of seed coat in imbibing soybean seeds observed by micro-magnetic resonance imaging. Ann Bot 2008;102:343–52. [25] Hong YS, Cho JH, Kim NR, Lee C, Cheong C, Hong KS, et al. Artifacts in the measurement of water distribution in soybeans using MR imaging. Food Chem 2009;112:267–72. [26] Mohorič A, Vergeldt F, Gerkema E, De Jager A, Van Duynhoven J, Van Dalen G, et al. Magnetic resonance imaging of single rice kernels during cooking. J Magn Reson 2004;171:157–62. [27] Mortensen M, Thybo AK, Bertram HC, Andersen HJ, Engelsen SB. Cooking effects on water distribution in potatoes using nuclear magnetic resonance relaxation. J Agric Food Chem 2005;53:5976–81.
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345 [28] Hong SW, Yan ZY, Otterburn MS, McCarthy MJ. Magnetic Resonance Imaging (MRI) of a cookie in comparison with time-lapse photographic analysis (TLPA) during baking process. Magn Reson Imaging 1996;14:923–7. [29] Takano H, Naito S, Ishida N, Koizumi M, Kano H. Fermentation process and grain structure of baked breads from frozen dough using freeze-tolerant yeasts. J Food Sci 2002;67:2725–33. [30] Bajd F, Sersa I. Continuous monitoring of dough fermentation and bread baking by magnetic resonance microscopy. Magn Reson Imaging 2011;29:434–42. [31] Emid S, Creyghton JHN. High resolution NMR imaging in solids. Physica B 1985;128: 81–3. [32] Gravina S, Cory DG. Sensitivity and resolution of constant-time imaging. J Magn Reson B 1994;104:53–61. [33] Hennig J, Nauerth A, Friedburg H. RARE imaging: a fast imaging method for clinical MR. Magn Reson Med 1986;3:823–33. [34] Mansfield P, Grannell PK. Diffraction and microscopy in solids and liquids by NMR. Phys Rev B 1975;12:3618–34. [35] Golonka P, Dryzek J, Kluza M. Bean cotyledon microporosity under hydration conditions. Nukleonika 2002;47:137–40.
345
[36] Oliveira AL, Colnaghi BG, da Silva EZ, Gouvea IR, Vieira RL, Augusto PED. Modelling the effect of temperature on the hydration kinetic of adzuki beans (Vigna angularis). J Food Eng 2013;118:417–20. [37] Khokhar S, Frias J, Price KR, Fenwick GR, Hedley CL. Psysico-chemical characteristics of Khesari dhal (Lathyrus sativus): changes in α-galactosides, monosaccharides and disaccharides during food processing. J Sci Food Agric 1996;70:487–92. [38] Shimelis EA, Rakshit SK. Effect of processing on antinutrients and in vitro protein digestibility of kidney bean (Phaseolus vulgaris L.) varieties grown in East Africa. Food Chem 2007;103:161–72. [39] Fernandes AC, Nishida W, da Costa Proenca RP. Influence of soaking on the nutritional quality of common beans (Phaseolus vulgaris L.) cooked with or without the soaking water: a review. Int J Food Sci Tech 2010;45:2209–18. [40] Ahmad MU, Tashiro Y, Matsukawa S, Ogawa H. Gelation mechanism of surimi studied by 1H NMR relaxation measurements. J Food Sci 2007;72: E362–7. [41] Tananuwong K, Reid DS. DSC and NMR relaxation studies of starch-water interactions during gelatinization. Carbohydr Polym 2004;58:345–58.
Contents lists available at ScienceDirect
Magnetic Resonance Imaging journal homepage: www.mrijournal.com
MR microscopy for noninvasive detection of water distribution during soaking and cooking in the common bean Urša Mikac a,⁎, Ana Sepe a, Igor Serša a, b a b
Jožef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia Department of Biomedical Engineering, Kyung Hee University, 1 Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446–701, Korea
a r t i c l e
i n f o
Article history: Received 10 April 2013 Revised 17 September 2014 Accepted 8 December 2014 Keywords: Common bean (Phaseolus vulgaris) Soaking Cooking Magnetic resonance microscopy
a b s t r a c t Magnetic resonance microscopy (MRM) was used to study water distribution and mobility in common bean (Phaseolus vulgaris) seed during soaking at room temperature (20 °C) and during the cooking of presoaked and dry bean seed in near-boiling water (98 °C). Two complementary MRI methods were used to determine the total water uptake into the seed: the T2-weighted 3D RARE method, which yielded an increased signal from regions of highly mobile (bulk) water and a suppressed signal from regions of poorly mobile (bound) water; and the 3D SPI method, which yielded an increased signal from regions of water restricted in motion and a suppressed signal from the bulk water regions owing to the short repetition time of the method. Based on these results, it can be concluded that during soaking water enters the bean through the micropyle, migrating below the seed coat. The raphe and hypocotyl are hydrated first, while the cotyledon tissue is hydrated next. It was also observed that the imbibition rate increases with an increasing soaking temperature. © 2015 Elsevier Inc. All rights reserved.
1. Introduction A variety of legumes play an important role in the human diet. Of legumes, common beans are widely grown and consumed in various regions of the world. They are high in starch, protein and dietary fiber and are an excellent source of minerals and vitamins [1]. Beans can be dried and stored for a very long time if kept in a cool and dry environment. Most of the legumes used for human food require hydration before cooking to cut down on the amount of cooking time and to reduce the amount of oligosaccharides, which cause flatulence [2]. The hydration of seeds before or during cooking is essential for protein denaturation, starch gelatinization and seed softening. A common procedure for assessing the processing quality of dry beans is to measure their total water uptake, which can be efficiently followed by measuring their weight increase at different hydration times. Different models of hydration introducing various hydration parameters have been proposed based on weight increase [3,4]. The microstructural characteristics of beans during soaking and cooking have been studied by way of a scanning electron microscope [5]. Soaking under different conditions, such as water temperature,
⁎ Corresponding author at: Jožef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia. Tel.: +386 1 477 3314; fax: +386 1 477 31 91. E-mail address: [email protected] (U. Mikac). http://dx.doi.org/10.1016/j.mri.2014.12.001 0730-725X/© 2015 Elsevier Inc. All rights reserved.
relative humidity, and salt concentration, has been studied for a number of bean varieties [6–10]. The effects of atmospheric pressure cooking and high-pressure cooking on the physiochemical and nutritional properties of different bean varieties have been investigated as well [11–13]. In addition, the impact of different storage conditions on bean quality has been examined [14,15]. The various parts of seeds have different water absorption properties. Measuring water distribution in seeds during imbibition and cooking in situ is a demanding task. Several methods have been employed to trace water imbibition in seeds, such as tracer measurement using dyes [16], scanning electron microscopy [17,18], neutron beam [19] and X-ray CT [20]. Magnetic resonance imaging (MRI) provides a very efficient and noninvasive way to follow water distribution in seeds. The method can detect both molecular mobility and localization. In addition, it is fast enough to follow changes during imbibition. MRI has been employed to study water uptake in seeds [21–25] and to follow changes during cooking and baking [26–30]. In our study, MRI was used to measure water distribution inside beans during soaking (hydration) and cooking. Two sets of experiments were performed. In the first set, beans were soaked before cooking, while in the second set they were cooked without presoaking. The water distribution in the beans was followed by MRI using two complementary pulse sequences: the T2-weighted 3D RARE (rapid acquisition with relaxation enhancement) sequence and the 3D SPI (single point imaging) sequence. The T2-weighted 3D RARE sequence enabled the detection of highly mobile water molecules, i.e., those in
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
regions with unrestricted water mobility and therefore relatively long spin–spin (T2) relaxation time. The 3D SPI sequence enabled the detection of bound water regions, i.e., the regions with restricted water mobility and therefore relatively short T2 relaxation time. The combination of both imaging methods enabled an efficient monitoring of water distribution and mobility in the bean seeds during soaking and cooking. The SPI method enabled the detection of bound water at low concentrations at the beginning of imbibition, while the RARE method enabled a more precise determination of water distribution in the seed at later stages, when the amount of water was higher and the T2 relaxation time was longer. 2. Materials and methods 2.1. Beans Fig. 1 shows a schematic presentation of the bean seed structure and a representative T2-weighted MR image in an identical slice. The relevant seed structures are: seed coat, plumule, cotyledon, hilum, micropyle and hypocotyl-radicle axis. A seed coat envelops and protects the embryo, which consists of two cotyledons (food storing structures) that are connected with each other along the radicle (primary root), hypocotyl (shoot system) and plumule (first leaves). The point where the bean is attached to its pod is called the hilum (a scar visible at the bean's exterior). At one end of the hilum is a small pore called the micropyle. Dry common bean seeds (Phaseolus vulgaris), a Slovenian autochthon variety “Savinjski sivček” with weight of 0.35 ± 0.03 g and water contents of 8.5 ± 0.5 %, were used in the study. The water content in the seeds was determined by the gravimetric method. The seeds were weighed first, then dried in an oven at 70 °C until completely dry and finally weighed again. The water content was then determined as the ratio of the water mass loss (the difference between the initial and the final mass of the seed) divided by the mass of the completely dry seed. 2.2. MR imaging of beans during soaking and cooking MRI experiments were performed with an Apollo (TecMag, Houston TX, USA) MRI spectrometer with a superconducting 2.35 T horizontal bore magnet (Oxford Instruments, Abingdon, UK) equipped with micro-imaging accessories (Bruker, Ettlingen, Germany) including radiofrequency (RF) coil inserts of various sizes and micro-imaging gradients with a maximum strength of 250 mT/m. In the MRI experiments, a bean seed was embedded in cotton wool and inserted into a glass tube filled with water. Thus, any eventual bean motion during soaking and cooking was prevented. The tube with the bean seed was then inserted into a 20 mm diameter RF coil, which was in turn inserted into an MR microscopy probe inside the magnet. The probe enabled sample thermoregulation by an air stream of controlled temperature that was blowing on the sample, thus maintaining the sample temperature constant throughout the imaging experiment. Air
PLUMULE
337
temperature was controlled by directing an air stream at a constant flow rate into in a glass Dewar tube containing a heater. The heater power was adjusted dynamically by a temperature controller that received temperature readings from a copper-constantan thermocouple sensor inside the RF coil. Such an experimental scheme allowed for a dynamic following of bean soaking and/or cooking and cooling without removing the bean from the magnet. Two sets of experiments were performed. In the first set, bean seeds were first soaked for 15 h at room temperature, then cooked for 1.5 h and finally cooled in the RF probe for another 3 h. In the second, the seeds were cooked for 3 h (without presoaking) and then cooled for another 3 h. Each type of experiment was repeated three times. The imbibition of water in the bean seed during soaking and cooking was followed by way of sequential MR imaging using the 3D SPI sequence [31,32] and the 3D RARE sequence [33]. The images were acquired every 60 min during the soaking in cold water (approximately 20 °C) and every 20 min during cooking in near-boiling water at 98 ± 1 °C. The 3D RARE images were acquired with an isotropic resolution of 266 μm, imaging matrix 64 × 64 × 64, inter-echo time TE = 1.64 ms, repetition time TR = 2 s, turbo factor 64 and scan time 140 s. The ordering of k-space lines in imaging using the RARE sequence was sequential, thus the corresponding images were T2-weighted. The signal intensity S of the RARE sequence is intricate [33], however, it can be estimated by the expression TEeff TR 1− exp − : SRARE ¼ ρ exp − T2 T1
ð1Þ
Here ρ is the spin density, TR the repetition time, T1 and T2 are the spin–lattice and spin–spin relaxation times, respectively, and TEeff is the effective echo time of the RARE sequence, i.e., the time between the excitation pulse and the echo tagged with a zero phase encode value. The TEeff was 54.12 ms, yielding to T2-weighting for the sample's regions with short T2 values. The signal intensity of the SPI sequence is on the other hand given by [32]
SSPI
3 2 TR 1− exp − 7 tp 6 T 6 1 7 sinφ; ¼ ρ exp − TR 5 T2 4 1−cosφ exp − T1
ð2Þ
where tp is the phase encoding time and φ is the RF excitation pulse flip angle. The spatial resolution of the 3D SPI images was equal to 500 μm isotropic. This is almost half of the resolution obtained with the 3D RARE method. The lower resolution of the SPI method was due to the hardware limitations associated with a limited gradient strength and a need for a short encoding time tp, which was only 0.15 ms. Other parameters of the 3D SPI sequence were: imaging matrix 64 × 64 × 16, RF excitation pulse flip angle φ = 20°, repetition time TR = 5 ms and scan time 330 s. To save imaging
HYPOCOTYLRADICLE AXIS
SEED COAT
PLUMULE SEED COAT
MICROPHYL COTYLEDON
COTYLEDON HILUM
Fig. 1. Bean seed anatomical structures: seed coat, plumule, cotyledon, hilum, micropyle, and hypocotyl-radicle axis. The structures are pointed out by arrows in a drawing (left) and in a representative T2-weighted image in a transverse slice across the bean seed (right).
338
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
time, a short TR was used, yielding to T1-weighting for the sample's regions with long T1 values. In addition, 2D T1 and T2 maps of the bean seeds were measured every 120 min during the soaking in water at room temperature and after the beans were cooked and cooled to room temperature for previously soaked and for un-soaked beans. The T1 map was calculated from images obtained by the spin-echo inversion-recovery (SE-IR) imaging sequence with parameters TE = 4.5 ms, TR = 7 s, and at 6 different inversion times TI varying in the range between 3 ms and 7 s. function for The T1 values were obtained from the h best fit ofthe model i to the experithe SE-IR image signal SðTIÞ ¼ a1 1−b1 exp − TTI1 mental data, where a1 is the proportionality constant and b1 ≅ 2. Similarly, the T2 map was calculated from images obtained by the standard multi-spin-echo (MSE) imaging sequence with the inter-echo time TE = 12 ms, number of echoes N = 60 and TR = 1 s. The T2 values of the map were obtainedfromthe best fit of the model function for the MSE signal SðiÞ ¼ a2 exp − iTE T 2 þ b2 to the experimental data, where i = 1 to N is the echo index and a2 and b2 are two model constants. For T1 and T2 maps, the slice thickness was 2 mm, field of view 17 mm, imaging matrix 64 by 64 and the in-plane resolution 266 μm. 2.3. Water uptake The study of water uptake was performed on ten bean seeds. Five seeds were soaked in water at room temperature (20 °C ± 2 °C) for 15 h. During soaking their weight was measured every hour. For weighing, the seed was taken out of water, wiped of excessive moisture with tissue paper, weighed and then immersed in water again. After the 15-hour soaking, the seeds were cooked for 1.5-hour in near-boiling water (98 °C ± 2 °C). During cooking the weights of the seeds were measured every 15 min. The final measurement of seed weight was performed after the seeds were cooled in water to room temperature. The remaining five seeds were cooked in near-boiling water for 3 h without presoaking. During cooking their weights were measured every 15 min. The final weight measurement was done after the beans were cooled in water to room temperature. The Mitscherlich model [4] was used to analyze water uptake measurements and to determine differences in water uptake at two different temperatures. The model is given by a formula mðt Þ t ¼ 1 þ α 1−β ; m0
ð3Þ
where m0 is a sample mass before soaking, m(t) is its mass after soaking for time t, α is a weight gain at an infinite soaking time, and β is a curve parameter. The value (1 - β) corresponds to the hydration rate, i.e., the larger the value of (1 - β), the faster the bean hydration. 3. Results The T1 and T2 maps as well as the RARE and SPI images of an identical transversal slice across each bean seed after an initial soaking for 2.4 h and after cooking for 1.5 h with 15 h presoaking are shown in Fig. 2a and b, respectively. The corresponding signal intensity profiles along the central line across the seed for three different times of soaking or cooking are provided in Fig. 2c. After 2.4 h of soaking (Fig. 2a), both T1 and T2 were the longest in the void between the cotyledons with the values of T1 ≅ 3000 ms and T2 ≅ 500 ms, indicating that the water in this region was relatively free and highly mobile. As expected, from Eqs. (1) and (2) and known relaxation times, the void region had the highest RARE and the lowest SPI image signal. The relaxation times T1 and T2 decreased significantly from the void region towards the outer edge of the bean causing a disappearance of the RARE signal and an increase in the SPI
signal. In the region below the hilum, where the relaxation times were T1 ≅ 1300 ms and T2 ≅ 32 ms, signals in both RARE and SPI images were high. The situation was quite different after the bean was cooked (Fig. 2b). Here, the void between the cotyledons was much smaller owing to cotyledon swelling. The relaxation times T1 and T2 were still the longest in the void between the cotyledons, but the values of T1 ≅ 1100 ms and T2 ≅ 120 ms were lower than after 2.4 h of soaking, indicating that molecules such as proteins, sugars, polysaccharides etc. were dissolved in the water. In the cotyledon region, the RARE and SPI signals increased after cooking with presoaking owing to an increased amount of water in the region and with it associated increased relaxation times T1 and T2. In the region next to the hilum, where the relaxation times were T1 ≅ 800 ms and T2 ≅ 40 ms, the RARE and SPI signals were still high. If a dry bean seed was cooked for 3 h without presoaking, the distribution of relaxation times was similar to the one obtained in the presoaked cooked bean, and consequently the signal intensities of the RARE and SPI images were also similar. However, the major difference in relaxation times was in the void between the cotyledons, where both T1 and T2 relaxation times of the water in this region were somewhat longer than that of the cooked presoaked bean. This finding indicates that fewer molecules were dissolved in the water of the void region (Fig. 2c). Fig. 2d depicts differences in the time courses of the T2 relaxation times in the void and in the cotyledon region during soaking and after the cooking of dry and presoaked beans. In the void region, the T2 of unbound water decreased during the first 10 h of soaking, while it remained constant during additional 5 h of soaking. The initial T2 decrease can be explained by an increased amount of dissolved substances from the bean tissue. In the cotyledon region, T2 increased during the first 10 h of soaking and remained constant during additional 5 h of soaking. After cooking an additional decrease of T2 was observed in the void region and a slight increase of T2 in the cotyledon region. The corresponding values of T2 for the cooked bean seed without presoaking in comparison to the cooked presoaked bean seed were higher in the void region and practically identical in the cotyledon region (Table 1). Here it should be noted that the T2 values measured by the MR imaging pulse sequence would be underestimated due to a diffusional loss of the signal during the read gradient [34]. 3.1. Soaking and cooking of a common bean seed Fig. 3 shows SPI and RARE images of a bean seed during 15-hour soaking and 1.5-hour cooking in three representative slices in the longitudinal direction across the seed. Two slices have an orientation perpendicular to the cotyledons; one is positioned across the hilum and the other across the center of the seed. The third slice is also positioned across the seed center but in the direction along the cotyledons. At 0.4 h of soaking, water had already entered the seed. At that time, the signal in SPI images was the highest in the micropyle region and was also relatively high in the radicle and hypocotyl regions as well as around the seed. The SPI signal in these regions increased with time, so that at 2.4 h the radicle and hypocotyl regions were the brightest. In addition, the SPI signal intensity increased in the superficial layer, which became thicker due to water penetration through the seed coat in the cotyledon tissue. At 2.4 h, water started to migrate into the cotyledons also from the gap between the cotyledons (Table 1). The void surrounding the hilum appears bright in the RARE images, as it is filled by highly mobile water. From that region, water migrated into the void between the bean coat and the cotyledons, causing changes in the seed coat shape during first 3.4 h of soaking. The thin superficial layer of the seed also gave a bulk water signal in the RARE images; however, it is difficult to distinguish it from the signal of the surrounding water. At 1.4 h of soaking, water filled the gap between the radicle and the cotyledon so that the radicle edge is visible; while
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
(a)
339
(b) T2 map
T1 map
T1 (ms) 2000
T2 (ms) 60
0
T1 map
T2 map
RARE
SPI
0
SPI
RARE
(c)
(d)
RARE (a.u.)
BEAN 60000 40000
T 2 void (soaking at 20
20000
300
o
C)
T2 void cooked un-soaked bean T2 cotyledon cooked pre-soaked bean
2000 0 3000 2000
T2 cotyledon cooked un-soaked bean
200
100
1
T (ms)
C)
T2 void cooked pre-soaked bean
4000
T2 (ms)
SPI (a.u.)
0 6000
o
T 2 cotyledon (soaking at 20
1000
T2 (ms)
0 200
0 5
150 100
10
15
t (h)
50 0
0
2
4
6
8
10
12
profile (mm)
Fig. 2. T1 and T2 maps and RARE and SPI images of an identical transverse slice across a bean seed obtained after the initial soaking for 2.4 h (a) and after cooking for 1.5 h with 15 h presoaking (b). Panel (c) depicts the signal intensity profiles along the central line (red line on the T1 map) for all four methods and for 2.4 h of soaking (red symbols) and cooking with presoaking (1.5 h cooking with 15 h presoaking, black symbols) and without presoaking (3 h cooking, open green symbols). Panel (d) depicts the time courses of the T2 relaxation times in the bean seed void and in the cotyledon during soaking at room temperature and after the cooking of dry and presoaked beans. Note the soaking made a difference in T2 only in the void region but not in the cotyledon region.
at 2.4 h water also started to fill the void between the cotyledons, the void then beginning to shrink after 7.3 h of soaking due to the swelling of the cotyledons.
After the soaking of the bean seed, it was cooked in near-boiling water at 98 °C. At higher temperatures the bulk water signal in the RARE images (Fig. 3) decreased due to an increase in the diffusional
Table 1 Summary of the various bean treatments, used imaging methods, and obtained results. Treatment methods Soaking Follow up
During soaking
During cooking
Results
Hydration
Indication
Released substances Starch gelatinization
Soaking + cooking
Cooking
- SPI, RARE: every 60 min - T1 and T2 maps: every 120 min - Weighing: every 60 min - SPI, RARE: every 20 min - T1 and T2 maps: after cooking and cooling to room T - weighing: every 15 min Slower hydration Faster hydration Same total weight and volume gain Same total weight and volume gain Cotyledon hydration from the outer and inner Cotyledon hydration only from the outer surface surface Intermediate Highest Lowest No Yes Yes
340
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
SPI
RARE
t = 0.4 h t = 1.4 h t = 2.4 h
t = 4.4 h t = 5.3 h
SOAKING
t = 3.4 h
t = 6.3 h
t = 8.3 h t = 10.3 h
t = 16.2 h t = 17.0 h
t = 18.8 h
COOLED
t = 15.5 h
COOKING
t = 14.3 h
Fig. 3. Time series of 3D SPI and 3D RARE images of a bean seed during the 15-hour soaking at room temperature followed by 1.5-hour cooking in near-boiling water; the last set of images was acquired when the seed was cooled back to room temperature. The images are shown in identical three representative slices across the bean seed in a longitudinal orientation; first two are in an orientation perpendicular to the cotyledons (one across the hilum and the other across the center of the seed), and the third is in orientation parallel to the cotyledons. A dark band on the right side of RARE images is an artifact due to the RF field inhomogeneity, while a bright region on the left side of SPI images is a Teflon plug signal.
and convectional motion of the water molecules as the temperature increased. An exception to this was the signal from the thin layer of water next to the seed surface where water molecules were less mobile, resulting in a higher signal. The images also showed a high water signal in the radicle and hypocotyl regions as well as in the gap between the cotyledons. No additional increase in seed volume during cooking was observed. After cooling the seed to room temperature, the free water signal increased, as the mobility of the molecules in the bulk water reduced again. Fig. 4 depicts the time evolution of the signal intensity profiles in the SPI and RARE images acquired during soaking and cooking. The profiles correspond to signals along the central line through the seed oriented perpendicularly to the cotyledons. During soaking, the SPI signal increased due to water migration into the seed, starting in the outer seed regions. At later points in the procedure, the SPI signal also increased in the cotyledon region, and the gap between the
cotyledons started shrinking. Simultaneously with the increase of the water signal in the void between the cotyledons, the signal from the inner side of the cotyledon tissue also increased. The RARE signal was high in the superficial layer of the bean seed. The layer could not be distinguished well from the free water surrounding the seed until the temperature increased, and at which time the RARE signal from the bulk water decreased. The RARE signal in the cotyledon tissue was low at all times and only slightly increased at later soaking and cooking times. After 4.4 h of soaking the RARE signal in the gap between the cotyledons increased owing to the migration of bulk water into that region. 3.2. Cooking of a common bean seed without presoaking Fig. 5 provides the SPI and RARE images of a bean seed during 3-hour cooking without presoaking in three representative slices across the
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
341
(a)
SOAKING
COOKING
signal intensity (a.u)
5x104 4x104 3x104 2x104 1x104
t = 0.5 h t = 1.7 h t = 3.2 h
6x104
signal intensity (a.u)
t = 0.4 h t = 2.4 h t = 4.4 h t = 5.3 h t = 7.3 h t = 9.3 h
6x104
0
5x104 4x104 3x104 2x104 1x104 0
0
5
10
0
5
10
profile (mm)
profile (mm)
SOAKING
COOKING
(b)
signal intensity (a.u)
4x104 3x104 2x104 1x104 0
t = 0.5 h t = 1.7 h t = 3.2 h
4x104
signal intensity (a.u)
t = 0.4 h t = 2.4 h t = 4.4 h t = 5.3 h t = 7.3 h t = 9.3 h
5x104
3x104
2x104
1x104
0 0
5
10
profile (mm)
0
5
10
profile (mm)
Fig. 4. Time series of signal intensity profiles during bean seed soaking and cooking. The profiles correspond to SPI (a) or RARE (b) signals along the line through the center of the bean seed in the direction perpendicular to the cotyledons (red line).
seed identical to those previously presented in Fig. 3. Initially, a high water signal is observed in the SPI images for the micropyle region and for the superficial layer of the seed. Similarly, in the RARE images, a high water signal is initially observed in the thin superficial layer as well as in the area around the hilum. In the initial RARE images, it can also be seen that the seed coat became wrinkled; this effect cannot be observed in SPI images due to the spatial resolution being too low. The SPI images acquired at 0.8 h of cooking have a signal increase to a quite noticeable degree in the radicle and hypocotyl regions; however, in the RARE images the signal increase in these regions is noticeable 0.4 h later (at 1.2 h). At cooking times of 1.2 h or more, the water signal in the RARE images increased in the superficial layer of the seed, which also
became thicker. In addition, a weak RARE signal is observed in the radicle and hypocotyl regions, while the RARE signal is considerably lower in the cotyledon region. The RARE images also show a high signal in some parts of the void between the radicle and the cotyledons. During cooling to room temperature, some water entered the gap between cotyledons and the void between the radicle and the cotyledons. 3.3. Volume and mass increase of a common bean seed during soaking and cooking Fig. 6a depicts the relative mass changes of a bean seed during soaking in water at room temperature (20 °C) and during cooking in
342
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
SPI
RARE
t = 0.5 h
COOKING
t = 0.8 h t = 1.2 h t = 1.4 h
COOLED
t = 3.0 h
t = 4.0 h
Fig. 5. Time series of 3D SPI and 3D RARE images of a bean seed during 3-hour cooking in near-boiling water and when the seed was cooled back to room temperature. The images are of three representative slices across the seed in longitudinal orientation, identical in orientation and position for each method; the first two are in an orientation perpendicular to the cotyledons (one across the hilum and the other across the center of the seed), and the third is in an orientation parallel to the cotyledons.
(a)
m / m0
2.0
1.5
o
T = 20 C o T = 98 C
1.0 0
5
10
15
t (h)
(b)
T = 20°C
T = 98°C 1.6
V(t)/V(0)
1.4 1.2
1.4 1.2
1.0
1.0
3.0
3.0
2.5
2.5
S(t)/S(0)
S(t)/S(0)
V(t)/V(0)
1.6
2.0 1.5 SPI RARE
1.0 0
5
10
t (h)
15
2.0 1.5 SPI RARE
1.0 0
1
2
3
t (h)
Fig. 6. Relative mass m(t)/m0 time dependence (a) and relative volume V(t)/V0 along with relative signal S(t)/S0 time dependences (b) during bean seed soaking at room temperature (20 °C) and during cooking in near-boiling water (98 °C). Relative masses were obtained by sequentially weighing the seed, while relative volumes and signals were obtained from the corresponding SPI and RARE images of the seed during soaking and cooking.
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
near-boiling water (98 °C) without presoaking. Seed soaking resulted in an exponential relative mass m/m0 increase that reached a plateau after 9 h. Seed soaking during cooking in near-boiling water was much faster and reached the plateau just after 2.1 h (Table 1). The data were compared to the Mitscherlich model (Eq.(3)), and the model parameters α and β were determined for soaking at room temperature (20 °C) and at near-boiling temperature (98 °C). The total weight gain is practically identical at both temperatures and yields α = 1.1 ± 0.1; however, the soaking rate is much slower at room temperature (1- β = 0.2 ± 0.05) than at the near-boiling temperature (1- β = 0.7 ± 0.05). Slightly different behavior is observed for the increase of the bean volume during its hydration (Fig. 6b). The bean volume stopped increasing after 7 h of soaking at room temperature and after 1.5 h of cooking. This is in agreement with the SPI signal time dependence: the SPI signal reached a plateau after 7 h of soaking and after 1.5 h of cooking. However, there is a significant delay between the SPI and RARE signal; the RARE signal reaches a plateau after 10.3 h of soaking and after 2.1 h of cooking. The differences among the various bean treatments observed by MRI methods and mass increase are summarized in Table 1.
4. Discussion In the study, changes in a bean seed during soaking and cooking were followed by two complementary MRI methods, RARE and SPI. The RARE method enabled fast image acquisition and was therefore suitable for tracking fast changes during soaking and particularly during cooking. The RARE images were T2-weighted so that they enhanced areas of highly mobile water, while areas of less mobile water gave practically no signal detectable by the RARE method. In contrast, the SPI method enhanced areas of water molecules with restricted mobility. A drawback of the SPI method was slow signal acquisition owing to phase encoding in all three spatial directions. Consequently, the SPI image resolution was reduced, and a fast repetition time was used in order to save imaging time and therefore still enable the tracking of changes in the seed during soaking and cooking. Owing to use of the short repetition time, the bulk water SPI signal was suppressed. SPI imaging is more sensitive than is RARE for the detection of water migrating into a seed, but unfortunately its spatial resolution is insufficient for distinguishing between different anatomical seed structures such as seed coat, the void between the seed coat and cotyledons, and the cotyledon edge. Therefore, it is not possible to determine accurately the migration path of water into the cotyledons; the question remains: did water migrate into the void between the seed coat and cotyledons, through the seed coat, or through the micropyle. The spatial resolution of RARE images is much better and allows a reliable discrimination among the listed seed anatomical structures. The RARE images reveal a high signal of bulk water that entered through the micropyle below the seed coat, suggesting that this water penetrated into the cotyledons. No signal was observed in the cotyledons in the initial RARE images, which implies that the water that penetrated the cotyledons was bound in the cotyledon tissue. This is in accordance with previous observations showing that at the beginning of cotyledon hydration water cannot be stored as free water in it but is absorbed only by the solid matrix of the cotyledon tissue [35]. In the RARE images a water signal was first observed in the gap between the cotyledons at approximately 2.4 h of imbibition. Simultaneously, a strong signal in the inner side of the cotyledons appeared in the SPI images. This supports our hypothesis that water enters the cotyledons immediately after it comes in contact with cotyledon tissue. Identical behavior was observed in kidney and adzuki beans [23], but not in soybeans, where no water penetration in the cotyledons from the void between the cotyledons was observed [24].
343
Our experiments confirmed that water enters the bean seed immediately after its immersion in water. As the initial water migration is very fast it is not possible to determine accurately its entry point and path. Nevertheless, our measurements indicate that most likely water enters the seed through the micropyle and that water penetration through the seed coat is delayed. Fast water uptake was also observed for pinto beans, regarding which it has been proposed that water entered through the hilum/micropyle region and that absorbed water softens the seed coat from inside making it permeable to water [21]. A different imbibition process was observed in kidney and adzuki beans. The water entry into these bean seeds was significantly delayed, and the lens was found to be the water entry point. In the case of kidney beans, the lens was first hydrated for 95 minutes, and water entered the bean after 3 hours of soaking. Regarding the adzuki beans, the lag time before the entry of water was even longer [23]. The MRI measurements of Pietrzak et al. indicated that water enters the soybean seed through the micropyle and the hilum [22]. For the soybean of the Japanese cultivar “Kuromame”, which has a thick seed coat, water uptake was also found to be slow, and it was shown that water entered the seed near the raphe [24]. Our measurements also showed that the seed coat delayed water penetration into the bean. The role of the seed coat in water uptake has been shown to be very important in legume seeds [18,21,24]. At the beginning of soaking, the seed coat delays water uptake in the embryonic axis and governs the direction of water penetration to the embryo. The slow and controlled hydration of a dry seed is essential to preventing soaking injuries of the seed. Eventually, the seed coat serves as a reservoir of water for the hydrating axis. The swelling at higher temperatures follows the same hydration behavior except that the cotyledons are hydrated only from the outer surface (Fig. 5). The reason is the much smaller amount of water that enters into the gap between the cotyledons at 98 °C. At 98 °C, water uptake is much faster than at room temperature, but no significant changes were found in the total weight and volume gain. With regard to water uptake, the same temperature dependence was observed in legume seeds [7,8,36], where the faster water uptake at an increased soaking temperature was attributed to higher water diffusivity resulting from lower fluid viscosity, larger seed pores and the greater water permeability of the seed coat. Our results are in good agreement with the theory that predicts higher water diffusivity at higher temperatures. Namely, the SPI images of the bean soaked at 98 °C showed that 0.8 h of soaking was enough for water to penetrate through the entire seed except the void between the cotyledons; whereas at room temperature, no water was observed in the cotyledons even after 1.4 h of soaking (Figs. 3 and 5). On the other hand, our measurements could not confirm increased water migration through the seed coat at higher temperatures, since the spatial resolution of the SPI images was too low to enable distinguishing between water that migrated through the seed coat into the void between the seed coat and cotyledons and water that migrated into the region through the micropyle. The spatial resolution of the RARE images was twice as good as that of the SPI images, however, their signal was low due to increased diffusional and convectional motions of water with increasing temperature and with it a decrease in the associated RARE image signal. The relaxation times of water measured at different points in the soaking and cooking procedures can provide information on water mobility. A decrease of the T1 and T2 of unbound water in the void between cotyledons during soaking and cooking is associated with an increased amount of dissolved substances in the water. On the other hand, the relaxation time of water in bean tissue is correlated with the amount of water in the tissue, as for example the longer relaxation time of the cotyledon tissue being associated with a higher amount of water in the tissue.
344
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345
The time dependence of water T2 in the void between the cotyledons (unbound water) during soaking at room temperature is an indication of bean tissue substances being dissolved. The results in Fig. 2d showed that the T2 of water in the void between the cotyledons decreased during the first 10 h of soaking and remained constant at longer soaking times. An additional decrease was observed after the cooking of the presoaked bean, while the T2 value after the cooking of the dry bean was considerably longer. The T2 changes of unbound water in the void between the cotyledons thus indicate that the total amount of dissolved substances is higher in the case of presoaked than in the case of dry bean cooking. Unfortunately, it is not possible to distinguish among different dissolved substances only on the basis of relaxation time measurements. Further analysis would be needed to confirm differences among the released amounts of dissolved substances and to be able to determine the nature of the dissolved substances. The findings are in agreement with previous studies that have demonstrated that presoaking removes more antinutritient and flatulent substances from bean tissues than solely cooking [37–39]. Therefore, for the most efficient removal of unwanted substances from beans, an optimal combination of soaking and cooking is essential. Soaking reduces certain unwanted heat-stable antinutritient compounds (tannins, raffinose, stachyose, phytates) and the flatulencecausing oligosaccharides, the so-called α-galactosides. Cooking, on the other hand, reduces certain heat-sensitive antinutritients (lectins, protease inhibitors, saponins). Therefore cooking without presoaking removes most of the heat-sensitive antinutritional factors, but the heat-stable and soluble compounds are removed to a limited extent. In addition, soaking shortens cooking time, which is an important factor affecting the protein quality of cooked beans. The correlation between the cotyledon tissue T2 time dependence and time dependence of the total bean mass during soaking in cold water can be observed in Figs. 2d and 6. Both values increased during the first 10 h of soaking and then reached a plateau. Therefore, the increase of T2 during the soaking in cold water can be attributed to the increase in the amount of water in the cotyledons. After cooking, an additional increase of T2 without changes in total bean weight was observed for both presoaked and dry beans (Fig. 2d). The additional increase at higher temperatures can be attributed to starch gelatinization and/or protein denaturation at higher temperatures [40]. As shown previously, T2 increases after starch gelatinization [41] and protein denaturation [40], i.e., the processes associated with heat treatment. A comparison of the graphs in Fig. 6 shows that the seed volume and the SPI signal have approximately identical time dependence and reach a plateau earlier than the seed mass and the RARE signal, which again have approximately identical time dependence. The delay can be explained as water still entering the seed (increasing RARE signal) after the seed stopped expanding (SPI signal plateau). The RARE signal increase that occurred after the SPI signal reached the plateau is not uniform. It is more intense in the cotyledons, which implies that the water that enters the bean after it stopped expanding is mobile, and that it accumulates mainly in the cotyledon region. MRI measurements on kidney beans have also showed that signal intensity does not stop rising when the seed stops expanding [23]. The authors propose two possible explanations for such an effect: one is water replacing gases from small pores in the cells and the second is a molecular structure rearrangement allowing for the retention of more water, which would cause a different physiological stage of water absorption than the previous stages of seed hydration. Our measurements are in a better agreement with the second option. 5. Conclusions Two complementary 3D MRI methods, RARE and SPI, were used for tracking water hydration during the soaking and cooking of
common bean seed. The RARE method enabled the detection of free water and excelled in imaging resolution and signal acquisition speed; while the SPI method enabled the detection of bound water and had relatively low resolution and slow signal acquisition. By combining information from the two imaging methods it was possible to track the mobile and bound water that penetrated the bean seed and to determine the hydration dynamics of different anatomical parts of the seed. The study also demonstrated the great potential of MRI to study water imbibition processes in various seeds as also in many other water susceptible materials.
References [1] Molina MR, Bressani R, Elias LG. Nonconventional legume grains as proteinsources. Food Technol 1977;31:188–90. [2] Jood S, Mehta U, Singh R, Bhat CM. Effect of processing on flatus-producing factors in legumes. J Agric Food Chem 1985;33:268–71. [3] Deshpande SS, Cheryan M. Water-uptake during cooking of dry beans (Phaseolus- vulgaris L.). Plant Food Hum Nutr 1986;36:157–65. [4] Wood JA, Harden S. A method to estimate the hydration and swelling properties of chickpeas (Cicer arietinum L.). J Food Sci 2006;71:E190–5. [5] Berg T, Singh J, Hardacre A, Boland MJ. The role of cotyledon cell structure during in vitro digestion of starch in navy beans. Carbohydr Polym 2012;87:1678–88. [6] Kon S. Effect of soaking temperature on cooking and nutritional quality of beans. J Food Sci 1979;44:1329–35. [7] Hsu KH, Kim CJ, Wilson LA. Factors affecting water-uptake of soybeans during soaking. Cereal Chem 1983;60:208–11. [8] Joshi M, Adhikari B, Panozzo J, Aldred P. Water uptake and its impact on the texture of lentils (Lens culinaris). J Food Eng 2010;100:61–9. [9] Tang JM, Sokhansanj S, Sosulski FW. Moisture-absorption characteristics of laird lentils and hardshell seeds. Cereal Chem 1994;71:423–9. [10] Mnkeni AP, Gierschner K, Maeda EE. Effect of blanching time and salt concentration on pectolytic enzymes, texture and acceptability of fermented green beans. Plant Food Hum Nutr 1999;53:285–96. [11] Landa-Habana L, Pina-Hernandez A, Agama-Acevedo E, Tovar J, Bello-Perez LA. Effect of cooking procedures and storage on starch bioavailability in common beans (Phaseolus vulgaris L.). Plant Food Hum Nutr 2004;59:133–6. [12] Güzel D, Sayar S. Effect of cooking methods on selected physicochemical and nutritional properties of barlotto bean, chickpea, faba bean, and white kidney bean. J Food Sci Technol 2012;49:89–95. [13] Iyer V, Salunkhe DK, Sathe SK, Rockland LB. Quick-cooking beans (Phaseolus vulgaris L.): II. Phytates, oligosaccharides, and antienzymes. Plant Food Hum Nutr 1980;30:45–52. [14] Yousif AM, Deeth HC, Caffin NA, Lisle AT. Effect of storage time and conditions on the hardness and cooking quality of adzuki (Vigna angularis). Lebensm Wiss U Technol 2002;35:338–43. [15] Yousif AM, Batey IL, Larroque OR, Curtin B, Bekes F, Deeth HC. Effect of storage of adzuki bean (Vigna angularis) on starch and protein properties. Lebensm Wiss U Technol 2003;36:601–7. [16] Collins EJ. The structure of the integumentary system of the barley grain in relation to localized water absorption and semi-permeability. Ann Bot 1918;32: 381–414. [17] McDonald MB, Vertucci CW, Roos EE. Soybean seed imbibition: water absorption by seed parts. Crop Sci 1988;28:993–7. [18] McDonald MB, Vertucci CW, Roos EE. Seed coat regulation of soybean seed imbibition. Crop Sci 1988;28:987–92. [19] Nakanishi TM, Matsubayashi M. Nondestructive water imaging by neutron beam analysis in living plants. J Plant Physiol 1997;151:442–5. [20] Lammertyn J, Dresselaers T, Van Hecke P, Jancsok P, Wevers M, Nicolai BM. MRI and X-ray CT study of spatial distribution of core breakdown in ‘Conference’ pears. Magn Reson Imaging 2003;21:805–15. [21] Heil JR, McCarthy MJ, Özilgen M. Magnetic resonance imaging and modeling of water-uptake into dry beans. Lebensm Wiss U Technol 1992;25:280–5. [22] Pietrzak LN, Frégeau-Reid J, Chatson B, Blackwell B. Observations on water distribution in soybean seed during hydration processes using nuclear magnetic resonance imaging. Can J Plant Sci 2002;82:513–9. [23] Kikuchi K, Koizumi M, Ishida N, Kano H. Water uptake by dry beans observed by micro-magnetic resonance imaging. Ann Bot 2006;98:545–53. [24] Koizumi M, Kikuchi K, Isobe S, Ishida N, Naito S, Kano H. Role of seed coat in imbibing soybean seeds observed by micro-magnetic resonance imaging. Ann Bot 2008;102:343–52. [25] Hong YS, Cho JH, Kim NR, Lee C, Cheong C, Hong KS, et al. Artifacts in the measurement of water distribution in soybeans using MR imaging. Food Chem 2009;112:267–72. [26] Mohorič A, Vergeldt F, Gerkema E, De Jager A, Van Duynhoven J, Van Dalen G, et al. Magnetic resonance imaging of single rice kernels during cooking. J Magn Reson 2004;171:157–62. [27] Mortensen M, Thybo AK, Bertram HC, Andersen HJ, Engelsen SB. Cooking effects on water distribution in potatoes using nuclear magnetic resonance relaxation. J Agric Food Chem 2005;53:5976–81.
U. Mikac et al. / Magnetic Resonance Imaging 33 (2015) 336–345 [28] Hong SW, Yan ZY, Otterburn MS, McCarthy MJ. Magnetic Resonance Imaging (MRI) of a cookie in comparison with time-lapse photographic analysis (TLPA) during baking process. Magn Reson Imaging 1996;14:923–7. [29] Takano H, Naito S, Ishida N, Koizumi M, Kano H. Fermentation process and grain structure of baked breads from frozen dough using freeze-tolerant yeasts. J Food Sci 2002;67:2725–33. [30] Bajd F, Sersa I. Continuous monitoring of dough fermentation and bread baking by magnetic resonance microscopy. Magn Reson Imaging 2011;29:434–42. [31] Emid S, Creyghton JHN. High resolution NMR imaging in solids. Physica B 1985;128: 81–3. [32] Gravina S, Cory DG. Sensitivity and resolution of constant-time imaging. J Magn Reson B 1994;104:53–61. [33] Hennig J, Nauerth A, Friedburg H. RARE imaging: a fast imaging method for clinical MR. Magn Reson Med 1986;3:823–33. [34] Mansfield P, Grannell PK. Diffraction and microscopy in solids and liquids by NMR. Phys Rev B 1975;12:3618–34. [35] Golonka P, Dryzek J, Kluza M. Bean cotyledon microporosity under hydration conditions. Nukleonika 2002;47:137–40.
345
[36] Oliveira AL, Colnaghi BG, da Silva EZ, Gouvea IR, Vieira RL, Augusto PED. Modelling the effect of temperature on the hydration kinetic of adzuki beans (Vigna angularis). J Food Eng 2013;118:417–20. [37] Khokhar S, Frias J, Price KR, Fenwick GR, Hedley CL. Psysico-chemical characteristics of Khesari dhal (Lathyrus sativus): changes in α-galactosides, monosaccharides and disaccharides during food processing. J Sci Food Agric 1996;70:487–92. [38] Shimelis EA, Rakshit SK. Effect of processing on antinutrients and in vitro protein digestibility of kidney bean (Phaseolus vulgaris L.) varieties grown in East Africa. Food Chem 2007;103:161–72. [39] Fernandes AC, Nishida W, da Costa Proenca RP. Influence of soaking on the nutritional quality of common beans (Phaseolus vulgaris L.) cooked with or without the soaking water: a review. Int J Food Sci Tech 2010;45:2209–18. [40] Ahmad MU, Tashiro Y, Matsukawa S, Ogawa H. Gelation mechanism of surimi studied by 1H NMR relaxation measurements. J Food Sci 2007;72: E362–7. [41] Tananuwong K, Reid DS. DSC and NMR relaxation studies of starch-water interactions during gelatinization. Carbohydr Polym 2004;58:345–58.