• Changes in the physical state of cellular water and its interrelations with carbohydrate metabolism were studied during preplanting storage of tulip bulbs (Tulipa gesneriana‘Apeldoorn’).
• Magnetic resonance imaging, light and scanning electron microscopy and high-performance anion exchange chromatography with pulsed amperometric detection were used to follow time-dependent changes during bulb storage at 17 or 20°C (nonchilled) or 4°C (chilled).
• No visible differences in scale structure and central bud development were observed microscopically between chilled and nonchilled bulbs. However, the scales of the chilled bulbs exhibited higher water content, faster starch degradation and increased concentrations of sucrose and ethanol-soluble fructan. Quantitative measurements of magnetization transfer (MT) indicated a smaller fraction of a solid or a restricted-mobility proton pool in the scales of the chilled bulbs. By contrast, the MT effect was significantly higher in the central bud of the chilled than in the nonchilled bulbs.
• Degradation of storage polysaccharides to low-molecular-weight sugar molecules during release from dormancy could be accompanied by local release of water molecules tightly bound to the polysaccharide granules into the bulk water, or by an influx of free water molecules due to increased osmotic potential caused by the raised sugar concentration, or by a combination of both effects.
Dormancy, manifested during the annual life cycle of geophytes, is an essential condition for their normal development. In tulip, low temperature (2–10°C for 8–15 wk) is required for ‘dormancy release’ and normal flowering, but dormancy completion cannot be determined visually and is genotype-specific. Without a long period of low temperature, tulip growth is very slow and small plants with abnormal or aborted flowers are the result (De Hertogh & Le Nard, 1993).
Proton MRI, however, detects signal only from mobile protons that have sufficiently long T2 relaxation times. The less mobile protons associated with macromolecules and membranes in biological tissue have T2 values too short to be detected directly by conventional MRI. However, they may be detected indirectly, because of their possible coupling to mobile or ‘liquid’ protons. Therefore, the saturation of the macromolecular proton pool by radio-frequency (RF) irradiation can be transferred to the liquid protons and measured as Magnetization Transfer Effect (MTE). This technique is widely used in medical MRI (Wolff & Balaban, 1989), and can be applied for the quantitative measurement of the amount of magnetization transfer (MT) from bound to free water in living plant tissues (Bendel et al., 2001).
The aim of the present study was to study the interrelationships between changes in water status (measured by MRI) and carbohydrate metabolism within tulip bulbs during the cold-induced transition from dormancy to active growth.
Materials and Methods
Bulbs of Tulipa gesneriana L., cv. Apeldoorn, 10–11 cm in circumference, were obtained from a commercial source in the Netherlands for parallel studies in Israel and the Netherlands. In September, the bulbs had reached ‘Stage G’ (a developmental stage at which chilling may be applied), and after the first measurements (time 0), the bulb population was divided into two parts. In both Israel and the Netherlands, one part was placed at 4°C and 60% relative humidity for 4 months (chilled bulbs), while the second part was kept at 20°C and 60–65% relative humidity (nonchilled bulbs). Six intact bulbs per treatment were sampled, tagged, and studied by MRI at the beginning of storage treatments and after 4, 8 and 12 wk of storage at each temperature.
For carbohydrate analysis in the USA, tulip bulbs were transported from the Netherlands at 17°C with arrival in Ithaca in October. Immediately after receipt, storage treatments at 4°C (chilled) or 17°C (nonchilled) were begun.
During storage, developmental and histological analyses of the bulbs were performed once a month. Bulbs were dissected into their parts – scales, basal plate, and central bud (developing shoot), and fresh material sections were stained with IKI (iodine in KI solution) for starch and with toluidine blue for carbohydrates and polysaccharides (Ruzin, 1999). In addition, each organ was fixed in FAA (5 : 5 : 90 mixture of glacial acetic acid: formalin (40%): ethanol (70%) and later embedded in paraffin (Ruzin, 1999). Paraffin sections, 12 µm thick, were stained with AGS (Joel, 1983) or with PAS reaction and PAS with alcian blue (Ruzin, 1999). The sections were examined with a stereoscope (Zeiss Stemi, 2000-C, Zeiss, Germany), magnification 16–50, or with light microscope Olympus BX50, magnification 50–100.
Scanning electron microscopy (SEM) was performed in fresh material by JEOL (Jsm-5410LV, Tokyo, Japan), at low vacuum.
Dry matter and water content
Once a month, 10 bulbs from each temperature treatment were sampled, measured, and divided into the storage scales, basal plate and central bud (developing shoot). The f. wt of each organ was measured individually. Water content was determined by drying the dissected bulbs in a forced oven at 70°C until constant weight was attained.
In these studies, the established MRI applications of spin-lattice relaxation time (T1), spin-spin relaxation time (T2), and spin density imaging were complemented by measurement of the degree of Magnetization Transfer (MT). Coupling between the macromolecular protons and the mobile or ‘liquid’ protons allows the spin state of the macromolecular protons to influence the spin state of the liquid protons through exchange processes. In this case, the selective saturation of macromolecular spins can be transferred to the liquid spins, depending on the rate of exchange between the two spin populations, and hence can be detected with MRI (Henkelman et al., 2001). MT experiments can assess the presence and abundance of a solid or restricted-mobility proton pool (Bendel et al., 2001). MT images were obtained at the Weizmann Institute in Rehovot, Israel, on a 4.7 T superconducting, horizontal-bore instrument (Biospec 47/30, Bruker, Karlsruhe, Germany), as described previously (Bendel et al., 2001). Under most practical conditions, the free water pool is much larger than the solid or restricted-mobility proton pool, and the magnetization of the latter is practically unobservable because of its short T2. Under such conditions, the strength of the MT effect increases with increasing size of the solid or restricted-mobility proton pool (Bendel et al., 2001).
MT images were derived from fast spin echo scans, with and without preceding off-resonance irradiation. A single slice was excited in each scan. Four echoes per excitation were acquired, with the following parameters: effective (echo time) TE = 22 ms and (repetition time) TR = 3 s; (field of view) FOV = 5 cm; 256 × 256 matrix; slice thickness = 3 mm; two averages. The procedure for each bulb consisted of acquiring irradiation-on and irradiation-off images for two perpendicular planes across the center of the bulb. Off-resonance irradiation was applied during the entire recovery delay in the form of a train of 3.5-ms pulses separated by 0.2-ms gaps. The irradiation strength (ω1/2π) was 510 Hz, and the frequency offset (Δωo/2π) was 10 kHz. Images displaying the quantity , where Ms and Mo are the image intensities in the presence and absence of the off-resonance irradiation, respectively, were calculated on a pixel-by-pixel basis. It was verified that this protocol caused no measurable MT effect on a phantom containing an aqueous solution of 5 mM NiCl2.
The scans for the T1, T2, and spin density (SD) images were conducted on a 0.47 T scanner located at Wageningen University, in the Netherlands. This system consisted of a 0.47 T electromagnet (Bruker, Karlsruhe, Germany) interfaced to a SMIS spectrometer (Surrey Medical Imaging Systems, Guildford, UK); it used a custom-engineered probe containing a 50-mm-diameter rf coil and gradient coils with strengths up to 500 mT/m (Doty Scientific Inc., Columbia, SC, USA).
Proton relaxation can be characterized either by T1 and T2 (relaxation times) or by R1 (= 1/T1) and R2 (= 1/T2) (relaxation rates). Fast relaxation corresponds to short relaxation times and to high relaxation rates. The use of T- or R-based images enables different contrasts to be obtained. While the use of the T-value is more popular, the use of the R-value is often more appropriate, because when several mechanisms contribute to the relaxation, the R-values from the respective mechanisms can simply be added to obtain the overall relaxation rate. The two approaches to the characterization of relaxation are therefore completely equivalent, and the choice between them is a matter of semantics or convenience (van der Toorn et al., 2000)
In the present study, R2 and spin-density maps were obtained from a multiecho sequence described previously (Edzes et al., 1998). Forty-eight echo images with an interecho delay of 4.7 ms were acquired; the operating parameters were: TR = 1.5 s, 128 × 128 matrix, FOV = 60 mm, two averages, slice thickness = 3 mm. The parameter images were generated by a pixel-by-pixel fit of the data to a single-exponential decay. R2 images represent the rate constant of this fitted decay, and the spin-density images the signal extrapolated to TE = 0. A 0.5-ml vial containing 1 mM MnCl2 was imaged alongside the bulb, and the reported spin-density values in the bulbs are relative to those in the reference tube.
R1 maps were generated from a series of fast spin-echo images (eight echoes per excitation, effective TE = 12.2 ms); a range of repetition times was used, TR = 0.2, 0.35, 0.55, 0.9, 1.5, 2.5, and 5 s and other parameters were: FOV = 60 mm, 128 × 128 matrix, eight averages, slice thickness = 3 mm. The pixel-by-pixel R1 values were derived by fitting the data to a single-exponential recovery function.
Mean values of R1, R2, SD, and MTE parameters were calculated by defining regions-of-interest (ROI), which expanded from ‘seed points’ within the storage scales or the bud, and connected pixels within a narrowly defined intensity range, of typically ±10% around the mean. The mean parameter values calculated for each bulb were then averaged over the population in each group.
During storage, bulbs were analyzed for carbohydrates biweekly using a combination of HPLC and colorimetric methods to assess all the major carbohydrate pools in the bulbs. Bulbs were dissected into scales, basal plate, and central bud. Tissues of the middle portion of the third scale (counting from the outside) and the entire central bud were frozen in liquid nitrogen, freeze-dried, and ground to a powder. Tissue samples (50 mg) were extracted at 70°C with 80% ethanol (three extractions of 3 ml each, 1 h per extraction). Tissue suspensions were centrifuged at 4000 g for 10 min after each extraction, and the supernatants were combined. The extracts were passed through ion exchange columns consisting of 1 ml each of Amberlite IRA-67 (acetate form) (Sigma) and Dowex 50 W (hydrogen form) (Sigma) to remove ions. The extracts were then evaporated to dryness at 55°C, and dissolved in 10 ml of distilled water. After appropriate dilution, samples were subjected to high-performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD) with a Dionex DX-500 series chromatograph, equipped with a Carbopac PA-1 column, a pulsed amperometric detector and a gold electrode (Dionex, Sunnyvale, CA, USA). Carbohydrates were eluted at a flow rate of 1.0 ml min−1 at about 1600 psi with the following elution conditions: 200 mM NaOH (0–5 min); a linear gradient from 0 to 500 mM NaOAc in 200 mM NaOH (5–25 min); 500 mM NaOAc in 200 mM NaOH (25–28 min); a linear decreasing gradient from 500 to 0 mM NaOAc in 200 mM NaOH (28–29 min); 200 mM NaOH (29–32 min). The ethanol-soluble fraction contained glucose, fructose, sucrose, and a series of short-chain fructan polymers. The amounts of glucose, fructose, and sucrose were determined by comparison with calibration curves derived from standard authentic sugars. Since standards were not available for fructan oligosaccharides, quantification of each individual fructan polymer was not possible. Therefore, fructan was estimated by subtracting the amounts of glucose, fructose and sucrose (determined from HPAE-PAD) from total carbohydrate in the ethanol extract as determined by the phenol-sulphuric acid colorimetric method (Dubois et al., 1956).
The residue remaining after ethanol extraction was dried in an oven at 55°C overnight; it was then extracted with 5 ml of distilled water (adjusted to pH 8.0 with CaCO3) for 1 h at room temperature to solubilize water-soluble fructans. Supernatants were collected after centrifugation, and the water-soluble (long-chain) fructan was determined by the phenol-sulphuric acid method (Dubois et al., 1956). The residue remaining after water extraction was boiled for 30 min in 4 ml of Na-acetate buffer (pH 4.5). After cooling, 50 units of amyloglucosidase (in 1 ml of Na-acetate buffer, pH 4.5) were added to each sample and incubated for 48 h at 55°C. The amount of glucose released was determined by HPAEC-PAD using an aliquot of the digested sample. The amount of starch was calculated according to the amount of glucose released.
During the storage period, the tulip bulbs consisted of four to five storage scales, a basal plate, and a central bud (Fig. 1). The scales had a heterogeneous layer structure (Fig. 2a). The outer region of the scale comprised epidermal cells and a few layers of palisade cells, which did not contain starch granules. The inner region consisted of a spongy parenchyma with large, unordered cells. Large parenchyma cells (about 150 µm in diameter) were heavily packed with starch granules between 5 and 40 µm diameter (Fig. 2b).
The developing bud consisted of leaf primordia, flower stem, and flower parts: petals, stamens, and gynoecium (Fig. 2c). The flower stem contained many vascular bundles and compact parenchyma with small cells 10 µm in diameter filled with small starch granules 1–2 µm diameter (Fig. 2d). Smaller numbers of starch granules were observed in the cells of the developing leaves.
At the beginning of storage in September, the length of the central bud was 0.82 ± 0.3 cm. By December, the lengths of the central buds were 3.1 ± 0.11 cm and 3.2 ± 0.09 cm for chilled and nonchilled bulbs, respectively. After 12 wk of storage, no visible differences in bud development or in starch and carbohydrate contents were observed microscopically, between chilled and nonchilled bulbs.
Dry matter and water content
Measurements of the water content of the scales and central bud by destructive methods revealed significant differences between the two temperature treatments. After 12 wk of storage, the scales of the chilled bulbs contained more water than those of the nonchilled ones, while the opposite tendency was observed in the central bud. However, no differences in the f. wts of the scales and central buds, between chilled and nonchilled bulbs, were observed at the end of storage at 4 and 20°C (Table 1).
Table 1. F. and d. wts and relative water content of tulip scales and the central bud after 12 wk of storage at 4° or 20°C
Statistical analysis in rows, means followed by different letters are significantly different at P = 0.05 (n = 8).
F. wt, g
D. wt, g
Relative water content, %
R1, R2, and spin density images The images representing R1, R2 and spin density (SD), before and after 12 wk of storage at 4° and 20°C are shown in Fig. 3. The average values of these parameters are shown in Fig. 4, where the spin density values were normalized with respect to the signal intensity in the reference tube. The normalized SD (NSD), which is proportional to the total amount of mobile protons, that is mainly free water, increased significantly in the scales after storage at 4°C, but was unchanged in bulbs stored at 20°C. In the central bud, the spin density was considerably higher than that in the scales, but no differences in this parameter were observed between buds stored at 4°C and those at 20°C (Fig. 4a,b).
In the scales of the chilled bulbs, R1 initially decreased slightly, and then did not change during storage, whereas it increased significantly in the nonchilled bulbs; no differences were found in the central buds in bulbs stored at the two temperatures (Fig. 4c,d). R2 values in the scales increased during storage at both temperatures, which may indicate an overall loss of mobility of the water pool. This change was, however, more pronounced in the nonchilled bulbs. At the same time, R2 values in the bud increased significantly in the chilled bulbs, but did not change in the nonchilled bulbs (Fig. 4e,f).
The calculated MT images at the beginning and after 8 wk of storage at 4° and 20°C and the averaged values of the measured MT effects in the scales and in the central bud are presented in Fig. 5 and in Table 2. The MT effect in the scales of chilled bulbs after 8 wk of storage at 4°C was significantly lower than that in nonchilled bulbs, indicating a relatively smaller solid or restricted-mobility proton pool. By contrast, the MT effect in the central bud of the chilled bulbs was significantly higher than that in nonchilled bulbs.
Table 2. Values of MT effect in storage scales and the bud of chilled and nonchilled bulbs at time 0 and after 8 wk at 4° or 20°C. Means of five plants ± SE
Time 0 (before storage)
0.553 ± 0.029
0.313 ± 0.002
After 8 wk at 4°C
0.371 ± 0.036
0.562 ± 0.012
After 8 wk at 20°C
0.507 ± 0.016
0.421 ± 0.022
Starch was the most abundant nonstructural carbohydrate in tulip bulbs (about 60% and 25% of the d. wt in scales and central buds, respectively, at the beginning of storage). Tulip bulbs also contained fructans (about 10 and 5% of the d. wt in scales and central buds, respectively), and simple sugars such as sucrose, glucose, and fructose. More than 90% of the total fructan was soluble in 80% ethanol. The HPAE-PAD was able to resolve ethanol-soluble fructan into more than 15 individual polymers (data not shown).
During storage, the concentration of starch decreased in both scales and buds, and this decrease was accompanied by increasing concentrations of soluble fructan, sucrose, glucose and fructose (Fig. 6). The rate of starch breakdown was significantly higher in bulbs stored at 4°C than in those stored at 17°C. Starch concentration in the third scale of bulbs decreased gradually throughout the storage period, to about 40% of the d. wt at the end of 12 wk of storage. In the central bud, the starch concentration remained fairly constant until the eighth week, and then decreased sharply; this sharp decline was much more pronounced at 4°C than at 17°C.
The concentrations of sucrose and ethanol-soluble fructan increased gradually but significantly in scales of bulbs stored at 4°C, reducing sugars and longer chain (water soluble) fructan showed no discernable change (Fig. 6). Similarly, in the central bud, bulbs stored at 4°C showed very gradual increases in reducing sugars (Fig. 6). In the first 2 wk, there appeared to be a transient increase in short chain (ethanol soluble) fructan in 4°C-stored buds, with the pool remaining elevated but stable thereafter (Fig. 6). While the actual amounts of each fructan polymer could not be determined, because of the unavailability of standards, relative changes in their amounts could be estimated by comparing their peak areas. The comparison of peak areas for the first five fructan polymers (DP 3 to DP 8) indicated that the concentration of each fructan polymer increased in the central bud during storage at 4°C (data not shown).
In spite of the fact that no visible morphological, developmental or histological differences were observed between chilled and nonchilled bulbs, significant changes were measured in water relations and carbohydrate metabolism within bulb organs. The magnetization transfer (MT) effect clearly indicated the presence of a restricted-mobility proton pool in tulip bulb tissues (Fig. 5). Such a restricted-mobility proton pool includes water protons, but may also include those in the solid matrix, for example cellulose chains, starch, and other polysaccharides (Bendel et al., 2001). The assessment of the MT effect revealed significant differences between the water status in the chilled and the nonchilled bulbs at the end of the storage period (Fig. 5, Table 2). In the scales, the proportion of protons with restricted mobility was less in the chilled bulbs than in the nonchilled ones. In this particular system, the findings of the conventional MRI methods were consistent with the MT results. The slower R1 and R2 and the higher spin density in the scales of the chilled bulbs can be interpreted as indicating a smaller fraction of bound protons in contact with a larger fraction of free water. Solid-located protons, associated with starch molecules, may be liberated during starch breakdown as in the chilled bulbs (Fig. 6). This idea is also supported by the data showing significantly higher water content in the chilled bulbs after 12 wk of storage (Table 1). In the central bud, the proportion of restricted-mobility protons increased during the storage of both chilled and nonchilled bulbs, but was larger in the chilled bulbs (Table 2). The lower water content, as assessed by d. wet measurements (Table 1), and the increase in the R2 in the central buds of the chilled bulbs (Fig. 4f) was also consistent with the presence of a larger fraction of restricted-mobility protons. These data support the hypothesis that the accumulation of dry matter and water bonding in the cells of the central bud occur more intensively at low temperatures. It must be mentioned, however, that no differences were observed between central buds of the chilled and nonchilled bulbs, in their spin density and R1 values (Fig. 4).
Water in plant tissues exists in several pools, each with differing physical characteristics and degrees of mobility (Leopold & Vertucci, 1986). We can therefore assume that the evaluation of water status by destructive methods and/or by conventional MRI techniques will examine different and parallel processes that occur simultaneously within bulb organs. Thus, in the scales, starch degradation requires free water for hydrolysis, but at the same time could be accompanied by the release of bound water protons from starch granules. As a result of starch degradation, the concentration of soluble sugars and the sap viscosity increase. This process could reduce the water mobility, as observed by MRI. Relaxation time might also reflect differences in cell morphology, as in Zantedeschia tubers, where large cells exhibited longer relaxation times than small ones (Robinson et al., 2000). In tulip, the scale parenchyma comprises large cells containing large amounts of starch, whereas the central bud consists of small cells with thin cell walls, small vacuoles and a limited number of small starch granules (Fig. 2). Thus, the conventional MRI techniques should observe the integrated effects of all aspects of proton dynamics, while the MT effect provides a more direct measure of the behavior of a solid or restricted-mobility proton pool in tulip tissues and may reflect the transport of sugars.
In conclusion, the release from dormancy in bulbs is not visible morphologically, but is accompanied by degradation of starch to low-molecular-weight sugar molecules. This process could be accompanied by local release of tightly bound water molecules from the polysaccharide granules into the bulk water, by an influx of free water molecules attracted by the higher osmotic potential that arises from the increased sugar concentration, or by a combination of both effects.
The authors thank the Foundation of the Chief Scientist of the Israeli Ministry of Agriculture and Rural Development and the Flower Growers’ Association of Israel for financial support. We also acknowledge the Wageningen NMR Centre (Wageningen University, The Netherlands) for the funds and assistance to make the MR experiments possible.