Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
1
1
2
Chemical profiling of gramineous plant root exudates using high-resolution NMR and MS Teresa W-M. Fan a *, Andrew N. Lane b , Moshe Shenker c , John P. Bartley d , David Crowley e , and Richard M. Higashi f
a * Department of Land, Air and Water Resources, f Crocker Nuclear Lab., University of California, Davis, CA 95616, USA
b National Institute for Medical Research, Mill Hill, London NW71AA, UK c Department of Soil and Water Sciences, The Hebrew University of Jerusalem, Rehovot 76-100, Israel d Queensland University of Technology, GPO Box 2434, Brisbane 4001, Queensland, Australia. e Dept. of Soil and Env. Sciences, University of California, Riverside, CA 92521-0424, USA Metal ion ligands in root exudates of different plant types were examined using multinuclear 2-D NMR and high-resolution MS. Uptake of nutrient transition metals was related to the mugineic acid phytosiderophores while uptake of pollutant Cd appeared to be mediated by a different mechanism.
Root Exudate
Fe, Cu, Mn, Zn uptake
Cd uptake?
N N H OH
O O O OH OH OH e.g.
2
TITLE: COMPREHENSIVE CHEMICAL PROFILING OF GRAMINEOUS PLANT ROOT 2
EXUDATES USING HIGH-RESOLUTION NMR AND MS 3
4
Teresa W-M. Fana*, Andrew N. Laneb, Moshe Shenkerc, John P. Bartleyd, David Crowleye, and 5
Richard M. Higashif. 6
7 a* Department of Land, Air and Water Resources, University of California, One Shields Ave., Davis, 8
CA 95616, USA. Tel: 530-752-1450; Fax: 530-752-1552; E-mail: [email protected] 9 b Division of Molecular Structure, National Institute for Medical Research, Mill Hill, London 10
NW71AA, UK 11 c Department of Soil and Water Sciences, Faculty of Agricultural, Food and Environmental Quality 12
Sciences, The Hebrew University of Jerusalem, Rehovot 76-100 Israel. 13 d School of Physical Sciences, Queensland University of Technology, GPO Box 2434, Brisbane 4001, 14
Queensland, Australia. 15 e Dept. of Soil and Env. Sciences, University of California, Riverside, CA 92521-0424, USA. 16 f Crocker Nuclear Laboratory, University of California, One Shields Ave., Davis, CA 95616, 17
USA. 18
19
ABSTRACT 20
Root exudates released into soil have important functions in mobilizing metal micronutrients and for 21
causing selective enrichment of plant beneficial soil microorganisms that colonize the rhizosphere. 22
Analysis of plant root exudates typically has involved chromatographic methods that rely on a priori 23
knowledge of which compounds might be present. In the research reported here, the combination of 24
multinuclear and 2-D NMR with GC-MS and high-resolution MS provided de novo identification of a 25
number of components directly in crude root exudates of different plant types. This approach was 26
applied to examine the role of exudate metal ion ligands (MIL) in the acquisition of Cd and transition 27
metals by barley and wheat. The exudation of mugineic acids and malate was enhanced by Fe 28
deficiency, which in turn led to an increase in the tissue content of Cu, Mn, and Zn. The presence of 29
elevated Cd maintained at a free activity pCd of 8.8 (10-8.8 M), resulted in reduced phytosiderophore 30
production by Fe deficient plants. The buffer morpholinoethane sulfonate (MES), which is commonly 31
used in chelator-buffering nutrient solutions, was detected in the root exudate mixture, suggesting uptake 32
and re-secretion of this compound by the roots. The ability to detect this compound in complex mixtures 33
3
containing organic acids, amino acids, and other substances suggests that the analytical methods used 34
here provide an unbiased method for simultaneous detection of all major components contained in root 35
exudates. 36
37
KEYWORDS 38
Hordeum vulgare, Triticum aestivum, multinuclear and 2-D NMR, high-resolution MS, Fe 39
deficiency, metal sequestration, metal ion ligands, cadmium, phytosiderophores, mugineic acids. 40
41
INTRODUCTION 42
Plant root exudates can constitute up to 30-40 % of the plant’s photosynthetic productivity 43
Whipps, 1990; Sauerbeck et al., 1981; Oades, 1978; Samtsevich, 1965). They are known to play 44
important roles in a number of rhizospheric processes, including nutrient (e.g. Fe and Zn) acquisition 45
(Hopkins et al., 1998, Romheld, 1991; Zhang et al., 1991), plant-microbe associations (e.g. Rhizobium 46
symbiosis with leguminous roots) (Janczarek et al., 1997; Scheidemann and Wetzel, 1997), regulation 47
of plant growth (Einhellig et al., 1993), and determination of microbial community structure in the plant 48
rhizosphere (Grayston et al., 1995; Yang and Crowley, 2000). It is also expected that root exudate 49
components interact with both organic and inorganic contaminants to regulate their bioavailability and 50
transport in the soil environment. Moreover, root exudation may be an important mediator in altering 51
the species composition of the rhizosphere microflora that function in nutrient transformations, 52
decomposition and mineralization of organic substances, and formation of soil organic matter, all of 53
which are related to soil quality (Hodge and Millard, 1998). This latter aspect is particularly in need of 54
clarification, in view of the ever-increasing threat of global environmental change and soil pollution 55
caused by anthropogenic activity. However, the relationship between changes in exudate composition 56
and microbial community structure and function in soils is poorly understood. The lack of 57
comprehensive knowledge of exudate chemistry has been a major barrier to such mechanistic 58
understanding. 59
The slow progress in exudate characterization can be largely attributed to the lack of analytical 60
approaches that are comprehensive, require neither sample fractionation nor a priori knowledge, and 61
that are capable of rigorous structure identification. Using a combination of multi-nuclear and 2-D 62
NMR with GC-MS, we have previously demonstrated a comprehensive determination of barley root 63
exudates (Fan et al., 1997b) based on known spectroscopic databases (Fan, 1996) and literature 64
information on phytosiderophores (Ma and Nomoto, 1993). 65
4
However, for totally unknown compounds or those not represented in the databases, the above 66
approach is not always sufficient to deduce the complete structure(s). To supply the necessary 67
information, we have now added high-resolution mass spectrometry to the previous methods to obtain 68
molecular formula information. This greatly facilitated the determination of the complete structures of 69
various exudate components not in the databases. We also demonstrated that root exudate profiles 70
were highly dependent on plant species and genotypes of the same species as well as growth conditions. 71
Moreover, we observed a significant attenuating effect of Cd treatment on root exudate profiles and yet 72
an enhancement of transition metal uptake for both barley and wheat plants. 73
74
RESULTS AND DISCUSSION 75
76
Exudate Analysis 77
Roots exudates of wheat, barely, and rice plants were analyzed by a combination of NMR and 78
GC-MS methods as described previously (Fan et al., 1997b). Using this approach, comprehensive 79
profiles of the components were obtained directly from crude exudates without sample fractionation. 80
Illustrated in Figure 1 are results of 2-D 1H total correlation spectroscopy (TOCSY), 1H-13C 81
heteronuclear single quantum coherence spectroscopy (HSQC), and 1H-13C heteronuclear multiple 82
bond correlation spectroscopy (HMBC) for the root exudate of the wheat amphiploid (AgCS) under 83
the low Fe-high Cd (↓Fe - ↑Cd) treatment. From 1H TOCSY analysis, covalent correlations between 84
pairs of protons that are up to 4 bonds apart were traced, while the HSQC and HMBC experiments 85
revealed carbon coupling to its directly attached protons and to protons 2-4 bonds away, respectively. 86
Two dominant components of the AgCS exudate were identified from these spectra, i.e. 2’-87
deoxymugineic acid (2’-DMA) and MES (the buffer included in the plant growth medium but not during 88
exudate collection). The phytosiderophore, 2’-DMA, has been reported in common wheat (Triticum 89
aestivum L.) root exudates (Murakami et al., 1989), but its exudation by the AgCS hybrid has not 90
been characterized previously. As for MES, it is thought to be an impermeant buffer (but see below) 91
and is not expected to be present in root exudates[DC1]. 92
2’-DMA was assigned based on the 1H and 13C chemical shifts, proton covalent (scalar) 93
connectivity acquired from the TOCSY spectrum (Fig. 1A), the proton to carbon covalent connectivity 94
obtained from the HSQC (Fig. 1B) and HMBC spectra (Fig. 1C), as well as molecular ion data 95
acquired from electrospray MS (ES-MS). All proton connectivities expected for 2’-DMA (H-2 to H-96
3 to H-4, H-1’ to H-2’ to H-3’, and H-1” to H-2” to H-3”) were present in Fig. 1A, so were the one-97
5
bond carbon to proton correlations in Fig. 1B. In addition, scalar correlations were evident in the 98
HMBC spectrum (Fig. 1C) for the three carboxylate carbons to their closest protons (C-1 to H-2, C-99
4’ to H-3’, and C-4” to H-3”) and between protonated carbons and their neighboring protons that are 100
2 (e.g. C-2” to H-3”), 3 (e.g. C-1” to H-3”), and 4 bonds (e.g. C-1’ to H-4) apart. Moreover, two 101
mass ions (m/z 304 and 326) were evident in the full scan ES-MS spectrum (Fig. 2) of the same 102
exudate, which corresponded to the molecular ion (M) and M+Na of 2’-DMA. Taken together, these 103
NMR and MS data provided unequivocal identification of 2’-DMA in crude AgCS root exudates. 104
The presence of MES was deduced from the NMR (Fig. 1A-C), ES-MS (Fig. 2), and high-105
resolution fast atom bombardment (FAB) mass spectral data (data not shown). As shown in Fig. 2, the 106
ES-MS spectrum of the crude AgCS exudate was abundant in the mass ions of 196, 218, and 413, in 107
addition to 304 and 326. The m/z 218 and 413 ions were related to the m/z 196 ion based on their 108
daughter ion patterns (data not shown), and could correspond to the Na adduct of the 196 monomer 109
and dimer, respectively. From the FAB data of m/z 196, a molecular formula of C6H14O4NS was 110
inferred with a mass error of 5.8 ppm. This information greatly assisted NMR in deducing the MES 111
structure as described below. 112
The 1H TOCSY data (Fig. 1A) revealed two sets of –CH2–CH2– groups with one set (A) having 113
nonequivalent and the other (B) equivalent geminal protons. The integrated area of set A resonances 114
was roughly twice that of set B resonances. In accordance with the molecular formula, set A protons 115
should have originated from two chemically equivalent –CH2–CH2– groups while set B protons were 116
derived from one –CH2–CH2– group. These three –CH2–CH2– groups accounted for all six of the 117
carbons allowed in the molecular formula. In addition, the chemical shifts of both set A and set B 118
protons indicated that they were neighboring to heteroatoms such as N, O, or S. No other carbons 119
including those from the carboxylates were present in the structure based on the analysis of 1-D 13C 120
NMR and 2-D HMBC (Fig. 1C and data not shown), which is consistent with the molecular formula. 121
Moreover, no exchangeable (primary or secondary) amines were present based on the 1H TOCSY 122
spectrum recorded in H2O instead of D2O (data not shown), suggesting that the N is in a tertiary amine. 123
A quarternary amine was ruled out based on the required number (14) of protons in the molecular 124
formula. Altogether, the above structural clues allowed the MES structure to be deduced ab initio. 125
Further proof for this assignment was obtained from the 1H NMR analysis of the MES standard (data 126
not shown), which matched exactly with that acquired from the AgCS exudate. Thus, using the 127
combined NMR and MS approach, molecular structures of exudate components were identified from 128
crude exudates without the need for a priori knowledge or any sample fractionation. 129
6
Using a similar approach, the exudate profile for the Chinese spring wheat (CS, a parent of AgCS) 130
was examined. An example analysis of the CS exudates by TOCSY is shown in Figure 3. Assigned 131
components included Ile, Leu, Val, lactate (lac), Ala, acetate, 2’-DMA (peaks 1-9), malate, citrate, 132
glycinebetaine (GB), Gly, α- and β-glucose (Glc), and α-trehalose (see (Fan, 1996) for NMR 133
parameters for these compounds). Except for the sugars, these assignments were verified by the 1H-134 13C HSQC analysis (data not shown) and in part by the GC-MS analysis of the same exudate (see 135
below). In addition to components yielding sharp resonances, there were those whose cross-peaks 136
were relatively broad (indicated by arrows). The identities of these components are currently unknown, 137
although they are likely to be of macromolecular origin (e.g. proteins), as indicated by their chemical 138
shifts and broad peak widths. A polypeptide that appears to be related to Al resistance has been 139
reported in wheat root exudates (Basu et al., 1999) 140
141
Comparison of Root Exudate Profiles from Different Plant Species, Genotypes, and Growth 142
Conditions 143
Altogether, the above NMR spectroscopic analysis provided comprehensive profiles of exudate 144
components (both known and unknown to our database) excreted by barley, wheat, and rice plants, as 145
illustrated in Figure 4. By comparing these 1H NMR spectra, it is clear that exudate profiles varied with 146
plant species and genotypes as well as treatment conditions. For example, the barley exudate differed 147
from wheat exudates (CS and AgCS), most notably in the dominant mugineic acids present (3-epi-148
OHMA+MA for barley and 2’-DMA for wheat). A significant difference in the relative quantity of GB, 149
Ala, γ-aminobutyrate (GAB), Gly, Glu, and acetate was also noted, although the difference in acetate 150
may be complicated by a differential loss through volatilization during the freeze-drying process. The 151
rice exudate profile was different again from those of barley and wheat, in that except for Gly, the amino 152
acids and mugineic acids were present only at very low concentration. In addition, an unknown polyol 153
component was abundant in the rice exudate but absent in barley or wheat exudates. 154
Except for the exudation of 2’-DMA, the two wheat genotypes (CS vs. AgCS) exhibited rather 155
different exudate profiles (Fig. 4). The AgCS exudate was dominated by the medium buffer MES and 156
contained relatively little amino and organic acids, while the CS exudate was abundant in both amino 157
and organic acids. In addition, unknown component(s) represented by a group of resonances from 1.1-158
1.6 ppm in the AgCS exudate were absent in the CS exudate. The difference in the MES content was 159
presumably a result of the different treatment conditions, with MES included in the AgCS but absent in 160
CS growth media. However, it is unlikely that the abundance of MES in the AgCS exudate was purely 161
7
a result of contamination from the growth medium. This is because extensive rinsing of roots was 162
performed before exudate collection. Moreover, an identical MES treatment was used to culture barley, 163
yet MES was much less abundant in the barley exudate (Fig. 4). Instead, it is plausible that MES was 164
taken up by AgCS roots and then excreted along with other exudate components. The possible effects 165
of MES on the physiology of plants grown in buffered hydroponic media are unknown, but this finding 166
suggests that this question should be studied in more detail for experimental protocols that use MES and 167
other buffers to maintain the pH of plant growth media. 168
The same exudates from Fig. 4 were also analyzed by GC-MS and the results are shown in Table 169
1, along with those from the NMR spectroscopic analysis. The GC-MS analysis not only confirmed 170
many of the NMR assigned components described above, but also provided absolute quantification of 171
these components plus those that were below NMR detection limits. At the same time, the NMR 172
analysis (with the aid of GC-MS analysis) provided absolute quantification for components (e.g. 173
mugineic acids) that were not readily analyzed by GC-MS. Thus, the two approaches complemented 174
each other in providing richer information than that achievable by each method alone. 175
In terms of total amounts of identified exudate components on a per gram root weight basis, the 176
wheat hybrid AgCS ranked the highest (although a major portion was attributed to MES), followed by 177
barley, wheat CS, and rice. It is also notable that many quantitative differences in the exudate 178
components of the four plant types were evident from the GC-MS profiles, which is consistent with the 179
NMR characterization. Moreover, the abundance of the mugineic acid phytosiderophores in barley and 180
wheat exudates was noted. In contrast, the exudation of these phytosiderophores by Fe-deficient rice 181
roots was not apparent, which may be related to the NMR detection limit plus a much lower capacity of 182
rice roots in excreting these compounds (Kawai et al., 1988). This lower capacity for mugineic acid 183
exudation by rice roots has been correlated with their reduced ability to tolerate Fe deficiency (Mori et 184
al., 1991; Kawai et al., 1988; Marschner et al., 1986). 185
186
Effect of Combined Cd and Fe Deficiency Treatment on Root Exudate Profiles of Barley and 187
Wheat Plants 188
Utilizing the above approach, we examined the chemical profiles of both barley and AgCS wheat 189
exudates collected under a combination of Cd and Fe deficiency treatments. The effects of these 190
treatments on plant growth, metal uptake, total Fe-mobilizing capacity, and other physiological 191
parameters have been reported elsewhere (Shenker et al., submitted to Journal of Environmental 192
Quality, JEQ). Here, we compared the effect of these treatments on root metal content with that on 193
8
exudate composition, as shown in Figures 5 and 6, for barley and wheat, respectively. Although many 194
more exudate components have been analyzed (see Figure 4), the comparison is illustrated for abundant 195
metal ion ligands (MIL), since they are more likely to influence metal mobilization and uptake by plant 196
roots. The stability constants of these MIL with different metal ions vary considerably, and their actual 197
ability to mobilize various metal ions will depend on the pH and ionic composition of the soil solution, as 198
well as kinetic factors involved in the dissolution of soil minerals. 199
Under Fe deficient conditions (↓Fe - ↑Cd and ↓Fe - ↓Cd), the increased exudation of the 3-epi-200
OHMA, MA, 2’-DMA, and malate in barley (Figure 5 and Table 2) and 2’-DMA in wheat exudates 201
(Figure 6) was evident, as compared to Fe sufficient conditions (↑ Fe - ↑Cd and ↑Fe - ↓Cd). This 202
increase was in turn related to the elevated production of Fe mobilizing substances (FeMS) by barley 203
(data not shown) and wheat (Figure 6) under Fe deficiency. These results are consistent with the role of 204
these MIL in Fe acquisition. The exudation of other abundant MIL (i.e. lactate, Ala, GAB, and GB for 205
barley; MES for wheat) did not exhibit a clear response to Fe deficiency (Figure 5, Fig. 6, and Table 206
2), which suggests that they are not specifically involved in Fe acquisition. However, these MIL may 207
still function generally in chelating Fe for uptake.[DC2] 208
Accompanying the increased exudation of the phytosiderophores was the elevated content of Zn 209
and Cu in barley roots (Figure 5), Cu in wheat roots (Figures 6) as well as Zn, Cu, and Mn in both 210
shoots (Shenker et al., submitted to JEQ). This finding suggests that the MAs [DC3]may be important in 211
the acquisition of Zn, Cu, and Mn, in addition to Fe. It is also noteworthy that the exudation pattern of 212
malate was similar to that of the MAs by barley roots (Fig. 5). This result could be accounted by the 213
exudation of malate and MAs via a common anion channel. However, other abundant anions in root 214
tissues including citrate and malate (unpublished results) did not exhibit such exudation pattern, which 215
argues against a systemic leakage through such a channel. In fact, citrate was not exuded by barley or 216
wheat roots to any significant extent (data not shown). Regardless of the exudation mechanism(s), it is 217
possible that the exuded malate could serve a similar role as MAs in mediating the observed 218
accumulation of transition metals in barley and wheat tissues. 219
As for Cd acquisition, there was no statistically significant correlation of the above MIL to the Cd 220
content of roots (Figures 5 and 6) or shoots (data not shown) of barley and wheat. This result suggests 221
that, Cd acquisition may involve a mechanism different from that of transition metal acquisition. It is also 222
interesting to note that there appeared to be a weak correlation between MES concentration in wheat 223
exudates and root Cd content (Figure 6). It is possible that the presence of MES may interact with Cd 224
to affect the latter uptake into wheat. Cd accumulation could weaken membrane integrity or increase 225
9
membrane permeability via lipid peroxidation (Goyer, 1997; Stohs and Bagchi, 1995), which allows 226
Cd-bound MES to enter into root cells and later to be released to the collection solution down the 227
concentration gradient. This hypothesis is consistent with the higher Cd uptake and MES release by 228
AgCS wheat roots than by barley roots (cf. Figures 5 and 6; Table 1). For barley, the exudation of 229
Ser, Thr, and Glu appeared to be related to the high Cd treatment under Fe sufficient conditions (Figure 230
5 and Table 2). It is possible that these MIL may be involved in Cd acquisition[DC4]. 231
In conclusion, the combination of multinuclear and 2-D NMR with GC-MS and high-resolution 232
MS provided de novo identification of a number of components directly in crude root exudates of 233
different plant types. This approach was applied to examine the role of exudate MIL in the acquisition 234
of Cd and transition metals by barley and wheat. The results suggest that enhanced exudation of 235
mugineic acids and malate may be involved in transition metal but not Cd acquisition. Therefore, the 236
acquisition mechanism appears to differ between the essential and toxic pollutant metals. 237
238
EXPERIMENTAL 239
240
Plant Growth 241
The plant varieties used were Barley (Hordeum vulgare L., cv. CM-72), wheat (Triticum 242
aestivum L., cv. AgCS, a hybrid of Chinese spring wheat and Lophopyrum. sp) supplied by Dr. J. 243
Dvorak (Rommel and Jenkins, 1959), Chinese spring wheat (CS, a parent of AgCS), and rice (Oryza 244
sativa, L., cv M201). Barley and AgCS wheat seedlings (eight per treatment) were grown for 22 d in 245
nutrient solutions containing CdCl2 buffered by an excess of EDTA and HEDTA (Fan et al., 1997b; 246
Shenker et al., submitted JEQ) to produce Fe-sufficient (pFe = 17.2) and deficient (pFe = 18.2-18.4) 247
conditions in combination with free Cd activity (pCd) of 8.8 or 10.0. Other free metal activities were 248
pZn = 10.1, pCu =13.5; and pMn = 7.3 (for complete plant growth protocol refer to Shenker et al., 249
submitted JEQ). The four treatments are designated as: 1) low Fe/high Cd (↓Fe - ↑Cd); 2) high 250
Fe/high Cd (↑Fe - ↑Cd); 3) low Fe/low Cd (↓Fe - ↓Cd); and 4) high Fe/low Cd (↑Fe - ↓Cd). 251
Morpholinoethane sulfonate (MES) at 2 mM was also included to buffer the pH of the nutrient solution 252
at 6.2±0.2. Chinese spring (CS) wheat and rice (cv. M201) seedlings (four per treatment) were grown 253
similarly for a month except that the nutrient solution employed was half-strength modified Hoagland 254
solution (Epstein, 1972) containing full-strength (50 µM) Fe(III)-EDTA, 25 µM EDTA, and 10 µM 255
CdSO4 in the absence of MES (Fan et al., 2000). The calculated pMs in this nutrient solution were 256
pFe = 17.6, pMn = 7.3, pCu = 12.8, pZn = 9.9, pNi = 12.1, and pCd = 8.9 (Shenker et al., submitted 257
10
to JEQ). The pH of the growth media was maintained at 6.0±0.3 by daily titration with 0.1 M HCl or 258
KOH while the nutrient solution was renewed frequently to avoid nutrient depletion. 259
260
Root Exudate Collection 261
Roots were soaked in 1 mM Ca(NO3)2 for 10-20 min and rinsed extensively with doubly 262
deionized (DD) H2O to remove organic and inorganic constituents of the nutrient solution, including 263
metabolites, exudates, chelates, mineral elements, and buffer ions before exudate collection. Root 264
exudates were collected 2-3 times each in 150-200 ml DDH2O during peak diurnal release of 265
phytosiderophores (2-4 h after light onset) of the last 3-6 growth days. The collected solutions were 266
immediately filtered through glass microfiber filters (GF/C, Whatman, Maidstone, England) to exclude 267
root fragments and much of the microbial population, pooled, and lyophilized for storage. Prior to 268
NMR studies, the exudates were resuspended in sterile water, filter sterilized, and passed through 269
Chelex 100 resin (BioRad, Hercules, CA) to remove paramagnetic ions, as described previously (Fan 270
et al., 1997b). 271
272
NMR Analysis 273
The Chelex 100-treated exudates of barley and AgCS were acidified to pH 2, lyophilized, and 274
redissolved in D2O, whereas the CS and rice exudates were not acidified (pH values ranging from 7.0-275
7.8). NMR spectra were recorded at 25 °C on a Varian (Palo Alto, CA) 11.75 Tesla UnityPlus (all 1H 276
experiments) and a Bruker AM-400 spectrometer (1-D 13C analysis). The following 2-D 1H 277
experiments were conducted: 1H total correlation (TOCSY), double-quantum filtered correlation 278
(DQF-COSY), 1H-13C heteronuclear single quantum coherence (HSQC), and 1H-13C heteronuclear 279
multiple bond correlation (HMBC). Selected samples were dissolved in DD H2O containing 10% 280
D2O+0.9M HCl for Watergate TOCSY measurement (Piotto et al., 1992) at 5 °C which optimized the 281
detection of exchangeable protons (Fan et al., 1997a). Detailed acquisition parameters are described in 282
respective figure legends. The 1-D NMR spectra were apodized using an exponential function yielding 283
0.5-2 Hz line broadening, and zerofilled to 64 or 128 K points before Fourier transformation. The 2-D 284
NMR data were zerofilled or linear-predicted to 8 K x 2 K points and apodized with a mild Gaussian 285
function before 2-D Fourier transformation. 286
287
11
Mass spectrometry 288
Nanospray mass spectra were acquired on an LCQ “classic” quadrupole ion trap mass 289
spectrometer (ThermoQuest, Austin, TX) equipped with a nanospray source (Protana, Odense, 290
Denmark) operated at a spray voltage of 800V and a capillary temperature of 150°C. Lyophilized root 291
exudates were dissolved in 50% methanol, 1% formic acid and 2ul transferred to an Econo12 292
nanospray needle (New Objective Inc, Cambridge, MA). MS2 and MS3 analyses were performed at a 293
collision energy of 30% and an isolation width of 4 Da. 294
Accurate mass of exudate components were measured on a Finnigan 2001 Newstar FT/MS 295
system (ThermoQuest, Madison WI) at 3 .0 Tesla. The spectrometer was externally calibrated with 296
PMMA 640. The lyophilized sample was delivered to the mass spectrometer by flow injection at 5 ml / 297
min using a syringe pump (Cole Palmer, Vernon Hills, IL) to an Analytica API external source operating 298
at a capillary voltage of 5.5 kV. 299
300
GC-MS Analysis 301
All exudates were acidified in 0.1 N HCl before an aliquot was lyophilized and derivatized with N-302
methyl-N-[tert-butyldimethylsilyl]trifluoroacetamide (MTBSTFA) before analysis on a Varian 3400 gas 303
chromatograph interfaced to a Finnegan ITD 806 mass spectrometer (Fan et al., 1997b). 304
305
Exudate Quantification 306
All components of exudate collection solutions were quantified by their GC-MS response as 307
compared to that of the standard with known concentration, except for the mugineic acids (MAs), 308
glycinebetaine (GB), acetate, and MES (see Fig. 4.). The latter were quantified from their 1H NMR 309
responses by comparison with the methyl resonance of alanine at 1.47 ppm (either endogenous or 310
added) after corrected for signal saturation and number of protons. The concentration of the 311
endogenous Ala was determined from GC-MS analysis (Fan , 1997b). In addition, the FeMS data 312
were adapted from Shenker et al. (submitted to JEQ), which were acquired using the Cu-CAS (chrome 313
azurol S) colorimetric assay (Shenker et al., 1995). This assay measured the total capacity of Fe 314
mobilization in crude exudates. 315
316
Acknowledgements. 317
This work was supported by U.S. Department of Energy grant numbers DE-FG07-96ER20255, 318
U.S. E.P.A. grant #R825960010, and U.S. E.P.A. funded (#R819658) Center for Ecological Health 319
12
Research at Univ. of California-Davis. We thank Dr. Steven Howell of NIMR for recording the 320
nanospray ES-MS spectra. NMR spectra were recorded at the MRC Biomedical NMR Centre, Mill 321
Hill. 322
323
13
Table 1. GC-MS and NMR quantification of wheat and barley root exudates under low Fe plus Cd 323
treatment. 324
325
µmole/gm root dry wt.
Component Barley1 Wheat CS1 Wheat AgCs2 Rice1
lactate 3.51 1.70 –3 3.23
pyruvate 6.95 ND3 – ND
Ala 2.08 0.13 – 0.01
Gly 0.28 0.07 – 0.52
Val 0.32 0.07 – ND
Leu 0.16 0.07 – ND
Ile 0.08 0.05 – ND
succinate 0.26 0.03 – ND
GAB 1.22 ND – ND
Pro 0.17 0.01 – ND
fumarate 0.23 0.02 – ND
PO4-containing 0.10 0.02 – 1.24
Ser 0.38 0.05 – 0.06
Thr 0.36 ND – ND
malate 0.46 0.49 – 0.38
Phe 0.06 0.01 – ND
Asp 0.22 0.02 – 0.01
Glu 0.25 ND – ND
Asn 0.21 ND – ND
Gln 0.17 ND – ND
3-epi-OHMA 8.19 ND – ND
2'-DMA 0.87 5.64 25.37 ND
MA 2.13 ND ND ND
GB 1.54 0.22 ND ND
MES 2.43 ND 209.18 ND
Acetate ND 11.60 ND ND
14
Total 32.63 20.19 234.55 5.46 1Quantified by GC-MS except for 3-epi-OHMA, 2’-DMA. MA, GB, MES, and acetate, which were 326
analyzed by NMR using existing lactate as an internal standard. 327 2AgCs quantified by NMR with the addition of 1 mM Ala as an internal standard. 328 3ND, not detected; –, not determined. 329
330
15
Table 2. Interactive effect of Cd and Fe deficiency on barley exudate profiles1 330
331
Compound ↓Fe - ↑Cd S.D. ↑ Fe - ↑Cd S.D. ↓Fe - ↓Cd S.D. ↑Fe - ↓Cd S.D. ______________________________________________________________________________________________________________________
lactate 2.88 0.48 2.83 1.07 3.15 1.13 2.69 0.10
pyruvate 7.59 1.74 5.72 0.69 6.06 0.91 5.98 1.70
alanine 1.63 0.45 2.95 0.98 2.20 1.44 1.80 0.62
glycine 0.24 0.07 0.38 0.12 0.27 0.14 0.23 0.08
valine 0.28 0.06 0.30 0.11 0.33 0.14 0.21 0.06
leucine 0.15 0.02 0.17 0.04 0.16 0.05 0.12 0.03
succinate 0.28 0.06 0.21 0.06 0.29 0.13 0.15 0.04
GAB 1.03 0.19 1.35 0.36 1.16 0.61 0.74 0.19
fumarate 0.27 0.08 0.28 0.05 0.27 0.13 0.24 0.02
serine 0.36 0.03 0.49 0.14 0.35 0.16 0.31 0.09
threonine 0.38 0.09 0.53 0.07 0.32 0.15 0.31 0.10
malate 0.34 0.08 0.20 0.08 0.48 0.53 0.09 0.05
aspartate 0.19 0.05 0.27 0.11 0.14 0.05 0.11 0.05
glutamate 0.19 0.05 0.35 0.16 0.22 0.10 0.19 0.08
asparagine 0.17 0.03 0.19 0.06 0.15 0.08 0.13 0.05
glutamine 0.11 0.09 0.14 0.05 0.23 0.16 0.16 0.05
αglycerol-3-P 0.20 0.05 0.16 0.02 0.17 0.07 0.12 0.05
3-epi-OH-
MA
5.25 2.14 1.15 0.77 7.97 4.06 0.96 0.22
2'-DMA 0.49 0.29 0.00 0.01 0.62 0.36 0.01 0.01
MA 0.79 0.92 0.09 0.16 1.02 0.35 0.04 0.07
GB 1.14 0.32 1.56 0.56 1.52 1.00 0.82 0.34
MES 1.28 0.77 1.44 1.04 2.28 1.44 3.24 2.04
Total 25.94 21.45 30.36 19.13 ______________________________________________________________________________________________________________________
1Units in µmole/g root dry wt. 332
333
16
REFERENCES 333
Basu, U., Good, A. G., Aung, T., Slaski, J. J., Basu, A., Briggs, K. G.. and Taylor, G. J. 1999. A 23-334
kDa, root exudate polypeptide co-segregates with aluminum resistance in Triticum aestivum. 335
Physiologia Plantarum 106: 53-61. 336
Einhellig, F. A., Rasmussen, J. A.,, Hejl, A. M., and Souza, I. F. 1993. Effects of Root Exudate 337
Sorgoleone on Photosynthesis. Journal of Chemical Ecology 19: 369-375. 338
Epstein, E. 1972. Mineral Nutrition of Plants: Principles and Perspectives. New York:John Wiley & 339
Sons. 340
Fan, T. W.-M. 1996. Metabolite profiling by one- and two-dimensional NMR analysis of complex 341
mixtures. Progress in Nuclear Magnetic Resonance Spectroscopy 28: 161-219. 342
Fan, T. W. M., Higashi, R. M.,, Frenkiel, T. A., and Lane, A. N. 1997a. Anaerobic nitrate and 343
ammonium metabolism in flood-tolerant rice coleoptiles. Journal of Experimental Botany 48: 1655-344
1666. 345
Fan, T.W.-M., Pedler, J., Lane, A.N., Crowley, D., and Higashi, R.M. 1997b. Comprehensive 346
analysis of organic ligands in whole root exudates using nuclear magnetic resonance and gas 347
chromatography-mass spectrometry. Analytical Biochemistry 251: 57-68. 348
Fan, T.W.-M., Baraud, F., Higashi, R.M., eds. Eller, P.G. and Heineman, W.R. 2000. Genotypic 349
influence on metal ion mobilization and sequestration via metal ion ligand production by wheat. 350
In: Nuclear Site Remediation. American Chemical Society Symposium Series, vol. 778, pp. 351
417-431, Washington, D.C. 352
Grayston, S. J., Campbell, C. D.,, Vaughan, D., and Jones, D. 1995. Influence of root exudate 353
heterogeneity on microbial diversity in the rhizosphere. Journal of Experimental Botany 46: 27. 354
Goyer,R.A. 1997. Toxic and essential metal interactions. Annu. Rev. Nutr. 17: 37–50. 355
Hodge, A. and Millard, P. 1998. Effect of elevated CO-2 on carbon partitioning and exudate release 356
from Plantago lanceolata seedlings. Physiologia Plantarum 103: 280-286. 357
Hopkins, B. G., Whitney, D. A.,, Lamond, R. E., and Jolley, V. D. 1998. Phytosiderophore release 358
by sorghum, wheat, and corn under zinc deficiency. Journal of Plant Nutrition 21: 2623-2637. 359
Janczarek, M., Urbanik-Sypniewska, T., and Skorupska, A.. 1997. Effect of authentic flavonoids and 360
the exudate of clover roots on growth rate and inducing ability of nod genes of Rhizobium 361
leguminosarum bv. trifolii. Microbiological Research 152: 93-98. 362
Kawai, S., Takagi, S., and Sato, Y. 1988. Mugineic acid-family phytosiderophores in root-secretions 363
of barley, corn and sorghum varieties. Journal of Plant Nutrition 11: 633-642. 364
17
Ma, J. F. and Nomoto, K.. 1993. Two Related Biosynthetic Pathways of Mugineic Acid in 365
Gramineous Plants. Plant Physiology (Rockville) 102: 373-378. 366
Marschner, H., V. Römheld and M. Kissel. 1986. Journal of Plant Nutrition 9: 695-713. 367
Mori, S., Nishizawa, N., Hayashi, H., Chino, M., Yoshimura, E., and Ishihara, J. 1991. Why are young 368
rice plants highly susceptible to iron-deficiency? Plant and Soil 130: 143-156. 369
Murakami, T., Ise, K., Hayakawa, M., Kamei, S., and Takagi, S. Chemical Letters 1989, 2137-2140, 370
1989. 371
Oades, J. M. 1978. Mucilages at the root surface. Journal of Soil Science 29: 1-16. 372
Piotto, M., Saudek, V. and Sklenar, V.. 1992. Gradient-Tailored Excitation For Single-Quantum Nmr 373
Spectroscopy of Aqueous Solutions. Journal of Biomolecular Nmr 2: 661-665. 374
Rommel, R., and Jenkins, B. C.. 1959 Amphiploids in Triticeae produced at the University of 375
Manitoba from March 1958 to December 1959. Wheat Inf. Serv. 9-10: 23. 376
Römheld V. 1991. The Role of Phytosiderophores in Acquisition of Iron and Other Micronutrients in 377
Graminaceous Species - an Ecological Approach. Plant and Soil, 130:127-134. 378
Samtsevich, S. A. 1965. Active excretion of plant roots and their significance. Soviet Pl. Physiol. 12: 379
731-740. 380
Sauerbeck, D., Nonnen, S. and Allard, J.L. 1981. Consumption and turnover of photosynthates in the 381
rhizosphere depending on plant species and growth conditions. Landw. Forschung Sonderheft 382
37:207-216. 383
Scheidemann, P. and Wetzel, A. 1997. Identification and characterization of flavonoids in the root 384
exudate of Robinia pseudoacacia. Trees (Berlin) 11: 316-321. 385
Shenker, M., Hadar, Y., and Chen, Y. 1995. Rapid method for accurate determination of colorless 386
siderophores and synthetic chelates. Soil Science Society of America J. 59: 1612-1618. 387
Shenker, M., Fan, T.W.M., and Crowley, D. E. 2000. Phytosiderophores influence on cadmium 388
uptake by wheat and barley Plants. Journal of Environmental Quality, submitted. 389
Stohs, S.J. and Bagchi, D. 1995. Oxidative mechanisms in the toxicity of metal ions. Free Radical 390
Biology in Medicine 18: 321–336. 391
Whipps J.M. 1990. Carbon economy. pp. 59-97 in: J.M. Lynch, The Rhizosphere. John Wiley & 392
Sons. Chichester, England. 393
Yang, C.H. and Crowley, D.E. 2000. Rhizosphere microbial community structure in relation to 394
root location and plant iron nutritional status. Applied and Environmental Microbiology, 66: 395
345-351. 396
18
Zhang F., V. Römheld, and Marschner, H. 1991. Diurnal rhythm of release of phytosiderophores and 397
uptake rate of zinc in iron-deficient wheat. Soil Science and Plant Nutrition 37:671-678. 398
399
19
FIGURE LEGENDS 399
400
Figure 1. 2-D 1H-TOCSY, 1H-13C HSQC, and 1H-13C HMBC spectra of crude AgCS wheat 401
exudate. 402
The TOCSY spectrum (a) was acquired with 2.5 s interpulse delay, 5 kHz spectral width in both 403
dimensions, 8.2 kHz B1 field strength, 47.3 ms mixing time, and 0.41s and 0.072 s acquisition time in 404
F2 and F1, respectively. The HSQC spectrum (b) was obtained with 2.5 s interpulse delay, 5 kHz 405
spectral width in F2 and 25 kHz in F1, and 0.147 s and 0.022 s acquisition time in F2 and F1, 406
respectively. The HMBC spectrum (c) was recorded with the same parameters as the HSQC 407
spectrum except for 0.499 s acquisition time in F2. DMA: 2’-deoxymugineic acid (right structure); CA, 408
CB: A and B carbons of morpholinoethane sulfonate (MES, left structure), respectively. The pH of the 409
exudate sample was 2.5. 410
411
Figure 2. Full scan electrospray mass spectrum of crude AgCS wheat exudate. 412
The ES-MS spectrum was acquired in 50% methanol plus 1% formic acid from a LC-Q mass 413
spectrometer 414
415
Figure 3. 2-D 1H-TOCSY spectrum of Chinese Spring wheat root exudate. 416
The TOCSY spectrum was acquired using 1.5 s interpulse delay, 5 K Hz spectral width in both 417
dimensions, 8.54 kHz B1 field strength, 47.3 ms mixing time, and 0.41 s and 0.051 s acquisition time in 418
F2 and F1, respectively. Peaks 1-9: H-2’, H-2”, H-3, H-1”, H-1’, H-3’, H-4, H-3”, and H-2 of 2’-419
DMA, respectively; lac: lactate, GB: glycinebetaine, α,β-Glc: α,β-glucose. The pH of the exudate 420
sample was 7.5. 421
422
Figure 4. 1-D 1H NMR spectra of barley, wheat (CS and AgCS), and rice root exudates. 423
The 1-D spectra were acquired using 2 s acquisition time, 2 (rice) or 5 s (the rest) interpulse delay, 424
5000 Hz spectral width, and 512 (rice) or 128 (the rest) transients. The abbreviations are as those in 425
Figures 1 and 5. The pH values of barley and AgCS root exudates were 2.2 and 2.5, respectively 426
while those of rice and CS root exudates were 7.5. 427
428
Figure 5. Relation of barley root exudate components to root metal contents. 429
20
The root exudate components were analyzed as described in Materials and Methods while the root 430
metal content was acquired from Shenker et al., submitted to JEQ. 3-epi-OHMA, 3-epi-431
hydroxymugineic acid; 2’-DMA, 2’-deoxymugineic acid; MA Sum, 3-epi-OHMA+2’-DMA+ MA. 432
Two-way ANOVA test or Duncan’s multiple range analysis of variance from the GLM statistical 433
package of SAS was performed on the data. Different letters (italicized letters for exudate components 434
and normal letters for root metal content) indicate statistical difference at a p < 0.05 significance level. 435
436
Figure 6. Relation of wheat root exudate components to root metal contents. 437
The root exudate components were analyzed as described in Materials and Methods while the root 438
metal content and concentration of Fe mobilizing substance (FeMS) were acquired from Shenker et al., 439
submitted to JEQ. 2’-DMA, 2’-deoxymugineic acid; MES, morpholinoethane sulfonate. Statistical 440
analysis was performed as in Figure 5. 441
Page: 4 [DC1] Do we have a reference for this? Perhaps one of Parkers papers on chelator buffering. Otherwise not important unless the reviewer ask for more documentation. Page: 8 [DC2] This is highly speculative since lactate is a very weak ligand. Nor am I aware that alanine and GAB have high affinity for iron as compared to protons or other ions such as sodium and calcium. This statement could be controversial to those who use Geochem to speciate soil solutions. Eg Parker would strongly disagree with their importance. The only way to tell would be to do some empirical test and use them as soil extractants. Page: 8 [DC3] Again, the role of malate in metal uptake is speculative and there is no evidence that this compound would be effective. An article by Jones and Darrah suggest that malate would be of little importance: See: DL Jones, and PR Darrah, Role of root derived organic acids in the mobilization of nutrients from the rhizosphere. Plant and Soil 166:247-257, 1994. Page: 9 [DC4] This is an interesting idea, but is there any prior literature suggesting that amino acids bind Cd or are involved in its uptake?
N
OH O OH O
NH
OH O
OH23 4
1'
2'3'
1"
2"3"
3"-DMA3-DMA
2"-DMA2'-DMA
1'-D
MA
2-DMA
4-D
MA
1"-DMA
3'-DMAN OH2C
CH2S-O
O
O
A
B
A
AB
5.0 4.0 3.0 2.04.5 3.5 2.5
1H Chemical Shift (ppm)
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Fig. 1A
2"-DMA2'-DMA
3-DMA
1'-DMA
4-DMA3'-DMA
1"-DMA
N
OH O OH O
NH
OH O
OH23 4
1'
2'3'
1"
2"3"
3"-DMA
2-DMA
1'-DMA 3-DMA2"-DMA
2-D
MA3"-DMA
4-DMA2'-DMA
3'-DMA
1"-DMA
70 65 60 55 50 45 40 35 30 25 20
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
13C Chemical Shift (ppm)
1H C
hem
ical Sh
ift (pp
m)
Fig. 1B
N OH2C
CH2S-O
O
O
A
B
A
CA CB
CA
1'-DMA 3-DMA2"-DMA
2-D
MA3"-DMA
4-DMA2'-DMA
3'-DMA
1"-DMA
3"-DMA
2-DMA
4-DMA
2"-DMA
1"-DMA
3-DMA
2'-DMA
3'-DMA
1'-DMA
70 65 60 55 50 45 40 35 30 25 2013C Chemical Shift (ppm)
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1H C
hem
ical Sh
ift (pp
m)
Fig. 1C
N
OH O OH O
NH
OH O
OH23 4
1'
2'3'
1"
2"3"
4"4'1
N OH2C
CH2S-O
O
O
A
B
A
CA CB
CA
2'-DMA
2'-DMA+NaMES
MES+Na
100
90
80
70
60
50
40
30
20
10
0
M/Z
Rel
ativ
e A
bu
nd
ance
100 150 200 250 300 350 400 450 500
MES dimer+Na
Fig. 2
1.01.52.02.53.03.54.04.55.0
1H Chemical Shift (ppm)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1.0
Ala
lac
acetate
Valmalate
citrate
GB
βGlcαGlc
αTrehalose
123
45
6
7
89
Gly
Leu Ile
Fig. 3
1H Chemical Shift (ppm)5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0
AgCS Wheat Exudate
2'-DMA2'-DMA 2'-DMA
N OH2C
CH2S-O
O
O
A
B
A
MES
MES
Unknown
CS Wheat Exudate
acetate
Gly
glycinebetaine
lactateAla2'-DMA
2'-DMA2'-DMA
Valmalate
succinate
IleLeu
2'-DMA
Barley Exudate
lactateAla
glycinebetaine
GAB
GAB
Val
succinate
MAs
malate
Thr
2'-DMA
MA
epi-OHMAMAs
Glu
malate
R = H, 2'-DMAR = OH, MA
COO-
CC
NC
NH
COO-
CC
COO-
OHH
R'
HHH H
H HH H
1
23
4
1'
2'
3'
1"
2"
3"CH R
R = R’= OH,
3- epi-OH-MA
MES
5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0
1H Chemical Shift (ppm)
Rice Exudate
lactate
Gly
acetate
Fig. 4
Mn vs Alanine
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
bab
b
a
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
0
100
200
300
400
500alanine
Mn
a
a
aa
Cd vs Malate
0.0
0.2
0.4
0.6
0
50
100
150
200
250malate
Cd
Fe vs 3-epi-OHMA
0.0
2.0
4.0
6.0
8.0
10.0
0
20
40
60
80
1003-epi-OH-MA
Fe
Fe vs MA
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0
20
40
60
80
100MA
Fe
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
a
b b
a
b b
a
a
10
20
30
0
Fe vs Lactate
0.0
0.8
1.6
2.4
3.2
4.0
0
20
40
60
80
100lactate
Fe
Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Cd vs Threonine
0.0
0.2
0.4
0.6
0
50
100
150
200
250threonine
Cda
a
b b
b
a
bb
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
mo
le E
xud
ate/
g r
oo
t d
ry w
t.
g M
etal
/g r
oo
t d
ry w
t.
b
a
b
a
aa
b b
a
b b
a
a
a
a
a
a
a
b
a
b b
a
a
Fig. 5
Zn vs 3-epi-OHMA
0.0
2.0
4.0
6.0
8.0 3-epi-OH-MA
Zn
ab
b
a
ab
a
b b
a
Cu vs MA Sum
0
20
40
60
80
100
120MA sum
Cu
a
bb
a
b b
0.0
2.0
4.0
6.0
8.0
10.0 a
a
Fe vs 2'-DMA
Zn vs 2'-DMA
Cu vs 2'-DMA120
Cd vs 2'-DMA
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Mn vs MESMES
Mn
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
mo
le E
xud
ate/
g r
oo
t d
ry w
t.
g M
etal
/g r
oo
t d
ry w
t.
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
Fe vs FeMS
0
6
12
18
24
30
0
20
40
60
80
100Total FeMS
Fe
Fe / Cd Fe / Cd Fe / Cd Fe / Cd
a
b
c
a
bb
a
b
a
b
a
b
a
b
a
b
a
b
a
b
0
40
80
120
160
200
240
0
100
200
300
400
500
a
b
b
a
bb
a
b
a
b
0
6
12
18
24
30
0
20
40
60
80
1002'-DMA
Fe
c
a
c
b
0
6
12
18
24
30
0
8
16
242'-DMA
Zna
b
a
b
c
bab
a
0
6
12
18
24
30
0
10
20
30
2'-DMA
Cua
ab
0
6
12
18
24
30
0
100
200
300
400
500
6002'-DMA
Cd
a a
b b
Fig. 6