Plant and Soil 265: 225–242, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.
225
Nickel Localization and Response to Increasing Ni Soil Levels in Leaves of the Ni Hyperaccumulator Alyssum murale C. Leigh Broadhurst1,5 , Rufus L. Chaney1 , J. Scott Angle2 , Eric F. Erbe3 & Timothy K. Maugel4 1 Animal
Manure and Byproducts Laboratory, Animal and Natural Resources Institute, Bldg. 007, US Department of Agriculture Henry A. Wallace Agricultural Research Center, Beltsville, MD 20705 USA. 2 Agricultural Experiment Station, University of Maryland, College Park, MD 20742; USA. 3 Electron Microscopy Unit, Soybean Genomics and Improvement Laboratory, Plant Sciences Institute, USDA Beltsville. 4 Laboratory for Biological Ultrastructure, Department of Biology, University of Maryland, College Park. 5 Corresponding author∗ Received 22 May 2003. Accepted in revised form 10 February 2004
Key words: Alyssum, hyperaccumulator, nickel localization, phytoremediation, trichomes, SEM-EDX
Abstract We have previously developed phytoremediation and phytomining technologies employing Alyssum Ni hyperaccumulators to quantitatively extract Ni from soils. Implementation of these technologies requires knowledge of Ni localization patterns for the Alyssum species/ecotypes of interest under realistic growth conditions. We investigated Ni uptake and localization in mature Alyssum murale ‘Kotodesh’ and ‘AJ9’ leaves. Seedlings were grown in potting mix with an increasing series of NiSO4 addition (0, 5, 10, 20, 40, 80 mmol Ni kg−1 ), NiC4 H6 O4 addition (0, 5, 10, 30, 60, 90 mmol Ni kg−1 ), in Ni-contaminated soil from metal refining operations, and serpentine soil. Plants at Ni levels 0, 5, 10, 20 mmolkg−1 and in native soils grew normally. Plants at 40 mmolkg−1 exhibited the onset of phytotoxicity, and 60, 80, and 90 mmolkg−1 were demonstrably phytotoxic, but symptoms of phytotoxicity abated within 6 months. Cryogenic complement fractures were made from frozen hydrated samples. High-resolution scanning electron microscope (SEM) images were taken of one half. The other half was freeze-dried and examined with SEM and semi-quantitative energy dispersive x-ray analysis. Ni was highly concentrated in epidermal cell vacuoles and Ni and S counts showed a positive correlation. Trichome pedicles and the epidermal tissue from which the trichome grows were primary Ni compartments, but Ni was not distributed throughout trichomes. Palisade and spongy mesophyll and guard/substomatal cells contained lesser Ni concentrations but palisade mesophyll was an increasingly important compartment as Ni soil levels increased. Ni was virtually excluded from vascular tissue and trichome rays.
Introduction Over the past 8 years our research consortium led by the USDA Agricultural Research Service has developed and demonstrated commercially feasible phytoremediation and phytomining technologies that can potentially clean up Ni-contaminated soils (Chaney et al. 1999; Li et al. 2003a, b). The technology employs two Ni-hyperaccumulating species, Alyssum murale and Alyssum corsicum to quantitatively extract Ni from serpentine soils that are naturally ∗ FAX No: 301-504-5031. E-mail:
[email protected]
rich in Ni. These species are endemic to serpentine soils throughout Mediterranean Southern Europe. Unlike many serpentine-endemic species, A. murale and A. corsicum can grow prolifically and hyperaccumulate Ni in other soil types, such as limestone soils, organic soils, and loam. Successful development of hyperaccumulator species for large-scale phytoremediation/phytomining requires knowledge of Ni localization patterns for each genus/species/ecotype of interest, and for a variety of realistic growth conditions. Microscopy techniques that combine scanning electron microscopy (SEM) with energy dispersive x-ray analysis (EDX) have
226 been successfully utilized to identify sites of metal localization in Alyssum and other hyperaccumulator plants. It is clear from the limited previous research that a range of cell types and structures are involved in heavy metal compartmentation in hyperaccumulators, and localization data from one species may not be applicable to another. However, a wide range of techniques have been utilized rather than a standard procedure, therefore it is expected that localization data will be somewhat contradictory. Cryogenic SEM/EDX techniques that use frozen bulk samples may be ideal since there is minimal processing of the samples and cell types and structures can be easily identified (Echlin, 1986; Van Steveninck and Van Steveninck, 1991; Wergin et al., 1999). Similar techniques developed to identify pollutant-derived heavy metal localization in aquatic species are also relevant (Pedersen et al., 1981; Nott, 1991). Ni enrichment is typically greatest in the leaves of hyperaccumulators, therefore leaf tissue has been the subject of most previous investigations. Proton or nuclear microprobe analyses of Ni hyperaccumulators collected from native South African ultramafic soils found Ni concentrated in leaf mesophyll and epidermis in Berkheya coddii (Mesjasz-Przybylowicz et al., 2001a), leaf epidermis in Senecio coronatus (Mesjasz-Przybylowicz et al., 1994) and throughout the leaf but increasing in concentration from mesophyll to epidermis in Senecio anomalochrous (Mesjasz-Przybylowicz et al., 2001b). For B. coddii grown hydroponically at varying nontoxic levels of Ni, the greatest concentration of Ni was in the apoplast of the upper epidermis, particularly in the cuticle (Robinson et al., 2003). The proton-induced x-ray fluorescence mapping technique utilized by MesjaszPrzybylowicz et al. (2001b) for S. anomalochrous is particularly sensitive, and yielded a range of dry weight Ni concentrations from 1000 µg g−1 in vascular bundles to 9000 µg g−1 in epidermis. Heath et al. (1997) examined the epidermis of a dimethylglyoxime fixed critical point dried Thlaspi montanum ‘Siskiyouense’ leaf from native Oregon ultramafic soil. Energy dispersive x-ray analysis analysis showed that Ni was not in elongate epidermal cells or guard cells–it was only concentrated in subsidiary cells. Freeze-substitution and EDX analysis of Hybanthus floribundus and Stackhousia tryonii collected from native Australian ultramafic soils found that Ni is concentrated specifically in the leaf cuticle in S. tryonii. For H. floribundus, Ni was distributed
homogeneously in high concentrations in the vacuoles of epidermal cells and neighboring palisade and mesophyll cells (Bidwell, 2000). Psaras et al. (2000) presented x-ray maps of leaf epidermis which showed no Ni whatsoever in the trichomes or guard cells of eight hyperaccumulator species collected from native Greek ultramafic soils. Four of these were Alyssum: A. lesbiacum, A. smolikanum, A. heldreichii, A. euboeum. One leaf cross section (Thlaspi pindicum) showed that Ni was strongly concentrated in the upper and lower epidermis. Marmiroli et al. (2002) studied localization of Ni in the hyperaccumulator Alyssum bertolonii collected from native Tuscan ultramafic soil or normal garden soil and oven dried with EDX analysis. Alyssum bertolonii root cross sections showed strong Ni concentration in parenchyma and sclerenchyma. Stem cross sections showed Ni concentrated in epidermal tissues vs. interior tissues, but leaf data were inconclusive. There was no tissue-type specificity for Ni in the plants grown in garden soil. Krämer et al. (2000) examined Thlaspi goesingense (hyperaccumulator) and T. arvense (nonhyperaccumulator) plants grown hydroponically in Ni-rich solutions. Radioactive Ni was given to the plants for 1 d, and counts of the vacuolar isolate indicated that about 75% of the total Ni was in the vacuoles of T. goesingense, with considerably less in the vacuoles of T. arvense. However, the Ni concentrations used in the growth medium were low enough to be nontoxic to T. arvense, and the exposure was shortterm–insufficient time for Ni redistribution within the plant. Only one Ni hyperaccumulator study to date has used plants grown in Ni-rich soil under controlled conditions. Küpper et al. (2001) grew T. goesingense, A. lesbiacum, and A. bertolonii in compost with 0, 500, 2000, or 4000 mg kg−1 Ni for 64 or 84 days. Bulk frozen hydrated samples were examined by cryogenic SEM/EDX analysis. X-ray mapping of leaf cross sections from the 4000 mg kg−1 Ni treated A. lesbiacum and A. bertolonii plants showed that Ni was strongly localized in the epidermal cells. Nickel was mainly in epidermal cell interiors, with some in epidermal cell walls. There was very little Ni in the mesophyll cells or the vascular bundle. Thlaspi goesingense failed to survive at 4000 mg kg−1 Ni, so the 2000 mg kg−1 plant was examined. Nickel localization followed the same pattern as seen for Alyssum, however the total concentration in the shoot was about half that of either Alyssum species.
227 A characteristic feature of Alyssum (and Arabidopsis halleri) is the presence of numerous trichomes on leaf upper and lower epidermal sufaces. There is no consensus as to whether trichomes are a storage location for hyperacummulated metals (Krämer et al., 1997; Küpper et al., 2000; Psaras et al., 2000; Zhao et al., 2000; Marmiroli et al., 2002). It remains to be determined whether the interior of trichomes is a location for metal sequestration, or whether only the point of attachment is metal-rich, reflecting conditions at the interface between the trichome and the epidermis proper. We have designed and implemented an experimental program for identifying Ni localization sites in the leaves of Alyssum hyperaccumulator species. Plants are grown long-term in soils as opposed to short-term hydroponically, and the species of interest serve as their own controls. Preparatory procedures that can relocate or remove Ni from plant tissues are avoided.
Materials and methods Horticulture Alyssum murale ‘Kotodesh’ and ‘AJ9’ were started from seed 28/08/02 in a greenhouse at USDA Beltsville. Twenty-one days later ‘Kotodesh’ seedlings were transplanted to prepared and equilibrated 250 g pots containing Promix soil with an increasing series of NiSO4 ·6H2 O addition (0, 5, 10, 20, 40, 80 mmol Ni kg−1 ; these are designated in the text by NI 00, NI 5, NI 10, NI 20, NI 40, and NI 80 respectively.) Carbonates (half CaCO3 and half MgCO3 ) were added to each pot at amounts equimolar with NiSO4 . ‘Kotodesh’ and ‘AJ9’ Seedlings were also transplanted into calcareous mineral soil associated with nickel mining tailings from Inco Ltd. operations at Port Colborne, Ontario (Welland soil, Typic Epiaquoll; Canadian classification, Terric Mesisol with 30 gkg−1 CaCO3 added). The Inco soil typically yields 1% Ni in Alyssum whole shoots and 1.7% in leaves, and would be phytotoxic to normal plants (Kukier and Chaney, 2001). There were duplicate plants for each treatment. Based on preliminary results, a second group of ‘Kotodesh’ seedlings were started and transplanted after 21 days on 05/03/03 to prepared and equilibrated 250 g pots containing Promix soil with an increasing series of NiC4 H6 O4 ·5H2 O addition (0, 5, 10, 30,
60, 90 mmol Ni kg−1 ), and equimolar carbonates as above. ‘Kotodesh’ and ‘AJ9’ seedlings were also transplanted on 17/03/03 into natural Brockman variant serpentine soil from Josephine Co., Oregon (Typic Xerochrepts) with 10 wt% Promix added to improve drainage. The experiments were conducted in a greenhouse under controlled temperature and light conditions and ambient humidity. Photoperiod was 15/9 day/night. During this time high-intensity sodium and incandescent lights supplying 400 µmol m−2 s−1 supplemented sunlight if necessary. Daytime temperature was 24 ◦ C with cooling initiated at 27 ◦ C. Night temperature was 18 ◦ C with cooling initiated at 21 ◦ C. Plants were grown in freely drained plastic pots with saucers to prevent loss of leachate, and watered with deionized water. The plants received standard fertilization (Miracle Grow) once per month, and MgSO4 (1.5 gL−1 ) bimonthly for the NiSO4 series and Inco plants. The NiC4 H6 O4 plants received 1.5 g supplemental MgCO3 and the serpentine soil plants did not receive supplemental Mg. Magnesium supplementation was initiated because there are 7 species of Alyssum hyperaccumulators under cultivation in the greenhouse, and some showed signs of Mg deficiency, probably due to being adapted to high-Mg serpentine soils. Total metals analysis Two centimeter lengths of apical stem tip from three locations on each plant were harvested on 25/10/02 (NiSO4 and ‘Kotodesh’ Inco) and 17/06/03 (NiC7 H6 O4 and ‘AJ9’ Inco) for total metal analysis. The plants grown in the serpentine soil had not achieved enough biomass to be harvested without harming them. Only mature growth was used, and leaves were stripped from the stem. Harvested leaves were rinsed in deionized water to remove any adhering soil particles. Leaf samples for each treatment were dried for 24 h at 60 ◦ C, weighed, and ashed in a 480 ◦ C oven for 16 h. After cooling, the ash was digested with 2 mL concentrated HNO3 , swirled and taken to dryness. The sample was then dissolved in 10 mL 3 N HCl, filtered through Whatman #40 filter paper and brought to volume in a 25 mL volumetric flask using 0.1 N HCl (final concentration 1 N HCl). For quality control, reagent blanks and an in-house Alyssum standard were included. Calcium, Cu, Mg, Mn, Ni, and Zn were determined by inductively-coupled plasma atomic emission spectrometry (Perkin-Elmer Optima
228 4300 DV) using 40 mgL−1 yttrium as an internal standard. Sample preparation and electron microscopy At the USDA Beltsville Electron Microscopy Facility we have developed low temperature scanning electron microscopy techniques for a wide range of bulk hydrated botanical specimens, including whole leaves (Steere and Erbe, 1979; Wergin et al., 1999; Erbe et al., 2003). Bulk samples cannot be rapidly frozen beyond about 20 µm depth regardless of the cryogen utilized because the surface layer of ice that forms initially acts as an insulator (Robards and Sleytr, 1985). High pressure freezing, which has shown better freezing in some samples, can extend the depth of good freezing in some tissues. However, high pressure freezing would significantly damage our fresh leaves because we would have to dissect the leaves prior to freezing, and the samples are exposed to pressure and chemicals (i.e. hexadecene). Liquid nitrogen (LN2) has distinct advantages in terms of ease of use and consistency of results. For bulk tissues, no freezing technique has been shown to be effective in preventing ice crystal formation. Given that all bulk freezing methods are relatively poor, we have found over more than 20 years of experience that freezing on a precooled metal surface in a large volume of LN2 produces very consistent freezing throughout a given sample. We do not claim that our freezing technique is superior or even good, only that it is consistent (Yaklich et al., 1996, 1999, 2001). We also utilized this technique because it does not require any dissection of leaves, nor does it subject the samples to croypreservation treatments that could potentially leach or redistribute Ni, and most importantly break or damage trichomes. In the worst case scenario, our freezing technique would result in the movement of solutes within cells or to adjacent cells, not throughout the leaf sample. Our previous experience with handling delicate trichomes has shown that rapid plunge freezing on copper stubs allows structures and details to be observed that would not be present after chemical fixation (Wergin et al., 1999; Erbe et al., 2003). Mature leaves near the apical tip of a stem for each NiSO4 addition level were harvested on 20/10/02, after the plants had been growing in high-Ni soil for a month. The Ni level 80 plant was harvested again on 13/12/02 for a follow-up analysis. The plants with NiC4 H6 O4 addition levels 30 and and
60 mmol Ni kg−1 (designated in the text by NiAc 30 and NiAc 60) and ‘AJ9’ Inco were harvested on 12/06/03. We did not examine all the Ni acetate levels at this time because we were also examining other Alyssum hyperaccumulator species; these data will be reported elsewhere. Fresh plant tissue was placed onto copper metal plate sample holders containing methyl cellulose solution (Tissue Tek ). The sample holders were fashioned from 1.5 mm sheets of stock copper cut into individual 15 mm × 29 mm pieces. The plates were immediately plunged into a reservoir of LN2 which rapidly cools them to −196 ◦ C and firmly attaches the tissue. Flat samples used for SEM examination of trichome density and general physiology were mounted and frozen in the above manner. Cross fractured samples for SEM-EDX analysis study were prepared by a slightly different complement fracture method. A single mature leaf near the apical tip of a stem for each Ni addition level was harvested on 20/10/02. Leaf pieces were placed into Tissue Tek-filled slots (3 mm deep, 1 mm wide) machined in a copper cylindrical stub. The stub was plunge-frozen in LN2. Frozen leaf sections (∼ 6 × 6 mm) were fractured off under LN2, using precooled flat wafer-tipped forceps. Fractured pieces were placed in cryovials under LN2 and documented as exact complements of the leaf halves retained in the slotted stub, and stored at −196 ◦ C until analysis. Upon SEM observation, the slotted stub sample holder was transferred to the preparation chamber of an Oxford CT 1500 HF Cryotrans System and etched at −90 ◦ C to remove contaminating water vapor and tissue water. After reducing the specimen temperature below −110 ◦ C, samples were Pt sputter coated and placed on the precooled (−170 ◦ C) stage of a Hitachi S-4100 field emission scanning electron microscope. A series of high resolution photographs of one half of each complement fracture were taken at 2.0 kV. The other halves of the complement fractures were prepared for x-ray microanalysis by freeze drying in a custom apparatus attached to a vacuum evaporator. The frozen samples were placed into individual screen vials under LN2 and stacked into a welded fitting. The fitting was quickly coupled to the vacuum chamber via flexible hose, and placed into a styrofoam box preloaded with metal ballast cooled with LN2 and packed with dry ice pellets. Under vacuum, the entire apparatus slowly warmed from −196 ◦ C to dry ice temperature (−78 ◦ C), then upwards to
229
Figure 1. SEM-EDX guide photo of freeze dried leaf margin, Alyssum murale ‘Kotodesh’ greenhouse grown in natural serpentine soil. At the leaf margin the cell walls are thick enough to analyze independently. Letters a, b, c, denote locations of sequential spectra in keV vs. counts per 100 ls (a) epidermal cell wall; (b) epidermal cell interior, note vacuolar material; (c) first layer palisade mesophyll interior.
room temperature over the course of one week. Freeze dried samples remained under vacuum in the apparatus at room temperature until mounting for x-ray microanalysis.
X-ray microanalysis Freeze dried samples were transported in their screen vials to the Laboratory for Biological Ultrastructure at the University of Maryland for semiquantitative SEM-EDX analysis. We did not utilize frozen hydrated samples for EDX because significant problems remain with this technique with respect to (1) detec-
230 tion limits and (2) variability in water content among cell/tissue types–both of which can be alleviated by removing water–and (3) reduced x-ray spatial resolution in frozen hydrated bulk samples (Goldstein et al., 1981; Van Steveninck and Van Steveninck, 1991). Further, Al sample coating was not available for cryogenic samples, but proved necessary in order to analyze individual trichomes. Samples were individually mounted on standard aluminum stubs coated with silver adhesive paste. Each complement fracture sample was mounted with the aid of a binocular microscope, and handled with lightweight glass microrods. Samples were placed on edge to view the cross sections that correspond to the complement fracture planes. Before analysis a light coat of carbon was deposited on the samples at a pressure of 13 Pa followed by 20 nm of aluminum deposited at 1.3 × 10−3 Pa. The dual coating provides better dissipation of charge in biological samples, but Al has no detectable influence on x-ray signals, unlike Au or Pt (Reid et al., 1993). The samples were stored in a dessicator until analysis. We used an Amray 1820D scanning electron microscope (KLA Tencor Corp., AMRAY Division, Bedford, MA), an EDAX ECON 4 detector with an active area of 10 mm and an EDAX DX Prime analyzer (EDAX Inc., Mahwah, NJ). The samples were analyzed under the following conditions: 20 kV accelerating voltage, 20◦ tilt, 12 mm working distance, 17.66◦ takeoff angle. Each location was counted for 100 live seconds (ls). Beam spot was 3 to 10 µm depending on the application, but always less than a cell diameter when targeting a specific cell/location. Every location that was analyzed was also digitally photographed with the AMRAY instrument (Figure 1). Figure 1 illustrates the spatial resolution we were able to attain with the x-ray beam Modeling the excitation volume for Ni in a plant matrix with the EDAX software provided indicated that we could easily penetrate cell walls and excite the entire volume of a given cell, and this was borne out in the spectra obtained. We counted k lines for C, O, Mg, Si, P, S, Cl, K, Ca, Mn, and k and l lines for Ni. Data are reported in peak/background count ratios. Ratios at or below unity indicate the element concentration is below detection limits. Iron Cu, Co, and Zn were below detection limits in all samples so are not reported. Magnesium was often below detection limits because the background correction was highest circa 1.5 keV. The Alyssum leaves posed analytical challenges due to the irregular leaf surfaces and the large number
of trichomes. Surface trichomes and epidermal tissue extending into free space had a tendency to charge and often shifted or waved. Recoating and thicker coatings were utilized with these samples and subsequent samples in order to reduce charging with only partial success. In the time span of 100 ls, a location of interest may move completely out of the electron beam, thus potentially allowing Ni from another location to be detected in error. Each time we analyzed a narrow or small diameter feature such as a trichome ray or guard cell we checked for beam drift, and repeated the analysis if necessary. ZAF calculation to 100 wt% on an H and N free basis was utilized as a qualitative guide to Ni concentrations in the set of samples. The Cu K line from a polished stub was the internal calibration standard. The ZAF model is not designed for biological materials, however there is no truly suitable means of handling the high, heterogenenous concentrations of Ni (and Ca in trichomes) our bulk plant samples contained. The ZAF model is applicable freeze dried or freeze substituted biological materials with mineralcontaining deposits if peak/background counts are utilized, and a high degree of accuracy in light element analyses is not critical (Goldstein et al., 1981; Van Steveninck and Van Steveninck, 1991). Aluminum and Ag peaks from the sample mount material were removed from the spectrum prior to calculation. Approximately 2 wt% N was present in the samples, however the intensity of the N peak was below the background correction for the large C and O peaks, so it was not entered into the analysis program. Concave leaf surfaces or holes can trap a certain percentage of low energy x-rays, resulting in some analyses with low C and O. This does not affect the Ni peak/background counts, however the ZAF calculation yields a relative Ni concentration that is too high. In most cases low C and O can be corrected by changing the sample geometry or increasing the beam size, but on occasion a location simply cannot be used.
Results Plant growth A. murale plants at Ni levels 0, 5, 10, 20 mmol kg−1 and in Inco soil grew normally and were eventually transplanted to 500 g pots with the same Ni levels and allowed to grow indefinitely. Plants grown in soil with 40 mmol Ni kg−1 (NI 40) exhibited the onset
231 Table 1. Selected element concentrations in Alyssum murale leaves for Ni addition series (zero to 80 mmol Ni kg−1 ) in dry weight ppm. KD: ‘Kotodesh’; AJ: ‘AJ9’. Letters a and b represent analyses of duplicate plants/treatments. NI designates NiSO4 addition; NiAc designates NiC4 H6 O4 addition. Inco samples are plants grown in contaminated mineral soil from a Ni refinery area in Ontario. Alyssum std. 1, 2, 3 are duplicates of our laboratory high-Ni Alyssum standard Sample
Dry wt. (g)
Ni
Ca
Cu
Mg
Mn
Zn
KD NI 0a KD NI 0b KD NI 5a KD NI 5b KD NiAc 5 KD NI 10a KD NI 10b KD NiAc 10 KD NI 20a KD NI 20b KD NiAc 30a KD NiAc 30b KD NI 40a KD NI 40b KD NiAc 60a KD NiAc 60b KD NI 80a KD NI 80b KD NiAc 90a KD Ni Ac 90b KD Inco a KD Inco b AJ 00 (control) AJ Inco a AJ Inco b Alyssum std. 1 Alyssum std. 2 Alyssum std. 3
0.072 0.157 0.198 0.116 0.923 0.101 0.159 0.640 0.160 0.216 0.430 0.584 0.133 0.108 0.261 0.397 0.044 0.087 0.278 0.462 0.133 0.101 0.967 1.684 1.178 0.150 0.136 0.202
4.60 5.20 2640 4230 9240 10,900 9470 10,800 16,300 13,700 21,200 25,000 17,300 17,000 45,900 30,900 36,800 25,900 30,200 22,400 3040 1670 5.00 3350 3520 15,700 15,900 15,100
18,300 20,700 22,200 25,400 24,900 29,100 33,600 32,200 28,500 24,300 43,200 30,300 32,400 39,700 39,600 37,600 31,000 27,500 36,100 34,100 43,700 42,800 48,400 29,700 29,500 12,200 12,400 11,200
5.77 7.23 6.57 8.74 11.0 6.47 5.39 9.32 5.29 7.80 6.7 10.1 11.4 7.35 5.15 5.44 4.64 6.67 6.59 4.79 25.4 14.2 8.58 7.69 11.4 0.02 0.21 0.28
3890 2740 5080 4010 1560 4110 4550 3110 5340 2960 6680 4820 4140 5820 8000 5420 4210 6150 8670 5270 3220 2500 4690 5005 4130 2080 2120 2060
91.4 154 201 271 483 286 389 550 322 223 388 532 279 223 741 394 386 254 326 368 125 122 295 133 155 23.4 24.2 23.1
131 145 165 400 152 307 199 125 179 474 217 204 1110 452 272 135 266 297 224 85.4 1480 139 55.1 40.4 44.4 60.2 59.0 60.5
of phytotoxicity, with 20% of leaves chlorotic, and leaf habit more blade-like and flattened as opposed to curved. There was no significant reduction in biomass for NI 40, so these plants were also transplanted to 500 g pots. NI 80 was demonstrably phytotoxic, with severe chlorosis, 60–70% reduction in biomass, minaturization of leaves, and loss of prostrate habit. After 3 months the aerial parts of NI 80 plants were nearly dead, but we continued to water them. After 2–3 weeks they regrew healthy, normal size, nonchlorotic leaves. Eventually chlorosis returned to the plants but they are only 50–60% affected, and continue to grow slowly. Plants grown at NiC4 H6 O4 addition levels 0, 5 10, and 30 mmol Ni kg−1 grew normally without signs of
phytoxicity. Both NiC4 H6 O4 60 and 90 mmol Ni kg−1 levels were demonstrably phytotoxic, with severe chlorosis, 50–60% reduction in biomass, minaturization of leaves, and loss of prostrate habit. However, none of the NiC4 H6 O4 addition series plants died and have been allowed to grow indefinitely. By 01/11/03, chlorosis in the NiAc 60 plants was diminished to less than 10% of leaves. Leaf size, plant growth habit, and vigorous growth returned to normal in the NiAc 60 plants so that they became virtually indistinguishable from NiAc 5 and NiAc 10 plants. The NiAc 90 plants have remained 60% chlorotic, with extreme miniaturization of leaves and very slow growth. Plants grown in the Oregon serpentine soil were not chlorotic, but grew more slowly, achieving only 20–
232
Figure 2. (a) Lower epidermis, frozen hydrated Alyssum murale ‘Kotodesh’ leaf, NI 00 (control). Note all following figures are frozen hydrated leaf. (b) Upper epidermis, NI 00. (c) Lower epidermis, NI 80. Trichome density is greatly reduced in this chlorotic leaf. (d) Upper epidermis, NI 80. Leaf from second growth (regeneration) of plant. Trichome density remains reduced although this leaf is not chlorotic.
25% of the biomass of the NiAc 0, 5, 10, and 30 plants after 9 months. Trichomes and trichome density The leaves of all the Alyssum species/ecotypes we grew are covered with an overlapping network of stellate trichomes on both the upper and lower epidermis (Figure 2). Each trichome has 8 to 14 elongate rays, some of which bifurcate. The upper sides of the rays are covered with hemispherical nodules, but the underside is smooth. The trichomes are attached to the epidermis with a roughly 20 µm cylindrical pedicle which is also smooth. For A. murale ‘Kotodesh’ we have recently utilized transmission electron microscopy to determine that the epidermal compartment from which the trichome grows and the trichome it-
self consist of a single large cell (Broadhurst, Chaney, Angle, Erbe, Maugel, CA Murphy in preparation). The trichome densities for A. murale ‘Kotodesh’ and ‘AJ9’ were greater on the lower epidermis. Figures 2a and 2b show the lower and upper epidermis of control (NI 00) ‘Kotodesh’ plants. At strongly phytotoxic Ni levels (NI 80, NiAc 90), the trichome density was greatly reduced, with no overlapping trichomes (Figure 2c). The trichome density remained low when the NI 80 plants regenerated, although the new leaves had not yet developed severe chlorosis (Figure 2d). The low trichome density in the NI 80 and NiAc 90 plants remained unchanged after 9 to 12 months. Metal concentrations Total Ni levels in the whole leaves increased positively with Ni addition levels to a maximum of 36,800 ppm
233 (Table 1). Leaf Ni concentrations from Inco soil plants were comparable to those achieved with NI 5. However, the Ni concentration in the Inco soil is reported at 2.9 g kg−1 (Chambers et al., 1998; Kukier and Chaney, 2001), even higher than our NI 40 concentration (2.4 g kg−1). Manganese and Zn levels were elevated throughout both Ni addition series as compared to control and standards, but do not increase with increasing Ni. Semiquantitative SEM-EDX analysis Semiquantitative SEM-EDX analyses of the complement fractured leaves are given in Tables 2 and 3 and Figures 1 and 3–8. The entire data set represents a comprehensive survey of cell/tissue types. For each Ni level we analyzed upper epidermal, mesophyll, and lower epidermal tissue. When a sample permitted, we analyzed multiple locations of the same cell/tissue type. Additionally, individual samples had certain secondary features that could be accessed easily or had favorable geometry. We analyzed trichomes, trichome pedicles, vascular tissue, cell wall, and stomates in the samples that fortuitously presented analytical opportunities. It was not possible to accurately analyze these features in every sample because their presence is dependent on the leaf fracture and the sample geometry. The cryogenic complement fracture method cleanly shears cells, or removes discs of cell membrane so that cell interiors can be accessed (c.f. Figures 4 and 5). At all levels Ni was mainly concentrated in upper and lower epidermal cells, but was also somewhat distributed throughout the mesophyll. Upper and lower epidermal cell interiors often contained roughly spherical tissue masses which were almost certainly vacuoles. For example, in the NiAc 60 and Inco samples (Table 2), upper epidermal cells in which the x-ray beam was focused directly on a large vacuole are compared to cells in which the vacuole was small or mostly absent. One advantage of freeze-drying is that it clearly defines the vacuolar tissue. When the vacuolar tissue mass was not present in an epidermal cell interior, the Ni counts were reduced. We did not observe the vacuolar material in any cell type except epidermal. A positive correlation of elevated Ni and S counts was observed throughout the analyses (Table 2 and Figure 3–8), and was very robust for epidermal cell vacuoles (r = 0.97, P < 0.001). Since the possibility existed that additional S was provided by NiSO4 or MgSO4 we started the NiC4 H6 O4 series and
serpentine plants. The Ni and S correlation was also observed in the plants grown with nickel acetate and in serpentine soil. We have observed this Ni/S correlation in other Alyssum hyperaccumulator species as well. Palisade mesophyll cells directly adjacent to the upper epidermis tended to contain more Ni than the second layer. At most levels Ni was was not detected in spongy mesophyll (Table 3). Ni was present in all the guard cells and associated substomatal cells we analyzed, but there were not many guard cells accessible in the cross-sectioned samples. Nickel was virtually excluded from vascular tissue. The cell walls of A. murale ‘Kotodesh’ are so thin that in almost all cases they could not be analyzed independent of the cell interiors. The majority of analyses are of cell interiors; however, occasionally we directed the beam through an intact cell wall into the cell interior to determine if there was a dramatic difference in Ni counts. Without exception there was no significant difference in Ni counts by including cell wall. In order to further clarify this issue we looked at leaf margin epidermal cells, where the cell wall was thicker. Figure 1 shows spectra from the cell interior and cell wall, illustrating that Ni is mainly concentrated in the cell interior. We specifically focused on trichomes and trichome attachment points, analyzing the following locations: (1) trichome rays on the upper and lower surface, both proximal and distal to the attachment point. This included (1) nodules on the upper surfaces of trichome rays (see Figures 2 and 3); (2) trichome pedicles, both attached and broken free; (3) trichome rays broken open so that the interior could be accessed; (4) epidermal compartments from which the trichome grows (trichome basal compartment); (5) intact trichome rays in which we could be certain the x-ray beam completely penetrated and passed through the ray. This last type of trichome ray extended into free space so it could be determined that the beam penetrated through the entire trichome ray but could not have hit the leaf elsewhere. Beam drift was strictly controlled, and the appearance of Ag in the spectrum (from the silver paste mounting substrate) showed that the beam fully penetrated (i.e. ‘punched through’), which indicates that the excitation volume included the entire distal ray, not just the surface. We found that the trichome basal compartment, trichome pedicle, and the epidermal cells adjacent to the trichome basal compartment strongly concentrate Ni, but there was no appreciable Ni in the rays or nodules, or within an inner core inside the trichome. Ni x-ray counts de-
234
Figure 3. Entire leaf cross section at margin, oblique view. Selected EDX-SEM x-ray spectra to left of each image. Upper spectrum: NI 5 guard cell on lower epidermis. Lower spectrum: NI 40 trichome basal compartment that accumulated Ni and Mn.
Figure 4. Upper epidermis and several layers of palisade mesophyll. Spectra from uppermost down: NI 80 upper epidermal cell interior; NI 40 upper epidermal cell interior; NI 5 palisade mesophyll cell, first layer (directly adjacent to epidermis); NI 5 palisade mesophyll cell, second layer
Figure 5. Entire leaf cross section at margin, oblique view. Selected EDX-SEM x-ray spectra to left of each image. Upper spectrum: NI 5 guard cell on lower epidermis. Lower spectrum: NI 40 trichome basal compartment that accumulated Ni and Mn.
235
Figure 6. Entire leaf cross section at midrib. Upper spectrum: NI 20 epidermal cell interior. Cell is directly adjacent to trichome pedicle. Lower spectrum: NI 10 trichome surface nodule.
Figure 7. Spongy mesophyll and vascular tissue. Upper spectrum: NI 40 spongy mesophyll interior. Lower spectrum: NI 40 vascular tissue.
Figure 8. Lower epidermis showing trichome broken off at pedicle and broken rays, allowing access to trichome interior. Upper spectrum: NI 00 trichome pedicle. Lower spectrum: NI 10 trichome pedicle.
236 Table 2. Semiquantitative SEM-EDX analysis of Ni and other selected elements in various cell/tissue types in Alyssum murale freeze dried leaf. All samples are ‘Kotodesh’ except Inco (‘AJ9’). Data are peak/background count ratios for 100 ls. Values at or below unity indicate nondetection (–). Mean and standard error (> 3 analyses) for different cells/locations given when possible; number of analyses averaged in parentheses. Note peak and background counts for the suite of elements analyzed are given in the spectra. First layer palisade adjacent to epidermis. +CW indicates analysis included cell wall and cell interior; LE: lower epidermis; UE: upper epidermis, VT: vascular tissue X-Ray Spectral Line Cell Type/Location
Ni K
Ni L
SK
KK
Ca K
Mg K
6.0 4.9 15 14 8.4 – 18
1.1 4.9 1.8 – 1.5 18 7.6
17 ± 2.6 14 6.7 8.4 18 ± 4.4 15 16 ± 2.2 4.3 10 11 7.0 ± 0.25
3.2 ± 2.8 – 3.4 – 5.6 – 9.9 – 2.0 ± 0.18 1.7 ± 0.27 – – 6.7 ± 4.3 – 23 1.5 2.5 – 9.1 1.5 29 ± 4.5 –
15 ± 2.4 19 20 13 11 8.7 30 ± 4.2 – 8.9 ± 1.5 4.35 7.0
– – – 2.7 – – – 31 ± 4.4 8.3 ± 2.1 18 18
0 mmolkg−1 Ni (control)
– upper epidermal 1.5 – 1.2 palisade mesophyll – – 1.8 spongy mesophyll – – 3.1 lower epidermal – – 1.5 vascular tissue – – 1.4 trichome nodule 1.4 – – trichome pedicle – – 4.1 5 mmolkg−1 Ni upper epidermal (5) 6.6 ± 2.0 1.8 ± 0.5 6.3 ± 1.9 7.1 2.3 5.8 palisade mesophyll, 1st palisade mesophyll, 2nd – – 2.6 spongy mesophyll – – 3.5 lower epidermal (3) 7.8 ± 1.2 3.7 ± 1.3 5.7 ± 1.9 vascular tissue (2) – – 1.4 guard and substomatal cells (3) 2.5 ± 0.34 – 4.2 ± 1.5 trichome nodule (2) – – – trichome pedicle 11 1.3 3.4 trichome ray, proximala 6.1 1.7 3.5 trichome ray, distalb (3) – – 1.5 ± 0.08 10 mmolkg−1 Ni upper epidermalc (3) 2.5 ± 0.5 – 6.3 ± 1.3 upper epidermal+cw (2) 4.0 1.5 7.6 palisade mesophyll, 1st (2) 1.9 – 5.0 palisade mesophyll, 2nd – – 3.2 palisade mesophyll, 2nd +cw – – 2.5 spongy mesophyll – – 1.5 lower epidermal (3) 6.2 ± 0.38 – 11 ± 8.0 trichome nodule (3) – – – trichome pedicle, proximala (3) 7.5 ± 3.6 2.2 ± 1.7d 2.3 ± 0.90 trichome pedicle, distal 2.2 – – trichome ray – – 1.8
crease as one moves distally from the trichome point of attachment. Calcium is strongly concentrated in trichomes, particularly on the upper surface of rays and nodules. Trichome nodule analyses detected mainly C, O, and Ca. On occasion Mn was strongly concentrated along with Ni in trichome pedicles and basal compartments although there was no enrichment of Mn in the experimental soils. In the ‘Kotodesh’ plant grown in
1.2 2.4 3.5 – 1.5 2.8 1.4
– – – – – – – – – – –
serpentine soil, we observed a sequential enrichment of Mn, Mn + Ni, Ni, Ni + Ca, and Ca as we moved from the floor of a trichome basal compartment up through the compartment into the pedicle. We analyzed cells directly adjacent to some of the high-Mn cells in order to determine if there was a contamination problem, and to see if our sample preparation was redistributing metals. To the contrary we found that Mn was highly localized in specific areas of trich-
237
Table 2. Continued. X-Ray Spectral Line Cell Type/Location 20 mmolkg−1 Ni upper epidermal (3) palisade mesophyll, 1st palisade mesophyll, 2nd +cw spongy mesophyll + cw lower epidermal (3) trichome nodule trichome pediclee (2) trichome basal compart. UE (2) trichome ray, distala (3) 30 mmolkg−1 Ni (NiAc) upper epidermal (3) palisade mesophyll, 1st +cw (2) palisade mesophyll, 2nd +cw spongy mesophyll, adj. VT (4) spongy mesophyll + cw lower epidermal vascular tissue (3) trichome pedicle trichome basal compart. LEf (3) 40 mmolkg−1 Ni upper epidermal (4) p. mes. 1st adj. hi Mng (3) palisade mesophyll, 1st (2) palisade mesophyll, 1st + cw (2) palisade mesophyll 3rd spongy mesophyll (2) lower epidermal (4) vascular tissue (4) guard cell (4) trichome nodule trichome bsl. compart. UEg (3) 60 mmolkg−1 Ni (NiAc) UE, large vacuoles (5) UE, small or no vacuoles (5) UE, adj. high Mnh (2) epidermis, lateral margin palisade mesophyll, 1st (2) p. mes., 1st , adj. hi Mnh (2) vascular tissue (3) trichome bsl. compart. UEh (3)
Ni K
Ni L
SK
14 ± 2.5 1.5 – – 17 ± 2.4 – 5.0 21 –
3.3 ± 1.7 – – – 7.9 ± 1.0 – 1.2 5.6 –
11 ± 5.4 17 ± 8.2 5.2 8.9 2.1 6.7 8.2 8.2 7.5 ± 0.68 7.3 ± 2.5 – – 3.9 7.7 14 8.5 – 3.0 ± 0.32
14 ± 0.92 1.8 5.0 – 1.5 19 1.2 ± 0.14 4.7 23 ± 4.5
3.5 ± 0.73 11 ± 5.4 – 1.8 1.8 2.7 – – – – 1.3 6.8 – – 1.3 – 4.0 ± 3.4 8.1 ± 1.6
18 ± 4.1 7.1 ± 3.9 7.8 7.3 – 11 16 ± 2.7 4.1 ± 1.5 5.6 ± 0.70 – 3.5
5.2 ± 0.66 11 ± 0.1.0 10 ± 3.8 3.3 ± 1.7 7.0 ± 0.96 4.1 ± 0.52 2.7 9.4 4.2 1.8 9.0 2.9 – 3.74 20 3.5 9.2 2.8 5.6 ± 0.63 13 ± 5.7 5.8 ± 2.6 1.6 ± 0.9d 3.6 ± 1.1 10 ± 4.3 3.0 ± 0.44 5.3 ± 0.86 13 ± 3.8 – – 1.7 2.6 4.2 13
21 ± 1.3 7.5 ± 2.5 2.6 12 4.2 5.0 – 3.7 ± 0.74
4.8 ± 0.74 2.7 ± 1.4 – – 1.8 1.2 – –
9.6 ± 0.74 5.1 ± 0.77 4.3 8.8 2.2 2.6 1.3 ± 0.03 3.6 ± 1.8
KK
5.6 ± 2.1 12 14 17 ± 0.94 6.4 5.5 12 ± 7.9 3.8 3.9 ± 2.6
8.7 ± 2.1 9.7 ± 3.2 13 13 13 15 10 ± 2.1 10 ± 4.1
Ca K 2.1 ± 0.60 7.7 9.6 11 – 25 13 4.0 39 ± 10
Mg K
– 1.5 1.7 4.1 – 4.0 – – –
3.3 ± 0.92 2.0 ± 0.94 – 1.9 – – 2.6 – – – 5.4 2.5 1.7 ± 0.79 nd
– – – –
3.0 ± 0.74 11 ± 1.0 17 14 1.3 14 3.0 ± 1.7 1.6 ± 0.44 2.0 ± 0.59 35 2.2
– 2.2 ± 0.47 1.5 – 4.1 3.7 – – – 1.9 –
2.9 ± 0.89 2.1 ± 0.29 2.4 1.8 7.3 7.8 1.7 ± 0.35 2.7 ± 1.1
1.5 ± 0.46 1.3 ± 0.21 1.2 1.2 2.1 1.9 – –
238
Table 2. Continued. X-Ray Spectral Line Cell Type/Location
Ni K
Ni L
SK
KK
17 ± 1.7 1.6 ± 0.40 1.2 3.3 2.7 16 – 5.3 2.8
4.6 ± 2.9 9.9 ± 3.8 12 ± 5.1 – 3.2 ± 0.49 13 ± 1.5 – 2.5 9.0 1.8 3.8 8.9 1.3 2.0 5.9 7.2 6.0 6.0 – – 8.6 3.6 1.7 2.9 – – –
Ca K
Mg K
80 mmolkg−1 Ni upper epidermal (5) palisade mesophyll, 1st (3) palisade mesophyll, 3rd spongy mesophyll (2) spongy mesophyll + cw lower epidermal (2) vascular tissue guard cell trichome ray bf serpentine soil upper epidermal (3) epidermis, lateral margin (4) epidermis, lat. mar. cell wall (2) palisade mesophyll, 1st vascular tissue lower epidermal (2) tri. ped. c, prox. to UEi (2) Tri. ped. d, dist. to UEi (2) tri. bsl. compart. b, UEi (2) tri. bsl. compart. a, dist. to UEi Inco soil ‘AJ9’ UE, large vacuoles (3) UE, small vacuoles (2) palisade mesophyll, 1st (2) palisade mesophyll, 2nd spongy mesophyll, 1st spongy mesophyll, 2nd vascular tissue (5) trichome pedicle, LE
8.8 ± 3.2 6.4 ± 1.4 4 2.1
12 ± 2.2 7.2 ± 2.2 3.5 7.4 – 4.9 1.2 – 13 13
– 11 – – 19 –
– – – – – 2.0 – – 2.2 –
18 ± 5.2 5.0 2.0 – 2.1 – – 3.1
2.5 ± 2.2 6.9 ± 0.68 – 2.3 – 3.0 – 3.2 – 2.4 – 1.5 – – 1.5 3.3
8.4 ± 8.0 6.2 ± 2.4 3.0 3.5 6.9 2.9 2.6 2.5 10 11
2.7 ± 0.38 4.1 ± 1.0 1.8 4.2 1.4 3.0 1.2 2.1 27 11 ± 2.1 5.2 ± 1.7 2.4 7.3 3.2 5.8 65 120 9.0 2.6
1.3 ± 0.76d 2.2 ± 1.7d 1.8 4.5 3.7 1.3 5.5 – 5.5 – 6.5 – 7.5 ± 2.6 1.3 ± 0.31 2.9 1.5
– 1.7 ± 0.48 1.5 2.5 2.6 1.3 – – – – 1.6 ± 0.42 – 1.7 – 1.1 1.9 3.9 1.6 3.2 – 1.2 1.6 3.1 1.5 – – 1.7
a Sample was fractured so trichome interior was accessible to the beam. Proximal is nearest trichome pedicle.
bTrichome rays extended into free space and it was determined that the beam penetrated through the entire trichome ray but could not have hit any other cell types. Distal is near the end of the ray, farthest from the pedicle. c Epidermal cells contained very few vacuoles in this sample. d At detection limit. e Mn (1.5, 2.2 cts) detected at both trichome pedicle locations. f Mn (3.9 cts) detected at one trichome basal compartment location. g Mn (11, 8.3 cts) detected at both trichome basal compartment locations. The first layer palisade mesophyll cells directly adjacent to the trichome basal compartment did not contain Mn but contained Ni. h Mn (26, 22, 9 cts) detected at all 3 trichome basal compartment locations. The upper epidermal cells adjacent to the trichome basal compartment contained Mn (3.3, 4.8) and Ni. i a, b, c, d are sequential scans moving up a single trichome basal compartment towards the upper epidermis, across the epidermal line, then up the trichome pedicle away from the upper epidermis towards the trichome rays. Mn counts in the locations are (a) 21; (b) 2.7, 4.5 (c) 4.9; (d) nd.
239 Table 3. Summary of Ni localization patterns in Alyssum murale ‘Kotodesh’ leaf. Range of approximate concentration values in dry weight percent. nd: not detected. na: cell type/location not observed or accessible for analysis. First layer palisade mesophyll is adjacent to upper epidermis Plant Treatment Location
NI 5
NI 10
NI 20
NiAc 30
upper epidermis 2–5 1–3 14–23 lower epidermis 6–7 2–14 7–8 palisade mesophyll, 1st layer 4 1 1 P. mesophyll lower layers
