• Aluminium (Al) stress reduces plant growth. However, some species such as Norway spruce (Picea abies) seem to tolerate high Al concentrations. The aim of this study was to investigate characteristics possibly involved in Al tolerance in Norway spruce seedlings.
• Seedlings (10-d-old) were exposed to Al3+ concentrations of 0.5 and 5 mm for up to 168 h. The effect of Al stress on root growth, cell morphology and Al distribution, callose production, and peroxidase and chitinase activity was analysed.
• Root growth decreased after 1 d and 2 d with 5 and 0.5 mm Al, respectively. Callose concentration increased strongly after 6 h treatment with 5 mm Al. The activity of many peroxidase and chitinase isoforms decreased after 1–24 h exposure of both treatments. Several isoforms increased after 48–168 h exposure to 5 mm Al.
• We postulate that, with external Al concentrations 0.5 mm or lower, an increased production above constitutive levels of peroxidase or chitinase is not required for Al tolerance in young Norway spruce seedlings. High constitutive levels of peroxidase and chitinase in this species may be part of this Al tolerance.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Aluminium (Al) is the most abundant metal in the world and it is the third most abundant element in the earth's crust. Al generally becomes phytotoxic in acid soils (pH ≤ 4.5) and is one of the major factors limiting crop production worldwide. It is therefore of great importance to understand the mechanisms of Al resistance in plants.
Callose (β-1,3-glucan) is a plant polysaccharide that is a natural constituent of sieve plates, pit fields and developing cell plates (Verma & Hong, 2001). Increased synthesis of callose is a well known early indicator of Al injury in several plant species (Horst, 1995). In maize (Zea mays), both aluminium and callose accumulate in the zone 1–2 mm behind the root tip (Sivaguru & Horst, 1998). The Al induced growth inhibition might in part be due to a translocation of sugars from cellulose to callose production (Zhang et al., 1994; Kaneko et al., 1999). It is thought that Al increases intracellular calcium levels, which in turn lead to induction of callose synthesis and increased callose deposits in plasmodesmatal pores, causing impaired intercellular transport and cell-to-cell communication (Jones et al., 1998; Sivaguru et al., 2000).
Most of what we know about Al tolerance in plants comes from work on Al sensitive species such as Arabidopsis, maize, rice and wheat. In general, tree species are more tolerant to Al than herbs, grasses and cereals. For example, it was found there was no reduction in root growth in seedlings of Norway spruce (Picea abies), silver birch (Betula pendula) and Scots pine (Pinus sylvestris) at Al concentrations below 0.3, 3, and 6 mm, respectively (Göransson & Eldhuset, 1987, 1991). In the field, 3 yr of artificial Al addition, producing potentially toxic Al concentrations in the soil solution, had no impact on fine root growth in Norway spruce trees (De Wit et al., 2001). Norway spruce thrives on naturally acid soil where its growth leads to a further acidification of the soil. It is therefore likely that Norway spruce has developed adaptive mechanisms to tolerate high Al conditions; however, there have been few attempts to elucidate Al tolerance mechanisms in tree species, and such studies could yield valuable new information about Al tolerance mechanisms in plants in general.
The main goal of this study was to investigate fundamental characteristics possibly involved in Al tolerance mechanisms in Norway spruce seedlings. Young seedlings exposed to Al solutions for up to 1 wk were used to study the effects of Al on callose production, root growth, and root tip browning and swelling. Various microscopy staining techniques were used to identify cellular changes and callose deposits, and the expression of chitinase and peroxidase isoforms was examined by polyacrylamide gel electrophoresis – isoelectric focusing.
Materials and Methods
Norway spruce (Picea abies (L.) Karst.) seeds from the Norwegian Seed Station (lot no. 5225, Eidskog) were washed in sterile water and germinated on filter paper in the dark for 10 d at room temperature. After germination, the seedlings were placed on 1% sterile agar plates, and the roots were covered with filter paper soaked in 0.5 mm CaCl2. The seedlings were grown horizontally under axenic conditions at 20°C with a photosynthetic photon fluence rate of 10 µmol m−2 s−1 (Osram 58 W/30 fluorescent tubes). The experiment consisted of two Al treatments (0.5 mm and 5 mm Al supplied as AlCl3 in 0.5 mm CaCl2, pH 3.8 ± 0.1) and one control treatment (0.5 mm CaCl2, pH 3.8 ± 0.1) of seedlings 1–168 h before sampling. For simplicity, Al treatments will be referred to only as 0.5 and 5 mm Al, although Ca was also present. Each treatment was repeated three times with 15 seedlings per agar plate (replicate). All chemicals were from Sigma Chemical Co. (St. Louis, MO, USA) if not stated otherwise.
Measurements of root growth
To assess the impact of different Al concentrations, we measured the root length of each seedling 24 h before the start of the longest lasting Al treatment and subsequently at harvest. The Al treatments started 48, 24, 12, 6 or 1 h before harvest. The seedlings were photographed with a digital camera (Leica DC200, Wetzlar, Germany), and root length was estimated using image analysis software (Image Tool, UTHSCSA, San Antonio, TX, USA). Root growth is presented as the difference between length at harvest and length at the first measurement.
Root tips from five plants per treatment and sampling time were fixed in 2% (w : v) paraformaldehyde and 1.25% (v : v) glutaraldehyde, in 50 mm PIPES (l-piperazine-N-N′-bis [2-ethane sulfonic acid] buffer, pH 7.2, for 24 h at room temperature. The samples were then dehydrated in an ethanol series (70-90-96-4 × 100%) and embedded with LR White acrylic resin (TAAB Laboratories, Aldermaston, Berkshire, UK), which was polymerised at 60°C for 24 h. Semi-thin sections (1.5 µm) were cut with a diamond knife using an ultramicrotome (Ultracut E Reichert Jung, C. Reichert AG, Wien, Austria) and dried onto gelatine coated glass slides. For general anatomy and detection of starch, the sections were stained with Stevenel's blue (Del Cerro et al., 1980) and periodic acid-Schiff's carbohydrate stain (Hotchkiss, 1948; Nagy et al., 2000), respectively. For detection of Al3+, fresh root tips and root sections were stained with 1% (w : v) solochrome azurine (Denton & Oughton, 1993) for 2 min. The sections were mounted with immersion oil and visualized in a light microscope (Leica DMR, Wetzlar, Germany) using a DC 200 camera. For detection of callose deposits, the sections were stained with aniline blue (CI 42755) for 30 min (De Neergaard, 1997) and examined for callose specific fluorescence in an epifluorescence mode. Blue light of 450–490 nm was used as the excitation wavelength and induced fluorescence was visualized with a long bandpass filter (520 nm and above).
Determination of callose content
Synthesis of β-1,3-glucan (callose) was used as an indication of Al induced injury using the method described by Köhle et al. (1985). Root tip sections (10 mm) from five seedlings per treatment replicate and sampling time were stored in 80% ethanol until analysis. The excess ethanol was removed and 200 µl of 1 m NaOH added to the root sections, which were then macerated with a microfuge pestle, heated (80°C, 15 min) and the samples allowed to cool before centrifugation at 1000 g to remove particulates. Callose content was determined by the stepwise addition of the following to a microfuge tube: 71 µl of sample supernatant, 142 µl of aniline blue (0.1% w : v), 75 µl of 1 m HCl and 210 µl of 1 m glycine/NaOH buffer (pH 9.5). After mixing, the samples were heated (50°C, 20 min) and left to cool before reading on a Perkin Elmer LS-5B fluorimeter (Boston, MA, USA) using a 400 nm excitation, 510 nm emission, and a slit width of 15 nm. The measurements are expressed in relative fluorescence units per centimetre (RFU cm−1).
Protein extraction and visualization of peroxidase and chitinase activity
Proteins from roots were extracted in a 0.1 m citrate buffer (pH 5.0) with 3% (w : v) CHAPS (3-[3-(cholamidopropyl) dimethylammonio]-1-proanesulfonate) and 3 mm DTT (1,4-dithio-DL-threitol, VWR International, Stockholm, Sweden). Insoluble material was removed by centrifugation at 15 000 g for 1 min, and protein concentration was measured with the Bradford (1976) method, using BSA as the standard protein.
The different peroxidase and chitinase isoforms were separated by nondenaturing polyacrylamide gel electrophoresis with 10% isoelectric focusing (IEF) gels using broad range ampholytes (Ampholine pH 3.5–10; Amersham Biosciences, Uppsala, Sweden). We loaded equal amounts of protein for each sample (0.4 µg), and the IEF was run at 4°C and 20 W for 2.5 h using a Multiphor II Electrophoresis System (Amersham Biosciences).
Peroxidase activity was detected according to the method described by Kerby & Somerville (1989), using 3-amino-9-ethylcarbazole and hydrogen peroxide (VWR International) as substrates. Detection of chitinase activity was performed as described by Pan et al. (1991). After IEF, the gel was washed in 0.1 m sodium acetate buffer (pH 5.2) for 5 min, and then covered with a 7.5% polyacrylamide overlay gel (Bio-Rad Laboratories, Hercules, CA, USA) containing 0.04% (w : v) glycol chitin in 0.1 m sodium acetate buffer (pH 5.2). The gel sandwich was incubated under moist conditions at 37°C for 2.5 h, and the overlay gel developed for 5 min at room temperature in freshly prepared 0.01% (w : v) calcofluor white in 0.5 m tris-HCl (pH 8.0). The overlay gel was washed overnight in distilled water at 4°C, and the various chitinolytic isoforms were photographed under UV light.
Quantification of peroxidase and chitinase isoforms in Norway spruce roots grown at 0.5 mm and 5 mm AlCl3 was performed as follows. The absorbance values (OD) corresponding to the relative amount of peroxidase and chitinase isoforms above background were obtained with the GeneSnap 5.00.09 program using the Syngene Bio Imaging system (Syngene, Cambridge, UK). The images were in gray scale, saved in Adobe Photoshop 6.0, and processed with the gel-doc program LabImage Version 2.62a (Kapelan GmBH, Halle, Germany). Six peroxidase bands and 15 chitinase bands were detected and compared using the software. In each IEF run, an untreated control sample was loaded in lane 1, from this the basic peroxidase isoform band 1 from the top was used as an internal control to calibrate the relative amount of the different isoforms present within each gel run. For the chitinase activity gels, band 5 from the top was used as the internal control in the same manner. Within each gel all isoforms band values were divided by the internal control (thus the value of the internal control in each gel is adjusted to 1 and all the other isoforms bands are set relative to the internal control) to compare bands within and between IEF gel runs. The results are the average of three separate IEF gel runs for both the peroxidase and chitinase activity gel.
For all quantifications, the results are presented as means with standard errors (SE). Root length, callose concentration and optical density of enzyme isoforms were subjected to anova and tested for significant differences between preplanned comparisons of treatment means by the GLM (SAS™) procedure by using least significant difference (LSD) among the means at P = 0.05 (Sokal & Rohlf, 1995). The Pearson correlation test (Sokal & Rohlf, 1995) was used to check for product moment correlations between root length and callose concentration, and was computed with the Corr procedure (SAS™). The SAS software (SAS Institute Inc., Cary, NC, USA) was used for all statistical analysis.
Root growth and appearance
The roots from the control plants were white and had a healthy appearance throughout the experiment (Fig. 1a,b). The seedlings exposed to the lower Al concentration (0.5 mm) were similar in appearance to the control plants, apart from the occasional occurrence of brown coloured regions (Fig. 1c). By contrast, in the seedlings exposed to a higher Al concentration (5 mm), the root tips became strongly brown coloured and had a stunted appearance after 48 h (Fig. 1d).
After 24 h of exposure to 0.5 and 5 mm Al, we observed the formation of brown coloured spots characteristic of Al damage, just behind the meristematic region of the root apex, as well as in the epidermis of young and mature root regions (Fig. 2a). Using solochrome azurine, an Al-specific stain that produces a bluish-black coloured complex with Al (Denton & Oughton, 1993), it was evident that more Al was binding to the surface of the roots exposed to 5 mm Al than to those treated with 0.5 mm Al (Fig. 2a). At 0.5 mm, Al stained moderately after 24 h, and more strongly after 48 h. By contrast, the seedlings treated with 5 mm Al stained intensely after only 1 h of Al exposure. After 24 h, the 5 mm Al treated roots possessed increased diameters and the root tip had a swollen appearance. After 48 h this was also evident in the roots treated with 0.5 mm Al.
In plants exposed to 5 mm Al, a significant reduction (P < 0.04) in root growth was observed after 24 h and 48 h compared with the control and 0.5 mm Al treated plants (Fig. 2b). In the 0.5 mm Al treated plants, a significant reduction (P = 0.063) in root growth only became apparent after 48 h of exposure (Fig. 2b) compared with the control. At the end of the experiment (168 h), root growth had completely ceased in both Al treatments.
Cellular changes in the root tip following Al stress
In control plants, the root cap and cortical and meristematic regions consisted mostly of densely packed cells (Fig. 3a). After 48 h exposure to 5 mm Al, a lesion characteristic for Al toxicity appeared 1–2 mm behind the root apical meristem (Fig. 3b). Compared with the control plants there was an almost complete disappearance of starch in the 5 mm Al treated plants after 48 h (Fig. 3c,d). In the lesion area, as well as in the rest of the root tip, larger intercellular air spaces were formed. This lesion area consisted of partly collapsed and disintegrated cells with condensed nuclei, indicating cell death, along the edge of the lesion (Fig. 4a,b). This area also contained brown polyphenolic substances.
The Al-specific solochrome azurine stain was mainly localized in the cell walls of the peripheral cells of the root cap and in the mucilage surrounding root cap cells (Fig. 5a,b). In the root cortex, Al deposits were found on the cell walls and in the intercellular spaces of the outer cortical cells (Fig. 5c,d). Al was also found associated with cell walls at the edge of lesions and where intercellular airspaces were formed (Fig. 5e). In Al treated plants, the root cap was larger – or more pronounced – than in the control plants. There were also visual indications of an increased presence of mucilage in this region (data not shown).
Callose induction in the root tip following Al stress
There was a strong increase in root tip callose content in the seedlings treated with 5 mm Al (Fig. 6). Already after 6 h, there was a significant increase in callose concentration compared with the control and the 0.5 mm Al treatment (P = 0.004). The increase continued throughout the experiment until day 4, and the values were significantly higher compared with the control and the 0.5 mm Al treatment at all times (P < 0.0001). The callose concentration was negatively correlated with root length (P = 0.013). In the seedlings exposed to 0.5 mm Al, there were no significant changes in callose concentration with time, and there was no significant difference compared with the control at any time (Fig. 6). Using aniline blue stain and fluorescence microscopy, we observed callose specific fluorescence in the cortical cell walls and at the tips and interconnecting points of the root cap cells in the 5 mm treatment only (Fig. 7a,b). At higher magnification, the callose was observed as individual lumps at sites along the walls of the cortex cells (Fig. 7c). In the control plants, only a very faint callose specific fluorescence was detected in the cell walls (Fig. 7d,e).
Temporal changes in chitinase- and peroxidase isoforms in Al exposed roots
IEF gel analysis revealed the existence of six peroxidase and 15 chitinase isoforms in the root tips (Fig. 8a,b). The isoform profile in the roots changed after exposure to different Al concentrations. For the chitinases, the activity of each isoform band at the 168 h control did not differ significantly from the activity at the start of the experiment (Fig. 9a–o). In the 0.5 mm Al treatment, all chitinase bands had lower activities than the control at 1 h, and the difference was significant (P < 0.02) for all isoforms apart from band 10. The chitinolytic activities stayed close to the 1 h level until 48 h.
In the 5 mm Al treatment, chitinase activities were lower than in the control at 1 h, and the difference was significant (P < 0.05) for bands 1–11 (Fig. 9a–o). At 24 h or earlier the activity increased in all bands. At 168 h the activity was significantly (P < 0.05) higher than in the control for all bands apart from number 11 and 12. The activity of each chitinase band at any time was generally higher in the 5 mm than in the 0.5 mm Al treatment.
For the peroxidase controls, the activity of each isoform band at 168 h did not differ significantly from the activity at the start of the experiment, apart from band 5 (Fig. 10a–f). In the 0.5 mm treatment, all peroxidase bands had lower activity than the control at 1 h; the difference was significant (P < 0.05) for bands 2–5. The peroxidase bands 2–5 generally stayed at the 1 h level or lower until 48 h (number 1 until 24 h), and were lowest at 12 h. Bands 1 (highly basic) and 6 (highly acidic) reached the level of the corresponding control isoforms at 48 h.
In the 5 mm Al treatment, all peroxidase bands showed lower activities than the control after 1 h, and the difference was significant for bands 2 and 5 (Fig. 10a–f). Bands 2–5 stayed at the 1 h level or lower until 48 h, when they were lowest. The peroxidase bands 2–4 then increased up to 168 h, while band 5 kept constant up to 168 h. Band 1 increased after 12 h, while band 6 stayed at the control level until 48 h and then increased up to 168 h. The activities at 168 h were above (bands 1 and 6) or below (bands 2–5) those of the control; the differences were significant for bands 1, 3, 5, and 6 (P < 0.04). The level of each peroxidase isoform band at any time was in most cases not significantly different between the two Al treatments.
The initial and most dramatic symptom of Al toxicity is inhibition of root growth. Our study shows that root growth in Norway spruce was strongly reduced after exposure to 5 mm Al, while in the 0.5 mm Al treated plants reduced root growth only occurred after prolonged exposure. Compared with what many other plant species can tolerate, an Al concentration of 0.5 mm is considered as a high dosage and is greater than that typically found in solutions of acid soils (5–200 µm). However, our observations are consistent with previous studies on Al stress in Norway spruce. For example, in a study on Norway spruce nursery seedlings, Godbold & Kettner (1991) found no impact on root growth after 5 d at 0.4 mm Al, while 60% of the seedlings had ceased growing already after 2 d exposure to 0.8 mm Al. The reduced root growth may be related to the fact that Al alters cytoskeletal dynamics and signal transduction pathways which may eventually lead to inhibition of cell division and elongation (Jones & Kochian, 1995; Blancaflor et al., 1998; Frantzios et al., 2001). Additionally, when Al binds to and modulates the cell walls, the wall expansion and cell extension can be affected (Le Van et al., 1994; Zhu et al., 2003).
The browning reaction observed in the seedling root tip, right behind the apical meristematic zone, is a characteristic of Al injury caused by oxidative stress and the accumulation of phenolic substances (Richards et al., 1998). Al-induced root tip browning, the concomitant collapse in cell structure, accumulation of polyphenols, and the development of intercellular air spaces, are similar to the symptoms observed during the hypersensitive response (localized cell death) at the site of initial contact with a pathogen after infection (Nicholson & Hammerschmidt, 1992). We observed localized cell death in the area where the phloem starts to differentiate. This phenomenon may lead to decreased transport of carbohydrates from the shoot to the root apical meristem, and could partially explain the observed decline in starch content in the roots of Al treated seedlings as a lack of soluble carbohydrates may induce starch degradation.
It was evident from observing stained roots that Al was located primarily in the walls of the root cap cells, border cells and the outer cortical cells of Norway spruce roots, even at 5 mm concentrations. Al may be confined to the cell wall region, most likely due to an occupation of carboxyl groups in pectin (Hamel et al., 1998). An inhibition of Al diffusion through the cell walls prevents it from reaching the symplast and the vascular system at high concentrations. A greater cell wall Al binding capacity could help explain the higher resistance observed in Norway spruce compared with other, more Al sensitive species. This ability to retain Al may be due to a high cationic exchange capacity of the cell wall of Al tolerant species (Godbold & Jentschke, 1998). Another prominent feature of spruce roots is the large root cap cells, which have a high mucilage production. A high mucilage production has been hypothesized to increase the capacity of the roots to bind and exclude Al from sensitive growing areas (Zhu et al., 2003). In Al sensitive species, for example the Graminae (grasses), the root cap cells are small and produce less mucilage, something that could explain their reduced capacity to scavenge and detoxify Al (Crawford & Wilkens, 1997; Zhu et al., 2003).
Callose concentration in root tips increased only at 5 mm Al, and even then the increase in callose production was much lower than what is commonly observed in other species, for example maize, treated with much lower Al concentrations (DL Jones, unpublished). At exposure to lower Al concentration no significant change in the callose levels was detected. In species like soybean, Al concentrations as low as 10 µm induce callose production in root tips within minutes (Wissemeier et al., 1987). In 5-wk-old Norway spruce seedlings, callose in root tips has been observed after 3 h treatment with 170 µm Al (Jorns et al., 1991). Callose has also been measured in root tips from field grown spruce trees where the soil solution Al concentration ranged from less than 10 to 200 µm (Wissemeier et al., 1998). Corresponding to the biochemical measurements, we also detected Al induced callose production in the form of callose deposits in the cell walls and interstitial spaces of root cap and cortical cells at 5 mm Al. These callose deposits may lead to an occlusion of pores and plasmodesmata, which effectively block symplastic transport (Sivaguru et al., 2000). Callose occlusion would perhaps act as a barrier to Al influx into the cortex of the root, but is also a very probable factor in the root growth inhibition of Al (Sivaguru et al., 2000). We found a correlation in callose increase and root growth cessation at 5 mm Al. It should be noted that root growth in 0.5 mm Al decreased without any production of callose above that of the control. This contradicts the findings of others (Wissemeier et al., 1998) that callose increases in concentration before root growth decreases. It is uncertain whether stress protein and callose production in response to increased Al is symptomatic of susceptible or of tolerant species (Jan et al., 2001). Ezaki et al. (2001) found a lower callose content in more Al resistant Arabidopsis lines compared with the wild type. Our study strongly indicates that in 10-d-old Norway spruce seedlings, callose is not an important inducible factor in Al tolerance; increased callose production is triggered only at extreme Al stress.
Genes corresponding to peroxidases and other pathogenesis and stress related proteins were previously reported to be up-regulated upon Al exposure in wheat (Triticum aestivum L. cv.) and Arabidopsis (Ezaki et al., 1996, Hamel et al., 1998). Our results show that in 5-wk-old Norway spruce only very strong and prolonged Al stress would lead to an up-regulation resulting in increased activity of several peroxidase and chitinase isoforms. The initial fall in enzyme activity during the first 1–12 h may be due to reduced enzyme activity as a direct inhibitory effect of Al (as many peroxidases are known to be extra cellular), increased degradation or reduced expression of the corresponding gene(s). In barley (Hordeum vulgare), Tamás et al. (2003) report an increased activity of at least five anionic (acidic) and four cationic (basic) peroxidase isozymes in response to Al exposure. In the present work, one highly basic and one highly acidic peroxidase isoform, and four basic and three acidic chitinase isoforms increased in response to 5 mm Al treatment after 7 d. In other stresses, such as drought and pathogen infection, we have found a strong increase in similar chitinase and peroxidase isoforms within 4 d (Fossdal et al., 2001; Nagy et al., 2004). It is possible that Norway spruce may not need to produce peroxidases and chitinases above the already high constitutive levels as a response to Al concentrations at or below 0.5 mm. These proteins are involved not only in stress responses, but also in normal development and morphogenesis. For example, peroxidase activities occur in the cell wall, were these enzymes have been suggested to modulate cell wall rigidity and extensibility, thus reducing Al diffusion through cell wall (Hamel et al., 1998). Even though the main role of chitinases is in defence against pathogenic fungi, evidence is suggesting that some chitinolytic enzymes play a role in normal growth and development acting on arabinogalactan proteins (Passarinho et al., 2001), and affect cell growth and cell-to-cell communication (Wiweger et al., 2003). The high ability of Norway spruce to grow on acidic soils could be explained by a higher capacity to tolerate Al, maybe in part by a generally high constitutive level of these and other stress related proteins. This is, to our knowledge, the first report indicating that isoforms of peroxidases and chitinases are affected only by extremely strong Al exposure in Norway spruce.
Based on a total evaluation of our results, we postulate that with external Al concentrations lower than 0.5 mm, an increased production of callose, peroxidase or chitinase (above constitutive levels) is not necessary for the Al tolerance in young Norway spruce seedlings. Callose seems to be important only at very high Al concentrations. It is possible that the high constitutive levels of peroxidase and chitinase confer to this species cell wall properties that help to exclude Al, and that increased production of these proteins are necessary only when external Al increases to unusual levels. In addition, physical barriers in the rhizosphere, for example mucilage, root cap (border) cells, ectomycorrhiza, and cation exchange capacity may be so effective that the more sophisticated inducible mechanisms is not effectuated. Only if the external Al concentration exceeds a limit so that the external defence barriers break down, and Al gets access further into the roots, increased callose production and enzyme activity will be apparent. The ability to endure within cell Al may also be associated with the high content of phenolics in the vacuoles of epidermal cells in Norway spruce roots (Ofei-Manu et al., 2001), a possible accumulation of Al in complexes with phenolics deserves further studies.
We thank Inger Heldal and Helene Haug Pagander for excellent technical assistance. We are also grateful to Dr Harald Kvaalen for valuable discussions and for performing the statistical analysis. The Research Council of Norway provided the financial support for this work.