|Home | About | Journals | Submit | Contact Us | Français|
The analysis of herbarium specimens has previously been used to prospect for ‘new’ hyperaccumulators, while the use of foliar manganese (Mn) concentrations as a taxonomic tool has been suggested. On the basis of their geographic and taxonomic affiliations to known Mn hyperaccumulators, six eastern Australian genera from the Queensland Herbarium collection were sampled for leaf tissue analyses.
ICP-OES was used to measure Mn and other elemental concentrations in 47 species within the genera Austromyrtus, Lenwebbia, Gossia (Myrtaceae), Macadamia (Proteaceae), Maytenus and Denhamia (Celastraceae).
The resulting data demonstrated (a) up to seven ‘new’ Mn hyperaccumulators, mostly tropical rainforest species; (b) that one of these ‘new’ Mn hyperaccumulators also had notably elevated foliar Ni concentrations; (c) evidence of an interrelationship between foliar Mn and Al uptake among the Macadamias; (d) considerable variability of Mn hyperaccumulation within Gossia; and (e) the possibility that Maytenus cunninghamii may include subspecies.
Gossia bamagensis, G. fragrantissima, G. sankowsiorum, G. gonoclada and Maytenus cunninghamii were identified as ‘new’ Mn hyperaccumulators, while Gossia lucida and G. shepherdii are possible ‘new’ Mn hyperaccumulators. Of the three Myrtaceae genera examined, Mn hyperaccumulation appears restricted to Gossia, supporting its recent taxonomic revision. In the context of this present investigation and existing information, a reassesment of the general definition of Mn hyperaccumulation may be warranted. Morphological variation of Maytenus cunninghamii at two extremities was consistent with variation in Mn accumulation, indicating two possible ‘new’ subspecies. Although caution should be exercised in interpreting the data, surveying herbarium specimens by chemical analysis has provided an effective means of assessing foliar Mn accumulation. These findings should be followed up by field studies.
Manganese (Mn) is essential for plant growth and ubiquitous in soil. Unlike other micronutrients, the normal concentration of Mn in plant shoots spans a wide range (Graham et al., 1988). Although this is typically around 50–800 µg g−1 dry weight (d. wt), values exceeding 1000 µg g−1 have been recorded in species growing on ‘normal’ soils (Marschner, 2002; Foulds, 2003). Plant Mn-deficiency symptoms may be detected at foliar concentrations below approx. 10 µg g−1, while toxicity symptoms usually appear at around 1000–12 000 µg g−1 depending on the species (Reeves, 2006). Certain tolerant species have shoot tissue concentrations of approx. 1000–7000 µg g−1, both on high-Mn soils (>1 % Mn) and on soils with ‘normal’ Mn concentrations (e.g. 800–4000 µg g−1 Mn; Reeves, 2006).
Hyperaccumulators, comprising <0·2 % of angiosperms, possess an extraordinary ability to accumulate high shoot-tissue metal and metalloid concentrations by scavenging host substrates. This trait is characteristically manifested in plants occurring in their natural habitats. Mn hyperaccumulators form a subset of this group and, according to present taxonomic classification, around 11 predominantly woody species are known worldwide (Baker and Brooks, 1989; Proctor et al., 1989; Bidwell et al., 2002; Xue et al., 2004; Mizuno et al., 2008). They fit notional criteria of being able to accumulate foliar Mn to a minimum of 10 000 µg g−1 while growing in their natural habitats (Baker and Brooks, 1989). Most are native to the ultramafic substrates of New Caledonia. The distinctive ability of hyperaccumulators to bioaccumulate over a range of substrate concentrations is evident among these New Caledonian species, since soil Mn concentrations are not highly elevated (Jaffré, 1980). They include two subspecies of Maytenus fournieri (Celastraceae) Panch. & Sebert Loes, and two Virotia (Proteaceae) species formerly classified in the taxonomically similar Macadamia genus (Baker and Brooks, 1989; Willis, 2007).
In addition to the geographic proximity between eastern Australia and New Caledonia, there are taxonomic parallels between these New Caledonian Mn hyperaccumulators and some eastern Australian flora. The worldwide distribution of Macadamia is almost completely restricted to the east coast of Australia (Hardiner et al., 2000), while the eastern Australian Maytenus and Denhamia (Jessup, 1984) can be considered similar to the New Caledonian Celastraceae. The only known Australian Mn hyperaccumulator Gossia bidwillii (Benth.) N.Snow & Guymer belongs to the large Myrtaceae family (Bidwell et al., 2002; Snow et al., 2003) and occurs in sparse populations widely distributed along the eastern Australian seaboard. They are interspersed with the closely related Austromyrtus and Lenwebbia genera (Snow et al., 2003). Gossia bidwillii was classified as Austromyrtus bidwillii when initially described as a Mn hyperaccumulator (Bidwell et al., 2002), and its taxonomic resolution is still considered incomplete (Snow et al., 2003). Substrates on which Gossia, Austromyrtus, Lenwebbia, Macadamia, Maytenus and Denhamia occur in eastern Australia are generally high in available Mn (CSIRO, 1983; Fernando et al., 2006a, b).
It could be reasoned that ‘undiscovered’ Mn hyperaccumulators may exist among eastern Australian Gossia, Lenwebbia, Austromyrtus, Macadamia, Denhamia and Maytenus species. These genera are distributed over a large geographic area that includes very remote sites. Thus it was decided to analyse leaf fragments sampled from the Queensland Herbarium collection, in which the flora of eastern Australia is highly represented. Brooks et al. (1977) originally employed this approach to prospect for nickel (Ni) hyperaccumulators. Here, herbarium leaf tissue samples were analysed by ICP-OES (inductively coupled plasma–optical emission spectrometry) in ‘screening’ for foliar Mn and other elements. This study also addressed the use of foliar Mn concentrations as a taxonomic tool, particularly within the Myrtaceae as has been proposed by Bidwell et al. (2002).
Leaf fragments weighing 0·01–0·1 g d. wt were sampled from Queensland Herbarium specimens, and all species sampled are listed in Table 1. They are divided here into seven major groups labelled A–G for ease of reference. Replicates of each species were taken randomly from the collection, depending on availability of material. ‘Replicates’ represented different trees, and three to seven were initially collected. However, the seven species identified in the Results section as being of interest were subsequently re-sampled as different replicates, and their first and second datasets combined.
For each sample, approx. 0·03 g of leaf material was weighed into a Pyrex test-tube and ashed at 450 °C for 2·5 h. The resulting ash was dissolved in 6 mL 2 m HCl, and the solution centrifuged (5000 g, 20 min), and the supernatant analysed by ICP-OES for Mn, Ni, iron (Fe), aluminium (Al) and titanium (Ti). Quantification was carried out against standards prepared in 2 m HCl. It is unlikely that the herbarium specimens were soil-contaminated when collected in the field, since these species are foliated well above soil surfaces. However, foliar concentrations of Fe and Ti were used here to monitor for possible contamination on the basis that soils commonly contain Fe, and that, although Ti is present in soils, its accumulation in leaf tissues is unusual. Foliar Fe concentrations >500 µg g−1 d. wt can be considered toxic (Marschner, 2002), while according to unpublished data (R. D. Reeves, pers. comm.) foliar Ti concentrations exceeding 100 µg g−1 d. wt are likely to be due to soil contamination. Hence in this study, foliar threshold concentrations of 600 µg g−1 Fe and 100 µg g−1 Ti were taken as being indicative of soil contamination, thus necessitating exclusion of data collected from four of the approx. 350 samples that were analysed.
Leaf fragments from single specimens of the species listed in Table 1 were analysed for Mn, Ni and Al. On the basis of data obtained from the initial sampling, and also on the availability of replicates not previously taken, the herbarium collection was re-sampled as before. The additional replicates taken were for the following species: G. bamagensis, G. fragrantissima, G. gonoclada, G. lucida, G, sankowskiorum, G. shepherdii and M. cunninghamii. Data from both samplings are summarized in Table 2, while mean Mn values for each group are given in Table 3.
Seven species from two groups stood out as possible hyperaccumulators. Among the Group A Gossia species, G. bamagensis, G. fragrantissima, G. gonoclada, G. lucida, G. sankowskiorum and G. shepherdii were of interest, as was Maytenus cunninghamii from Group G. Replicates corresponding to each of these species contained one to five plants whose foliar Mn concentrations exceeded the Mn-hyperaccumulation threshold of 10 000 µg g−1, and/or more than half of the trees in each of these groups had foliar Mn concentrations above 5000 µg g−1.
The previously established variability of Mn hyperaccumulation by G. bidwillii (Fernando et al., 2007a) is evident in the present findings (Table 2). All replicates of Austromyrtus and Lenwebbia species tested here showed little evidence of any Mn hyperaccumulation, with foliar Mn concentrations <1000 µg g−1, and similarly for the Denhamia species, with the exception of a single plant of D. mooreii whose foliar Mn was 1870 µg g−1.
The Macadamia species were found to have high concentrations of either Mn or Al (Table 2), with maximum foliar concentrations of approx. 4000 µg g−1 Mn in M. ternifolia, and 6800 µg g−1 Al in M. grandis. Regression analysis indicated a negative correlation between mean foliar Mn and Al concentrations in the eight Macadamia species examined in this study (r2 = 0·57 P = 0·03). With the exception of the four Gossia species with elevated foliar Al concentrations to be discussed later, and the aforementioned Macadamia species, observed foliar Al concentrations in this study were within the range 20–800 µg g−1. The six G. fragrantissima trees initially tested had high foliar Mn and Ni concentrations (Fig. 1), including three above the threshold for Mn hyperaccumulation. The corresponding six Ni concentrations in the range 341–865 µg g−1 were all below the threshold value of 1000 µg g−1 for Ni hyperaccumulation. However, they well exceed ‘normal’ tolerance limits of approx. 50 µg g−1, based on moderately tolerant crop plants (Marschner, 2002). Notably, all other samples included in this study, with the exception of two individuals in the Maytenus group, were found to have Ni concentrations that ranged from undetectable levels up to a maximum of 26 µg g−1. One M. disperma sample and a M. cunninghamii sample had Ni concentrations of 63 and 207 µg g−1. Figure 2 shows Mn and unusually elevated Al concentrations in the four Gossia species, G. hillii, G. inophlia, G. lewisensis and G. macilwraithensis. Regression analyses of mean values showed no strong interrelationship between foliar Mn and Al concentrations in these four species (r2 = 0·57 P = 0·25). None of the individual replicates of these four Gossia species had foliar Mn concentrations above 5000 µg g−1.
Based on the initial screening of samples, seven species were re-sampled and analysed, their second Mn, Ni and Al data were combined with those initially obtained (Table 2), and the Mn data grouped into concentration ranges (Fig. 3A–G). Foliar Ni concentrations corresponding to the G. fragrantissima samples have been similarly combined and included (Fig. 3H). For each of these species, there were at least two trees whose foliar Mn concentrations were above 10 000 µg g−1. All nine G. bamagensis samples (Fig. 3A) had Mn concentrations above 10 000 µg g−1, including three extremely high values of 26 700, 31 700 and 37 500 µg g−1 Mn. Data showing Mn concentration groupings in 17 G. sankowskiorum samples, 43 M. cunninghamii samples and 12 G. fragrantissima samples (Fig. 3D, F and G) were approximately normally distributed around 10 000 µg g−1.
Nine G. sankowskiorum samples were above 10 000 µg g−1 Mn, and three each in the 2000–5000 and 5000–10 000 µg g−1 ranges. The M. cunninghamii Mn data showed 21 samples >10 000 µg g−1 Mn, 14 and six samples in the respective ranges 5000–10 000 µg g−1 and 2000–5000 µg g−1; and two samples <2000 µg g−1 Mn. Overall, the 43 M. cunninghamii samples examined represented plants from geographically different areas. They included a group of five from Weipa on the Cape York peninsula, for example, that all had foliar Mn >13 000 µg g−1, while a group of eight plants from Burke in north-west Queensland with comparatively narrow leaves all had Mn concentrations between 2800 and 9000 µg g−1.
The majority of G. gonoclada, G. lucida and G. shepherdii replicates (Fig. 3B, C, E) had foliar Mn concentrations >5000 µg g−1. Mn data corresponding to the 12 G. gonoclada samples included four above 10 000 µg g−1, five within the range 7000–10 000 µg g−1, one sample at 1850 µg g−1 and two outliers <100 µg g−1. Of the ten G. lucida replicates, there were three each above 10 000 µg g−1 Mn, and within the ranges 5000–10 000 µg g−1 and 2000–5000 µg g−1 Mn; as well as a single plant whose foliar Mn was 840 µg g−1. There were two replicates of the 13 G. shepherdii samples whose foliar Mn was >10 000 µg g−1, including an extremely high value of 31 200 µg g−1 Mn; five within the 5000–10 000 µg g−1 range, four in the 2000–5000 µg g−1 range and two <1000 µg g−1. Foliar Mn and Ni concentration ranges in 12 G. fragrantissima samples are summarized in Fig. 3G and H, and show that six exceeded the Mn-hyperaccumulation threshold. They included two extremely high Mn concentrations above 20 000 µg g−1. Five samples were between 5000–10 000 µg g−1, and one was 2400 µg g−1. The Ni data (Fig. 3H) shows that none of the G. fragrantissima samples were Ni-hyperaccumulative; however, they were all well above ‘normal’ upper limits of plant tolerance, and occurred within the range 217–865 µg g−1. Figure 4 shows 132 foliar Mn concentrations in the 18 Gossia species of Group A that excluded G. bidwillii. More than 50 % of these values exceeded 5000 µg g−1 Mn, including a small group of extreme outliers.
Mn hyperaccumulation by plants is a rare phenomenon, currently defined by a notional threshold concentration of 10 000 µg g−1 (d. wt). This study surveyed foliar Mn concentrations in approx. 350 herbarium samples from a selection of Australian east-coast flora. Gossia is a relatively new genus (Snow et al., 2003), of which 19 were examined here. It includes the only known Mn-hyperaccumulator G. bidwillii, which is the most geographically widespread (Snow et al., 2003). On the basis of the findings of this study, four other Gossia species, G. bamagensis, G. fragrantissima, G. sankowskiorum and G. gonoclada, were discovered to be Mn hyperaccumulators. In addition, G. lucida and G. shepherdii may also be ‘new’ Mn hyperaccumulators, although further investigation is required to conclusively assign their Mn-accumulative status. Of added interest is the finding that G. fragrantissima accumulated high foliar Ni concentrations, and was the only species in this survey noted for the trait. Maytenus cunninghamii was the fifth ‘new’ Mn hyperaccumulator discovered in this investigation.
The heterogeneity of Mn hyperaccumulation by G. bidwillii in its native habitats has been examined previously (Fernando et al., 2007), as have Ni and Zn hyperaccumulators by members of the Brassicaceae (Reeves and Baker, 1984; Baker et al., 1994; Macnair et al., 1999; Macnair, 2002). It was shown that wild populations of G. bidwillii hyperaccumulate Mn over a range of substrate Mn concentrations (Fernando et al., 2007). The frequency distribution of foliar concentrations of Mn in G. bidwillii, and Ni and Zn in the Brassicaceae is characteristically bimodal. They are usually normalized over a value exceeding the threshold, with a small outlying cluster at extremely high concentrations. Frequency plots of foliar Mn concentrations in the five ‘new’ Mn hyperaccumulators and the two ‘likely new’ hyperaccumulators in this investigation did not exhibit such behaviour (Fig. 3). Insufficient data may be a possible reason, especially for the six Gossia species of interest here, since 9–15 replicates were used for each plot. Data from 43 Maytenus samples were approximately normally distributed over the 5000–10 000 µg g−1 range, as were G. sankowskiorum and G. fragrantissima samples. Except for the non-normal distribution of G. bamagensis that occurred entirely above the 10 000 µg g−1 threshold, the remaining asymmetrical distributions were mostly spread across the 0–15 000 µg g−1 range. The overall frequency distribution of all Gossia species surveyed excluding G. bidwillii showed the highest frequency to be in the 0–5000 µg g−1 range, and extreme outliers between 20 000 and 40 000 µg g−1 (Fig. 4).
The natural distributions of the non-Mn-accumulator Lenwebbia prominens, the moderate accumulator Macadamia ternifolia and the hyperaccumulator Gossia bidwillii are in close proximity to each other near the Queensland–New South Wales border, and include high-Mn lateritic soils. Their greatly differing foliar Mn levels indicate determinants other than substrate composition affecting Mn uptake. These include genetic predisposition and factors such as the rhizosphere effects of soil microflora and/or root exudates that affect soil-Mn mobility. In describing the Mn-hyperaccumulative trait in G. bidwillii for the first time, Bidwell et al. (2002) also tested a similar species G. acmenoides from the same habitat in southern Queensland, and found foliar Mn in the latter to be approx. 2000 µg g−1 (d. wt). Both species were formerly in the genus Austromyrtus; and populations from which Bidwell et al. sampled for their study were included in this present survey. The current findings using herbarium material are in agreement with those of Bidwell et al. who tested field material. Interestingly, of the three Myrtaceae genera examined here, high foliar Mn concentration is a feature restricted to Gossia, hence supporting recent taxonomic revision (Snow et al., 2003). Considerable variability of this trait within the genus may be evidence of in situ adaptation of its species to their host soils, based on genetic factors restricted to Gossia. Leaf morphologies of the six Gossia species, newly classified here as definite or likely Mn hyperaccumulators, were diverse. There is no essential link between physiology and morphology, i.e. physiological similarity need not necessitate morphological similarity and vice versa. Evolution of the Mn-accumulative trait in Gossia may be similar to that proposed by D. O. Burge, Botany Department, Duke University (unpubl. res.) to describe evolution of Ni hyperaccumulation in Stackhousia tryonii. This model proposes hyperaccumulation as emanating from two possible phylogenetic patterns of Mn–plant relationships within a genus, i.e. Mn tolerance and low-level Mn accumulation as possible precursors to the evolution of Mn hyperaccumulation, or alternatively, the current Mn-accumulation trait may have evolved from ancestrally Mn-intolerant taxa.
The findings of this study confirmed Mn accumulation as a characteristic of the Australian Macadamia genus. Data obtained previously (Fernando et al., 2006b), from fresh field samples of two Macadamia species, M. integrifolia and M. tetraphylla, from southern Queensland agree with those obtained here from corresponding herbarium samples. All Australian Macadamia species were tested here and, interestingly, in contrast to the moderate-high Mn-accumulating M. integrifolia, M. ternifolia, M tetraphylla and M. jansenii, the three non-Mn-accumulating species, M. claudiensis, M. grandis and M. whelanii, had extremely elevated foliar Al concentrations in the range 2000–6800 µg g−1. Lateritic substrates predominate in eastern Australia, and their early soil-horizons commonly contain extremely high Al concentrations (CSIRO, 1983). It is well known that Al and Mn become similarly phytoavailable in low-pH soils, and can be problematic to agriculture in eastern Australia (CSIRO, 1983; Graham et al., 1988; White, 1997; Atwell et al., 2003). Moreover, a weak negative correlation between mean foliar Mn and Al concentrations was also observed in the four Gossia species, G. hillii, G. inophlia, G. lewensis and G. macilwraithensis. The reasons for these relationships are not immediately apparent, warranting further investigation involving field sampling of plants and soils. Interestingly, Jansen et al. (2003, 2001) have noted that ‘Al hyperaccumulators include mainly woody perennial taxa from tropical regions’, listing Myrtaceae and Proteaceae taxa among many such species. They propose a threshold of approx. 3000 µg g−1 for Al hyperaccumulation. By this criterion, the four Gossia species mentioned above, and Macadamia claudiensis and grandis may be regarded as Al hyperaccumulators. Earlier in this discussion it was pointed out that one of the ‘new’ Mn hyperaccumulators G. fragrantissima is also a strong Ni accumulator. The availability of Ni in eastern Australian soils is far less prevalent and more highly restricted than are Mn and Al. Hence it would be prudent to interpret foliar Ni data in conjunction with corresponding soil data, and generalizations cannot be made solely on the basis of foliar Ni concentrations obtained in this study.
There are seven Maytenus species native to Australia, and for reasons outlined in the Introduction, this survey was restricted to the five known east-coast species, all of which were found here for the first time to be strong Mn accumulators. Maytenus cunninghamii was assigned the status of Mn hyperaccumulator after re-sampling and further testing. This species has a wide geographic distribution, and pictorial and analytical data can be combined to delineate three possible subgroups. Although two of these, a distinctly narrow-leaved moderate Mn-accumulating taxon and a broad-leaved Mn-hyperaccumulating taxon are distinguishable from each other, there is an intermediate group with broadish leaves and variably high foliar Mn concentrations. Chemical and morphological differences may warrant future taxonomic re-evaluation of these apparently distinct populations of M. cunninghamii.
The majority of currently designated Mn hyperaccumulators are native to New Caledonia, and were originally identified as such by their foliar Mn concentrations in field-collected samples (Jaffré, 1977, 1979; 1981). From those data it was evident that among the genera that included Mn hyperaccumulators, and others such as Grevillea (Proteaceae), there were obvious traits of high Mn-accumulation within certain plant groups, including the New Caledonian Maytenus.
Threshold values defining hyperaccumulation have been derived by the 10-fold multiplication of a nominal ‘upper limit’ of tolerance, which, in the case of Mn, was taken as being approximately 1000 µg g−1 (Brooks, 1998b). While data from this present investigation and those of Brooks et al. (1981) and Jaffré (1977) may be insufficient to draw conclusive hypotheses regarding the notion of what constitutes Mn hyperaccumulation per se, they clearly demonstrate that related species can be grouped by their common trait of Mn tolerance. As can be seen here and in previous work, this trait is manifested as a range of high foliar Mn concentrations that can be regarded in normal physiological terms as being extraordinarily high. There has been long-standing discussion in the literature regarding the genetic difference, if any, between hyperaccumulation and tolerance. Although Macnair et al. (1999) have demonstrated these to be genetically independent characteristics in a Brassicaceae species, there is a lack of similar investigation of woody hyperaccumulators and of Mn hyperaccumulators. The nutritional status of Mn, its widespread occurrence in soils, together with the physiological ability of most plants to tolerate it well above their normal requirements and also over varying ranges, sets it apart from most other hyperaccumulated elements. It is not unreasonable to suggest that the definition of Mn hyperaccumulation be re-examined and consideration given to applying a threshold concentration range, rather than a single value. On the basis of the findings of this study and those of Jaffré (1977), a dry weight foliar concentration range 5000–7000 µg g−1 Mn is proposed as a reasonable threshold estimate for extreme plant tolerance or hyperaccumulation. In addition to notional threshold tissue concentrations, plant hyperaccumulation is recognized by other important features that should also be included in any overall assessment for the trait. In the present investigation, species classified as ‘new’ hyperaccumulators meet criteria of threshold concentration, as well as naturally occurring on Mn-rich substrates. Controlled-environment and field studies are needed to conclusively establish if the trait is constitutive within these species.
For several reasons, caution needs to be exercised when drawing hypotheses based on data obtained from herbarium samples. First, only very small fragments are sampled and taken as representing whole plants. Secondly, the age of leaves sampled is relevant to Mn accumulation. It is well known that Mn taken up in the xylem does not re-enter the phloem, leading to an increasing concentration gradient with leaf age (Lonergan, 1988). Best endeavours were made in this study to select tissue fragments from oldest available leaves. Thirdly, certain practices of tissue preservation for permanent herbarium storage, such as immersion in alcohol, etc. could be detrimental to Mn retention in the tissues. When selecting samples for this study, samples incorporating such preparations were avoided where possible. Soil contamination of the specimens tested here did not pose a major problem, as shown by the monitoring system employed. It can be assumed that here no artificial elevation of foliar Mn concentrations occurred as a result of potential confounding factors. The net effect of possible artefacts outlined above would be to artificially lower Mn concentrations and, as such, data presented here may be viewed as minimum values. There is also the possibility that some of these species have been wrongly classified; however, this is also a risk when sampling fresh tissue from the field.
In conclusion, it can be said that this survey has raised many questions regarding the uptake of Mn by a select group of Australian east-coast flora, and also the affinity of some of these groups for Al. As well as describing a group of ‘new’ Mn hyperaccumulators, this study has brought out possible new taxa, and has highlighted the need for re-evaluating the definition of Mn hyperaccumulation. The findings of this study also have potential benefits for the phytoremediation of Mn-contaminated soils. On the basis of data obtained here, sampling of fresh plant material and host substrates is now required in order to consolidate major findings of this investigation, and in doing so to pursue new leads for future research.
We thank Gillian Naylor at the Queensland Herbarium for her invaluable assistance. Preparation of this manuscript was financed by the David Hay Memorial Fund (University of Melbourne, Australia).
This paper is dedicated to the memory of George Batianoff who passed away on 20 February 2009, shortly before this paper went to press. George devoted the whole of his working life to increasing our knowledge and understanding of the native and introduced plants of eastern Australia.