Plant and Soil

, Volume 382, Issue 1, pp 219–236

As, Pb, Sb, and Zn transfer from soil to root of wild rosemary: do native symbionts matter?

  • Marie-Cécile Affholder
  • Anca-Diana Pricop
  • Isabelle Laffont-Schwob
  • Bruno Coulomb
  • Jacques Rabier
  • Andreea Borla
  • Carine Demelas
  • Pascale Prudent
Regular Article

DOI: 10.1007/s11104-014-2135-4

Cite this article as:
Affholder, MC., Pricop, AD., Laffont-Schwob, I. et al. Plant Soil (2014) 382: 219. doi:10.1007/s11104-014-2135-4


Background and aims

This is an in natura study aimed to determine the potential of Rosmarinus officinalis for phytostabilization of trace metal and metalloid (TMM)-contaminated soils in the Calanques National Park (Marseille, southeast of France). The link between rosemary tolerance/accumulation of As, Pb, Sb, and Zn and root symbioses with arbuscular mycorrhizal (AM) fungi and/or dark septate endophytes (DSE) was examined.


Eight sites along a gradient of contamination were selected for soil and root collections. TMM concentrations were analyzed in all the samples and root symbioses were observed. Moreover, in the roots of various diameters collected in the most contaminated site, X-ray microfluorescence methods were used to determine TMM localization in tissues.


Rosemary accumulated, in its roots, the most labile TMM fraction in the soil. The positive linear correlation between TMM concentrations in soil and endophyte root colonization rates suggests the involvement of AM fungi and DSE in rosemary tolerance to TMM. Moreover, a typical TMM localization in root peripheral tissues of thin roots containing endophytes forming AM and DSE development was observed using X-ray microfluorescence.


Rosemary and its root symbioses appeared as a potential candidate for a phytostabilization process of metal-contaminated soils in Mediterranean area.


Trace metals and metalloid multicontamination Arbuscular mycorrhizal fungi Dark septate endophytes Phytostabilization μXRF analyses 



Trace metals and metalloids


Arbuscular mycorrhizal


Dark septate endophytes


Bioconcentration factor


Contamination factor


Pollution load index


Trace metals and metalloids (TMM) are naturally present in the environment due to their occurrence in the Earth’s crust. However, for several years, high concentrations (in comparison with background levels) have been detected, mainly in area utilized for agricultural and/or industrial activities (He et al. 2005; Wuana and Okieimen 2011), and these may lead to major environmental issues. Depending on their speciation which influences their mobility and bioavailability (Kabata-Pendias 2004), TMM can be disseminated in the various ecosystem compartments including plants. Indeed, plants are able to assimilate and, for some species, to hyperaccumulate TMM (Zheng et al. 2007), making them interesting to manage contaminated sites. Thus, in phytoremediation processes, plant and their associated microorganisms can be used for TMM stabilization or extraction from soil (Pilon-Smits 2005). Amongst associated microbes involved in phytoremediation processes, root symbioses, especially those with arbuscular mycorrhizal (AM) fungi, have a significant effect on tolerance and absorption of TMM in plants (Gohre and Paszkowski 2006). In addition, these symbioses promote plant nutrition, particularly phosphorus, and therefore plant growth and health (Harrison 1999; Barea et al. 2002). Plants appear as well to form symbiotic associations with fungal endophytes, and among diverse groups of endophytic fungi, dark septate endophytes (DSE) have received much attention in the recent years (Li et al. 2011). DSE are ubiquitous root-associated fungi that form inter- and intracellular melanized hyphae which can aggregate and form groups of thick-walled nodules (called microsclerotia) in epidermis and cortex of roots. They can form mutualistic, mycorrhiza-like relationship though their relationships with their host are not clear and seem to go from parasitic to mutualistic according to the species involved and the environmental conditions (Jumpponen 2001). However, their presence in disturbed environments suggests a role in plant protection in case of abiotic stress, such as TMM stress (Regvar et al. 2010).

Phytoremediation may be part of the solution for dispersed pollution as a non-invasive and ecological alternative method (Moreno-Jimenez et al. 2011). It seems to be the most adapted choice for the particular case of the contamination from the past industrial activity of the former smelter factory of l’Escalette (southeast of Marseille, France), located in the Calanques National Park. In this smelter factory, which ceased its activity in 1925, galena was converted into lead and silver. As a consequence, this former industrial site is still polluted by TMM, especially As, Pb, Sb, and Zn (Testiati et al. 2013). Moreover, the occurrence of these trace elements was detected quite all over Marseilleveyre hills (Laffont-Schwob et al. 2011). Today, neither the area affected by its activity nor even a part of this area has been reclaimed, and in this context, the containment of pollution of this site must be a first priority. Since the polluted area is located on a protected zone for its plant species richness, non-intrusive solutions have to be considered. Autochthonous plant species may be evaluated for heavy metal stabilization or extraction (Rabier et al. 2007). Previous studies (Cala et al. 2005; Sekeroglu et al. 2008; El-Rjoob et al. 2008; Testiati et al. 2013) showed that Rosmarinus officinalis is tolerant to very high TMM concentrations and could be used for phytostabilization of amended soils polluted with trace elements (Madejon et al. 2009). Moreover, Affholder et al. (2013) indicated a limited health risk of rosemary consumption as medicinal (herbal tea) or aromatic (dry leaves) herbs, from plants grown on a highly contaminated soil in the Calanques hills, since this plant species does not accumulate high concentrations of As, Pb, Sb, and Zn in its aerial parts. These criteria are interesting for a phytostabilization process.

Besides, as more than 80 % of plant species (Harrison 1999), it has been proven that rosemary forms root symbioses with AM fungi (Estaún et al. 1997; Sánchez-Blanco et al. 2004; Turrini et al. 2010; Sánchez-Castro et al. 2012), particularly with the genus Glomus (Turrini et al. 2010; Sánchez-Castro et al. 2012).

Thus, this study aimed at determining the in natura ability of rosemary to transfer TMM in its roots, namely As, Pb, Sb, and Zn which are major contaminants from past activity of the smelting factory of l'Escalette, and to study the involvement of root symbiosis in tolerance and accumulation abilities of this species. Therefore, TMM accumulation in rosemary roots spontaneously grown in soils presenting a gradient of contamination, localization of TMM in root, and pattern of adsorption was determined. Moreover, root colonization rate by AM fungi and/or DSE was estimated.

Materials and methods

Study area

The study area, located in the Calanques hills, in the peri-urban area of Marseille (southeast of France), was characterized by a matorral vegetation dominated by R. officinalis, Cistus albidus, Quercus coccifera, and Pistacia lentiscus under Mediterranean climatic conditions.

Eight sites were selected for this study, with similar soil composition, vegetation, and climatic conditions (Fig. 1). The sampling areas were chosen along a double transect from the former smelter factory of l'Escalette to the Garenne valley following the direction of the prevailing wind and from the latter to the sea through the Mounine valley. Seven sites were selected along a suspected contamination gradient: G0 on the site of the factory (close to the horizontal creeping chimney exit); G1, G2, and G3 in the Garenne valley; G4 and G5 in the Mounine valley; and G6 in Sormiou cove which is far away from the Escalette and the urban area and could be considered as low contaminated according to the mapping of soil element concentrations conducted in an extended area around the factory site (unpublished results). The last site S3, in a never industrialized part of the Calanques hills, was considered as a reference site since the soil was least contaminated on the mapping previously conducted cited above.
Fig. 1

Geographic localization of the eight sampling sites (plain dots) and the former smelting factory (star-shaped dot) on the Marseilleveyre massif in the National park of Calanques (Southeast France)

Activities of the factory of l’Escalette lasted from 1851 to 1925, and during this period, around 30,000 t/year of silver-galena was treated by pyrometallurgical processes (Raveux 2002). During this industrial activity, approximately 20,000 m3 of slag was deposited on the former factory site, but slags were also scattered along the coast as several main deposits and as road fill. The study area is submitted to the Mediterranean climate with average minimal and maximal temperatures of 7.1 and 24.1 °C, respectively, and a mean annual rainfall of 554 mm on 56 days (reference period 1971–2000). Dominant winds are north and northeast and can reach 120 km/h. The geological underground is limestone and bare rocks dominate. The soils of the selected sites are stony, and their thicknesses vary from place to place but are generally less than 50 cm. Soils are alkaline with an average pH (ISO 10390 2005) between 7.8 and 8.1, belonging to the typical pH range of soils from calcareous areas. Soil fertility is low with total organic carbon contents (ISO 10694 1995) varying from 3.6 to 14.2 %, total Kjeldahl nitrogen (ISO 11261 1995) contents from 0.28 to 0.72 %, assimilable phosphorus (ISO 11263 1994) from 0.010 to 0.057 g P/kg, and cation exchange capacity (CEC, ISO 22036 2008) from 15 to 42 cmol+/kg (Table 1).
Table 1

Soils characteristics of the studied sites


Soil physicochemical parametersa

CEC (cmol+/kg)

N %

C %

P (g/kg)



31.6 ± 6.6 ac

0.72 ± 0.18a

9.6 ± 5.7ab

0.057 ± 0.014

7.9 ± 0.2ab


15.7 ± 1.1bd

0.29 ± 0.04bd

9.4 ± 1.9a

0.014 ± 0.041

8.1 ± 0.08ab


41.8 ± 8.5c

0.55 ± 0.10 ac

12.6 ± 3.1a

0.030 ± 0.010

7.8 ± 0.3a


24.6 ± 6.2acd

0.44 ± 0.05c

11.1 ± 1.5a

0.012 ± 0.002

8.1 ± 0.06b


41.7 ± 15.7 ac

0.66 ± 0.26 ac

14.2 ± 6.1a

0.013 ± 0.006

7.9 ± 0.1a


42.4 ± 13.9c

0.54 ± 0.15 ac

11.2 ± 1.9a

0.013 ± 0.006

7.9 ± 0.1ab


27.3 ± 1.6a

0.43 ± 0.13 cd

9.8 ± 2.9a

0.010 ± 0.002

7.9 ± 0.1a


14.0 ± 8.4d

0.28 ± 0.09d

3.6 ± 2.8b

0.013 ± 0.005

8.1 ± 0.1ab

In the same column, values followed by the same letter are not significantly different (Mann–Whitney–Wilcoxon test, p < 0.05)

CEC cationic exchange capacity (in cmol+/kg), N% total Kjeldahl nitrogen (in percentage), C% total organic carbon (in percentage), P assimilable phosphorus (in g P/kg)

aMean ± standard deviation (n = 5).

Plant and soil sampling

Sample collection was done as previously reported by Affholder et al. (2013) on five rosemary individuals of the same population on each site. In order to obtain representative samples from each site, an area of 100 m2 (10 × 10 m) was delimited on the eight selected sites, i.e., G0, G1, G2, G3, G4, G5, G6, and S3. The plant cover was over 60 % on all the selected areas. The five rosemary individuals were selected according to a cross pattern inside the delimited area, spaced by around 2 m, and were with similar sizes, i.e., heights and collar diameters, and same phenological stage. Therefore, in each of the eight sites, a total of five plant/soil couples were taken. Roots were collected for TMM analysis and symbiosis detection. Soil samples were collected from the top 15 cm (after removal of the litter) in the rhizospheric area of the plants. In addition, on the contaminated site G0 (as previously defined by Affholder et al. 2013), the whole root network of one rosemary individual was sampled in order to determine TMM adsorption and localization in primary and lateral roots. G0 site was the best candidate to ensure TMM localization and quantification in rosemary since it is the most polluted one.

Fresh plant and soil samples were stored in clean plastic bags for transport to the laboratory.

Soil TMM analysis

Soil samples were sieved on site to 2-mm mesh, air-dried at room temperature in laboratory, and then ground (RETSCH zm 1000 with tungsten blades) to pass through a 0.2-mm titanium sieve, before analyses.

Pseudo-total TMM

Soils were mineralized (in triplicates) in a microwave mineralizer (Milestone Start D) using aqua regia (1/3 HNO3 + 2/3 HCl). A ratio of dry soil/aqua regia solution corresponding to 1/20 w/v was used. The mineralization products were filtered with a 0.45-μm mesh, and the TMM levels were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) and graphite furnace-atomic absorption spectroscopy (GF-AAS) (Thermo Scientific ICE 3000) for As and Sb. Quality assurance–quality controls and accuracy were checked using standard soil reference materials (CRM049–050, from RTC, USA) with accuracies within 100 ± 10 %.

Exchangeable TMM

A 0.05 M Ca(NO3)2 solution was used as extractant and was prepared following the protocol of the Community Bureau of Reference (BCR) and Quevauviller (1998). A ratio of dry soil/Ca(NO3)2 solution corresponding to 1/10 w/v was used. The mixture was prepared into a PTFE (Teflon) tube (triplicates per soil sample) and stirred at room temperature on an orbital shaker (Fisher Scientific Bioblock SM30B) at 125 rpm for 2 h. The tubes were then centrifuged for 10 min at 8,000 rpm (JP SELECTA, Médifriger BL-S), and the supernatants were collected and filtered to 0.45 μm. The resulting solutions were acidified to 1 % with supra-pure nitric acid and stored at 4 °C until analysis by ICP-AES or GF-AAS.

Plant TMM analysis

Root samples were thoroughly washed, three times under tap water then three times with a dental spray using Milli-Q water. Samples were dried at 40 °C during 1 week and afterwards ground at 0.2 mm (RETSCH zm 1000 blender).

Pseudo-total TMM

About 0.5 g dry matter of each root sample (triplicates) was digested in microwave mineralizer system (Milestone Start D) with an acidic mixture (2/3 HNO3 + 1/3 HCl). A ratio of dry plant/acid mixture solution corresponding to 1/18 w/v was used. After filtration (0.45 μm), mineralisates were analyzed for Pb and Zn contents by ICP-AES and for As and Sb by GF-AAS. Standard plant reference materials (DC 73349) from China National Analysis Centre for Iron and Steel (NCS) were analyzed as a part of the quality assurance–quality control protocol (accuracies within 100 ± 10 %). Bioconcentration factors (BCFs), i.e., ratios of root concentration vs. soil pseudo-total (BCFtotal) or exchangeable concentrations (BCFexch), were calculated.

Adsorbed TMM at the root surface

Adsorbed TMM extraction was carried on the selected rosemary roots, sampled on the site G0, considering different root diameters (Φ) gathered as follows: taproot (Φ ranging from 10 to 15 mm), thick primary root (Φ ranging from 5 to 10 mm), medium secondary root (Φ ranging from 2.5 to 5 mm), and thin secondary roots (Φ ranging from 1 to 2.5 mm). The dried root samples were immersed in a 20-mmol/L solution of NaEDTA and stirred for 15 min at 125 rpm, following the protocol proposed by Xiong et al. (2011). A ratio of dry root/NaEDTA solution corresponding to ca. 1/20 w/v was used. After stirring, the solution was recovered, filtered to 0.45 μm, and then stored at 4 °C until analysis. In order to determine the percentage of adsorbed TMM, the treated roots were rinsed, dried, and ground for remaining (i.e., absorbed) TMM quantification.

Symbiosis colonization rate

The selected roots have been prepared and stained by adapting the technique of Phillips and Hayman (1970) on rosemary. Previously, the roots were thoroughly cleaned with distilled water in order to remove soil particles. The washed roots were stored at room temperature in ethanol 60 % until proceeding. Then, roots were cleared in KOH 20 % for 50 min at 90 °C. After three rinses with distilled water, roots were soaked in 2 % hydrochloric acid for 5 min. Fungal structures are highlighted by staining in 0.05 % trypan blue in lactophenol. The roots were impregnated with dye and placed in a water bath at 80 °C for 5 min. The excess of dye was removed, and the roots were soaked in glycerol before the slide preparation. The mycelium morphology of DSE was observed at ×400 magnification with an optical microscope. The location of DSE structures including partitioned melanized hyphae and microsclerotia was observed and their colonization percentage estimated. AM structures such as arbuscules, vesicles, and hyphae were examined. The percentages of colonization of AM fungi and DSE were estimated independently according to the method proposed by Zhang et al. (2010). On each microscope slide, 10 fragments of stained roots were placed to estimate the proportion of roots colonized by AM fungi and/or DSE. For each rosemary individual, 10 slides were prepared. Thus, the current results correspond to the examination of 100 root fragments per individual.

TMM localization in roots

Fragments of dried root of various diameters were taken from the whole rosemary sampled on site G0, i.e., one collected on the primary root (root 1), one on a secondary root (root 2), and one on a thinner secondary root (root 3). These fragments were included in an epoxy resin (EPON), using De Jong et al. (2013) technique, and cross sections of 200-μm thickness were cut with a wire saw.

TMM localization in roots was performed by X-ray microfluorescence spectroscopy (μXRF) with a XGT7000 spectrometer (Horiba Jobin Yvon) equipped with an X-ray tube constituted by a rhodium source, running at an accelerating voltage of 30 kV, a current of 1 mA, and in partial vacuum. X-rays emitted from the irradiated sample were detected by an energy dispersive X-ray spectrometry (EDXS) detector equipped with a high-purity silicon, cooled by liquid nitrogen. Elemental mapping (128 px2 meaning a pixel size of 4 μm) showed the distribution of Ca, Fe, K, S, Pb, and Zn in the root sections with a counting time of 20 × 1,000 s. Maps were done on whole root section for both secondary and thinner root sections. For the primary root, only one quarter was mapped since the section was too large for a complete observation. Then, to confirm mapping observations, some X-ray fluorescence spectra were done on selected areas with a counting time of 1,000 s. For primary root, spectra were obtained in the whole section mapped. Concerning secondary and thinner roots, two spectra were performed: one on a part where TMM were detected and the other one beside the root in order to exclude a background noise.

Statistical analysis

Statistical analyses were performed using the R software. Comparisons of means were performed using the nonparametric Mann–Whitney–Wilcoxon test with a p ≤ 0.05.

Significant linear correlations were determined with the same level by the Pearson test available from the Hmisc software.


TMM in soil

The results of the TMM pseudo-total concentrations in soils showed clearly the influence of the former industrial site of l'Escalette on diffuse pollution in its surroundings (Table 2). Indeed, the furthest sites (G4, G5, G6, and S3) presented a significantly lower contamination than those localized closer to the factory ruins (G0, G1, G2, and G3). In the case of important contamination (G0 to G3), relative standard deviation exceeding 50 % was observed between samples from the same site. A wide range of pseudo-total TMM concentrations was covered by whole sites (Table 2) since the minimum and maximum concentrations measured were 4 and 2,738 for As, 21 and 19,646 for Pb, 1.5 and 759 for Sb, and 46 and 5,843 for Zn. S3 could be considered as the reference area for this study since TMM soil concentrations in this site were very close to the local background of contamination in the Calanques hills, which was estimated at <10, 40, <50, and 53 mg/kg for As, Pb, Sb, and Zn, respectively (Affholder et al. 2013). Thus, TMM concentration values from S3 were used to calculate contamination factors (CF = average TMM soil concentration on a site / average TMM concentration of the local background). These factors gave a contamination level per element and allowed to estimate a multicontamination level by calculating pollution load index (\( \left(\mathrm{PLI}\ {=}^4\sqrt{{\mathrm{CF}}_{\mathrm{As}}\times {\mathrm{CF}}_{\mathrm{Pb}}\times {\mathrm{CF}}_{\mathrm{Sb}}\times {\mathrm{CF}}_{\mathrm{Zn}}}\right) \)) (Rashed 2010) (Table 3). Concerning G0, CF values showed contamination above 200 for As and Pb, >100 for Sb, and >40 for Zn. These factors decreased slightly in site G2 for all TMM, and then, CF values were divided by a factor higher than 2 on G1 and G3. Sites G4 and G5 presented similar CF values between 10 and 15 for As and Pb and <5 for Sb and Zn. Finally, G6 appeared slightly contaminated with CF <2 for Pb, Sb, and Zn but still ca. 10 for As. Concerning PLI values, results showed an important decreasing multicontamination following this order: G0 > G2 > G3 > G1 > G4 ≥ G5 > G6.
Table 2

As, Pb, Sb, and Zn average (in bold, ±standard deviation, n = 5), minimal, and maximal concentrations in soils of the eight studied sites, for pseudo-total (total) and exchangeable (Ca(NO3)2) TMM fractions in milligrams per kilogram



Trace elements (mg/kg)
























1,134 ± 934

0.66 ± 0.44

9,210 ± 6,193

0.77 ± 0.59

319 ± 256

1.47 ± 0.82

2,723 ± 1,874

3.10 ± 2.95





















306 ± 156

0.09 ± 0.05

3,031 ± 1,835

0.56 ± 0.73

112 ± 60.9

0.21 ± 0.08

1,017 ± 539

0.91 ± 1.01





















949 ± 543

0.34 ± 0.17

8,939 ± 4,475

0.73 ± 0.32

284 ± 158

0.79 ± 0.55

2,821 ± 1,802

3.83 ± 3.78





















390 ± 503

0.16 ± 0.06

3,445 ± 4,367

0.29 ± 0.51

119 ± 159

0.30 ± 0.37

1,523 ± 2,182

2.52 ± 4.65





















53.1 ± 37.5

0.08 ± 0.05

605 ± 250

0.73 ± 0.55

10.5 ± 4.2

0.03 ± 0.01

295 ± 87.2

0.57 ± 0.40





















53.3 ± 14.2

0.08 ± 0.07

480 ± 247

1.0 ± 0.15

9.6 ± 2.8

0.01 ± 0.01

293 ± 114

1.4 ± 2.2





















48.1 ± 9.9

0.009 ± 0.016

62.0 ± 15.0

0.0005 ± 0.0011

4.0 ± 3.1

0.003 ± 0.003

100 ± 18.4

0.20 ± 0.14





















4.9 ± 0.69

0.06 ± 0.04

42.9 ± 17.3

0.008 ± 0.004

3.1 ± 1.8

0.009 ± 0.011

66.0 ± 23.3

0.16 ± 0.29










Min minimal, Max maximal, Aver average, QL quantification limit (≤0.010, ≤0.0015, and ≤0.016 mg/kg for exchangeable As, Pb, and Zn, respectively)

Table 3

As, Pb, Sb, and Zn contamination factors (CF) and pollution load index (PLI in bold) calculated for all studied sites


Contamination factors
















































Exchangeable TMM in soils were low as the maximum concentrations analyzed were 1.42, 1.80, 2.75, and 10.8 mg/kg for As, Pb, Sb, and Zn, respectively (Table 2). Thus, there were low percentages of exchangeable TMM, considered as the bioavailable fraction (Meers et al. 2007), since ratio between exchangeable and pseudo-total fractions was generally below 1 %, whatever the element (Table 2). On another hand, a significant and positive correlation (Pearson test, p ≤ 0.05) of 0.66 was observed between the exchangeable concentrations and pseudo-total concentrations of As and Pb, 0.84 for Zn, and 0.93 for Sb. Thus, the average exchangeable fractions seemed to be correlated with concentrations of pseudo-total TMM.

TMM in roots

TMM concentrations in studied roots were between 0.22 and 85, 0.8 and 1,983, 0.02 and 35, and 5.4 and 803 mg/kg of dry weight (DW) for As, Pb, Sb, and Zn, respectively (Fig. 2). The highest TMM concentrations were found in the roots of rosemary individuals from highly contaminated sites (G0 to G3). Indeed, for Pb, Sb, and Zn in the four most contaminated sites (G0, G1, G2, and G3), the average root concentrations were significantly higher (Mann–Whitney–Wilcoxon test, p ≤ 0.05) than those observed in roots collected in the less contaminated sites (G4, G5, G6, and S3). For As, only the average TMM root concentrations of G0, G1, and G2 sites were significantly higher than those of the five others (Table in Supplementary Material). Thus, TMM contents in roots were influenced by soil contamination level. Figure 2 shows TMM concentrations in roots in function of TMM pseudo-total concentrations in soil, and in this figure, each point represents one rosemary individual. There is a high discrepancy between each soil sample from the same site, especially in the most contaminated sites. However, a significant linear correlation (Pearson's test, p ≤ 0.05) was demonstrated for As, Pb, Sb, and Zn with a correlation coefficient of 0.52, 0.56, 0.52, and 0.7, respectively. Thus, TMM concentrations in the roots of rosemary individuals were linearly correlated with the pseudo-total concentrations of soils.
Fig. 2

Relationships between root concentrations and pseudo-total soil concentrations for As, Pb, Sb, and Zn (logarithmic scale, μg/g DW), over the eight sampling sites. Values for samples from the same site are represented by dots with same form and color. DW dry weight

Data of TMM concentration in roots of different diameters collected on the rosemary individual sampled on site G0 showed a tendency to a larger storage in the finest roots for As, Pb, Sb, and Zn than in the thickest ones (Fig. 3). However, the low number of replicates (n = 3) for this experiment did not allow to highlight TMM root concentrations depending on the root diameter with statistically significant difference.
Fig. 3

As, Pb, Sb, and Zn root concentrations (μg/g DW, n = 3) according to root diameters i.e., taproot, thick primary root, medium secondary root, and thin secondary root from the selected rosemary individual sampled on site G0

Root bioconcentration factors (BCF)

Root BCFtotal values calculated with soil pseudo-total concentrations were low for all elements since they were far below 1. Indeed, the average values of root BCFtotal for all sites were between 0.02 and 0.23, 0.04 and 0.26, 0.03 and 0.17, and 0.08 and 0.25 for As, Pb, Sb, and Zn, respectively. For these TMM, the maximal average values were obtained on the reference site, S3. However, despite the low variability of the soil contamination on this site, the root BCFtotal values were widely dispersed (between 30 to 77 % of variation). No significant differences were found for the average values of root BCFtotal between the studied sites and also among the various elements.

Root BCFexch values calculated from soil exchangeable concentrations showed average values between 24 and 246, 49 and 17,632, 6.3 and 115, and 110 and 350 for As, Pb, Sb, and Zn, respectively. The average values of root BCFexch were all distinctly above 1, with some particularly high values for Pb in contaminated sites (G3). Yet, as for BCFtotal, no significant differences were found for the average values of root BCFexch between the studied sites and also among the various elements.

Ab- and adsorption of TMM at the root level

Figure 4 shows the percentage of As, Pb, Sb, and Zn absorbed into roots regarding their diameter. The low number of replicates (n = 3) did not allow to determine significant difference between TMM quantities absorbed in roots regarding to root diameters. However, results suggested that all studied elements were not similarly treated in roots. Average TMM absorbed fractions ranged from 46 to 70 %, 30 to 51 %, and 49 to 67 %, for As, Pb, and Zn, respectively. Thus, for these three elements, the absorbed proportion was highly variable and could be lower, equal, or higher than the adsorbed part. But for Sb, absorption seemed to be the major way of sorption by roots since average absorbed fractions varied between 91 and 96 %.
Fig. 4

Percentage of As, Pb, Sb, and Zn absorbed by roots according to root diameters, i.e., taproot, thick primary root, medium secondary root, and thin secondary root from the selected rosemary individual sampled on site G0

Symbiosis colonization percentage

Figure 5 shows a significant positive correlation (Pearson test, p ≤ 0.05) between the percentages of AM colonization of rosemary roots in function of TMM pseudo-total concentrations in soil, over the eight studied sites. Thus, a higher AM colonization was observed in the most contaminated soils. Strong correlations were observed for As, Pb, Sb, and Zn with high correlation coefficients (>0.6).
Fig. 5

Arbuscular mycorrhizal (AM) colonization rates (in percent) in function of pseudo-total soil concentrations in As, Pb, Sb, and Zn (logarithmic scale, μg/g DW) for all rosemary individuals on the eight sampling sites. Values for samples from the same site are represented by dots with same form and color. DW dry weight

In Fig. 6, the percentages of rosemary root colonization by DSE in function of TMM soil pseudo-total concentrations are shown along six sites of the transect. No DSE were detected in S3 root fragments while DSE occurrence was observed in G0. However, the quality of the root preparation did not allow to include the observation data in the statistical analysis. The most important percentage of DSE colonization (24.5 %) calculated by site was obtained for rosemary individuals that had grown on the site G2, the most contaminated site. On the other sites, except on G4 (average rate of 5 %), DSE colonization was similar with an average percentage ca. 10 %. In this case also, a significant positive correlation (Pearson test, p ≤ 0.05) between DSE colonization and TMM soil contamination was highlighted with correlation coefficients higher than 0.5 for As, Pb, Sb, and Zn.
Fig. 6

Dark septate endophyte (DSE) colonization rates (in percent) in function of pseudo-total soil concentrations in As, Pb, Sb, and Zn (logarithmic scale, μg/g DW) for all rosemary individuals on the six sites along the transect. Values for samples from the same site are represented by dots with same form and color. DW dry weight

TMM localization in roots

Localization and pictures of the analyzed root samples collected on the selected rosemary individual on the site G0, a section of primary root (root 1) and two sections of secondary roots, with different diameter (root 2 for the largest and root 3 for the thinnest), were presented in Fig. 7. It also shows μXRF mapping results for the three studied root samples, where elements detected by μXRF are represented by white dots. Only Pb and Zn were detected in this analysis, the other elements being probably below detection limits of the device. The maps produced by X-ray microfluorescence spectroscopy highlighted the presence of Pb and Zn in the peripheral tissues of the lateral roots (roots 2 and 3). On the primary root, none of these elements were detected.
Fig. 7

Localization and pictures of the three different root samples collected on a rosemary individual from site G0 and treated for micro-X-ray fluorescence spectrometry analyses and μXRF maps obtained on each root sample for two elements: Pb and Zn (white dots represent detected elements)

In order to confirm μXRF mapping observation, X-ray fluorescence spectra were recorded on each of the three root samples by analyzing the whole mapped section for root 1 and only the selected areas located inside the marked squares for roots 2 and 3 (Fig. 8). The spectrum of the root 1 showed the presence of Zn and confirmed the absence of Pb (Fig. 8), while the presence of both elements was confirmed by spectra from the selected areas in roots 2 and 3. A background effect could be rejected since spectra obtained after analysis of areas outside the roots did not show peaks corresponding to Pb nor Zn (Figure in Supplementary Material).
Fig. 8

Micro-X-ray fluorescence analysis spectra from whole mapped area of a primary root (root 1 and spectrum 1) and from selected areas on secondary root, located inside marked squares (roots 2 and 3 and spectra 2 and 3, respectively). Spectra 2 were recorded in a zone of the roots 2 and 3 where Pb and Zb were detected during mapping


Diffuse soil contamination

The results of the TMM pseudo-total concentrations in soils clearly showed the influence of the former industrial site of l’Escalette on diffuse contamination in Calanques Hills. Indeed, contaminations of the farthest sites (G4, G5, G6, and S3) were significantly lower than those located close to the former smelter factory (G0, G1, G2, and G3) (Tables 2 and 3). Furthermore, in the case of highly contaminated areas, the high heterogeneity of TMM soil concentrations was characteristic of anthropogenic contamination, particularly contamination related to atmospheric deposits from industrial activities (Yaylali-Abanuz 2011; Testiati et al. 2013).

PLI order of values (similar to CF for As, Pb, Sb, and Zn) showed that, for the four most contaminated sites, the distance was not the only criterion affecting the level of contamination. Indeed, a higher contamination level was found on the site G2 than on the site G1, while the latter was closer to the former factory. Two phenomena might be at the origin of the very high contamination measured on G2 (similar to that measured in G0, located close to the old creeping chimney exit). On one hand, this site could be exposed to vortex phenomena promoting contaminated particle deposition as G2 was located in the Garenne valley directly in the axis of the prevailing wind from the former factory and the chimney exit and where the valley presented a narrowing and a sharp altitude increase. In addition, soil of G2 was a colluvial soil formed by erosion and leaching of upstream sites, where contamination particles were likely to accumulate.

In soils from each site, TMM exchangeable fraction was low (Table 2). Exchangeable TMM correspond to the weakly adsorbed part on solid phase of soil, mainly constituted of exchangeable ions within soil solution (Morel 1997), which can be considered as phytoavailable TMM. An important result in our study was that a significant positive linear correlation was observed between the exchangeable and pseudo-total concentrations of As, Pb, Sb, and Zn in soils, meaning that TMM exchangeable concentrations increased with pseudo-total TMM concentrations (results not shown). Thus, the risk of transfer of these elements to the biocenosis seemed to be more important in the heavily contaminated sites.

Rosemary as a good candidate for phytostabilization

Linear correlation between TMM concentrations in soil and concentrations in root (Fig. 2) matched the definition of an indicator species according to Baker (1981). It also suggested a major passive root absorption for Zn and at least a partial passive absorption for As, Pb, and Sb. The maxima of TMM concentration in roots were 100, 50, 10, and 10 times higher for Pb, As, Sb, and Zn, respectively, than those found in aerial parts (results not shown). Moreover, the TMM concentrations measured in rosemary roots reached high values, highlighting that this species accumulated preferentially TMM in roots than in aerial parts. The ability of rosemary root to accumulate TMM was determined using root BCF. Following Baker and Brooks (1989), a BCF higher than 1 for one element means that the plant species is able to hyperaccumulate this element. Thus, as BCFtot calculated with soil pseudo-total concentrations were lower than 1 for all TMM and all sites, rosemary could not be considered as a hyperaccumulator species. However, the nonsignificant differences between BCFtot obtained from all sites whatever the element, except for S3, showed that TMM accumulation in rosemary roots was not dependent on contamination level. Moreover, BCF values were of the same order of magnitude for the four elements, highlighting that rosemary did not accumulate preferentially a TMM than another. Moreno-Jimenez et al. (2011) obtained similar BCF values for the same species. For soil pseudo-total concentrations in the same range of values, they calculated root BCFtot of 0.025 and 0.1 against 0.06 to 0.03 and 0.13 to 0.17 in the present case for As and Zn, respectively. Unfortunately, to our knowledge, there is no data on Pb and Sb analysis in rosemary roots in the literature. However, root BCFexch were largely above 1 for all TMM, and this highlighted that rosemary was able to accumulate in its root the most labile fraction of TMM in soil, which presented the largest risks for environment.

Insight in the mechanism involved in TMM storage in rosemary root

TMM partition in rosemary root system

The maps and spectra obtained by X-ray microfluorescence spectroscopy analyses highlighted that Pb and Zn were stored preferentially in peripheral tissues and in thin roots (Figs. 7 and 8). These results seemed to be confirmed by the results of TMM concentrations in roots regarding the different root diameters (Fig. 3). Indeed, a global trend was observed, showing higher total TMM concentrations in both fine and medium roots (corresponding to secondary roots) than in thick roots (primary root) and in taproots. Vamerali et al. (2009) showed similar results in poplar and willow, i.e., a higher concentration of As, Cu, Pb, and Zn in fine than in thicker roots of these species. These observations could be related to a higher concentration of tannins in fine roots as it has been highlighted by Chin et al. (2009) in Symphytum officinale. Indeed, the tannins are able to bind TMM and thus increase their adsorption on the surface of roots. However, as shown in Fig. 4, in the case of rosemary, the absorbed TMM fraction seemed not to be affected by root diameter. Despite this, the results showed that an important part of As, Pb, and Zn was adsorbed since, for instance, until 70 % of Pb could be adsorbed on rosemary roots growing in contaminated site. This low absorption could be related to a mechanism of resistance by rosemary, promoting the adsorption of these elements on the surface of roots. Indeed, previous studies have shown that the presence of TMM in soil, including Pb, resulted in an increase of phenolic compound production (Kovacik and Klejdus 2008; Pawlak-Sprada et al. 2011) which could be able to bind TMM on root surfaces (Michalak 2006). To confirm this hypothesis, a similar study with rosemary roots from a non-contaminated site would have been required to compare the patterns of absorption and adsorption of these elements in both cases, i.e., in low and heavily contaminated environments, if the low level of elements in non-contaminated soils is compatible with the level of detection of the method.

Antimony seemed to have a different behavior. Indeed, this element was mainly absorbed (over 90 %), which was also observed in this study for Cu and Fe (results not shown). Usually, Sb and As behaviors are similar (Wilson et al. 2010). However, in this case, both elements did not seem to be absorbed by rosemary in the same way. Little is known about the absorption pathway of Sb in plant roots, but it is likely that both forms of Sb are not absorbed in the same way. Sb(III) can pass through the membrane aquaporins passively. For Sb(V), which is an anion, this kind of passive transport is excluded, and its absorption requires a carrier (Tschan et al. 2009). However, in discharges of lead smelters, the dominant form of Sb in soil seems to be Sb(V) as described by Ettler et al. (2010), suggesting an active absorption of this element. Nevertheless, the hypothesis that this observation is related to a not well-adapted extractant for lixiviation of Sb bond on root surfaces could not be dismissed.

A contribution of root symbionts

Results highlighted that root colonization rates by AM fungi and DSE increased with soil contamination. It seemed that the fungi strains present in soils in the surroundings of l’Escalette former factory were adapted and resistant to high levels of TMM contamination (Figs. 5 and 6). Although similar results have been identified in the literature for AM fungi (Whitfield et al. 2004), contradictory data exist. Thus, several studies showed a decrease in mycorrhizal colonization with increasing TMM concentrations in soil (Gildon and Tinker 1983; Del Val et al. 1999; Andrade et al. 2004; Yang et al. 2008; Turrini et al. 2010). Mechanisms used by AM fungi to increase plant tolerance to TMM depend on the studied species and the strategy response of plants to tolerate TMM. It seems that for accumulator plants with a high TMM translocation potential towards the aerial parts, the presence of AM fungi increases this phenomenon, whereas in plants that exclude TMM, the opposite phenomenon is observed and amplified by the presence of AM fungi (Rabie 2005), allowing improving plant capacity to extract or stabilize TMM (Gohre and Paszkowski 2006). Concerning DSE, similar results have also been highlighted on Salix caprea, in which a positive correlation between DSE colonization rates and concentrations of Pb in soil was shown (Regvar et al. 2010). In addition, Li et al. (2011) showed a significant increase of Pb, Zn, and Cd concentrations in the roots of Zea mays and a decrease in the translocation of these elements to aerial parts in the presence of DSE.

In the case of rosemary, we hypothesized that its association with fungal species involved in root symbioses could promote accumulation in roots and limit the transfer into the aerial parts, playing the role of a filter. This might be due to adsorption and/or absorption in fungal compartments. AM and DSE hyphae may adsorb significant amounts of TMM. However, X-ray microfluorescence methods require thick cross sections of non-stained roots on which the microscopic observation of fungal structures is not easy. Thus, it was not possible to distinguish if TMM were or not inside fungal structures. According to Gohre and Paszkowski (2006), metal-tolerant fungi may have a greater affinity for TMM than roots have (2.4 times). Hyphal cell walls are constituted by compounds able to adsorb TMM on their surface, like melanin for DSE (Gadd 1993) and chitin (Gohre and Paszkowski 2006) and glomalin (Driver et al. 2005) for AM fungi. For example, glomalin, which is a glycoprotein, is able to extract and sequester large amounts of TMM including Cu, Pb, and Zn (Gonzalez-Chavez et al. 2004; Chern et al. 2007; Vodnik et al. 2008; Cornejo et al. 2008). Results concerning the sequestration capacity of TMM by glomalin are contrasting. However, Cornejo et al. (2008) and Chern et al. (2007) agree on the fact that up to 4.8 mg Cu and 188 mg of Pb could be sequestered by 1 g of glomalin. Concerning AM vesicles, they also play an important role since these fungal structures allow the detoxification and storage of toxic elements. Their involvement in the immobilization of Pb has been demonstrated by Chen et al. (2005), who showed a correlation between the number of fungal vesicles and Pb root uptake. It is noteworthy that a considerable amount of fungal vesicles was found in the roots of rosemary exposed to high contamination and that a significant correlation of 0.34 was highlighted between Pb concentration in rosemary roots and AM vesicle percentage (Pearson test, p ≤ 0.05).

AM and DSE structures are located in the cortex of thin roots, where TMM and particularly Pb and Zn were heavily concentrated in rosemary root. Thus, the involvement of these fungal structures in rosemary root accumulation ability seems significant. Moreover, AM fungi and DSE could also be involved in rosemary tolerance to high TMM contamination levels. Indeed, previous studies highlighted that, in the case of rosemary, these symbioses improved water status of the plant, promoting resistance to water stress (Sánchez-Blanco et al. 2004), and increased the survival and growth of rosemary through a replanting process (Estaún et al. 1997). More generally, AM fungi and DSE are able to improve plant growth, health, and tolerance to contamination promoting phosphorus nutrition (Haselwandter and Read 1982; Harrison 1999; Barea et al. 2002). Concerning AM fungi, glomalin excretion in soil, following hyphae turnover (Driver et al. 2005) but also past fungal growth and slow turnover of this protein once in soil (Purin and Rillig 2007), helps to stabilize soil aggregates (Wright and Upadhyaya 1998; Bedini et al. 2009). This enhances soil structure favoring soil aeration and water drainage which optimize plant development (Oades 1984).


Developing non-intrusive ecotechnologies for TMM containment is a challenging target with an increasing demand in protected areas like those of the National Park of Calanques. The use of native plant species with their native symbionts offers a promising perspective to manage TMM phytostabilization objectives and ecological and biological conservation constraints. We demonstrated the ability of rosemary, a perennial with an extensive root network, to accumulate TMM in its roots. Indeed, accumulations of As, Pb, Sb, and Zn in rosemary roots were linearly correlated to their pseudo-total concentrations in soil. Despite concentrations could be important in the roots of rosemary individuals subjected to high levels of contamination, the root BCFtotal were low (<0.3) for these four elements, which highlighted that rosemary was not able to hyperaccumulate TMM in its root system. However, although this type of extraction (i.e., pseudo-total soil concentration) was questionable because it is not necessarily representative of the phytoavailable fraction, BCFexch values, calculated with exchangeable fractions in soils, were greater than 1 meaning that rosemary was able to significantly accumulate the most labile TMM in soils. Indeed, exchangeable TMM were representative of the fraction which was the most likely to leach during rainfall events and/or to be dissolved in case of physicochemical changes in the soil (pH, Eh, complexing conditions, etc.), promoting expansion of TMM contamination into the environment (groundwater, biocoenosis, etc.).

Among the TMM detected in roots, a significant proportion of As, Pb, and Zn seemed to be mainly adsorbed on root surfaces and non-absorbed into roots. This was not the case for Sb, which seemed mainly absorbed. For As, Pb, and Zn, strong adsorption on roots could lead to the limited accumulation in aerial parts, previously highlighted by Affholder et al. (2013), and which is reassuring since rosemary aerial parts are edible. From an ecological point of view, the containment of TMM in root parts is also of interest lowering the risk of transfer to the food web. Root symbiotic role in rosemary tolerance to TMM and in TMM accumulation increase in root was difficult to prove with this in situ study. However, the positive correlation observed between the root colonization rates and the TMM contamination levels, as well as the storage of TMM in peripheral tissues of secondary roots, fully agrees with this hypothesis. Endomycorrhizal fungi developed in the cortical parenchyma and rhizoderm of rosemary secondary roots. Thus, this study suggested that rosemary in association with its symbiotic microorganisms could be an interesting candidate for a phytostabilization process of TMM-contaminated soil in the surroundings of l'Escalette former smelter factory or for other equivalent contaminated sites in the Mediterranean area. However, further studies are necessary (i) to elucidate TMM tolerance mechanisms in rosemary, (ii) to determine more precisely symbiont role in rosemary tolerance/accumulation and, (iii) to characterize genetical diversity of tolerant indigenous fungal strains.


The authors thank Laetitia De Jong for her advice for root sample inclusion, Perrine Chaurand for her technical assistance in μXRF analyses, Laurent Vassalo for his analytical assistance in TMM measurements, and Estelle Dumas for her help at designing Fig. 1. This study was funded by the French Research National Agency (ANR Marséco 2008 CESA 018) and financially amended by the National Innovative Cluster on Risks Management.

Supplementary material

11104_2014_2135_Fig9_ESM.jpg (399 kb)
Figure in Supplementary Material

(JPEG 399 kb)

11104_2014_2135_MOESM1_ESM.doc (34 kb)
Table in Supplementary Material(DOC 33.5 kb)

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Marie-Cécile Affholder
    • 1
  • Anca-Diana Pricop
    • 1
    • 2
  • Isabelle Laffont-Schwob
    • 2
  • Bruno Coulomb
    • 1
  • Jacques Rabier
    • 2
  • Andreea Borla
    • 2
  • Carine Demelas
    • 1
  • Pascale Prudent
    • 1
  1. 1.Aix Marseille Université, CNRS, Laboratoire de Chimie de l’Environnement (LCE)Marseille cedex 03France
  2. 2.Aix Marseille Université, Avignon Université, CNRS, IRD, Institut Méditerranéen de Biodiversité et d’Ecologie marine et continentale (IMBE)Marseille cedex 03France

Personalised recommendations