Plant biomass and shoot analyses
Investigating plant growth and Ni accumulation from soils ranging from typically to low ultramafic provided interesting insights concerning plant growth and the acquisition of Ni and other nutrients by O. chalcidica. Remarkably, the highest plant shoots biomass was obtained from the two soils with the more pronounced ultramafic characteristics (soils S6 and LS, Fig. 1), despite the very different plant availability of essential nutrients such as P, N (Fig. S2, supplementary information) and Fe (Table 2). This reflects specific adaptation strategies of O. chalcidica, where plant growth seems to be favoured by typically ultramafic soil characteristics. Shoots Ni accumulation from all soils resulted to be over the hyperaccumulation threshold of 1000 µg g− 1 in dry shoots (Baker and Brooks 1989; van der Ent et al. 2013) and comparable with previous experimental work on O. chalcidica (Bani et al. 2015a; Bani et al. 2015b; Rosenkranz et al., 2019; Tognacchini et al., 2020). As for the shoot biomass, also the highest Ni shoot uptake in O. chalcidica was obtained from the more distinctly ultramafic soils S6 and LS (Fig. 2), suggesting that specific soil ultramafic properties might promote Ni shoot uptake more than Ni availability itself. A special remark should be done concerning the significantly higher P plant uptake in LS (Fig. 2), which might indicate very efficient P acquisition strategy or favourable soil conditions for P uptake in LS soil, which had the lowest P availability (Table 2). Possibly, P uptake was favoured in O. chalcidica by certain soil characteristics of LS, such as higher pH or very low concentrations of soluble Fe in pore water.
Soil pore water analyses
The significant increase in Ni pore water (PW) concentrations in planted pots from T0 to T3 in soils S3, S4, S5 and S6, and marginally S1 and S2 (Fig. 3) is a clear evidence of plant-induced Ni solubilization from insoluble Ni pools to soil water solution. Conversely, the decrease in PW Ni concentrations in soil LS along the experiment suggests either that no Ni mobilization occurred, or a stronger effect of plant uptake over solubilization. The strong Fe solubilization observed in planted pots along the experiment in all soils except LS (Fig. 3) further confirms rhizosphere processes involved in metal mobilization; the high correlation (Pearson´s r = 0.81) between Ni and Fe solubilization also suggests a co-mobilization of those metals from the same soil fractions, possibly of Ni associated with soil Fe oxides, which was shown to be one of the main source of labile Ni in ultramafic soils (Álvarez-López et al. 2021; Chardot et al. 2005; Massoura et al. 2004). As for Ni, also the limited Fe solubilization in LS is reflecting the different geochemistry and rhizosphere processes for this soil. Furthermore, the significant increase in DOC in planted pots along the experiment (Fig. 3) is likely due to root-related increase of soluble organic compounds and a possible indication of root exudation by O. chalcidica in all soils. The increased DOC in pore water also seems to promote Ni and Fe soil solubilization in planted pots, as suggested by the high positive correlation between ΔDOC /ΔNi and ΔDOC/ΔFe in PW at T3 (Pearson´s r = 0.96 and r = 0.86 respectively, Fig. S1 supplementary information). A further indication that plant-derived DOC is involved in Ni and Fe solubilization, is suggested by the fact that lacking Ni and Fe PW mobilization in soil LS was associated with negligible DOC increases. Similarly to our results, an increased Ni solubility in rhizosphere of ultramafic soils compared to bulk soil was observed in Wenzel et al. (2003) and Álvarez-López et al. (2021) on, respectively, the Ni hyperaccumulator species Noccaea goesingensis and Odontarrhena serpyllifolia. In both studies, increased Ni solubility was associated with higher DOC and pH in soil and a positive correlation between soil Ni availability and DOC in the rhizosphere was also observed. Puschenreiter et al. (2005) also reported higher concentrations of oxalic and citric acid in the rhizosphere compared to bulk soil in of Thlaspi goesingense from a natural ultramafic site. Based on chemical speciation analysis (MINTEQA2), Wenzel et al. (2003) suggested the formation of Ni-organic complexes in the rhizosphere of N. goesingensis and that root exudated organic ligands might form stronger complexes with Ni then organic compounds from bulk soil. This might explain the enhanced Ni present in PW, seemingly in complexed form with organic ligands from root exudation.
The pH increase of pore water samples from rooted pots along the experiment (Fig. 3) suggests a significant root-induced soil alkalinization, as observed in previous studies on hyperaccumulator plants (Álvarez-López et al. 2021; Kukier et al. 2004; Luo et al. 2000; Puschenreiter et al. 2005; Singer et al 2007; Wenzel et al. 2003; Wenzel et al. 2004). Wenzel et al. (2003) hypothesised that the pH increase in the rhizosphere of N. goesingensis could be related with the release of hydroxyl ions during mineral dissolution of Mg and Ni-bearing orthosilicates. From a hydroponic test with O. chalcidica (Tognacchini et al. 2022, unpublished) a substantial pH increase in a sampling solution was also observed within two hours of root exposure, which implies other mechanisms of pH increase besides mineral dissolution. As previously proposed in literature, possible beneficial effect of pH increase in the rhizosphere might be the stabilization of metal-organic ligands (Li et al. 2003; van der Ent et al. 2016), thus keeping in solution metals as Ni and Fe associated with soluble organic compounds. Because of the very low P concentrations in PW (results not shown) and partially unclear results, a thorough discussion about P geochemistry and mobilization is unfortunately limited. Despite this, a general trend in PPW increase in planted pots was observed, which would suggest a plant-promoted solubilization. Exception, again, is made for soil LS, where a significant P depletion occurred, which might be reflected in the enhanced plant uptake observed in plant shoots of O.chalcidica.
Soil Ni availability and plant uptake
Another main research question we wanted to investigate was whether a linear relation between a soil Ni gradient and plant uptake could be observed. From our results, the increasing Ni availability from soil S1 to soil S6 (NiPW, NiDTPA and NiSr(NO3)2 ) was not entirely reflected in shoot Ni uptake in O. chalcidica. Plant shoots Ni seems to follow a Ni availability gradient only limited to some experimental soils (S1, S2, S3 and S6). Especially, the fact that the highest Ni uptake in O. chalcidica shoots were obtained with very contrasting methods for determining soil Ni availability (NiPW, NiDTPA, NiSr(NO3)2) from soils S6 and LS is clearly showing the limitations in predicting shoot Ni uptake based on soil availability assessments alone. Previous studies conducted on ultramafic soils from the same location show a weak correlation of Ni concentration in Noccaea goesingensis shoots with soil available Ni assessed by Sr(NO3)2 extraction, DTPA extraction and DGT (diffusive gradients in thin films) assessment (Noller 2017; Puschenreiter et al. 2019). Bani et al. (2014) observed that Ni accumulation in Odontarrhena chalcidica was independent from soil DTPA-extractable Ni. On the contrary, Centofanti et al. (2012) showed that Ni accumulation in Alyssum corsicum was dependent upon the solubility of the Ni mineral present in the growth substrate. While in our results a linear relation could be observed in specific range of Ni availability (e.g. from soil S1 to S3), this is not valid anymore considering a larger Ni range and this could partially explain contradictory literature results. Surprisingly, from our experiment shoot Ni uptake in O.chalcidica resulted to be highly predictable from the soil pseudo-total Ni and pH of soil and PW, as confirmed by the very high positive correlations (respectively: Pearson´s r = 0.87, Pearson´s r = 0.72 and Pearson’s r = 0.75; see Fig. S1 in supplementary information). This aspect would suggest that soil total Ni pools and pH might be better predictors of shoot Ni concentrations than the Ni plant-available fractions. From this observation, it could be deduced either that Ni mobilization processes from non-available Ni pools might play a central role in hyperaccumulation, or/and that Ni uptake might be regulated by soil pH. As already underlined in literature, soil pH seems to have a central role in plant Ni uptake (Everhart et al. 2006; Ghafoori et al. 2022; Li et al. 2003; Kukier et al. 2004). In particular, in several studies a higher Ni accumulation by Alyssum (synonymous Odontarrhena) species was observed as soil pH was raised and thus soil available Ni (NiDTPA, NiSr(NO3)2 and Ni biosensor) declined (Everhart et al. 2006; Kukier et al. 2004; Li et al. 2003). Since increasing soil ultramafic properties (and total Ni) were associated with increasing pH (Table 2 ), the apparent linear relation of shoot uptake with total soil Ni might be in fact a bias related to the effect of soil pH. Another relevant aspect is the opposite responses to soil alkalinization of extraction-based Ni availability assessments (DTPA and Sr(NO3)2) and Ni in PW. The reduction in DTPA and Sr(NO3)2 extractable Ni after plant growth (Fig. 4) cannot be justified by plant uptake or PW removal alone, which accounts for approximately half of the Ni loss and it is seemingly a combined effect of immobilization due to pH increase. In fact, it was previously investigated that availability of Ni to plants is mainly controlled by soil pH (Echevarria et al. 2006). In contrast with our results, Álvarez-López et al. (2021) measured a significantly higher available Ni (Sr(NO3)2 and DGT) in rhizosphere soil of O. serpillifoilia compared to bulk soil, showing that rhizosphere processes in ultramafic soil induced Ni mobilization. Being a field study, in Álvarez-López et al. (2021) the long-term rhizosphere effect might justify the different results obtained compared to our pot experiment. Also, considering the soil extractable Ni fractions only (NiDTPA and NiSr(NO3)2; Fig. 4) we would have concluded that rhizosphere processes have caused Ni immobilization, while PW analyses resulted to be crucial in observing Ni solubilization. Our results might clarify that even if Ni availability decreases with typical assessments as soil extractions, Ni concentrations might at the same time be increasing in soil solution in complexed form with organic ligands. Contrasting literature results can be found regarding the capability of hyperaccumulator plants to access larger Ni pools than non-accumulators. For example, it was shown that the hyperaccumulators Odontarrhena spp. (Alyssum murale) accesses the same Ni pool as non-hyperaccumulators (Massoura et al. 2004; Shallari et al. 2001;). Similarly, it was observed that the hyperaccumulator plant Noccaea caerulescens takes up Cd from the same pool as other plants (Gérard et al. 2000; Hamon et al. 1997; Hutchinson et al. 2000). In contrast, Chardot-Jaques et al. (2013) observed enhanced dissolution of a Ni-bearing mineral in the rhizosphere of the Ni hyperaccumulator Bornmuellera emarginata (syn. Leptoplax emarginata). Although in the present experiment indication of active Ni solubilization in the rhizosphere of O. chalcidica was observed, it cannot be stated if this indicates a mobilization from larger Ni pools which are unavailable to non-hyperaccumulators. Rhizosphere Ni mobilization in soil pore water could support the hypothesis that hyperaccumulators rely on metal mobilization from less available metal fractions; however, a clear link with Ni shoot uptake was not observed and high Ni accumulation in O. chalcidica occurred in soil LS without signs of solubilization. Especially, the absence of a clear link between rhizosphere processes and plant Ni uptake suggests that although root mobilization takes place, it is not representing a strategy to enhance Ni plant uptake. As also suggested by Álvarez-López et al. (2021), Ni mobilization seems to be a consequence of mobilization targeting other nutrients. In particular, considering our results, we could hypothesize that a specific nutrient acquisition strategy based on rhizosphere alkalinization and formation of metal complexes with organic ligands, might have developed in O. chalcidica as adaptation to ultramafic soils conditions.