Casting a wide cross-species transcriptomics net: convergent evolution of nickel hyperaccumulation.

Nearly 800 plant species are able to accumulate certain metals or metalloids in their shoots to levels twoto three-orders of magnitude higher than normally found in other plants in the same habitat (Reeves et al., 2018). This is an astounding trait since in plants, as well as in practically all other types of organisms, the normal physiological range of essential minerals such as zinc (Zn) – one of the hyperaccumulated metals – is rather narrow, with thresholds of deficiency and toxicity differing by only about a factor of 10 (Kr€amer, 2010). Perhaps even more amazing, hyperaccumulated elements include cadmium (Cd) and arsenic (As), environmental pollutants with no known biological function and highly toxic to practically all organisms. Thus, metal hyperaccumulating plants possess metal hypertolerance too, meaning they can thrive in the presence of soil metal concentrations that far exceed the capacity of most other plant species to cope with metal/metalloid overload. Molecular mechanisms of metal hyperaccumulation and hypertolerance have attracted considerable interest mainly for three reasons. First, metal hyperaccumulators are attractive models to understand the evolutionary processes underlying adaptation to harsh environments. Second, analysis of an extreme deviation can identify key nodes in the metal homeostasis network, which accomplishes the proper distribution of essential minerals to thousands of metalrequiring proteins and other target sites in plants. Third, the knowledge gained could potentially be applied to engineer biofortified crops enriched in micronutrients and to promote the use of hyperaccumulators for the phytoremediation of metalcontaminated sites or the phytoextraction (phytomining) of valuable metals. By far the most common hyperaccumulated metal is nickel (Ni), probably because Ni is highly abundant in the Earth’s crust and Ni-rich serpentine soils therefore occur frequently. Still, the mechanistic understanding of Ni hyperaccumulation has been lagging behind Zn and Cd hyperaccumulation (Van der Pas & Ingle, 2019). In this issue of New Phytologist, Garc ıa de la Torre et al. (2021; pp. 994–1006) report two important advances. The ability to hyperaccumulate metals or metalloids is found in dozens of families scattered across the phylogenetic tree, a pattern reminiscent of other complex adaptations including C4 photosynthesis and Crassulacean Acid Metabolism (Edwards, 2019). Thus, metal hyperaccumulation most likely evolved many times independently, raising the fascinating biological question as to whether convergent evolution selected similar genetic changes that enable hyperaccumulation (Verbruggen et al., 2013). Garc ıa de la Torre et al. addressed this by performing transcriptome studies on pairs of related hyperaccumulator and nonhyperaccumulator species from a wide range of families. Their data suggest that, indeed, convergent mechanisms exist. Furthermore, for one of the candidate mechanisms apparently shared by phylogenetically distant Ni hyperaccumulators, they obtained direct evidence by genetic intervention, a first in Ni hyperaccumulation research.

Nearly 800 plant species are able to accumulate certain metals or metalloids in their shoots to levels two-to three-orders of magnitude higher than normally found in other plants in the same habitat (Reeves et al., 2018). This is an astounding trait since in plants, as well as in practically all other types of organisms, the normal physiological range of essential minerals such as zinc (Zn)one of the hyperaccumulated metalsis rather narrow, with thresholds of deficiency and toxicity differing by only about a factor of 10 (Kr€ amer, 2010). Perhaps even more amazing, hyperaccumulated elements include cadmium (Cd) and arsenic (As), environmental pollutants with no known biological function and highly toxic to practically all organisms. Thus, metal hyperaccumulating plants possess metal hypertolerance too, meaning they can thrive in the presence of soil metal concentrations that far exceed the capacity of most other plant species to cope with metal/metalloid overload. Molecular mechanisms of metal hyperaccumulation and hypertolerance have attracted considerable interest mainly for three reasons. First, metal hyperaccumulators are attractive models to understand the evolutionary processes underlying adaptation to harsh environments. Second, analysis of an extreme deviation can identify key nodes in the metal homeostasis network, which accomplishes the proper distribution of essential minerals to thousands of metalrequiring proteins and other target sites in plants. Third, the knowledge gained could potentially be applied to engineer biofortified crops enriched in micronutrients and to promote the use of hyperaccumulators for the phytoremediation of metalcontaminated sites or the phytoextraction (phytomining) of valuable metals. By far the most common hyperaccumulated metal is nickel (Ni), probably because Ni is highly abundant in the Earth's crust and Ni-rich serpentine soils therefore occur frequently. Still, the mechanistic understanding of Ni hyperaccumulation has been lagging behind Zn and Cd hyperaccumulation (Van der Pas & Ingle, 2019). In this issue of New Phytologist, Garc ıa de la Torre et al. (2021;pp. 994-1006) report two important advances. The ability to hyperaccumulate metals or metalloids is found in dozens of families scattered across the phylogenetic tree, a pattern reminiscent of other complex adaptations including C 4 photosynthesis and Crassulacean Acid Metabolism (Edwards, 2019). Thus, metal hyperaccumulation most likely evolved many times independently, raising the fascinating biological question as to whether convergent evolution selected similar genetic changes that enable hyperaccumulation (Verbruggen et al., 2013). Garc ıa de la Torre et al. addressed this by performing transcriptome studies on pairs of related hyperaccumulator and nonhyperaccumulator species from a wide range of families. Their data suggest that, indeed, convergent mechanisms exist. Furthermore, for one of the candidate mechanisms apparently shared by phylogenetically distant Ni hyperaccumulators, they obtained direct evidence by genetic intervention, a first in Ni hyperaccumulation research.
'. . . for one of the candidate mechanisms apparently shared by phylogenetically distant Ni hyperaccumulators, they obtained direct evidence by genetic intervention, a first in Ni hyperaccumulation research.'

Cross-species transcriptome analyses
With the exception of two Sedum species, Sedum alfredii and Sedum plumbizincicola, most molecular research on metal hyperaccumulation has, to date, been performed in the two Brassicaceae models Arabidopsis halleri and Noccaea caerulescens. A combination of comparative transcriptome studies, quantitative trail locus (QTL) analyses in progeny of crosses between hyperaccumulating and nonhyperaccumulating species or accessions, and targeted RNA interference experiments established a general scheme of Zn and Cd hyperaccumulation (Verbruggen et al., 2009;Kr€ amer, 2010;Corso & Torre, 2020). A key feature is the altered expression of several metal homeostasis genes encoding metal transporters or enzymes involved in the synthesis of metal ligands. Many of the identified changes, including the elevated expression of ZIP (Zrt1/IRT1-related proteins) transporter genes, HMA (Heavy metal ATPase) genes and NAS (nicotianamine synthase) genes, were found in both model systems. This suggests convergent evolution of hyperaccumulation at least within the Brassicaceae family. Garc ıa de la Torre et al. cast a much wider net. They sampled leaves of Ni hyperaccumulators from five distinct families, growing on serpentine soils in France, Cuba and New Caledonia, the latter two representing biodiversity hotspots where many endemic Ni hyperaccumulating species are found. For comparison, related nonhyperaccumulator species or accessions were collected in the This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. same areas. RNA was sequenced and from the resulting reads, reference transcriptomes were generated and annotated. In order to allow functional comparison across lineages, clusters of orthologous groups (COGs) were inferred that contain orthologous genes differentially expressed between related hyperaccumulator and nonhyperaccumulator species in at least three different families. Altogether 26 such COGs were identified. The most robustly up-regulated COG among them corresponds to ferroportin (FPN)/IREG (iron-regulated) transporters. They are named after the only known cellular iron (Fe) export system in mammals, FPN. In humans, it is a key component of Fe homeostasis, controlling the release of absorbed Fe from enterocytes into the plasma (Coffey & Ganz, 2017). Other known FPN substrates include Zn, cobalt (Co) and Ni. FPN/ IREG transporters had been implicated in Ni hyperaccumulation before as they were identified in intra-specific comparisons to be constitutively up-regulated in Ni hyperaccumulating accessions relative to nonhyperaccumulating accessions of N. caerulescens (Halimaa et al., 2014) and Senecio coronatus (Meier et al., 2018). Garc ıa de la Torre et al. have now found strong evidence for a functional role of FPN/IREG genes across diverse families and thus for convergent evolution.

The plant transformation challenge
Corroboration of a hypothesized gene function is most convincing when phenotypes of wild-type individuals can be compared to those of respective loss-of-function mutants. Whenever phenomena are studied that are not expressed by the model systems of molecular genetics, this calls for genetic interventions to obtain such mutants. A prerequisite for the use of RNA interference, artificial microRNAs or CRISPR-Cas is a successful plant transformation. The lack of reliable, highly efficient transformation protocols has, for a long time, hampered research on major crop plants and hyperaccumulators. The study by Garc ıa de la Torre et al. now makes a significant step forward by developing a method to study gene function via knockdowns in a hyperaccumulator species. Thus far, few successful transformations of metal hyperaccumulating species have been reported (Liu et al., 2017;Ahmadi et al., 2018), none for N. caerulescens. Using Rhizobium rhizogenes root transformation, Garc ıa de la Torre et al. obtained several lines of a Ni hyperaccumulating N. caerulescens accession carrying an amiRNA construct targeting the IREG2 gene. This represents a major achievement, yet such lines are less informative than transformants with a stable knockdown of IREG2 in leaves would be. Silencing of IREG2 in roots caused a reduction of Ni accumulation. This is cautiously interpreted as evidence for IREG2-mediated vacuolar sequestration of Ni (Fig. 1), which is consistent with the demonstrated localization of IREG2 in the tonoplast and the transport activity for Ni ions shown by heterologous expression in yeast. Nickel accumulation in leaves, however, was not affected in the transgenic lines. Directly analysing the significance of elevated IREG2 expression in leaves of diverse Ni hyperaccumulators, for example via targeted mutagenesis by CRISPR-Cas, will depend on further advances in plant transformation.

Convergent mechanisms besides FPN/IREG
The reference transcriptomes assembled by Garc ıa de la Torre et al. represent a valuable resource for future studies. Some of those may be guided by other identified COGs containing genes differentially expressed in Ni hyperaccumulators. Several up-regulated COGs indicate a functional role of specialized metabolites such as flavonoids. An obvious hypothesis is that Ni hyperaccumulators synthesize higher concentrations of low molecular weight metal ligands. They are essential for different aspects of metal hyperaccumulation (Clemens, 2019;Van der Pas & Ingle, 2019). Following uptake into root cells, Ni ions have to be bound in the cytosol by molecules that are small enough to maintain symplastic mobility of Ni so that it can reach cells surrounding the xylem (Fig. 1). Inside the xylem, at least some of the Ni is assumed to be complexed with ligands too. Cytosolic ligands are then again needed in leaf cells. Inside vacuoles, Ni is probably associated with organic acids. Candidate Ni ligands, besides organic acids, include histidine and nicotianamine. However, available evidence suggests that their role is restricted to particular plant families (Van der Pas & Ingle, 2019). Thus, the existence of additional Ni ligands has to be postulated.
Another lead derived from the COG analysis points to specific cell wall modifications in Ni hyperaccumulators. Besides the vacuoles of leaf cells, the leaf apoplast is another storage site for the huge amounts of metal accumulated aboveground (Fig. 1), making alterations of the cation binding capacity of cell wall polymers a likely adaptation. Finally, the COG results highlight a possible driver of genetic variation selection acts upon. Tracing the molecular evolution of metal hyperaccumulation includes the identification of mechanisms underlying altered regulation of metal homeostasis and other genes. Copy number expansions and cis-regulatory mutations are known factors, exemplified by HMA4 and its crucial role in Zn and Cd hyperaccumulation (Hanikenne et al., 2008). While comparable insights are missing for Ni hyperaccumulators, the identification of several COGs corresponding to class I retrotransposons suggests a possible reason for the propensity that some taxa may have to evolve metal hyperaccumulation. Ultimately, the degree of genetic convergence can only be determined by identifying the mutations and genomic changes causal for the higher expression of FPN/IREG genes in diverse Ni hyperaccumulators.