**5. Conclusions**

Strategies for controlling the accumulation of specific metabolites are commonly based on switching o ff structural or regulatory genes of the biosynthetic pathway. However, in order to avoid or reduce downstream e ffects on derived pathways, a di fferent approach is to interfere with compound transport to the site of accumulation (organ, cell type, subcellular compartment). In this review we reported data showing how PA reduction can be achieved with mutations in di fferent types of transporters that control PA transport to the vacuole (MRP), or by modifying Pi availability for PA synthesis through mutations in transporters involved in Pi loading and organ/intracellular distribution (SULTR) or by Pi acquisition and mobilization during seed development (PHT1;4) (Figure 6).

**Figure 6.** Rice transporters identified to modulate PA homeostasis. Modified from [99].

These types of *lpa* mutants are potentially advantageous over other *lpa* mutants in structural biosynthetic genes to achieve a triple goal: (i) Increased bioavailability of mineral cations, which are no longer chelated by PA, (ii) less PA is excreted into the environment with manures, and hence there is reduced impact on water eutrophication, (iii) increased P use efficiency as the seeds are loaded with less P which remains in the straw and may potentially contribute to reducing the demand for P fertilizers, hence increasing crop sustainability [99].

From an agronomic point of view, mutants in *MRP* and *PHT1;4* genes show compromised yields, seed setting, seed germination, and/or seedling growth, etc., so that they are not very attractive for breeders [21,22,37]. However, good field performance has been demonstrated for *MRP* mutants of common bean, which is due to the presence of duplicated *MRP* gene(s) able to complement the mutated seed-specific copy [32]. So far, *sultr3;3* and *sultr3;4*/*spdt* mutants, besides Arabidopsis, have been isolated and described only in rice and barley. Interestingly, these mutants do not display negative pleiotropic effects as observed for the other transporters described above, hence they potentially represent a valuable tool to simultaneously achieve seed biofortification and a more sustainable crop while reducing the environmental impact of the crop cultivation. It would be very interesting to verify the role and function of *sultr3;3* and *sultr3;4*/*spdt* genes in legume crops as well, since contrasting data on function and subcellular localization have been reported for Arabidopsis mutants and a possible species-specific and/or development-specific behavior has been proposed [100].

Finally, the finding that both phosphate and putative sulfate transporters produce similar *lpa* phenotypes suggests the "existence of a multilevel coordination in the regulation of the two ions in which currently unidentified key elements are actively cross-talking between the two signaling pathways" [101,102]. The availability of such mutants from different crops may help towards understanding this cross-talk and identifying new players.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2223-7747/9/1/69/s1. Figure S1. PA-MRP proteins alignment. See Table 1 for the correspondence with genes accession numbers. The Clustal W alignment (cost matrix Blosum, gap open cost 10, gap extend cost 0.1) of Geneious 11.0.2 software was used. Figure S2. Phylogenetic tree of characterized crop PA-MRP proteins, listed in Table 1. Phylogenies were constructed with the Geneious Tree Builder tool, using the Jukes-Cantor distance model, Neighbor-Joining tree build method. Figure S3. Phylogenetic tree of SULTR3;3 and SULTR3;4 proteins, listed in Table 2. Phylogenies were constructed as described in Figure S2. Figure S4. SULTR3;3 proteins alignment. See Tables 1 and 2 for the correspondence with genes accession numbers. The method used is described in Figure S1 legend. Figure S5. SULTR3;4 proteins alignment. See Tables 1 and 2 for the correspondence with genes accession numbers. The method used is described in Figure S1 legend.

**Funding:** This research received no external funding.

**Acknowledgments:** We thank Fabio Nocito, Università degli Studi di Milano (Milan, Italy), for fruitful discussions.

**Conflicts of Interest:** The authors declare no conflict of interest.
