Minimally altering a critical kinase for low-phytate maize

Nutritional security is of vital importance for combating malnutrition and catering to increasing energy demands. Phytic acid is considered an anti-nutrient, which sequesters important metal ions, limiting their bioavailability. The lpa mutants of maize contain reduced phytate, thus increase its nutritive value. But low phytate is accompanied by negative pleiotropic effects. This article discusses the importance of lpa2 gene amongst available options, for precise DNA editing to simultaneously improve nutrition and avoid pleiotropic effects.

are usually involved in interactions with many other proteins. Null expression or silencing of a protein results in disruption of protein-protein interactions and signaling or downstream metabolic networks associated with such protein-protein interactions. Disruption of protein-protein interaction networks comprises a potential reason for the diverse negative pleiotropic effects linked with natural or engineered mutants. Raboy has enlisted four potential target areas, which include (a) inhibition of the synthesis of myo-inositol and inositol-3-phosphates, (b) inhibition of synthesis of inositol-6-phosphate (phytic acid), (c) inhibition of transport and storage of phytic acid in the cell, and (d) expression of phytase encoding transgenes 10 . Given the public aversion to genetically modified foods in certain countries, the last approach is limited in practical value, hence the first three may be considered for a broader applicability. Genome editing gives an opportunity to modify specific gene(s) to achieve desirable effect 11 . In this regard, it is essential to prioritize target gene(s), for achieving precise engineering of physiological processes with minimal negative pleiotropic effects associated with changes in a particular gene sequence(s). One possible strategy to avoid off-target effects is to mutate a protein through gene editing to result in a variant that has reduced or null enzymatic activity. The edited protein would thus block the concerned metabolic reaction. However, the protein would still be fully expressing to fulfill its function of interacting with other proteins. Targeting the upstream reactions of a metabolic pathway would result in a snow-ball effect, influencing all the other downstream processes. Hence, inhibition of myo-inositol and inositol-3-phosphates would disrupt not only the inositol metabolism, but all other allied processes associated with it. Inositol phosphates are involved in a range of cellular functions like membrane transport, cell division, cytodifferentiation, regulatory role in signal transduction (lipid signalling) etc 12 . On the other hand, blocking the transport of phytic acid in cellular compartments has been described as a viable strategy. Shi et al. silenced embryo-specific expression of an ABC transporter protein that transports phytic acid from cytoplasm to vacuole 13 . The authors report that while the null lpa1 mutant has attenuated seed germination, the lpa1-1 point mutant displays normal germination percentage. However, the seed weight of point mutant is lesser than normal maize. Landoni et al. have demonstrated changes in physical properties of lpa1 maize, including modifications in density, starch properties, fiber content etc., which has implications for the practical utility of lpa1 maize 14 . The available evidence shows that phytic acid can be packaged into cell organelles via multiple transporter proteins of the ABC Multiple Drug Resistance associated Protein (MRP) type. In Arabidopsis, the phytic acid is known to be packaged in at least three compartments: protein storage vacuole of embryo, endoplasmic reticulum and vacuolar compartments of chalazal endosperm 15 . The storage in last two compartments is transient. Due to multiple transport compartments, silencing of one transporter protein in one compartment may not be a viable strategy. Further, the accumulation of phytic acid in cytoplasm has the potential to lead to toxicity, as phytic acid at defined concentrations has been demonstrated to possess cytotoxicity in human cell lines 16 . Either or combination of the above phenomenon may be responsible for the decreased seed weight of lpa1 mutants.
Being the most downstream enzyme in the metabolic pathway of phytic acid formation, Inositol Phosphate Kinase 1 or IPK1 is a potential gene target for achieving reduction in phytate content. IPK1 catalyzes the phosphorylation of inositol phosphates to higher phosphorylated forms like IP 6 (Phytic acid). It is also involved in protein-protein interactions with other proteins. IPK1 can thus be mutated, such that the phosphorylation of different inositol phosphates does not take place, but the protein remains intact for protein-protein interactions to take place. Figure 1 shows the current understanding of phtyic acid formation and transport in cell. Studies have shown that apart from inositol phosphates, inositol pyrophosphates also play an important role in cellular metabolism 17,18 . Phytic acid is further converted to higher phosphorylated forms by Inositol hexakisphosphate kinase (IP6K) enzyme. This leads to the formation of inositol pyrophosphates IP 7 (PP-IP 5 ) and IP 8 (2(PP-IP 4 ) 19 . The available literature suggests that the function of IP 7 and IP 8 can be carried out by inositol pyrophosphates PP-IP 4 and 2(PP)IP 3 19 . Saiardi et al. have shown that a null mutant of IPK1 (ipk1Δ) in yeast results in accumulation of pyrophosphates PP-IP 4 and 2(PP)IP 3 besides IP 5 , yet resembles wild-type cells in morphology 19 . Abnormal vesicular morphology in case of null mutations of other inositol phosphate kinases is attributed to the loss of inositol pyrophosphates, which does not happen in case of ipk1Δ, thus explaining its similar morphology to wild type cells. Hence, loss of phytic acid would not impair the function mediated by IP 7 and IP 8 , since PP-IP 4 and 2(PP) IP 3 impart functional redundancy by complementation [ Fig. 1(f,g)]. In view of the above and multiple transport routes of phytic acid along with protein-protein interactions of IPK1, it appears to be the critical kinase which can be targeted without introducing multiple negative plieotropic effects, as observed with other mutants. The IPK1 enzyme can be minimally altered by disrupting its catalytically active site, so that the protein is fully expressing, but unable to convert IP 5 to phytic acid. The desired minimal alteration in IPK1 would only affect phytic acid formation, leaving the upstream and downstream process, as well as the protein-protein interactions intact.
In order to evaluate the prospects of minimally altering Zea mays IPK1, a computational model of the protein was made using PSI-BLAST (Position-Specific Iterative Basic Local Alignment Search Tool) based structure prediction 21 . The structure was refined by side-chain repacking 22 . The refined structure contains 13 α-helices and 15 β-sheets, with 92.4% residues in Rama-favoured region and no poor rotamers. The Class (C), Architecture (A), Topology (T), superfamily (H) analysis 23,24 of the modeled Zea mays IPK1 structure showed it to contain the structure typically found in inositol phosphate kinases. The substrate inositol pentkisphosphate (IP 5 ) was docked to the IPK1 model using a rigid docking algorithm 26,27 , which was further refined 28,29 . Both IP 5 and cofactor Adenosine triphosphate (ATP) bind in a cleft formed by four β-sheets from residues 195-202, 205-212, 278-286, 292-299 and two α-helices 260-273 & 310-326. Analysis of the docked structure using PDBsum web server 30,31 showed Alanine 5, Histidine 205, Threonine 207 and Cysteine 208 to be closely interacting with IP 5 (Fig. 2A). The proteins that interact with IPK1 via protein-protein interactions include acid phosphatase, inositol-pentakisphosphate 2-kinase, inositol polyphosphate multikinase, succinate-CoA ligase, inositol 3-kinase, ABC MRP4 transporter and a metal ion binding protein (Fig. 2B). Mutation of the IPK1 protein at key amino acids that result in destabilization of the protein in its active site or hinder interactions with substrate or cofactor will result in a protein, functional for protein-protein interactions but non-functional for phytic acid formation.
www.nature.com/scientificreports www.nature.com/scientificreports/ In the present case, Histidine 205 is implicated to be important for protein stability. Various mutations at His205 position have the potential to destabilize the protein, thereby hindering its function of phosphorylation (Fig. 2C). Similarly, alanine mutants of other interacting residues have the potential to inhibit phosphorylation by IPK1.
Cowieson et al. have postulated a term 'phytate-free nutrition' to emphasize the importance of reducing the content of phytic acid in feed 20 . The authors mention the fact that prominence is being given to formulation of phytate-free diets rather than accommodating phytase enzyme in the diet, as majority of phytic acid is digested  8 , which would not be formed in the event of absence of IP 6 . IPK1 itself, rather than a single transporter protein or an upstream enzyme, appears to be the most promising target for low-phytate maize. in ruminants, but recalcitrant phytic acid does not get digested. If phytic acid content is reduced in the first place, mineral bioavailability would be enhanced. At the same time, efforts must be directed to minimize any possibilities for off-target effects of phytate reduction on plants. The above study is a step towards prioritization of target genes for dephytinization of maize to enhance its nutritional value.

Methods
Generation of protein model. The DNA sequence of lpa2 encoding for inositol phosphate kinase 1 was taken from National Center of Biotechnology Information. It was modelled through Bioserf (available on PSI-PRED webserver) 21 . The obtained model was refined through GalaxyRefine program 22 . The modelled structure was analysed for class, architecture, topology, superfamily through CATH/Gene3Dv4.2 program 23,24 . Docking of substrate and cofactor with protein. The substrate inositol pentakisphosphate (IP 5 ) and cofactor Adenosine triphosphate (ATP) were downloaded from ZINC database 25 . The substrate and cofactor were docked on to the modelled protein using Patchdock algorithm 26,27 and refined using Firedock webserver 28,29 . The interacting residues were identified using PDBsum 30,31 . protein interactions and prediction of stability at key residues. The protein-protein interactions of IPK1 were obtained using STRING database 32 . The stability of protein with mutation at key residues was determined using Site Directed Mutator program 33 .