Lauren E. Pope

The Arabidopsis (Arabidopsis thaliana) genome contains nine β-amylase (BAM) genes, some of which play important roles in starch hydrolysis. However, little is known about BAM2, a plastid-localized enzyme reported to have extremely low catalytic activity. Using conservation of intron positions, we determined that the nine Arabidopsis BAM genes fall into two distinct subfamilies. A similar pattern was found in each major lineage of land plants, suggesting that these subfamilies diverged prior to the origin of land plants. Moreover, phylogenetic analysis indicated that BAM2 is the ancestral member of one of these subfamilies. This finding, along with the conservation of amino acids in the active site of BAM2, suggested that it might be catalytically active. We then identified KCl as necessary for BAM2 activity. Unlike BAM1, BAM3, and BAM5, three Arabidopsis BAMs that all exhibited hyperbolic kinetics, BAM2 exhibited sigmoidal kinetics with a Hill coefficient of over 3. Using multi-angle light scattering, we determined that BAM2 was a tetramer, whereas BAM5 was a monomer. Conserved residues from a diverse set of BAM2 orthologs were mapped onto a homology model of the protein, revealing a large, conserved surface away from the active site that we hypothesize is a secondary carbohydrate-binding site. Introduction of bulky methionine for glycine at two points on this surface reduced catalytic activity significantly without disrupting the tetrameric structure. Expression analysis indicated that BAM2 is more closely coexpressed with other starch degradation enzymes than any other BAM, suggesting that BAM2 may play an important role in starch degradation in plants.

Ferroptosis is a regulated nonapoptotic cell death process characterized by iron-dependent lipid peroxidation. Peroxidation of polyunsaturated fatty acid-containing phospholipids (PUFA-PL) is necessary for the execution of ferroptosis. Glutathione peroxidase 4 (GPX4) suppresses ferroptosis by reducing lipid hydroperoxides to lipid alcohols. GPX4 may be a useful target for drug development, highlighting the need to identify factors that govern GPX4 inhibitor sensitivity. In this study, we found that reduced GPX4 expression was sufficient to induce ferroptosis in multiple adherent (2D) cancer cell cultures. However, lower GPX4 protein levels did not consistently affect tumor xenograft growth in mice. Culturing cells as spheroids (3D) was sufficient to reduce sensitivity to pharmacologic GPX4 inhibition. Mechanistically, growth in 3D versus 2D conditions upregulated expression of the monounsaturated fatty acid (MUFA) biosynthetic gene stearoyl-CoA desaturase, altering the ratio of MUFA-PLs to PUFA-PLs in a direction favoring ferroptosis resistance. Similar shifts in MUFA-PL:PUFA-PL ratios were observed in xenograft tumors. Thus, lipidome remodeling in 3D growth conditions and in vivo may limit GPX4 inhibitor efficacy. SIGNIFICANCE: Changes in lipid composition can affect induction of ferroptosis, explaining why sensitivity of cancer cells in tissue culture does not reliably translate to more complex models and suggesting potential ferroptosis sensitization strategies.

The β-amylase family in Arabidopsis thaliana has nine members, four of which are both plastid-localized and, based on active-site sequence conservation, potentially capable of hydrolyzing starch to maltose. We recently reported that one of these enzymes, BAM2, is catalytically active in the presence of physiological levels of KCl, exhibits sigmoidal kinetics with a Hill coefficient of over 3, is tetrameric, has a putative secondary binding site (SBS) for starch, and is highly co-expressed with other starch metabolizing enzymes. Here we generated a tetrameric homology model of Arabidopsis BAM2 that is a dimer of dimers in which the putative SBSs of two subunits form a deep groove between the subunits. To validate this model and identify key residues, we generated a series of mutations and characterized the purified proteins. 1) Three point mutations in the putative subunit interfaces disrupted tetramerization; two that interfered with the formation of the starch-binding groove were largely inactive, whereas a third mutation prevented pairs of dimers from forming and was active. 2) The model revealed that a 30-residue N-terminal acidic region, not found in other BAMs, appears to form part of the putative starch-binding groove. A mutant lacking this acidic region was active and did not require KCl for activity. 3) A conserved tryptophan residue in the SBS is necessary for activation and may form -bonds with sugars in starch. 4) Sequence alignments revealed a conserved serine residue next to one of the catalytic glutamic acid residues, that is a conserved glycine in all other active BAMs. The serine side chain points away from the active site and towards the putative starch-binding groove. Mutating the serine in BAM2 to a glycine resulted in an enzyme with a Vmax similar to that of the wild type enzyme but with a 7.5-fold lower KM for soluble starch. Interestingly, the mutant no longer exhibited sigmoidal kinetics, suggesting that allosteric communication between the putative SBS and the active site was disrupted. These results confirm the unusual structure and function of this widespread enzyme, and suggest that our understanding of starch degradation in plants is incomplete.