Catalytic function of PLA2G6 is impaired by mutations associated with infantile neuroaxonal dystrophy but not dystonia-parkinsonism

Abstract

Background: Mutations in the PLA2G6 gene have been identified in autosomal recessive neurodegenerative diseases classified as infantile neuroaxonal dystrophy (INAD), neurodegeneration with brain iron accumulation (NBIA), and dystonia-parkinsonism. These clinical syndromes display two significantly different disease phenotypes. NBIA and INAD are very similar, involving widespread neurodegeneration that begins within the first 1-2 years of life. In contrast, patients with dystonia-parkinsonism present with a parkinsonian movement disorder beginning at 15 to 30 years of age. The PLA2G6 gene encodes the PLA2G6 enzyme, also known as group VIA calcium-independent phospholipase A(2), which has previously been shown to hydrolyze the sn-2 acyl chain of phospholipids, generating free fatty acids and lysophospholipids.

Methodology/principal findings: We produced purified recombinant wildtype (WT) and mutant human PLA2G6 proteins and examined their catalytic function using in vitro assays with radiolabeled lipid substrates. We find that human PLA2G6 enzyme hydrolyzes both phospholipids and lysophospholipids, releasing free fatty acids. Mutations associated with different disease phenotypes have different effects on catalytic activity. Mutations associated with INAD/NBIA cause loss of enzyme activity, with mutant proteins exhibiting less than 20% of the specific activity of WT protein in both lysophospholipase and phospholipase assays. In contrast, mutations associated with dystonia-parkinsonism do not impair catalytic activity, and two mutations produce a significant increase in specific activity for phospholipid but not lysophospholipid substrates.

Conclusions/significance: These results indicate that different alterations in PLA2G6 function produce the different disease phenotypes of NBIA/INAD and dystonia-parkinsonism. INAD/NBIA is caused by loss of the ability of PLA2G6 to catalyze fatty acid release from phospholipids, which predicts accumulation of PLA2G6 phospholipid substrates and provides a mechanistic explanation for the accumulation of membranes in neuroaxonal spheroids previously observed in histopathological studies of INAD/NBIA. In contrast, dystonia-parkinsonism mutations do not appear to directly impair catalytic function, but may modify substrate preferences or regulatory mechanisms for PLA2G6.

Figure 1. Recombinant human PLA2G6 protein catalyzes the hydrolysis of fatty acids from PC and LPC.

(A) Purified recombinant protein preparations from cells transfected with an empty expression vector, WT PLA2G6 or S519A PLA2G6 were added to in vitro catalytic assays. Free fatty acids (FFA) released from ¹⁴C-labeled 1-palmitoyl lysophosphatidylcholine were separated on TLC and detected using a phosphorimager. Incubation of substrate with WT PLA2G6 enzyme produces robust release of fatty acids compared to control preparations from vector-transfected cells. Catalytic activity is abolished by mutation of S519 in the lipase catalytic site. (B) Quantitation of catalytic activity for the ¹⁴C-labeled phospholipid substrate 1-palmitoyl-2-oleyl phosphatidylcholine. (C) Quantitation of activity for ¹⁴C-labeled LPC lysophospholipid substrate. Fatty acid release in catalytic assays was quantitated from the phosphorimager screen in photostimulated luminescence (PSL) units, and the graphs indicate the rate of hydrolysis for each substrate, measured by the increase in PSL units per min of incubation time in each assay. Since the PLA2G6 protein concentrations, radiochemical specific activities and substrate concentrations were the same in both assays, the results indicate that the catalytic rates for the two substrates are similar.

Figure 2. Disease-associated PLA2G6 mutations examined in this study and their relationship to functional domains of the PLA2G6 protein.

Illustrated functional domains in the PLA2G6 protein (encoded by transcript variant 1) include the ankyrin repeat regions (numbered regions between amino acids 150–382) and the GXSXG lipase catalytic site (S519). Shaded regions indicate a nucleotide binding domain centered at amino acid 485, and a calmodulin binding region (amino acids 747–759). The locations of amino acid changes resulting from mutations associated with INAD/NBIA are indicated above the diagram and mutations associated with dystonia-parkinsonism are indicated below the diagram.

Figure 3. PLA2G6 mutations that cause NBIA/INAD significantly disrupt the ability of PLA2G6 to hydrolyze PC and LPC.

Phospholipase and lysophospholipase catalytic activities of mutant PLA2G6 proteins were compared to wild type (WT) PLA2G6 in assays with PC and LPC substrates, respectively. Initial experiments examined the group of INAD-associated mutations A341T, G517C, G638R, and Y790X. Phospholipase (A) and lysophospholipase (B) activities are shown as percent of WT activity measured when equal concentrations of protein were added to assays measuring release of fatty acids from radiolabeled substrate. Bars indicate mean plus standard deviation (n = 3) from representative experiments. The specific activities of all mutations were significantly different from WT (p<0.05, unpaired t-test). Similar results were obtained in at least two experiments with independent transfection and purification of recombinant proteins.

Figure 4. The effects of additional mutations illustrate the significance of genotype distinctions between INAD/NBIA and dystonia-parkinsonism.

The INAD/NBIA-associated R741W mutation (A, B) was investigated because mutation of the same residue to a glutamine (R741Q) has been identified in dystonia-parkinsonism. The V691del mutation was examined given its association with an R632W mutation in an INAD patient with compound heterozygous mutations (C, D), in contrast to homozygous R632W mutations associated with dystonia-parkinsonism. Phospholipase (A,C) and lysophospholipase (B,D) activities are shown as percent of WT activity measured when equal concentrations of protein were added to assays measuring release of fatty acids from radiolabeled substrate. Bars indicate mean plus standard deviation (n = 3) from representative experiments. Asterisks indicate mean activities that were significantly different from WT (p<0.05, unpaired t-test). Similar results were obtained in at least two experiments with independent transfection and purification of recombinant proteins.

Figure 5. PLA2G6 mutations associated with dystonia-parkinsonism do not impair phospholipase and lysophospholipase activity.

Purified recombinant protein was produced for WT PLA2G6 protein and for PLA2G6 proteins containing missense mutations associated with dystonia-parkinsonism, Equal amounts of WT or mutant proteins were added to enzymatic assays utilizing (A) PC or (B) LPC as substrate. Relative rates of fatty acid release are shown for WT and each mutant protein (mean percent WT + standard deviation, n = 3 independently prepared protein preparations for each of WT and 3 mutants). Asterisks indicate mean activities that were significantly different from WT (p<0.05, unpaired t-test). Although an initial experiment with n = 1 recombinant protein preparations indicated decreased activity of all three mutant proteins relative to WT, results similar to those shown in panels A and B were observed in all subsequent experiments, which included at least 3 additional independent protein preparations.