Substantial Decrease in Plasmalogen in the Heart Associated with Tafazzin Deficiency

Abstract

Tafazzin is the mitochondrial enzyme that catalyzes transacylation between a phospholipid and a lysophospholipid in remodeling. Mutations in tafazzin cause Barth syndrome, a potentially life-threatening disease with the major symptom being cardiomyopathy. In the tafazzin-deficient heart, cardiolipin (CL) acyl chains become abnormally heterogeneous unlike those in the normal heart with a single dominant linoleoyl species, tetralinoleoyl CL. In addition, the amount of CL decreases and monolysocardiolipin (MLCL) accumulates. Here we determine using high-resolution ³¹P nuclear magnetic resonance with cryoprobe technology the fundamental phospholipid composition, including the major but oxidation-labile plasmalogens, in the tafazzin-knockdown (TAZ-KD) mouse heart as a model of Barth syndrome. In addition to confirming a lower level of CL (6.4 ± 0.1 → 2.0 ± 0.4 mol % of the total phospholipid) and accumulation of MLCL (not detected → 3.3 ± 0.5 mol %) in the TAZ-KD, we found a substantial reduction in the level of plasmenylcholine (30.8 ± 2.8 → 18.1 ± 3.1 mol %), the most abundant phospholipid in the control wild type. A quantitative Western blot revealed that while the level of peroxisomes, where early steps of plasmalogen synthesis take place, was normal in the TAZ-KD model, expression of Far1 as a rate-determining enzyme in plasmalogen synthesis was dramatically upregulated by 8.3 (±1.6)-fold to accelerate the synthesis in response to the reduced level of plasmalogen. We confirmed lyso-plasmenylcholine or plasmenylcholine is a substrate of purified tafazzin for transacylation with CL or MLCL, respectively. Our results suggest that plasmenylcholine, abundant in linoleoyl species, is important in remodeling CL in the heart. Tafazzin deficiency thus has a major impact on the cardiac plasmenylcholine level and thereby its functions.

Figure 1

Structures of plasmenylcholine and diacyl phosphatidylcholine (PC). Structures are drawn with the most abundant acyl chain species for those lipids in the mammalian heart.

Figure 2

(A) ³¹P NMR spectra of phospholipids in the heart of the wild type (WT, blue trace) and tafazzin-knockdown (TAZ-KD, red trace) mice. A magnified region on the right shows inversion of the spectral asymmetry, indicating a drastic decrease in plasmenylcholine (−0.163 ppm) and a counterbalancing increase in diacyl PC (−0.174 ppm) in due to the TAZ-KD. The signal from the minor component plasmanylcholine (see the main text) overlaps with the signal from diacyl PC (Figure S6). A magnified region on the upper left shows a large decrease in CL (0.745 ppm) along with increases in 2-MLCL (1.137 ppm) and 1-MLCL (1.009 ppm). An increase in PG (0.845 ppm) is also seen. *A part of the CL decrease is masked by contributions from the resonance of a phosphate group in the diacyl half of MLCLs (Figure S5). (B) A spectrum of the standard plasmenylcholine (green trace) overlaid with the WT mouse heart phospholipid spectrum (blue trace).

Figure 3

³¹P NMR identification of lysophospholipids in the mouse heart, using different solubilizing detergent systems of SDS and cholate. (A) ³¹P NMR of lyso PC and lyso-plasmenylcholine (lyso-plsCho) in SDS micelles yields similar chemical shift values. (B) Use of cholate micelles results in distinct signal separation for lyso PC and lyso-plsCho, thus unambiguously identifying the presence of lyso PC and the absence of a detectable amount of lyso-plsCho in the WT and TAZ-KD mouse heart. (C) The absence of a detectable amount of lyso PE in the WT and TAZ-KD mouse heart is evident from the measurements in SDS micelles. Lyso-plasmenylethanolamine (lyso-plsEtn) shows the chemical shift close to that of CL and the diacyl half of MLCLs at 0.745 ppm in SDS micelles. (D) Use of cholate micelles enabled to identify the absence of a detectable amount of lyso-plsEtn in the WT and TAZ-KD mouse heart.

Figure 4

The biosynthesis pathway for plasmalogens (1-O-alk-1′-enyl-2-acyl-GPE (plasmenylethanolamine) and 1-O-alk-1′-enyl-2-acyl-GPC (plasmenylcholine)) in the cell.

Figure 5

Quantitative Western blot of proteins in the mouse heart that are indicative of (i) the amount of peroxisomes (Pex19p, PMP70, catalase), (ii) a degree of oxidative stress (catalase), (iii) a degree of feedback regulation of the plasmalogen level (Far1), and (iv) a degree of plasmalogen-selective lipid degradation (iPLA₂β). (A) The ratio of expression levels (TAZ-KD/WT) in the TAZ-KD and WT mouse heart on cytosolic Pex19p (N=3) and peroxisomal integral membrane protein PMP70 (N=3). Measurements to determine the ratio were conducted on the postnuclear supernatant (PNS), while fractionation by ultracentrifugation (see Materials and Methods) results in detection of Pex19p exclusively in the supernatant fraction (S), and PMP70 exclusively in the membrane pellet fraction (P) that includes peroxisomes. (B) The TAZ-KD/WT ratio on catalase determined for the PNS (N=9). Furthermore, fractionation was performed to determine the TAZ-KD/WT ratio in each of the S (N=3) and P fractions (N=3). The Western blotting image shows quantities of catalase in the PNS, S and P fractions of the WT and TAZ-KD mouse heart homogenate. Distributions of cytosolic (in S) and peroxisomal (in P) catalase in the WT and TAZ-KD samples are illustrated as pie charts. (C) The Western blotting image showing greatly enhanced expression of Far1 in the TAZ-KD mouse heart, where the large decrease in plasmenylcholine was observed by ³¹P NMR. The fractionation experiment resulted in detection of Far1 dominantly in the supernatant. The TAZ-KD/WT ratio on the PNS is shown (N=10). (D) The Western blotting image showing quantities of iPLA₂β. The fractionation experiment resulted in detection of the protein in the supernatant fraction. The TAZ-KD/WT ratio determined from the PNS is shown (N=10).