Lysophospholipid micelles sustain the stability and catalytic activity of diacylglycerol kinase in the absence of lipids

Abstract

There has been a renewal of interest in interactions of membrane proteins with detergents and lipids, sparked both by recent results that illuminate the structural details of these interactions and also by the realization that some experimental membrane protein structures are distorted by detergent-protein interactions. The integral membrane enzyme diacylglycerol kinase (DAGK) has long been thought to require the presence of lipid as an obligate "cofactor" in order to be catalytically viable in micelles. Here, we report that near-optimal catalytic properties are observed for DAGK in micelles composed of lysomyristoylphosphatidylcholine (LMPC), with significant activity also being observed in micelles composed of lysomyristoylphosphatidylglycerol and tetradecylphosphocholine. All three of these detergents were also sustained high stability of the enzyme. NMR measurements revealed significant differences in DAGK-detergent interactions involving LMPC micelles versus micelles composed of dodecylphosphocholine. These results highlight the fact that some integral membrane proteins can maintain native-like properties in lipid-free detergent micelles and also suggest that C(14)-based detergents may be worthy of more widespread use in studies of membrane proteins.

Figure 1

Steady-state kinetic analysis of DAGK in TDPC, LMPG, and TDPC at pH 6.5 and 30 °C. The steady-state kinetic data in LMPC, LMPG, and TDPC were collected by measuring DAGK activity (A) over a range of MgATP concentrations (0–8 mM) while keeping the concentration of DBG at a near-saturating concentration (20 mM) or (B) over a range of DBG concentrations (0–25 mM) while keeping the MgATP concentrations constant at a near saturating concentration (3 mM). Each data point is the average of three measurements, and the resulting curves represent best fits by a modified form of the Michaelis-Menten equation in which a Hill coefficient was applied to the variable substrate concentration.

Figure 2

CD spectra of DAGK in micelles at pH 6.5 and 45 °C. Far-UV (A) and near-UV (B) CD spectra of DAGK were collected to assess the secondary structure and aromatic side chain order of DAGK, respectively, in LMPC (black), LMPG (blue), and TDPC (red). For comparison, spectra of DAGK in DPC are also included (yellow), as well as for DM (near-UV only, green). Analysis of this far-UV data led to the following estimates of secondary structure. LMPC: 79 ± 9% α-helix, 0% β-sheet, 21 ± 4% random coil; LMPG: 83 ± 8% α-helix, 0% β-sheet, 17 ± 4% random coil; TDPC: 79 ± 8% α-helix, 0% β-sheet, 20 ± 4% random coil; DPC: 69 ± 7% α-helix, 4 ± 1%, β-sheet, 27 ± 4% random coil.

Figure 3

Irreversible loss of DAGK enzymatic activity with time during incubation at elevated temperatures. The stabilities of DAGK in LMPC (black), LMPG (blue), TDPC (red), and DM (green) were examined. The samples were prepared at pH 6.5 and incubated at either 45 °C (A) or 70 °C (B). Time point aliquots were removed and subjected to the standard DM/CL/DHG DAGK assay at 30 °C. The resulting data points were fit using Solver in Microsoft Excel to obtain the half-life for retention of DAGK.

Figure 4

Assessment of helix stability in DAGK by monitoring the temperature dependence of the far-UV CD spectrum of DAGK in LMPC, LMPG, and TDPC micelles at pH 6.5.

Figure 5

Use of near-UV CD to assess aromatic side chain order in DAGK as a function of temperature at pH 6.5. The ellipticity values at 268, 274, and 286 nm correspond to absorbance maxima for phenylalanine, tyrosine, and tryptophan side chains, respectively,

Figure 6

800 MHz ¹⁵N-TROSY spectra of DAGK in LMPC, LMPG, DPC, and TDPC micelles at 45 °C. The samples in TDPC and LMPG contained 10 mM Bis-Tris, 2 mM magnesium chloride, 0.5 mM EDTA, and 10% (v/v) D₂O, pH 6.5. The samples in DPC and LMPC contained 250 mM imidazole, 0.5 mM EDTA, and 10% (v/v) D₂O, pH 6.5.

Figure 7

2-D strips plots showing NOEs to the tryptophan indole NH protons from 3-D ¹H-¹⁵N-NOESY-TROSY spectra collected for U-¹⁵N-DAGK in DPC (A) and LMPC (B) micelles at 800 MHz and 45°C. Associated with each set of strips is a 1-D NMR spectrum for pure DPC and LMPC showing resonance assignments. In the strip plots for the DPC case (A) the indole NH proton diagonal peaks appear upfield of the rest of the spectrum rather than in the expected 9.5–10.5 PPM range because the peaks are “aliased” as a result of being outside of the spectral window of the TPPI/States-based NMR experiment used to acquire this data (39). In the LMPC case (B) the peaks appear at the expected chemical shifts because they fell within the observation sweep width. In the LMPC case, side chain-perdeuterated DAGK was used, which is why NOEs between the NH indole protons to their two nearest neighbors on the indole rings are not observed at 7.2–7.8 PPM, unlike the DPC case where DAGK was not deuterated and these NOEs are quite pronounced.