Expression, Purification, and Characterization of Human Diacylglycerol Kinase ζ.

Diacylglycerol kinase ζ (DGKζ) phosphorylates diacylglycerol (DG) to generate phosphatidic acid. The dysfunction of DGKζ has been linked to several diseases, such as cardiac hypertrophy, ischemia, and seizures. Moreover, much attention has been paid to DGKζ, together with DGKα, as a potential target for cancer immunotherapy. However, DGKζ has never been purified and, thus, neither its enzymatic properties nor its structure has yet been reported, hindering our understanding of the catalytic mechanism of DGKζ and the development of a reasonable structure-based drug design. In the present study, we generated a full-length DGKζ using a baculovirus-insect cell expression system for enzymological and structural studies. Full-length DGKζ remained soluble and was purified to near homogeneity as a monomer with yields suitable for protein crystallization (0.63 mg/1 L culture). Enzymatic characterization showed that the purified DGKζ is in a fully functional state. The K m values for adenosine triphosphate (ATP) and DG were 0.05 mM and 1.5 mol %, respectively, and the EC50 for the activator phosphatidylserine was 8.6 mol %, indicating that its affinity for ATP is moderately higher than those of DGKα and DGKε, and its affinities for DG and phosphatidylserine are comparable to those of DGKα/DGKε. We further confirmed that the purified enzyme could be concentrated without any significant aggregation. Circular dichroism revealed that DGKζ is comprised of 25% α-helices and 18% β-strands. This is the first successful purification and characterization of the enzymatic and conformational properties of DGKζ. The purification of DGKζ allows detailed analyses of this important enzyme and will advance our understanding of DGKζ-related diseases and therapies.

Introduction

Diacylglycerol kinase (DGK) phosphorylates diacylglycerol (DG) to produce phosphatidic acid (PA).[1−4] DG is well known to be an activator of conventional and novel protein kinase C (PKC), Ras guanyl nucleotide-releasing protein, chimaerins, and Unc-13.[5−7] Moreover, PA has been reported to regulate the activities of many physiologically important enzymes such as Ras GTPase-activating protein, phosphatidylinositol-4-phosphate kinase, Raf-1 (C-Raf) kinase, atypical PKC, and mammalian target of Rapamycin.[8,9] Thus, DGK appears to participate in a wide variety of physiological and pathological events by controlling the balance between two lipid second messengers, DG and PA.

To date, 10 mammalian DGK isozymes (α, β, γ, δ, ε, ζ, η, θ, ι, and κ) have been identified. These DGK isozymes are divided into five groups (type I (α, β, and γ), II (δ, η, and κ), III (ε), IV (ζ and ι), and V (θ)) according to their structural features.[1−4] Type IV DGK isozymes (ζ and ι) commonly contain two C1 domains at the N-terminus, a MARCKS-like domain, four ankyrin repeats, and a PDZ domain at the C-terminus in addition to a catalytic domain (CD).[10]

DGKζ is a versatile enzyme. Topham et al. revealed that reducing nuclear DG levels through the action of DGKζ attenuated cell proliferation.[11] Liu et al. reported that DGKζ constitutes a downstream component of the leptin signaling pathway in the hypothalamus.[12] In NIE-115 neuroblastoma cells, DGKζ promotes neurite outgrowth.[13] We previously showed that retinoic acid-induced neurite outgrowth was attenuated by a deficiency of DGKζ, which produced 16:0/16:0-PA species in Neuro-2a neuroblastoma cells.[14] In that study, we observed the role of DGKζ in the morphological changes at the initial/early stages of neuronal differentiation. Rac1 is essential for retinoic acid-induced neurite extension of Neuro-2a cells.[15] DGKζ-derived PA activates p21-activated protein kinase 1, which induces the release of Rac1 from Rho guanine nucleotide dissociation inhibitors.[16] DGKζ-mediated synaptic conversion of DG to PA is involved in the maintenance of dendritic spines.[17] In addition, DGKζ attenuates the hypertrophic signaling cascade and resultant cardiac hypertrophy in response to Gq protein-coupled receptor agonists.[18] DGKζ in hippocampal neurons is involved in global ischemia[19] and kainate-induced seizures.[20] Moreover, we now recognize the importance of DGKζ as a physiological negative regulator of T-cell receptor signaling and T-cell activation. DGKζ-deficient T cells were hyperresponsive to T-cell receptor stimulation both ex vivo and in vivo.[21,22] Therefore, it is thought that selective inhibitors of DGKζ enhance T-cell receptor signaling and, consequently, can be promising anticancer drugs through cancer immunity.[23−26]

However, despite their physiological and biomedical importance, no structural and enzymatic properties of DGKζ have been determined, hindering our understanding of the catalytic mechanism of DGKζ and the development of a reasonable structure-based medicine design. In the present study, a full-length human DGKζ expressed in the soluble form in the baculovirus–insect cell expression system was successfully purified to near homogeneity as a monomer by a series of column chromatographies. The purified protein was fully active. Interestingly, the enzymatic parameters of DGKζ are different from those of other DGK isozymes.

Materials and Methods

Expression of DGKζ in Insect Cells Using a Baculovirus Expression Vector System

Expression of human full-length DGKζ (DGKζ-FL) and its catalytic domain (CD) alone (DGKζ-CD) was performed as previously described.[27] The constructs of DGKζ-FL[10] (aa 1–928 (UniProt accession ID: DGKZ Q13574-2)) or DGKζ-CD (aa 276–629) with an N-terminal His × 6 tag were amplified by polymerase chain reaction and inserted into the pOET3 vector (Oxford Expression Technologies, Oxford, U.K.) at SalI/NotI sites.

Purification of DGKζ Expressed in Insect Cells

Purification of DGKζ-FL and DGKζ-CD using nickel-nitrilotriacetic acid (Ni-NTA) agarose (Qiagen, Venlo, The Netherlands) and a Superdex 200 column 16/60 (GE Healthcare (Chicago, IL)) was performed as previously described.[27] Collected fractions were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) followed by Coomassie blue staining and Western blot analysis using the anti-DGKζ antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Protein quantification was performed by the Bradford assay or using the absorbance of a 1% solution of the protein in a 1 cm pathlength cell at 280 nm (E1% = 11.4).

DGKζ Activity Assay in Vitro

The activity of DGKζ-FL was analyzed using the octyl-β-d-glucoside mixed micelle assay followed by the ADP-Glo kinase assay (Promega, Madison, WI) as previously described.[28,29] To determine the kinetic constants, the activity assay was performed under a series of concentrations of adenosine triphosphate (ATP) (20–200 μm) and DG (0–5.4 mol %). For each reaction, 14.4 ng of purified DGKζ was added, and the assays were performed in triplicate for each ATP and DG concentration. The Km value was obtained by fitting the kinase activity of DGKζ with the Michaelis–Menten equation using Prism 5 (GraphPad Software, La Jolla, CA).

Circular Dichroism Spectroscopy

Circular dichroism spectra were recorded under ambient conditions between 190 and 250 nm on a Jasco J-805 spectrometer (Jasco Corporation) using a cell with a path length of 0.2 mm, a scan speed of 60 nm/min, and a bandwidth of 1 nm. DGKζ for the spectrometry was prepared at 0.33 mg/mL (3.1 μm) in 20 mM Tris–HCl buffer, pH 7.4, 200 mM NaCl, 3 mM MgCl2, 0.5 mM dithiothreitol, and 5% glycerol. Five spectra were averaged and the spectrum obtained for the buffer was subtracted. Spectral data were analyzed using the program K2X[30] suited in the DICHROWEB platform (http://dichroweb.cryst.bbk.ac.uk/html/home.shtml).[31]

Statistical Analysis

Statistical comparisons were performed using a two-tailed t-test or a one-way analysis of variance followed by a Tukey’s test.

Results

Expression and Purification of Full-Length DGKζ

It is difficult to express soluble and active DGK in bacterial expression systems. Therefore, we chose the baculovirus–Sf9 cell system to express DGKζ. Human DGKζ was reported to exist in two alternatively spliced forms.[10,32] We selected human DGKζ1,[10] which is ubiquitously expressed.[10,32] The catalytic domain is an essential target for inhibitor development. Thus, we first attempted to express human DGKζ-CD (aa 276–629) in Sf9 cells. The cDNA construct of DGKζ-CD (Figure S1A) with an N-terminal His × 6 tag was expressed in Sf9 insect cells and purified using Ni-NTA affinity chromatography (Figure S1B). However, DGKζ-CD was not highly soluble (64% soluble) (Figure S1B) and was recovered in flow-through fractions in Ni-NTA affinity chromatography. Therefore, the amount of Ni-NTA-purified DGKζ-CD was low (85 μg/L of Sf9 cell culture) (Figure S1C).

We previously reported that the full-length DGKα expressed in Sf9 cells was soluble.[27] We therefore tried to produce human DGKζ-FL (aa 1–928) (Figure A) using the baculovirus expression system. Western blot analysis showed that 96% of DGKζ-FL was soluble after cell lysis (Figure B). DGKζ was purified using Ni-affinity chromatography from the cell lysis supernatant and eluted in the fractions containing 50 and 100 mM imidazole (Figure C). Next, to further purify DGKζ-FL, size-exclusion chromatography on a Superdex 200 column was performed after concentration. DGKζ-FL was eluted as a single peak at the molecular mass of 125 kDa based on a calibration curve obtained with molecular mass standard proteins (Figure D). The calculated molecular mass of His × 6-tagged DGKζ-FL is 106 kDa (DGKζ (∼104 kDa) + His × 6 tag (∼2 kDa)).[10,33] Therefore, it is likely that DGKζ-FL exists as a monomer in solution. Overall, DGKζ-FL was purified to near homogeneity (Figure E), and the yield was ∼0.63 mg per 1 L of Sf9 cell culture.

Figure 1

Expression of DGKζ-FL in baculovirus-infected insect cells and purification. (A) Schematic domain architecture of DGKζ-FL. (B) Immunoblot analysis of the solubility of DGKζ-FL expressed in Sf9 cells. Cell lysates were separated into a 25 000g supernatant (Sup) and pellets (Ppt) and subjected to SDS-PAGE (7.5%) followed by immunoblot analysis using anti-DGKζ antibody. (C) SDS-PAGE (7.5%) analysis of fractions from Ni2+-affinity purification; separated proteins were stained with Coomassie blue staining. (D) Elution profile of DGKζ-FL from size exclusion chromatography; fraction numbers (elution volume) used for the following SDS-PAGE analysis are labeled. Inset shows the calibration of the gel-filtration column using protein standards of known molecular weight (thyroglobulin (670 kDa), γ-globulin (158 kDa), ovalbumin (44 kDa), and myoglobin (17 kDa)). The partition coefficient (Kav) was calculated from the formula Kav = (VE – V0)/(VT – V0), where VE is the retention volume of each sample, VT is the total column volume (120 mL), and V0 is the void volume of the column (44 mL). Kav was plotted against the molecular weight of the proteins, and linear regression analysis was conducted. (E) SDS-PAGE (7.5%) analysis of DGKζ-FL purified using size-exclusion chromatography.

Enzymatic Properties of Purified DGKζ

To examine whether the purified DGKζ-FL is catalytically active, we performed the octyl-β-D-glucoside mixed micellar assay followed by a luminescence-based assay that measures the adenosine diphosphate (ADP) produced in a kinase reaction.[28,29] DGKζ purified via size-exclusion chromatography was found to show strong kinase activity, and its specific activity was 3.6 nmol PA/min/μg. This value is comparable to those obtained from DGKα (type I isozyme) purified from porcine thymus (2.4 nmol PA/min/μg),[34] DGKα from the baculovirus–insect cell expression system (2.0 nmol PA/min/μg),[27] and DGKε (type III isozyme) purified from the baculovirus–insect cell expression system (3.8 nmol PA/min/μg).[35] Moreover, no significant changes in the activity were observed after storage of the purified DGKζ at −80 °C for at least 1 month.

We also determined the kinetic parameters of DGKζ for ATP and DG to assess its catalytic properties. An ATP-dependent increase in the kinase activity was detected (Figure A), and the Km value was determined to be 0.050 ± 0.005 mM (Vmax: 4.05 ± 0.17 nmol PA/min/μg) (Table ). The Km values for ATP obtained with DGKα purified from porcine thymus, expressed in COS-7 cells, and purified from the baculovirus–insect cell expression system were 0.1,[34] 0.24,[27] and 0.1–0.25 mM,[28,29] respectively. DGKε purified from bovine testis showed a Km value for ATP of 0.09–0.10 mM.[36] We recently reported that the Km value for ATP of DGKη (type II isozyme) expressed in COS-7 cells was 0.052 mM.[37] Taken together, the Km value of DGKζ (type IV isozyme) for ATP is lower than those of DGKα (type I isozyme) and DGKε (type III isozyme) and comparable to that of DGKη (type II isozyme).

Figure 2

Enzyme kinetics of purified DGKζ with ATP and DG. (A) ATP and (B) DG activity dependencies of the purified DGKζ as measured by the luminescence-based assay. DGKζ activity was plotted as a function of (A) ATP concentration (mM) or (B) DG concentration (mol %). Measured luminescence values were converted into the amount of ADP produced (nmol) based on the ATP-to-ADP conversion curve measured separately with a known concentration of ATP (50 μm to 1 mM). Data are represented as the mean ± SD for triplicate measurements.

Table 1

DGKζ Enzymatic Kinetic Parameters with ATP and DGa

substrate	Km	Vmax
ATP	0.050 ± 0.005 mM	4.05 ± 0.17 nmol/(min μg)
DG	1.49 ± 0.12 mol %	2.12 ± 0.06 nmol/(min μg)

Mixed micellar assay was performed to test the enzymatic activity over a series of substrate concentrations. The kinetic parameters were obtained as shown in Figure . Data shown are mean ± standard deviation (SD) for triplicate measurements.

The activity also increased in a DG concentration-dependent manner (Figure B), and the Km value was 1.49 ± 0.12 mol % (Vmax: 2.12 ± 0.06 nmol PA/min/μg) (Table ). This value is comparable to those from our previous studies with DGKα purified from porcine thymus (3.3 mol %),[34] from the baculovirus–insect cell expression system (1.1 mol %), and expressed in COS-7 cells (1.9–3.4 mol %).[28,29] The Km value for DG of DGKε purified from bovine testis was 2.4 mM.[36] On the other hand, the value of DGKη expressed in COS-7 cells was 0.14 mol %.[37] Thus, the Km value of DGKζ for DG is comparable to those of both DGKα and DGKε and higher than that of DGKη.

DGK is generally activated by an anionic phospholipid phosphatidylserine (PS).[38] PS was reported to increase the activity of DGKζ in insect Sf21 cell lysates.[39] Thus, we next examined the effect of PS on the activity of purified DGKζ (Figure ). As shown in Figure , we confirmed that DGKζ activity increased in a PS-dependent manner. The ED50 value for PS was 8.6 mol % (Figure ). The PS ED50 values for DGKα purified from porcine thymus and expressed in COS-7 cells were 16[27] and 9 mol %,[37] respectively. The ED50 value of DGKη for PS was 8.5 mol %.[37] Therefore, The ED50 value of DGKζ for PS is comparable to those of both DGKα and DGKη.

Figure 3

Effect of PS on purified DGKζ activity: purified DGKζ was incubated with various concentrations of PS as indicated for 5 min. Concentrations of the nonvaried assay components were 0.2 mM ATP and 5.4 mol % DG. Data are represented as the means ± SD for triplicate measurements.

DGKζ was reported to show no DG species selectivity.[10,32] We analyzed the activities of purified DGKζ in the presence of several DG species, 12:0/12:0-, 16:0/16:0-, 16:0/18:1-, 18:1/18:1-, 18:0/18:0-, 18:0/20:4-, and 18:0/22:6-DG, as the substrate. When the activity with 12:0/12:0-DG was set to 100%, activities with 16:0/16:0-, 16:0/18:1-, 18:1/18:1-, 18:0/18:0-, 18:0/20:4-, and 18:0/22:6-DG were all 78–108% (Figure ). Thus, we confirmed that this isozyme does not have selectivity for a particular DG species (Figure ).

Figure 4

Effects of various DG species on the activity of purified DGKζ. The purified DGKζ was incubated with 12:0/12:0-DG, 16:0/16:0-DG, 16:0/18:1-DG, 18:1/18:1-DG, 18:0/18:0-DG, 18:0/20:4-DG, or 18:0/22:6-DG for 30 min. The concentrations of the assay components were 2.7 mol % DG, 0.2 mM ATP, and 13 mol % PS. Data are represented as the means ± SD for triplicate measurements. Statistical significance was determined by Student’s t test. Data shown are representative of three separate experiments.

Structural Characterization of the Purified DGKζ

We also revealed that the DGKζ solution could be concentrated using a centrifugal filter without any significant loss of the protein. Using the concentrated DGKζ (0.33 mg/mL (3.1 μm)), we attempted to understand the secondary structure composition using circular dichroism spectroscopy from 190 to 250 nm. The circular dichroism spectrum of DGKζ and following analysis using K2X[30] suited in the DICHROWEB platform (http://dichroweb.cryst.bbk.ac.uk/html/home.shtml)[31] indicates that DGKζ is well-folded and contains α-helical (25%) and β-strand (18%) structures (Figure A and Table ), further demonstrating that the expression of a full-length DGKζ in the baculovirus-infected insect cells is suitable for producing a natively folded and active form of DGKζ. The secondary structure composition of DGKζ was also assessed using the secondary structure prediction server self-optimized prediction method with alignment (SOPMA, https://omictools.com/sopma-tool), which predicted the α-helix and β-sheet contents to be 30 and 18%, respectively (Figure B and Table ). The values obtained from the experimental calculations and the theoretical predictions were mostly in agreement with each other.

Figure 5

Secondary structure of purified DGKζ: (A) Circular dichroism spectrum of DGKζ measured under ambient conditions between 190 and 250 nm on a Jasco J-805 spectrometer; DGKζ was prepared at 0.33 mg/mL (3.1 μm) in 20 mM Tris–HCl buffer, pH 7.4, 200 mM NaCl, 3 mM MgCl2, 0.5 mM dithiothreitol, and 5% glycerol. The analysis of the circular dichroism spectrum using the program K2X[30] suited in the DICHROWEB platform[31] showed the presence of both α-helical (25%) and β-strand (18%) structures. (B) Theoretical secondary structure analysis of DGKζ using SOPMA software.

Table 2

Secondary Structures of DGKζ

secondary structure	DICHROWEB (%)	SOPMA (%)
α-helix	25	30
β-strand	18	18

Discussion

DGKζ is a lipid kinase that regulates a wide variety of cellular processes. Particularly, DGKζ has recently attracted much attention as a novel therapeutic target for cancer immunotherapy. However, no enzymatic or structural information on DGKζ is available, hindering our understanding of the catalytic mechanism of DGKζ and the strategy design of a reasonable structure-based drug development. In the present study, we purified a full-length form of DGKζ using the baculovirus–insect cell expression system. Moreover, we revealed its enzymatic and structural properties. In contrast to DGKζ-FL, the yield of DGKζ-CD was very low (Figure S1).

Using size-exclusion chromatography, DGKζ-FL eluted in a relatively sharp peak and remained as a monomer (Figure D,E). Such a production of DGKζ-FL in a soluble and monomeric form using the baculovirus–insect cell expression system is useful for the preparation of a DGKζ sample suitable for protein crystallization screening. DGKζ-FL was eluted at the molecular mass of 125 kDa. This value is larger than the calculated molecular mass (106 kDa), reflecting a multidomain architecture of DGKζ where each domain is connected by a flexible linker (Figure A).

Spectral data analysis using the program K2X[30] suited in the DICHROWEB platform[31] suggested that the secondary structures of DGKζ are comprised of 25% α-helices and 18% β-strands, which are mostly in agreement with theoretical predictions (30% α-helices and 18% β-strands) (Figure and Table ). When compared with other mammalian DGK isozymes (DGKα and DGKε), the α-helical content of DGKζ is higher than that of DGKα (19%[27]) and lower than that of DGKε (29%[35]). The β-strand content of DGKζ is lower than those of both DGKα (27%[27]) and DGKε (22%[35]). DGKζ and DGKα have long regulatory regions in addition to the catalytic and C1 domains, whereas DGKε contains only C1 domains and the CD. Theoretical predictions (SOPMA) indicate that the C1 domains and CDs in DGKζ (C1 domains: 36%; CD: 36% (Figure )), DGKα (C1 domains: 22%; CD: 37%), and DGKε (C1 domains: 35%; CD: 32%) are relatively α-helix-rich. Therefore, the α-helix content of DGKε is likely to be higher than those of DGKζ and DGKα. On the other hand, the ankyrin repeat is known to be α-helix-rich.[40] DGKζ possesses four ankyrin repeats that contain 54% α-helices (Figure B). The ankyrin repeats would provide a higher α-helix content in DGKζ compared with DGKα.

Enzymatic characterization of DGKζ, which was carried out for the first time, reveals that the Km values to ATP (0.050 mM) and DG (1.49 mol %) (Figure and Table ) are moderately lower than or comparable to those obtained using DGKα purified from porcine thymus,[34] expressed in COS-7 cells,[28,29] expressed in the baculovirus–insect cell expression system,[27] and DGKε purified from bovine testis. Unlike DGKα and DGKε, DGKη has a high affinity for ATP (Km: 0.05 mM) and DG (Km: 0.14 mol %).[37] In addition to different properties in DGK activities, the 1-MGK and 2-MGK activities of DGKζ are also different from those of other isozymes. Although type I (DGKα, β, and γ), II (DGKδ, η, and κ), and III (DGKε) isozymes have substantial 2-monoacylglycerol kinase (MGK) activity and the type V (DGKθ) isozyme possesses 1-MGK activity, DGKζ and ι (type IV) have only negligible 1-MGK and 2-MGK activities.[41] Therefore, with regard to enzymatic properties, there are many kinds of DGK isozymes, suggesting that each DGK isozyme functions in different environments, which include distinct ATP and DG concentrations.

Deletion of DGKα or DGKζ in mouse models results in T cells bearing a hyperresponsive phenotype through the attenuation of Ras guanyl nucleotide-releasing protein 1 and enhanced T-cell activity against malignancy.[21,22,42,43] Therefore, both DGKα and DGKζ function as immunosuppressors in T cells. The functional similarities of DGKα and DGKζ suggest that they coordinate with and complement each other. Different enzymatic properties between DGKα and DGKζ would contribute to the cooperative and complementary actions of these isozymes in T cells.

In summary, the present study demonstrates that the production of DGKζ-FL using the baculovirus–insect cell expression system is a very useful approach to obtain DGKζ preparations for future functional and structural studies. First, DGKζ was purified to near homogeneity, and the purified DGKζ was soluble and monomeric, and was concentrated without any significant loss. These properties are essential for protein crystallization. Second, the obtained yield of DGKζ, 0.63 mg per 1 L cell culture, is sufficient to start crystal screening. Third, the purified DGKζ is catalytically sufficient. Fourth, the Km values for ATP and DG of DGKζ are different from those obtained for other DGK isozymes. Therefore, there are many enzymatically distinct DGK isozymes, suggesting that each DGK isozyme acts under different circumstances in cells. For example, DGKζ and DGKα, which act as physiological negative regulators of T-cell activation, are thought to function cooperatively and complementarily to each other. The successful purification of DGKζ in the present study permits detailed analyses of this important enzyme and will advance our understanding of DGKζ-related diseases and therapies.