Introduction

Phosphatidylcholine (PC) is the main phospholipid in eukaryotic cell membranes, and serves as a precursor to a number of phospholipids and second messengers, such as phosphatidylserine, sphingomyelin, phosphatidic acid (PA), lyso-PC, and diacylglycerol (DAG)1,2,3,4,5. Inhibition of PC synthesis leads to arrested cell growth and apoptosis6,7,8. In eukaryotic cells, the production of PC is mainly achieved through de novo synthesis by the Kennedy pathway, which is a multi-stepped process and the last of which is the production of PC catalyzed by two highly homologous enzymes, cholinephosphotransferase-1 (CHPT1) and choline/ethanolamine phosphotransferase-1 (CEPT1)9,10. In CHPT1 and CEPT1, choline phosphate from CDP-choline is transferred to the free hydroxyl of a DAG to produce a PC (Supplementary Fig. 1, ref. 11,12). Another closely related enzyme, ethanolamine phosphotransferase-1 (EPT1), catalyzes the transfer of ethanolamine phosphate from CDP-ethanolamine to DAG and produces phosphatidylethanolamine (PE) (Supplementary Fig. 1, ref. 13). All three enzymes are found in the membranes of endoplasmic reticulum or Golgi complexes and critical for cell functions14.

Although CHPT1, CEPT1, and EPT1 are phosphotransferases, they are classified into the superfamily of CDP-alcohol phosphatidyltransferases (CDP-AP, InterPro IPR000462), which are ubiquitous in both prokaryotic and eukaryotic organisms. CDP-APs share a signature sequence motif, D1xxD2G1xxAR…G2xxxD3xxxD4, and catalyze the displacement of cytidine monophosphate (CMP) from a CDP-linked alcohol with another alcohol. In prokaryotic cells, the CDP-linked alcohol is frequently a CDP-DAG and hence the name phosphatidyltransferase. Structures of several bacterial CDP-APs have been reported15,16,17, which define a common structural fold composed of six TMs and have the CDP-AP signature motif located on TM2 and TM3 that coordinate either one or two Mg2+. The structures of eukaryotic CDP-APs remain elusive. Since most eukaryotic CDP-APs have 10 TMs and utilize different substrates from their prokaryotic relatives, we pursued their structure and examined their functions to gain a better understanding of the mechanisms.

Results

Function of CHPT1

Full-length Xenopus laevis CHPT1 (xlCHPT1, UniProt accession number Q4KLV1), which is ~68% identical and ~83% similar to both human CHPT1 and CEPT1, was expressed and purified (Methods, Supplementary Figs. 2a and 3). xlCHPT1 was purified as a homodimer and we examined its enzymatic activity and substrate preference. We first examined CDP-choline and 1,2-sn-diacylglycerol as substrates and measured the enzymatic activity by following the production of CMP (Methods and Supplementary Fig. 2b). The initial rate of CMP production at different concentrations of CDP-choline was determined and plotted in Fig. 1a, and the data are well-fit with a Michaelis-Menten equation with a KM of 18.0 ± 1.9 µM and a Vmax of 72.8 ± 1.7 nmol/min/mg. These values are similar to human CEPT112,18. Next, we examined CDP-ethanolamine and 1,2-sn- diacylglycerol as substrates, and obtained a KM of 603.0 ± 61.6 µM and a Vmax of 55.6 ± 1.6 nmol/min/mg (Fig. 1b). Thus, xlCHPT1 has a clear preference for CDP-choline. We also examined the effect of divalent cations, and found that xlCHPT1 is active in the presence of either Mg2+ or Mn2+, but not in Ca2+ or Zn2+ (Fig. 1c). These functional properties of xlCHPT1, i.e., preference to CDP-choline and Mg2+ over CDP-ethanolamine and Ca2+, respectively, are consistent with previous reports of human CEPT112.

Fig. 1: Function of purified xlCHPT1.
figure 1

a–b Initial rate of reaction in different concentrations of CDP-choline (a) or CDP-ethanolamine (b). c Relative activity of xlCHPT1 in the presence of different cations. Each symbol or bar is the average of three independent measurements, and error bars are standard errors of the mean (s.e.m.). The solid lines in a, b are fit of the data point to a Michaelis–Menten equation.

Full size image

Overall structure of xlCHPT1

We determined two structures of xlCHPT1 by single-particle cryo-electron microscopy (cryo-EM). The structure of xlCHPT1 in complex with CDP-choline and Mg2+ has an overall resolution of ~3.7 Å, and that of xlCHPT1 in complex with CDP and Mg2+, ~3.2 Å resolution (Fig. 2a, b, Supplementary Figs. 4–5). Both density maps are of sufficient quality to allow de novo model building of residues 20–383 (Supplementary Fig. 6a). The first 19 and the last 18 residues are not resolved, although they are predicted to form α-helices in a structural model generated by AlphaFold (Supplementary Fig. 7 and ref. 19). Since the main chain atoms of the two structures are almost identical, we will focus on the structure of xlCHPT1 in complex with CDP and Mg2+ in all the figures except in panels that show CDP-choline.

Fig. 2: Overall structure of xlCHPT1.
figure 2

a Density map of dimeric xlCHPT1. Density of the two monomers is colored in cyan and pink, and the density of lipids is colored in yellow. b Model of dimeric xlCHPT1. Two monomers are shown as cartoons in cyan and pink, and substrates and lipids are shown as yellow sticks. Two Mg2+ are shown as green spheres. c Structure of xlCHPT1 monomer in two orientations. d Membrane topology plot of an xlCHPT1 monomer.

Full size image

The structure of xlCHPT1 is a homodimer, and each xlCHPT1 protomer has 10 transmembrane helices. Based on the “positive-inside” rule20, both the N- and C-terminus are located on the cytosolic side (Supplementary Fig. 6b). The first six TMs carve out a cone-shaped enclosure in the membrane with a prominent slit between TM5 and TM6. The wide end of the cone opens to the cytosolic side and the opening is partially covered by a well-defined two-stranded β-sheet, one strand from the N-terminus (β1) and another from the loop between TM4 and 5 (β2, Fig. 2c, d). The N-terminus preceding TM1, residues 20–62, is well-structured with three secondary structural elements, a short α-helix (α1), followed by a β-strand that forms part of the two-stranded β-sheet, and a long amphipathic helix (α2) that wraps around the perimeter of the cone. TMs7-10, which are present only in eukaryotic CDP-APs, form a structured bundle and are related to TMs3-6 by a pseudo twofold rotational symmetry (Supplementary Fig. 8). This internal repeat is not recognizable from the amino acid sequence, and we will discuss its significance in comparison to bacterial CDP-APs. Also located close to the wide end of the cone are the CDP-AP signature motif, the two Mg2+, and one CDP-choline (Figs. 2c and 3a).

Fig. 3: Mg2+ and CDP-choline-binding sites.
figure 3

a Overview of CDP-choline and Mg2+-binding sites in two orientations. xlCHPT1 is shown as cartoon, CDP-choline as a stick, and Mg2+ as green spheres. b–c Density maps of CDP or CDP-choline with Mg2+. Density map is shown as transparent blue surface, Mg2+ as spheres, and CDP or CDP-choline as sticks. d–f Coordination of Mg2+, CDP, or CDP-choline in xlCHPT1. Binding site residues and CDP/CDP-choline are both shown as sticks but with a different color scheme. g Relative activity of wide type and mutants of xlCHPT1. Each symbol is an average of three independent measurements. Error bars are standard errors of the mean (s.e.m.). h The catalytic site, which is marked as a star, and the proposed catalytic residues, E129 and H133, are shown as sticks.

Full size image

Mg2+ and CDP-choline-binding sites

Density for two Mg2+ are prominent in both cryo-EM maps, and the Mg2+ are coordinated by side chain carboxylates of Asp111, Asp114 from TM2 and Asp132, Asp136 from TM3 (Fig. 3b–d and Supplementary Fig. 6). The side chain carboxylates of Asp111 and Asp132 interact with both Mg2+ simultaneously, forming two bidentate bridges. We define the Mg2+ that interacts with Asp114 as Mg1 and the other one as Mg2 (Fig. 3d). The four aspartate residues are part of the conserved CDP-AP motif, and we validated their functional impact: mutating any of the aspartate to alanine reduces the enzymatic activity to <20% of that of the wild type xlCHPT1 (Fig. 3g and Supplementary Fig. 9). Densities for CDP or CDP-choline are clearly recognizable, with the cytidine ring coordinated by the side chain of Asn64 and by the backbone nitrogen of Ser124 (Fig. 3b, c, e, f). The diphosphate group in CDP is coordinated by Mg1, the side chain guanidinium group of Arg119, and the side chain hydroxyl group of Tyr34. The quaternary ammonium of CDP-choline is stabilized by the aromatic rings of Trp50 and Phe186 and by the side chain carboxylate of Glu43 (Fig. 3f). The importance of residues at the CDP-choline binding site is demonstrated by functional analysis: point mutations to these residues reduce enzymatic activity by 50–80% (Fig. 3g and Supplementary Fig. 9). Residues involved in the CDP-choline binding site are conserved in human homologs (Supplementary Fig. 3 and 10a–c).

Catalytic site and DAG entryway

Located in the vicinity of the bound Mg2+ and CDP-choline, Glu129 and His133 on TM3 are highly conserved in eukaryotic CDP-AP enzymes but are absent in most of the bacterial CDP-APs (Fig. 3h, Supplementary Fig. 3 and 10d). The two residues do not have direct interactions with Mg2+ or CDP-choline but His133 is ~7 Å away from the β-phosphate of CDP. A conserved histidine paired with glutamate or aspartate is commonly found in the active site of enzymes that catalyze nucleophilic attack by a hydroxyl group, in which the histidine facilitates extraction of the proton from the hydroxyl group and the glutamate or aspartate stabilizes the protonated histidine21,22,23,24. We propose that Glu129 and His133 in xlCHPT1 are part of the catalytic center so that the free hydroxyl group of a DAG substrate is within hydrogen bond distance of His133. Mutations of His133 or Glu129 to alanine almost completely abolish the enzymatic activity of xlCHPT1 (Fig. 3g). To further explore this, we manually placed a 1, 2-sn-dioleoylglycerol into the enclosure with its hydroxyl group between His133 and the phosphate, and we noticed that the glycerol backbone of DAG could be coordinated by residues Asp136 and Gln175, with the two acyl chains extending through the slit between TM5 and 6 and thus partially buried in the hydrophobic core of the membrane (Supplementary Fig. 11). In both density maps, there are tubular shaped non-protein densities resembling two acyl chains at the entrance of the slit, suggesting that the slit is wide enough to accommodate two acyl chains (Supplementary Fig. 6c). In addition, TM6 does not pack tightly against the rest of the protein and thus could afford movement to allow unimpeded entry of DAG into the proposed catalytic site (Fig. 4a, b).

Fig. 4: Dimerization interface.
figure 4

a–b Dimeric xlCHPT1 structure is shown as cartoon in two orientations, and lipids are shown as yellow sticks. The green and orange boxes mark the region with close-up views shown in c & d. c Residues at the dimer interface. Two monomers are shown in cyan and pink. d Lipids trapped at the dimer interface. Density of lipids is shown as a transparent gray surface, and partially resolved lipid molecules are shown as sticks.

Full size image

Dimerization interface

xlCHPT1 forms a homodimer, and the dimer interface is mediated mainly by residues from TM7 and TM9 of each protomer (Fig. 4a, b). The aromatic rings of Phe336 and Tyr340 on TM9 form stacking interactions, and the hydroxyl group of Tyr340 could form a hydrogen bond with the side chain of Gln339 (Fig. 4c). In addition, Lys284 from TM7 could form a hydrogen bond with the side chain carbonyl of Asn342 from TM9. The dimer interface spans only ~10 Å at the luminal side of the membrane, leaving a large gap between the two monomers that are filled with lipids (Fig. 4d). TM9 is a short helix with three and half helical turns, and it is preceded by an extended amphipathic strand with its hydrophobic side packed against TM7 and TM8 and the hydrophilic side exposed to solvent. Because of this construction, the lipid bilayer at the dimer interface is significantly thinner with a hydrophobic core of ~25 Å (Fig. 4a and Supplementary Fig. 6b).

Discussion

Here we report the structures of a eukaryotic CDP-AP, which reveal structural features that are not present in prokaryotic homologs. Interestingly, these features are predicted accurately by the AlphaFold (Supplementary Fig. 7a–d). The predicted structure deviates from the experimental structure with a root mean squared distance (r.m.s.d.) of 0.96 Å for Cα atoms. The deviations are mainly from the intracellular side of TM2 and TM3, incurred likely due to the presence of Mg2+ in the experimental structure, and from the amphipathic helix, which may have intrinsic motions in the plain of the membrane (Supplementary Fig. 7e).

Several prokaryotic CDP-APs structures were reported previously, and all but one form homodimers15,16,17,25,26,27,28. When an xlCHPT1 monomer, which has 10 TMs, is aligned to a prokaryotic homolog, phosphatidyl inositol phosphate synthase from Renibacterium salmoninarum (RsPIPS, pdb code 5D92), which has 6 TMs, TM1-3 of xlCHPT1 are aligned with that of RsPIPS with a root mean squared distance (r.m.s.d.) of 1.9 Å, while TM4-6 and the amphipathic helix preceding TM1 have larger deviations, 3.1 Å, but follow a similar arrangement. Similar differences in alignments are observed when xlCHPT1 is aligned to other bacterial CDP-APs (Supplementary Fig. 12). In both eukaryotic and prokaryotic CDP-AP structures, the CDP-AP signature motif, D1xxD2G1xxAR…G2xxxD3xxxD4, is located to TM2 and TM3. The spatial arrangement of these conserved residues is preserved between bacterial CDP-APs and xlCHPT1, and seems optimized for the coordination of two Mg2+ and the CDP group. The structure of xlCHPT1 reveals that His133 and Glu129 may participate in the catalysis, and we propose that in xlCHPT1 and other eukaryotic CHPT1/CEPT1, His133 and Glu129 should be included in the signature motif, D1xxD2G1xxAR…G2ExxD3HxxD4, which facilitate extraction of a proton from DAG and thus nucleophilic attack on CDP-choline (Fig. 5c, Supplementary Fig. 3).

Fig. 5: Substrate entry and reaction mechanism.
figure 5

a–b Proposed entry pathways for both substrates. CDP-choline and DAG are shown as stick and spheres, respectively. xlCHPT1 is shown as cartoon. Proposed active site residues H133 and E129 are shown as spheres. c Proposed catalytic mechanism of CHPT1. E129 and H133 activate the 3-hydroxyl on DAG to enhance nucleophilic attack on the phosphate group of CDP-choline.

Full size image

We speculate that the relatively smaller variations of TM1-3 between prokaryotic and eukaryotic CDP-APs are due to their common functional role in supporting the coordination of Mg2+ and CDP group, while the larger deviations in the alignment of TM4-6 and the amphipathic helix and the additional two-stranded β sheet reflect adaptation of eukaryotic enzymes to different substrates. For example, RsPIPS and SaPgsA do not have the two-stranded β sheet, and the amphipathic helix wraps loosely around TMs 2 and 5 to allow a larger substrate, CDP-DAG, to enter the binding site (ref. 17,27), while the inositol phosphate substrate enters from the large cytosolic opening and is accommodated by additional positively charged residues at the cytosolic side of the enzyme (Supplementary Fig. 13b–d). In xlCHPT1, the two-stranded β-sheet partially covers the large cytosolic opening of the enzyme and seems to leave an entrance just enough for a CDP-choline or CDP-ethanolamine. In addition, the amphipathic helix wraps tightly around TM2 and TM5, and the two helices are more closely packed. However, there is a large slit between TM5 and TM6 that could allow lateral entrance of DAG from the cell membrane (Fig. 5a, b, Supplementary Fig. 13c).

We noticed that when an xlCHPT1 protomer is aligned to a dimeric prokaryotic CDP-AP, TM7-10 of xlCHPT1 align with TM3-6 of the neighboring protomer (Supplementary Fig. 8b). While this observation reinforces the notion that a core of six TMs is the minimum functional unit of a CDP-AP, it also provides an explanation for the internal pseudo-two-fold symmetry observed between TM3-6 and TM7-10 in xlCHPT1. It is likely that eukaryotic CDP-APs have evolved from prokaryotic ancestors by a gene duplication or fusion event followed by a loss of the first two TMs in the second unit. A search of InterPro database for CDP-APs domain architectures shows that a small fraction of bacterial CDP-APs are indeed composed of two tandem CDP-AP units (represented by InterPro B3T1S3).

In summary, structural and functional analysis of xlCHPT1 enhances our understanding of substrate recognition and catalysis in eukaryotic CDP-APs. Our study also establishes a framework for further investigations that will reveal the binding site of DAG, the mechanisms of catalysis, and regulation of enzymatic activity by associated proteins and lipids.

Methods

Cloning, expression, and purification of Xenopus laevis CHPT1

The xlCHPT1 gene (NCBI accession number NP_001089575.1) was codon-optimized and cloned into the pFastBac-Dual expression vector for production of baculovirus29. The P3 virus was used to infect Hi Five (Trichoplusia ni) cells at a density of 3 × 106 cells/ml, then the cells were harvested after 48–60 h. Cell membranes were prepared by a hypotonic/hypertonic wash protocol as previously described30. Briefly, cells were lysed in a hypotonic buffer containing 10 mM 4-(2-hydroxyethyl)−1-piperazineethanesulfonic acid (HEPES) pH 7.5, 10 mM NaCl, and 2 mM β-mercaptoethanol (BME), 1 mM phenylmethylsulfonyl fluoride (PMSF), and 25 μg/ml DNase I. After ultracentrifugation at 55,000 × g for 10 min, the pelleted cell membranes were resuspended in a hypertonic buffer containing 25 mM HEPES pH 7.5, 1 M NaCl, 2 mM BME, 1 mM PMSF, and 25 μg/ml DNase I, and were centrifuged again at 55,000 × g for 20 min. The pelleted cell membranes were resuspended in 20 mM HEPES pH 7.5, 150 mM NaCl, 20% glycerol, and flash frozen in liquid nitrogen for further use.

Purified membranes were thawed and extracted in 20 mM HEPES pH 7.5, 150 mM NaCl, 2 mM BME, 4 mM MgCl2, 1 pill of inhibitor cocktail tablet (Roche), and then solubilized with 1.5% (w/v) lauryl maltose neopentyl glycol (LMNG, Anatrace) at 4 °C for 2 h. After solubilization, cell debris was removed by centrifugation in 55,000 × g, 45 min, 4 °C, xlCHPT1 was purified from the supernatant with a cobalt-based affinity resin (Talon, Clontech). The C-terminal His6-tag was cleaved with tobacco etch virus protease at room temperature for 30 min. The protein was then concentrated to around 5 mg/ml (Amicon 100 kDa cut-off, Millipore), and loaded onto a size-exclusion column (SRT-3C SEC-300, Sepax Technologies) equilibrated with 20 mM HEPES, 150 mM NaCl, 2 mM BME, 4 mM MgCl2 and 0.1% (w/v) LMNG, 0.01 (w/v) cholesteryl hemisuccinate (CHS, sigma). For the sample used in cryo-EM, the size-exclusion column was equilibrated with 20 mM HEPES, 150 mM NaCl, 2 mM BME, 4 mM MgCl2, and 0.02% (w/v) GDN. Ligands were added after the size-exclusion chromatography. For the xlCHPT1-Mg2+-CDP-choline sample, 10 mM CDP-choline was added. For the xlCHPT1-Mg2+-CDP sample, 10 mM CDP and 0.5 mM DAG were added.

xlCHPT1 mutants were generated using the QuikChange method (Stratagene) with primers (Supplementary Table 2) and the entire cDNA was sequenced to verify the mutation. Mutants were expressed and purified following the same protocol as wild type.

Cloning, expression, and purification of yeast pyrimidine 5′-nucleotidase SDT1

The yeast SDT1 gene (NCBI accession number NM_001181089.1) was codon-optimized and cloned into the pMCSG28 vector containing a C-terminal His-tag for expression. The expression protocol is similar to the previously described31. Transformed BL21 E.coli cells were grown in Luria-Bertani (LB) media until reaching an optical density of 0.6–0.8 and then induced with 0.5 mM IPTG. Cells were then grown at 20 °C overnight, spun down at 4000 × g for 15 min, and resuspended in a solution of 20 mM HEPES pH 7.5, 150 mM NaCl, and 10% glycerol. To purify SDT1, cell pellet in 20 mM HEPES pH 7.5, 150 mM NaCl, and 10% glycerol with cocktail inhibitors was sonicated on ice and centrifugated in 55,000 rpm for 30 min at 4 °C. SDT1 with his-tag was purified from the supernatant with a cobalt-based affinity resin (Talon, Clontech). The eluted protein was then concentrated to around 5 mg/ml (Amicon 50 kDa cutoff, Millipore), and loaded onto a size-exclusion column (SRT-3C SEC-300, Sepax Technologies) equilibrated with 20 mM HEPES, 150 mM NaCl, 2 mM BME, 4 mM MgCl2. The purified SDT1 protein was concentrated to ~1 mg/mL, then stored at −80 °C after aliquoted and fast frozen in liquid nitrogen.

Cryo-EM sample preparation and data collection

Cryo grids were prepared on the Thermo Fisher Vitrobot Mark IV. Quantifoil R1.2/1.3 Cu grids were glow-discharged using the Pelco Easyglow. Concentrated xlCHPT1 (3.5 μL) was applied to glow-discharged grids. After blotting with filter paper (Ted Pella) for 3.5–4.5 s, the grids were plunged into liquid ethane cooled with liquid nitrogen. For cryo-EM data collection, movie stacks were collected using EPU (Thermo Fisher Scientific) on a Titan Krios at 300 kV with a Quantum energy filter (Gatan), at a nominal magnification of ×81,000 and with defocus values of −2.0 to −0.8 μm. A K3 Summit direct electron detector (Gatan) was paired with the microscope. Each stack was collected in the super-resolution mode with an exposing time of 0.175 s per frame for a total of 50 frames. The dose was about 50 e- per Å2 for each stack. The stacks were motion-corrected with Relion 3 and binned (2 × 2) so that the pixel size was 1.07 Å32. Dose weighting was performed during motion correction, and the defocus values were estimated with Gctf33.

Cryo-EM data processing

For xlCHPT1-Mg2+-CDP data set, a total of 5,734,696 particles were automatically picked in RELION 3.1 with template picking from 8100 images34, and imported into cryoSPARC35. After three rounds of two-dimensional (2D) classifications, 21 classes (containing 734,414 particles) were selected out of 200 2D classes for ab initio three-dimensional (3D) reconstruction, which produced one good class with recognizable structural features and three bad classes that did not have structural features. Both the good and bad classes were used as references in three rounds of heterogeneous refinement (cryoSPARC) and yielded a good class at 4.06 Å from 465,752 particles. Then nonuniform refinement (cryoSPARC) was performed with C2 symmetry and an adaptive solvent mask, CTF refinement yielded a map with an overall resolution of 3.2 Å.

For xlCHPT1-Mg2+-CDP-choline data set, a total of 3,859,534 particles were automatically picked in RELION 3.1 with template picking from 5510 images and imported into cryoSPARC. After three rounds of 2D classification, 15 classes (containing 1,229,270 particles) were selected out of 200 2D classes for ab initio three-dimensional 3D reconstruction, which produced one good class with recognizable structural features and three bad classes that did not have structural features. Both the good and bad classes were used as references in three rounds of the heterogeneous refinement (cryoSPARC) and yielded a good class at 4.15 Å from 381,720 particles. Then nonuniform refinement (cryoSPARC) was performed with C2 symmetry and an adaptive solvent mask, CTF refinement yielded a map with an overall resolution of 3.68 Å. Resolutions were estimated using the gold-standard Fourier shell correlation with a 0.143 cutoff36 and high-resolution noise substitution37. Local resolution was estimated using ResMap38.

Model building and refinement

The structural models of xlCHPT1 were built de novo into the density map starting with poly-alanine, and sidechains were then added onto the model based on the map. Model building was conducted in Coot39. Structural refinements were carried out in PHENIX in real space with secondary structure and geometry restraints40. The EMRinger Score was calculated as described41.

Enzyme-coupled enzymatic assay

xlCHPT1 activity was measured using an absorbance-based coupled-enzyme assay (Supplementary Fig. 2b). All reaction assays were done in a buffer with 20 mM HEPES, pH 7.5, 150 mM NaCl, 0.02% GDN, 5 mM MgCl2. In single point activity assay, final concentrations of CDP-choline and 1,2-sn-diacylglycerol were 0.2 mM and 0.25 mM, respectively. Reactions were initiated with the addition of 10 μg purified xlCHPT1 protein. After reaction at 37 °C for 15 mins, reactions were stopped by heating up at 95 °C for 5 min. After removing the precipitated protein with centrifugation, purified yeast pyrimidine 5′-nucleosidase SDT1 was added into the supernatant, then react at 37 °C for 20 mins. Finally, phosphate dye (Sigma, MAK030) was added to the reaction solution and A650 nm was measured in 96 plates. When CDP-choline or CDP-ethanolamine concentrations are varied, DAG concentration is fixed at 250 μM. The initial rate versus different concentrations of CDP-choline and CDP-ethanolamine can be fit with a Michaelis–Menten equation.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.