2024 Volume 47 Issue 1 Pages 23-27
Mammalian type 2 carnitine parmitoyltransferase (EC 2.3.1.21), abbreviated as CPT2, is an enzyme involved in the translocation of fatty acid into the mitochondrial matrix space, and catalyzes the reaction acylcarnitine + CoA = acyl-CoA + carnitine. When rat CPT2 was expressed in Escherichia coli, its behavior was dependent on the presence or absence of i) its mitochondrial localization sequence and ii) a short amino acid sequence thought to anchor it to the mitochondrial inner membrane: CPT2 containing both sequences behaved as a hydrophobic protein, while recombinant CPT2 lacking both regions behaved as a water soluble protein; if only one region was present, the resultant proteins were observed in both fractions. Because relatively few protein species could be obtained from bacterial lysates as insoluble pellets under the experimental conditions used, selective enrichment of recombinant CPT2 protein containing both hydrophobic sequences was easily achieved. Furthermore, when CPT2 enriched in insoluble fraction was resuspended in an appropriate medium, it showed catalytic activity typical of CPT2: it was completely suppressed by the CPT2 inhibitor, ST1326, but not by the CPT1 inhibitor, malonyl-CoA. Therefore, we conclude that the bacterial expression system is an effective tool for characterization studies of mammalian CPT2.
Mitochondria are organelles in eukaryotic cells that represent the main site of energy conversion; in addition, many metabolic reactions, such as the tricarboxylic acid (TCA) cycle, oxidative phosphorylation and fatty acid oxidation, take place in the mitochondrial matrix. Mitochondria have a double membrane structure and the inner mitochondrial membrane is particularly impermeable, because it is important to maintain an electrochemical gradient across this membrane so that it can be used to drive ATP synthesis. To enable translocation of molecules across the inner mitochondrial membrane, specific transporters are located in the inner mitochondrial membrane. These transporters show structural similarity to each other and are thought to derive from a common ancestral protein; hence, they are referred to as the mitochondrial solute carrier family, classified as SLC25 (for recent reviews).1–4)
The translocation of fatty acids across the inner mitochondrial membrane is somewhat more complex, involving the “carnitine shuttle system” (Supplementary Fig. S1). Central to this system is the carnitine/acylcarnitine carrier, which is a member of the mitochondrial solute carrier family, designated SLC25A20, but in addition fatty acid translocation requires three enzymes: i) acyl-CoA synthetase, ii) carnitine palmitoyltransferase 1 (CPT1), and iii) CPT2 (for reviews of CPT1 and 2).5–9) In this system, fatty acid is first converted to acyl-CoA by acyl-CoA synthetase, then acyl-CoA is converted to acylcarnitine by CPT1. The acylcarnitine molecule thus formed is translocated across the inner mitochondrial membrane via the carnitine/acylcarnitine carrier, and the acylcarnitine delivered into the mitochondrial matrix space is then reconverted into acyl-CoA by CPT2. Three isozymes (a–c) are known for CPT1, but no isozymes have been identified for CPT2. CPT1 is anchored to the outer mitochondrial membrane by its N-terminal “membrane spanning regions,”10,11) but CPT2 is thought to attach to the inner surface of the inner mitochondrial membrane via its “putative membrane interacting region”6,8,9) (for comparison of amino acid sequences of human and rat CPT1a, b, and 2, see Supplementary Fig. S2). The membrane spanning region of CPT1 is essential for its catalytic activity,12) but detailed studies on the relationship between structure and function in CPT2 have not yet been carried out. In the present study, to understand its structural and functional properties, we expressed CPT2 in Escherichia (E.) coli and examined the properties of the resultant recombinant protein and various derivatives.
Plasmid vector pColdIII and E. coli strain DH5α were purchased from TaKaRa Bio Inc. (Shiga, Japan). E. coli strain BL21(DE3) was purchased from Thermo Fisher Scientific Inc. (Waltham, MA, U.S.A.). Polyclonal antibody against CPT2 (code GTX33117) and secondary antibody against rabbit antibodies (code NA34-1ML) were purchased from GeneTex (Irvine, CA, U.S.A.) and Cytiva (Marlborough, MA, U.S.A), respectively. Pre-stained protein standard, broad range (code P7718) was purchased from New England Biolabs (Ipswich, MA, U.S.A.). CPT2 specific inhibitor, ST1326, was purchased from Merck (code 870853, Darmstadt, Germany). Carnitine hydrochloride, L-[N-methyl-3H] (code ART0293) was purchased from American Radiolabeled Chemicals Inc. (St. Louis, MO, U.S.A.).
Preparation of Expression VectorsFull-length cDNA encoding rat CPT2 was prepared by RT-PCR based on its nucleotide sequence (NCBI Reference Sequence: NM_012930.1) using primers shown in Supplementary Fig. S3. To facilitate subsequent recombination experiments, intrinsic restriction sites for BamHI and NdeI present in the original cDNA encoding rat CPT2 were disrupted without changing the encoded amino acids, and new restriction sites for NdeI and BamHI were introduced into the 5′ and 3′ ends of the cDNA, respectively. Rat CPT2 cDNAs lacking a mitochondrial localization signal (amino acids Met1 to Leu25) and lacking the putative membrane interacting region (amino acids Asn179 to Asn208) were both obtained by overlap extension PCR.13) Primers used for these purposes are also listed in Supplementary Fig. S3. After construction of individual expression vectors, their nucleotide sequences were confirmed.
Preparation of Bacterial LysatesA 100 mL aliquot of bacterial culture was centrifuged at 1710 × g, 4 °C for 20 min, and the pellet was suspended in 10 mL Tris-EDTA (TE) (10 mM Tris-Cl pH 7.5, 1 mM ethylenediaminetetraacetic acid (EDTA)). The bacterial suspension was sonicated using a Branson Sonifier, model 450, at output level 10, for 8 min on ice. This was then centrifuged at 1710 × g, 4 °C for 20 min, to remove unbroken cells, and the cellular lysate obtained was termed the whole lysate. A 1 mL aliquot of whole lysate was subjected to further centrifugation at 16100 × g, 4 °C for 10 min, and the supernatant obtained was termed the supernatant fraction. The pellet obtained was washed with 1 mL TE and resuspended in 50 µL TE; this was termed the insoluble fraction.
Sodium Dodecyl Sulfate (SDS)-Polyacrylamide Gel Electrophoresis (PAGE) and Western BlottingSDS-PAGE was performed using the standard procedure of Laemmli with 7.5% polyacrylamide gels. Antibody against CPT2 and secondary antibody were used at 2000-fold dilutions, and immunoreactive protein bands were visualized using enhanced chemiluminescence (ECL) Western blotting detection reagents (GE Healthcare, Chicago, IL, U.S.A.).
Measurement of Enzyme Activity of Bacterially Expressed CPT2The activity of CPT2 was measured using L-[methyl-3H] carnitine hydrochloride as a tracer, essentially as described previously for measurement of CPT1 activity.14–16) Briefly, the reaction was initiated by addition of 50 µL sample solution containing 30 µg protein sample of crudely purified CPT2 to the prewarmed reaction solution (450 µL), to make final concentrations of 150 mM KCl, 1 mM EDTA, 0.25 mM reduced glutathione, 1.3 mg/mL fatty acid-free bovine serum albumin (BSA), 0.5 mM L-carnitine, 50 µM palmitoyl-CoA, 2 mM NaCN, 9.25 kBq L-[methyl-3H] carnitine and 50 mM N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (Hepes) buffer, pH 7.4. After incubation at 30 °C for 20 min, the reaction was terminated by the addition of 200 µL 3 M HC1. The palmitoyl L-[N-methyl-3H] carnitine formed was extracted into 500 µL n-butanol, and the butanol phase was back-extracted with 200 µL water saturated with butanol. Then, 200 µL of the final organic phase was taken for liquid scintillation counting. The activity of rat CPT1b expressed in COS7 cells, which is sensitive to malonyl-CoA, was also measured as a control.
Mass SpectrometryCoomassie-stained protein bands around 74, 40, and 37 kDa were excised and digested in-gel with trypsin as described previously.17) The digested peptides were desalted using GL-Tip SDB (GL Sciences, Tokyo, Japan) desalting tips, evaporated in a SpeedVac concentrator and dissolved in 0.1% trifluoroacetic acid and 3% acetonitrile. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis was performed on an EASY-nLC 1200 UHPLC instrument connected to an Orbitrap Fusion mass spectrometer through a nanoelectrospray ion source (Thermo Fisher Scientific). Raw data were directly analyzed against the Swiss-Prot database restricted to E. coli sequences plus the rat CPT2 sequence using Proteome Discoverer 2.5 (Thermo Fisher Scientific) with the Sequest HT search engine.
Statistical AnalysisStatistical analysis of multiple groups was performed using Dunnett’s test with R software (version 4.1.2, The R Foundation for Statistical Computing). p < 0.05 was considered significant. All data are presented as the mean ± standard deviation (S.D.).
We first constructed an expression vector of full length rat CPT2, including its mitochondrial localization signal and putative membrane interacting domain, using pColdIII in the E. coli strain, DH5α. Although DH5α strain is mainly used for the purpose of recombinant experiments,18) and not optimized for expression of foreign protein, to check whether the expression vector is properly constructed, and whether the antibody properly recognize the target CPT2, bacterial lysates prepared from DH5α cells containing either pColdIII (“vector”) or pColdIII/rCPT2 (“full-length”) were subjected to SDS-PAGE; and subsequent Coomassie staining (Fig. 1A, upper panel) or Western blotting using anti-CPT2 antibody (Fig. 1A, lower panel) were performed. As a result, an immunoreactive protein band corresponding to rat CPT2 (74 kDa) was clearly observed in the whole lysate (W) prepared from DH5α cells containing the pColdIII/rCPT2 plasmid, but not from DH5α cells with the empty pColdIII vector, indicating that CPT2 was successfully expressed in DH5α cells. To seek the physicochemical properties of the expressed protein, the whole lysate prepared from DH5α transformed with the pColdIII/rCPT2 plasmid was centrifuged in a conventional “microfuge” (at 16100 × g, 4 °C for 10 min), and the resulting supernatant (S) and pellet (P) were subjected to SDS-PAGE and Western blotting. As a result, a protein showing the expected molecular size of full length CPT2 was found to be selectively pelleted to a greater extent than most other bacterial proteins (see lane “rat CPT2/full-length/P”). It should be noted that two bacterial protein bands showing apparent molecular sizes of 40 and 37 kDa were also selectively enriched in the pellet fraction. To confirm that the 74 kDa protein represents expressed CPT2, and to identify the molecular species of the protein bands at 40 and 37 kDa, these three protein bands were dissected from the gel and subjected to mass spectrometry. The results identified the majority of these protein bands as rat CPT2, E. coli outer membrane porin C (OmpC), and outer membrane protein A (OmpA), respectively (for detailed results, see Supplementary Tables I–III). It is currently unclear why these three protein species were selectively obtained under the experimental conditions used. However, it is possible that they show similar physicochemical properties, such as hydrophobicity.
The upper and lower panels represent the results of Coomassie staining and Western blotting using anti-CPT2 antibody, respectively. Lane M at edge of the gel represents molecular weight standards (New England BioLabs, P7718). Lanes W, S, and P in upper and lower photograph (A) represent 21 or 10-µg aliquots of protein samples of whole lysate, supernatant and pellet, respectively, obtained from DH5α strain harboring pColdIII vector (vector) or pColdIII vector encoding full-length rat CPT2 (full-length). In the upper photograph (A), a protein band seemed to reflect the expressed CPT2 (74 kDa), and those of two protein bands showing apparent migration of 40 and 37 kDa, selectively precipitated by centrifugation, were highlighted with arrows. In the photograph (B), in addition to the transformant of BL21(DE3) with sham vector (vector) or with pColdIII vector encoding full-length rat CPT2 (full-length), with those lacking amino acid sequence Met1 to Leu25 (Δ1–25), Asn179 to Asn208 (Δ179–208) or both (Δ1–25, Δ179–208) were also subjected to the analysis. Arrangements of loaded protein samples in the photograph (B) are as shown in photograph (A), but amounts of protein samples loaded for Coomassie staining and Western blotting were 10 and 3 µg, respectively.
To examine whether these results are specifically observed with DH5α cells or constantly observed with the other cells, the E. coli strain, BL21(DE3), which was established for efficient expression of foreign proteins,19) was transformed with the empty pColdIII vector or with pColdIII/rCPT2 plasmid, and protein expression was induced either by incubation at lower temperature (at 15 °C) and addition of IPTG, and protein samples prepared from the bacteria were subjected to SDS-PAGE and subsequent Coomassie staining or Western blotting. As a result, expression level of full length CPT2 in BL21(DE3) strain was much remarkable than that observed in DH5α strain (Fig. 1B, “rat CPT2,” “full length”), and it was effectively recovered as pellet by microfuge, and remarkable amounts of full length CPT2, being approximately 70% of total proteins in the pellet fraction, was obtained (Fig. 1B, “rat CPT2,” “full length,” “P”).
To understand why CPT2 shows such unique behavior, we focused on its primary structure. Newly translated CPT2 protein has a mitochondrial localization signal at its N-terminus (Met1 to Leu25). In addition, CPT2 has a unique additional sequence (Asn179 to Asn208; “CPT2-specific insert”) that is not observed in the CPT1 isozymes 1a or b (for homology alignment of amino acid sequences of CPT1 and 2, see Supplementary Fig. S2). As these sequences seemed to confer hydrophobic properties on the CPT2 protein, we next tested the effect of deleting these regions on the behavior of CPT2. Accordingly, we prepared three expression vectors of CPT2: CPT2 lacking its mitochondrial localization signal (designated “Δ1–25”); CPT2 lacking its putative membrane interacting region (designated “Δ179–208”); and CPT2 lacking both regions (designated “Δ1–25, Δ179–208”). When BL21(DE3) strain was transformed with the individual expression vectors and their respective lysates were subjected to SDS-PAGE and Western blotting (Fig. 1B), the effects of deleting the individual regions were apparent. Namely, deleting either the mitochondrial localization signal (Δ1–25) or the putative membrane interacting region (Δ179–208) reduced the amount of CPT2 protein in the insoluble fraction. Surprisingly, in the protein sample where both regions were deleted (Δ1–25, Δ179–208), almost no protein was observed in the pellet fraction, although immunoreactive protein bands were observed in the whole lysate (W) and supernatant (S) samples. These results clearly indicate that the unique behavior of bacterially expressed rat CPT2 is dependent on its mitochondrial localization signal and putative membrane interacting region. As a result, contents of CPT2 protein in pellet fraction of the protein sample expressing full length rat CPT2 (lane “P,” sample of “full length”) became dominant, and hence, we concluded that it was selectively precipitated and concentrated.
Finally, we asked whether bacterially expressed full-length CPT2, partially purified by simple centrifugation as described above, shows catalytic activity. As a comparison, we also expressed rat CPT1b in COS7 cells as previously described,14–16) and used lysates of these cells to represent CPT1b. As shown in Fig. 2A, CPT1b expressed in COS7 cells was inhibited by malonyl-CoA (CPT1 inhibitor) in a dose-dependent manner, and more than 92% suppressed by 10 µM malonyl-CoA, but it showed a high level of resistance to ST1326 (CPT2 inhibitor), i.e., it showed more than 88% activity even in the presence of 3 µM ST1326. In contrast, CPT2 showed an opposite pattern of sensitivity to these inhibitors (Fig. 2B): it showed a high degree of resistance to malonyl-CoA, i.e., it showed more than 93% activity even in the presence of 10 µM malonyl-CoA, but it was strongly suppressed by ST1326 in a dose-dependent manner, and more than 79% suppressed by 3 µM ST1326. These results apparently show that CPT2 expressed in E. coli, and selectively enriched by conventional centrifugation, not only retained its catalytic activity, but was also specifically inhibited by ST1326. Therefore, the bacterial expression system was concluded to be appropriate for expression studies of mammalian CPT2. It should be mentioned that we did not make additional treatment of CPT2 protein for its solubilization in the reaction mixture subjected to the analysis of catalytic activities. However, certain parts of CPT2 protein were expected to be dissolved in the reaction mixture, because protein concentration of partially purified CPT2 in the reaction mixture used for activity measurement was approximately 23-fold diluted than that in the suspension of whole lysate.
Data represent mean value ± S.D. *** p < 0.001; NS, not significant; versus “none” by Dunnett’s test (n = 3). Individual data points are also indicated by small open circles.
When studying recombinant proteins that are naturally embedded in or associated with a membrane, particular care must be taken because such proteins often require the membrane to form (or keep) their active conformations. In the case of CPT2, the main structural difference with CPT1a and 1b is the specific insertion of a short amino acid sequence, at around position 180-210 in the amino acid sequence; a possible interaction of CPT2 with the mitochondrial inner membrane via this region has been hypothesized.6,8,9) Moreover, the crystal structure of CPT2 shows that this region protrudes from the main body of the protein20–22) and, in the present study, we have shown that the presence or absence of this region has a marked effect on the physicochemical properties of the CPT2 protein.
As for the catalytic activities of bacterially expressed CPT2, two preceding papers were published.23,24) The group of McGarry achieved bacterial expression of CPT2, in a similar manner to that in the present study, and mainly reported different behavior of the rat and human enzymes.23) On the other hand, Cook and his colleagues reported the addition effects of phospholipids to the bacterially expressed CPT2.24) However, both groups did not mention the unique behavior of bacterially expressed full-length mammalian CPT2, which could be easily and selectively precipitated by conventional centrifugation, as clearly shown in the present study.
Although for a long period there were few reports on the functional characterization of CPT2 as stated above, recently several papers were published by the group of Zierz,25–27) mainly to characterize the mutants discovered in patients with CPT2 deficiency. They expressed histidine (His)-tagged CPT2 in E. coli BL21-Gold (DE3) cells and purified it using Ni-NTA affinity chromatography, gel filtration and ion-exchange chromatography on Q-Sepharose HP. The authors added 0.01 g/L n-octyl-β-D-glucopyranoside (β-OG) to stabilize CPT2, and the purified enzyme was stored in this solution. Otherwise, the experimental procedure used was conventional. However, as stated in this report, we found bacterially expressed CPT2, carrying its N-terminal mitochondrial localization signal, to be significantly enriched in the fraction of insoluble proteins, and hence, its partial purification was easily achieved. CPT2, with its N-terminal mitochondrial localization signal and putative membrane-interacting region is likely to have hydrophobic properties due to these two regions, thereby making the protein easily and selectively precipitable from aqueous solutions. However, the recovered protein was not markedly denatured, because it showed appropriate enzyme activity when resuspended in reaction medium. Importantly, we succeeded in demonstrating different sensitivities of CPT1b and CPT2 to the inhibitors malonyl-CoA and ST1326. Therefore, protein samples prepared as outlined above represent convenient experimental tools for improving our understanding of CPT2 and possibly other proteins involved in the carnitine shuttle system.
The authors declare no conflict of interest.
This article contains supplementary materials.
