- Email: [email protected]
CTP:phosphoethanolamine cytidylyltransferase
Biochimica et Biophysica Acta 1348 Ž1997. 91–99
Chapter VII
CTP:phosphoethanolamine cytidylyltransferase Bellinda A. Bladergroen, Lambert M.G. van Golde
)
Department of Veterinary Basic Sciences, DiÕision of Biochemistry, Graduate School of Animal Health and Institute of Biomembranes, UniÕersity of Utrecht, P.O. Box 80176, 3508 TD Utrecht, The Netherlands Received 9 April 1997; accepted 20 May 1997
Abstract CTP:phosphoethanolamine cytidylyltransferase ŽET. catalyzes the conversion of phosphoethanolamine into CDP-ethanolamine. Immunogold electron microscopy studies have demonstrated that, in hepatocytes, ET is localized predominantly in areas of the cytoplasm that are rich in rough endoplasmic reticulum ŽRER.. Within these areas the enzyme shows a bimodal distribution between the cisternae of the RER and the cytosolic space. Studies on the substrate specificity of ET have shown that it can utilize both CTP and dCTP as substrates, but not other trinucleotides. In addition, the enzyme shows a very pronounced specificity for phosphoethanolamine. Under most conditions ET contributes significantly to the overall regulation of the CDP-ethanolamine pathway. Reversible binding of the enzyme to the endoplasmic reticulum could potentially play a key-role in metabolic channeling of phosphatidylethanolamine synthesis. ET has been purified from rat liver. Convincing evidence has been provided that ET and CTP:phosphocholine cytidylyltransferase ŽCT., the analogous enzyme in the CDP-choline pathway, are separate activities that reside on different proteins. The gene coding for yeast ET has been cloned. The deduced amino acid sequence contained a region in the N-terminal half with significant similarities to the conserved catalytic domain of both yeast and rat CT. The human cDNA for ET was also cloned recently. The predicted amino acid sequence of human ET shows a high degree of similarity Ž36% identity. to that of yeast ET, but the human protein is longer than the yeast protein, especially at the C-terminal region. Interestingly, both yeast and human ET have a large repetitive sequence in their N-terminal and C-terminal half. q 1997 Elsevier Science B.V. Keywords: CTP:phosphoethanolamine cytidylyltransferase; Phosphatidylethanolamine; CDP-ethanolamine; CTP:phosphocholine cytidylyltransferase; Phosphatidylcholine; Hepatocyte
1. Discovery, occurrence and cellular localization 1.1. DiscoÕery and occurrence As is the case for most enzymes involved in the biogenesis of phospholipids, the credit for the discovAbbreviations: ET, CTP:phosphoethanolamine cytidylyltransferase; RER, rough endoplasmic reticulum; CT, CTP:phosphocholine cytidylyltransferase; PE, phosphatidylethanolamine; PC, phosphatidylcholine ) Corresponding author. Fax: q31 30 2535492.
ery of CTP: phosphoethanolamine cytidylyltransferase ŽET. goes to Eugene Kennedy and his associates. It was about four decades ago that Kennedy and Weiss elucidated the CDP-ethanolamine pathway for the synthesis de novo of phosphatidylethanolamine ŽPE.. In their famous 1956 paper w1x they reported that particulate fractions of rat liver catalyzed the incorporation of phosphoethanolamine into PE and that this conversion specifically required the addition of CTP. They proved that CDP-ethanolamine was the immediate precursor of PE and that this novel intermediate was formed from phosphoethanolamine and
0005-2760r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 5 - 2 7 6 0 Ž 9 7 . 0 0 1 1 3 - 6
92
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
CTP in a reaction catalyzed by a cytidylyltransferase activity: phosphoethanolamineq CTP ° CDP y ethanolamineq PPi The hydrolysis of pyrophosphate, catalyzed by ubiquitous intracellular pyrophosphatases, drives this reaction, which would otherwise be freely reversible, in the direction of synthesis of CDP-ethanolamine w1–3x. ET has been found in all eukaryotic cells investigated so far, including lower eukaryotes such as yeast, where the CDP-ethanolamine pathway plays only an auxiliary role w4x. However, in most cell types the activity of ET is substantially less than that of CTP:phosphocholine cytidylyltransferase Ž CT. , the analogous enzyme in the biosynthesis of phosphatidylcholine ŽPC.. 1.2. Subcellular organization Until recently it was generally accepted that ET of mammalian cells was a soluble enzyme that, in contrast to CT, did not associate with cellular membranes. This conclusion was based on classical differential centrifugation studies as well as on enzyme-release measurements from digitonin-permeabilized
hepatocytes Žsee Ref. w5x and references cited therein.. However, it is not always certain that enzymes, classified as soluble from such studies, are indeed soluble in situ. The possibility cannot be excluded that in the cell they are weakly bound to membranes or to other structural elements, from which they could be released during the mechanical fractionation of the tissue or during cell permeabilization employing cholesterol-sequestering agents such as digitonin. Van Hellemond et al. w6x performed immunogold electron microscopy studies on the cellular localization of ET in rat liver using a polyclonal affinity-purified antibody against the enzyme. Their results showed that ET was not randomly distributed in hepatocytes. As illustrated in Fig. 1 the ET label was concentrated particularly in areas that contained cisternae of the rough endoplasmic reticulum ŽRER.. Other cellular organelles, including nuclei, mitochondria and plasma membranes, were only marginally labeled. Doublelabel experiments for ET and established markers for either soluble or integral endoplasmic reticulum proteins suggested a bimodal distribution of ET between the cisternae of the RER and the cytosolic space w6x. These findings suggested that ET might reversibly interact with membranes of the endoplasmic reticulum, which could bring the enzyme into close prox-
Fig. 1. Immunogold electron microscopy study on the localization of ET in rat hepatocytes. ET label is strongest in RER-rich areas of the cytoplasma Žsolid star.. Less labeling is observed in other cytoplasm regions Žopen star.. Almost no gold markers are present in the nucleus Žn., mitochondria Žm., and plasma membranes Žarrows.. Reprinted with permission ŽRef. w6x..
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
imity of ethanolaminephosphotransferase, an established integral microsomal protein w7x that catalyzes the final step in the CDP-ethanolamine pathway. The virtual absence of ET in rat liver nuclei is interesting in view of recent studies on the subcellular localization of CT. Until a few years ago it was unanimously accepted that CT was localized in both the cytosolic and endoplasmic reticulum compartment of all animal cells Žfor review see Ref. w8x.. However, intriguing immunofluorescence studies of Kent and her colleagues have shown that CT is predominantly present in the nucleus of CHO cells and in several other cell lines w4,9,10x. In addition, these investigators showed that activation of PC synthesis in HeLa cells treated with oleate resulted in translocation of inactive soluble nuclear CT to the nuclear membrane Žreviewed in Ref. w4x.. Houweling et al. w11x recently employed a combined approach of indirect immunofluorescence microscopy, immunogold electron microscopy, and biochemical techniques to localize CT in CHO cells and in hepatocytes. Although they confirmed the abundant nuclear localization of CT in CHO cells, they provided strong evidence that CT is not predominantly a nuclear enzyme in all cells, as had been suggested by Kent and colleagues. In rat hepatocytes CT appeared to be —like ET—primarily a cytoplasmic enzyme w11x, despite the fact that rat liver CT possesses the nuclear localization sequences w4x. The significance of the fact that at least in some cells a substantial portion of CT is located in the nuclei w4,8–11x is still unclear.
2. Substrate specificity and regulation 2.1. Substrate specificity The original studies of Kennedy and Weiss had already shown that CTP could not be replaced as a substrate for ET and CT by ATP, ITP, UTP or GTP w1x. Further investigations by Kennedy and his associates w2x showed that ET and CT did not only catalyze the formation of CDP-ethanolamine and CDP-choline, respectively, but also that of the corresponding deoxyribonucleotides dCDP-ethanolamine and dCDP-choline. Later investigations with a mutant cell line derived from CHO cells, in which the specific activity of CT was reduced to 1–10% of wild type,
93
provided convincing evidence that the formation of CDP-choline and dCDP-choline was catalyzed by a single enzyme as the pool sizes of both CDP-choline and dCDP-choline were both strongly depressed w12x. Likewise it is highly likely that one single enzyme ŽET. is responsible for the formation of both CDPethanolamine and its deoxy analog. The biological function of dCDP-choline and dCDP-ethanolamine remains an enigma w3x. The presence of relatively high levels of these deoxyribonucleotides in rapidly metabolizing cells, may simply reflect the high levels of dCTP in the cytoplasma of such cells. On the other hand, it cannot be excluded that dCDP-ethanolamine and dCDP-choline are selectively utilized, e.g. for the formation of distinct pools or distinct molecular species of PE and PC, respectively w3x. Studies on the specificity of purified ET for phosphorylated bases with a varying degree of N-methylation, which will be discussed in more detail in Section 3.2, showed that ET displays a high specificity for phosphoethanolamine w13x. 2.2. Role of ET in the regulation of the CDP-ethanolamine pathway Whereas the overall regulation of PC synthesis via the CDP-choline pathway has been subject of many intense studies w4,8x, there is much less information on the control of the CDP-ethanolamine pathway. The evidence available so far strongly suggests, however, that these two pathways are under independent metabolic and hormonal control Ž for review see Ref. w14x.. Several studies have addressed the question which of the individual steps in the CDP-ethanolamine pathway could provide an important contribution to the overall metabolic regulation of the process. On the basis of theoretical considerations, Infante w15x suggested that both the reaction catalyzed by ethanolamine kinase Žcholinerethanolamine kinase. and that catalyzed by ET could be important regulatory steps in the CDP-ethanolamine pathway in rat liver. Sundler ˚ w16x demonstrated that exposure of hepand Akesson atocytes to increasing concentrations of ethanolamine resulted in an enhanced entry of labeled glycerol into PE that was accompanied by a considerable increase of the pool size of phosphoethanolamine. As the pool size of CDP-ethanolamine remained constant, these
94
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
observations strongly suggested that the reaction catalyzed by ET should indeed be considered as a potentially important regulatory site in PE biosynthesis via the CDP-ethanolamine route. Further evidence for such a role of ET came from studies by Tijburg et al. w17x who showed that exposure of hepatocytes to phorbol ester led to an increased rate of incorporation of labeled ethanolamine into PE, increased activity of ET, and a reduction in the pool size of phosphoethanolamine. Interestingly, the phorbol ester treatment also caused an enhanced conversion of CDPethanolamine into PE. The last enzyme in the CDPethanolamine pathway, ethanolaminephosphotransferase, is thought to be at equilibrium in vivo and as such is in itself not an important regulatory enzyme. The effect was probably due to the increased amount of cellular diacylglycerol that was generated by the treatment of the cells with phorbol ester. Further evidence for a potentially important role of diacylglycerol supply to the regulation of PE synthesis came from a subsequent study on the effect of glucagon on PE synthesis in hepatocytes w18x. Exposure of hepatocytes to glucagon or cyclic AMP analogs decreased the rate of incorporation of labeled ethanolamine into PE, but did not affect the activity of ET or the pool size of phosphoethanolamine. The inhibitory effect of glucagon on PE synthesis could, however, be attributed to a diminished supply of diacylglycerol resulting in a decreased conversion of CDP-ethanolamine into PE. At least in liver there is very little experimental evidence suggesting important control of PE synthesis at the site of ethanolamine kinase. Experiments of ˚ w16x and Houweling et al. w19x Sundler and Akesson with hepatocytes showed that at low ethanolamine concentrations in the medium the rate of PE synthesis was controlled by the supply of ethanolamine, whereas at higher ethanolamine concentrations Ž ) 30–50 mM. phosphoethanolamine started to accumulate indicating that ET had become rate-limiting. It is, however, important to emphasize that the contribution of each step to overall metabolic regulation of the pathway may differ among tissues. For example, there is evidence from studies with isolated perfused hamster heart that ethanolamine kinase could contribute significantly to the overall control of the CDP-ethanolamine pathway, particularly at high ethanolamine concentrations in the medium w20x.
Summarizing, it can be stated that regulation of PE synthesis can take place at multiple sites in the CDP-ethanolamine pathway, although under most conditions the supply of CDP-ethanolamine and diacylglycerol appear to be principal factors governing the rate of this process. 2.3. Potential key-role of ET in metabolic substrate channeling It has been well established that the subcellular organization of CT is a major determinant in regulating the activity of this enzyme and that of the CDPcholine pathway as a whole w4,8x. The soluble form of CT appears to represent an inactive reservoir. From this reservoir the enzyme can rapidly and reversibly translocate to cellular membranes Žendoplasmic reticulum membranes w8x andror the nuclear envelope w4x. where it becomes activated by interactions with lipids. A comparable translocation mechanism for regulation of the activity of ET, the putative key-regulatory enzyme in the CDP-ethanolamine pathway, seemed less likely as it was until recently generally accepted that ET of mammalian cells was a fully soluble enzyme that did not bind significantly to membranes. In addition, the activity of ET is not stimulated by interactions with lipids w5x. However, as mentioned in a preceding section, the immunogold electron microscopy studies by van Hellemond et al. w6x demonstrated that ET was present predominantly in RER-rich areas of hepatocytes, and that within these areas the enzyme partitioned between ER membranes and the cytosol. Although difficult to prove experimentally, reversible binding of ET to the ER could be an essential step in metabolic channeling of PE synthesis as proposed schematically in Fig. 2. The early observation by Sundler w21x that phosphoethanolamine formed from exogenous ethanolamine did not freely equilibrate with the endogenous liver phosphoethanolamine pool already suggested that the product of the ethanolamine kinase reaction may be specifically channeled towards ET. As ethanolaminephosphotransferase, the next and final enzyme in the pathway, is an integral microsomal enzyme, reversible binding of ET to the membrane could indeed play a key-role in the topodynamic regulation w22x of the CDPethanolamine pathway as a whole Ž Fig. 2. . The con-
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
cept of compartmentalization of the aqueous precursors in the CDP-ethanolamine pathway finds also support in a recent study of Shiao and Vance w23x. These investigators presented evidence for the operation of an ethanolamine-PE cycle in CHO cells. Via this cycle the ethanolamine moiety of PE synthesized from labeled ethanolamine was continuously released from PE and recycled back to PE via the CDPethanolamine route. Importantly, the labeled ethanolamine released from PE was re-incorporated into PE without dilution if the medium did not contain unlabeled ethanolamine. If, however, the cells were incubated with unlabeled ethanolamine in the medium, the specific radioactivity of the intracellular ethanol-
95
amine pool and that of PE decreased as a result of dilution by the exogenous ethanolamine. Earlier studies by George et al. w24x had provided evidence for the compartmentalization of the enzymes of the CDP-choline pathway in cultured glioma cells. Only choline that was taken up by the high affinity uptake system was converted into PC, whereas phosphocholine and CDP-choline that were administered to the cells by electropermeabilization did not enter into PC. It would be of great interest if this compartmentalization could be demonstrated for other cell types and by other techniques of introducing phosphocholine, CDP-choline, and their ethanolamine analogs into the cells.
Fig. 2. Potential key-role of ET in compartmentalization of the CDP-ethanolamine pathway. The bimodal distribution of ET between RER membranes and the cytosolic compartment in RER-rich areas could provide the possibility of substrate channeling by topographic linking of ethanolamine kinase and ethanolaminephosphotransferase. Abbreviations: Eth, ethanolamine; P-Eth, phosphoethanolamine; PE, phosphatidylethanolamine; DAG, diacylglycerol; Eth-K, cholinerethanolamine kinase; ET, phosphoethanolamine cytidylyltransferase; EPT, ethanolaminephosphotransferase. The dashed line encircles the proposed metabolon of the CDP-ethanolamine pathway.
96
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
3. Purification Already in 1975 ET had been purified about 1000-fold from a post-microsomal fraction of rat liver w25x. More recently a new procedure was developed to purify the rat-liver enzyme to apparent homogeneity w5x. This procedure included the following steps: ammonium sulphate precipitation, basic anion exchange chromatography on DEAE cellulose, dye– ligand pseudo affinity on Matrex gel red A, hydrophobic interaction chromatography on octyl-Sepharose CL-4B, and dye–ligand pseudo affinity chromatography on Matrix gel blue A. The procedure resulted in a 1400-fold purification of ET with a final specific activity Ž6.5 mmolrmin per mg protein. that was a factor of five higher than that reported by Sundler w25x. Polyacrylamide gel electrophoresis under denaturing and reducing conditions showed the presence of a single protein with an estimated molecular mass of 49.6 kDa. The purified protein was identified as the ET enzyme based on the following observations: Ž i. it also showed one single protein band on gel under native polyacrylamide gel conditions Žthis protein carried ET enzyme activity and showed one single 49.6-kDa band on SDS gel. ; Ž ii. an affinity-purified polyclonal antibody raised against ET was strongly immunoreactive against the 49.6-kDa protein band on SDS gel; and Žiii. this antibody had also the ability to precipitate ET activity from rat liver cytosol. 3.1. Properties of the purified enzyme Superose 12 gel filtration revealed a molecular weight of the native enzyme of 100 000–120 000 which agreed with the earlier estimate by Sundler w25x. This may indicate that the enzyme occurs as a dimer. ET has one sharp pH optimum at pH 7.8 and a broader one, with lower maximal activity, around pH 6–6.5 w13,25x. Interestingly, two apparent K m values of 70 and 185 mM were found for phosphoethanolamine w13x. This could imply that ET contains a second binding site for phosphoethanolamine. Kinetic studies by Sundler had shown that the reaction catalyzed by ET probably proceeds via an ordered sequential reaction with CTP being the first substrate to bind to the enzyme and CDP-ethanolamine the last product to be released w25x.
The requirement of ET for a divalent cation is absolute: Mg 2q Ž5–10 mM. was most efficient, and the activity of ET was reduced to 50% if Mg 2q was replaced by optimal concentrations Ž 2–4 mM. of Mn2q. Other divalent cations were much less effective ŽCa2q . or completely inactive w25x. 3.2. Substrate specificity of purified ET The observations of Vermeulen et al. w5x provided convincing evidence that, unlike ethanolamine and choline kinase w26x, ET and CT of rat liver are separate activities residing on different proteins: Ž i. CT activity could not be detected in purified ET, whereas pure CT did not exhibit any ET activity; Ž ii. antibodies raised against ET did not cross-react with CT Žand vice versa.; Žiii. both cytosolic and purified ET were not markedly affected by the presence of a variety of exogenous lipids Žthis is in contrast to the well-known high dependence of the activities of cytosolic and purified CT on the presence of exogenous lipids w8x.; and Živ. the subcellular localization of ET and CT are distinct, as discussed in a previous section. These conclusions are obviously in line with recent studies showing that ET and CT are coded by different genes w27,28x Ž see Section 4. . As already predicted by the early experiments of Kennedy et al. with crude fractions from rat liver w2x, purified ET also utilized dCTP as a substrate w13x. Substitution of CTP by dCTP hardly affected the K m value for the nucleotide Ž 50 mM for CTP vs. 54 mM for dCTP.. However, the Vmax value decreased by a factor of 4.4 when CTP was substituted by dCTP. Vermeulen et al. w13x also studied the specificity of pure ET for phosphorylated bases with a varying degree of N-methylation. Increasing the number of N-methyl groups on phosphoethanolamine led to a strong increase in the apparent K m value and to an even more pronounced decrease of the Vmax of ET. Phosphomonomethylethanolamine, phosphodimethylethanolamine and phosphocholine appeared to be weak competitive inhibitors of the reaction catalyzed by ET when phosphoethanolamine was used as a substrate, with Ki values of 7.0, 6.8 and 52.9 mM, respectively. Vermeulen et al. calculated the specificity constants k catrK m of ET for these various phosphobases and compared the values with the corresponding data published for CT w29x. The data
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
demonstrated that increasing the number of N-methyl groups on phosphoethanolamine drastically decreased the specificity constants of ET by factors of about 10, 350 and 2.10 5, respectively. Decreasing the number of N-methyl groups on phosphocholine had a similar but smaller effect on the specificity constants of CT Žby factors of about 8, 40 and 570, respectively. . These data indicate that both ET and CT can tolerate a difference of one N-methyl group when compared to their regular substrates phosphoethanolamine and phosphocholine, respectively. While ET strongly favors phosphomonomethylethanolamine above phosphodimethylethanolamine, the reverse is true for CT. However, it is clear that both ET and CT show a very pronounced specificity for their regular substrates phosphoethanolamine and phosphocholine, respectively.
4. Cloning of the gene Min-Seok et al. w27x recently reported the isolation and partial characterization of the gene coding for ET of Saccharomyces cereÕisiae. These workers isolated several mutants of S. cereÕisiae that were unable to utilize extracellular ethanolamine for PE synthesis. Two of these mutants carried recessive chromosomal mutations in the same gene and they were both deficient in ET. By screening a genomic library, they were able to isolate three overlapping clones of different sizes that complemented the mutation. These shared a 2.8-kb DNA region comprising an open reading frame of 969 bp in length that appeared to code for ET. This was proven by expression studies in E. coli. It is now clear that mammalian CT contains at least four structural domains Žfor review see Refs. w4,8x.. Near the N-terminal is a nuclear localization signal, followed by a catalytic domain Ž residues 75– 235. in which CTs from yeast and rat are 65% identical. This is followed by the lipid-binding amphipathic helical domain Žresidues 239–298. and the phosphorylation domain of the enzyme Ž residues 315–367.. These two domains are involved in the regulation of CT. The predicted amino acid sequence of yeast ET contained one region Ž residues 7–160. with significant similarities to the conserved catalytic domain of both yeast and rat CT Ž55 amino acids
97
identical with rat CT and 52 with yeast CT. . As suggested by Min-Seok et al. w27x it is likely that this region of the yeast ET sequence also encompasses the catalytic domain of the enzyme. Another rather conserved region in yeast ET appeared to be a repeat of the first half of the conserved region in the Nterminal part of the enzyme. This repetition had not been observed for yeast and rat CT. The amino acid sequence of yeast ET lacks a hydrophobic domain that is long enough to span a bilayer and, in contrast to CT, no amphipathic helical domains could be predicted for yeast ET w27x. Short stretches of hydrophobic amino acids occur after residue 229 of yeast ET, but it is uncertain whether these could be involved in possible binding of ET to membranes. Interestingly, the expression studies of Min-Seok et al. w27x also provided preliminary evidence that yeast ET, like rat liver ET w6x, may associate with the endoplasmic reticulum. Very recently Nakashima et al. w28x reported the cloning of a human cDNA for ET by complementation in vivo of a yeast mutant of which the ECT1 gene was disrupted. The deduced protein encoded by the cDNA consists of 389 amino acids with a calculated molecular mass of approx. 43.8 kDa. The predicted amino acid sequence shows a high degree of similarity Ž36% identity. to that of yeast ET w27x, but the human protein is longer than the yeast protein in both the N- and C-terminal regions. Interestingly, as in yeast ET, there is a large repetitive sequence in the N- and C-terminal halves of human ET. As suggested by Nakashima et al., this could imply the presence of two catalytic domains in human and yeast ET. Whether this is related to the earlier mentioned suggestion from kinetic studies with ET purified from rat liver that the enzyme might contain two binding sites for phosphoethanolamine w13x, remains to be investigated.
5. Future developments The progress in our understanding of ET since it was first reported in 1956 has been much less impressive than that of the simultaneously discovered CT. The major reason is obviously the fact that the CDPcholine pathway has traditionally received much more attention from researchers than the analogous CDP-
98
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
ethanolamine route. However, evidence is accruing for specific metabolic functions of PE in mammalian cells, e.g. as a substrate for the production of second messengers in signal transduction w30x and its role as donor of the ethanolamine residue to the phosphoethanolamine bridge that links the glycosylphosphatidylinositol anchor to surface glycoproteins of eukaryotes w31x. This will undoubtedly intensify interest in the study of overall PE metabolism and its regulation in eukaryotic cells. As there is evidence from a number of studies that the CDP-ethanolamine pathway can play an important role in the biosynthesis of PE and its plasmalogen analog in several mammalian tissues Žfor review see Ref. w14x. , it is to be expected that ET and its regulation will attract increasing interest. However, in this context it is important to emphasize that the CDP-ethanolamine pathway is not the only route leading to PE synthesis. As discussed in another contribution to this special issue w32x, there is convincing evidence that decarboxylation of phosphatidylserine is another major contributor to PE biogenesis in mammalian cells. The specific roles of each of these two biosynthetic pathways and their relative contribution to overall PE synthesis are not yet precisely known but might be different among different mammalian tissues. The availability of the cDNA w28x, complemented with that of purified ET and specific antibodies against the enzyme, will undoubtedly provide impetus to further studies on the properties of ET and the mechanisms by which the enzyme is regulated in mammalian cells. It is obvious that studies will be performed to pinpoint various domains in the structure of the protein, with identification of the catalytic domain, the domain via which the enzyme may bind to the endoplasmic reticulum, and domains that could otherwise play a role in regulation of ET activity. It would be of interest to investigate whether ET, like CT, can be reversibly phosphorylated in the mammalian cell and whether such phosphorylationrdephosphorylation mechanism could contribute to regulating the activity of the enzyme. Another interesting question is whether regulation of ET can also occur at a pretranslational level, as it has been shown that the expression of CT mRNA can be modulated under certain conditions, e.g. in the regenerating liver w33x. It would also be desirable to perform ET knockout experiments in mice, as such an approach might
provide useful information on the role and importance of ET and the CDP-ethanolamine pathway in cellular phospholipid metabolism and function.
Acknowledgements It is an enormous pleasure to dedicate this review to Professor Eugene Kennedy. After opening up the field of research on phospholipid metabolism in the early fifties, Gene Kennedy has remained a global leader in this field for a period of more than four decades. For many investigators he has been a highly appreciated and stimulating teacher and an experienced guide in this intriguing but complex area of research. One of the authors of this review Ž Bert van Golde. had the excellent opportunity of spending one year as a research associate in Gene Kennedy’s laboratory in Boston, where he performed, under Gene’s stimulating guidance, the initial studies on the membrane-derived oligosaccharides in Escherichia coli. He would like to take this opportunity of formally thanking Gene again for his warm hospitality and for providing a friendly and scientifically highly stimulating atmosphere in the laboratory. In such an atmosphere, in which each of Gene’s co-workers had almost unlimited access to him to discuss ups and downs in their project, it was almost impossible not to perform good research.
References w1x E.P. Kennedy, S.B. Weiss, The function of cytidine coenzymes in the biosynthesis of phospholipids, J. Biol. Chem. 222 Ž1956. 193–214. w2x E.P. Kennedy, L.F. Borkenhagen, S.W. Smith, Possible metabolic functions of deoxycytidine diphosphate choline and deoxycytidine diphosphate ethanolamine, J. Biol. Chem. 234 Ž1959. 1998–2000. w3x E.P. Kennedy, The biosynthesis of phospholipids, in: J.A.F. Op den Kamp, B. Roelofsen, K.W.A. Wirtz ŽEds.., Lipids and Membranes. Past, Present and Future, Elsevier Science Publishers, Amsterdam, 1986, pp. 171–206. w4x C. Kent, Eukaryotic phospholipid biosynthesis, Annu. Rev. Biochem. 64 Ž1995. 315–343. w5x P.S. Vermeulen, L.B.M. Tijburg, M.J.H. Geelen, L.M.G. Van Golde, Immunological characterization, lipid dependence, and subcellular localization of CTP:phosphoethanolamine cytidylyltransferase purified from rat liver.
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
w6x
w7x
w8x
w9x
w10x
w11x
w12x
w13x
w14x
w15x
w16x
w17x
w18x
w19x
Comparison with CTP:phoshocholine cytidylyltransferase, J. Biol. Chem. 268 Ž1993. 7458–7464. J.J. Van Hellemond, J.W. Slot, M.J.H. Geelen, L.M.G. van Golde, P.S. Vermeulen, Ultrastructural localization of CTP:phosphoethanolamine cytidylyltransferase in rat liver, J. Biol. Chem. 269 Ž1994. 15415–15418. J.E. Vance, D.E. Vance, Does rat liver Golgi have the capacity to synthesize phospholipids for lipoprotein secretion?, J. Biol. Chem. 263 Ž1988. 5898–5909. D.E. Vance, Glycerolipid biosynthesis in eukaryotes, in: D.E. Vance, J.E. Vance ŽEds.., Biochemistry of Lipids, Lipoproteins and Membranes, Elsevier Science Publishers, Amsterdam, 1996, pp. 153–182. J.D. Watkins, C. Kent, Immunolocalization of membraneassociated CTP:phosphocholine cytidylyltransferase in phosphatidylcholine-deficient Chinese Hamster Ovary cells, J. Biol. Chem. 267 Ž1992. 5686–5692. Y. Wang, T.D. Sweitzer, P.A. Weinhold, C. Kent, Nuclear localization of soluble CTP:phosphocholine cytidylyltransferase, J. Biol. Chem. 268 Ž1993. 5899–5904. M. Houweling, Z. Cui, C.D. Anfuso, M. Bussiere, ` M.H. Chen, D.E. Vance, CTP:phosphocholine cytidylyltransferase is both a nuclear and cytoplasmic protein in primary hepatocytes, Eur. J. Cell Biol. 69 Ž1996. 55–63. J.D. Esko, M.M. Wermuth, C.R.M. Raetz, Thermolabile CDP-choline synthetase in an animal cell mutant defective in lecithin formation, J. Biol. Chem. 256 Ž1981. 7388–7393. P.S. Vermeulen, M.J.H. Geelen, L.M.G. van Golde, Substrate specificity of CTP:phosphoethanolamine cytidylyltransferase purified from rat liver, Biochim. Biophys. Acta 1211 Ž1994. 343–349. P.S. Vermeulen, M.J.H. Geelen, L.B.M. Tijburg, L.M.G. van Golde, The CDP-ethanolamine pathway in mammalian cells, in: R.W. Gross ŽEd.., Advances in Lipobiology, vol. 2, JAI Press, Greenwich, CT, 1997, pp. 287–322. J.P. Infante, Rate-limiting steps in the cytidine pathway for the synthesis of phosphatidylcholine and phosphatidylethanolamine, Biochem. J. 167 Ž1977. 847–849. ˚ R. Sundler, B. Akesson, Regulation of phospholipid biosynthesis in isolated rat hepatocytes. Effect of different substrates, J. Biol. Chem. 250 Ž1975. 3359–3367. L.B.M. Tijburg, M. Houweling, M.J.H. Geelen, L.M.G. van Golde, Stimulation of phosphatidylethanolamine synthesis in isolated rat hepatocytes by phorbol 12-myristate 13-acetate, Biochim. Biophys. Acta 922 Ž1987. 184–190. L.B.M. Tijburg, M. Houweling, M.J.H. Geelen, L.M.G. van Golde, Inhibition of phosphatidylethanolamine synthesis by glucagon in isolated rat hepatocytes, Biochem. J. 257 Ž1989. 645–650. M. Houweling, L.B.M. Tijburg, W.J. Vaartjes, L.M.G. van Golde, Phosphatidylethanolamine metabolism in rat liver after partial hepatectomy. Control of biosynthesis of phosphatidylethanolamine by the availability of ethanolamine, Biochem. J. 283 Ž1992. 55–61.
99
w20x C.R. McMaster, P.C. Choy, Newly imported ethanolamine is preferentially utilized for phosphatidylethanolamine biosynthesis in the hamster heart, Biochim. Biophys. Acta 1124 Ž1992. 13–16. w 21 x R . Sundler, B iosynthesis of rat liver phos phatidylethanolamines from intraportally injected ethanolamine, Biochim. Biophys. Acta 306 Ž1973. 218–226. w22x B.N. Kholodenko, H.V. Westerhoff, Metabolic channeling and control of the flux, FEBS Lett. 320 Ž1993. 71–74. w23x Y.-J. Shiao, J.E. Vance, Evidence for an ethanolamine cycle: differential recycling of the ethanolamine moiety of phosphatidylethanolamine derived from phosphatidylserine and ethanolamine, Biochem. J. 310 Ž1995. 673–679. w24x T.P. George, S.C. Morash, H.W. Cook, D.M. Byers, F.B.St.C. Palmer, M.W. Spence, Phosphatidylcholine biosynthesis in cultured glioma cells: evidence for channeling of intermediates, Biochim. Biophys. Acta 1004 Ž1989. 283–291. w25x R. Sundler, Ethanolaminephosphate cytidylyltransferase. Purification and characterization of the enzyme from rat liver, J. Biol. Chem. 250 Ž1975. 8585–8590. w26x T.J. Porter, C. Kent, Purification and characterization of cholinerethanolamine kinase from rat liver, J. Biol. Chem. 265 Ž1990. 414–422. w27x R. Min-Seok, Y. Kawamata, H. Nakamura, A. Ohta, M. Takagi, Isolation and characterization of ECT1 gene encoding CTP:phosphoethanolamine cytidylyltransferase of Saccharomyces cereÕisiae, J. Biochem. 120 Ž1996. 1040–1047. w28x A. Nakashima, K. Hosaka, J. Nikawa, Cloning of a human cDNA for CTP-phosphoethanolamine cytidylyltransferase by complementation in vivo of a yeast mutant, J. Biol. Chem. 272 Ž1997. 9567–9572. w29x H. Jamil, D.E. Vance, Substrate specificity of CTP:phosphocholine cytidylyltransferase, Biochim. Biophys. Acta 1086 Ž1991. 335–339. w30x J.J. Mukherjee, T. Chung, D.K. Ways, Z. Kiss, Protein kinase C a is a major mediator of the stimulatory effect of phorbol ester on phospholipase D-mediated hydrolysis of phosphatidylethanolamine, J. Biol. Chem. 271 Ž1996. 28912–28917. w31x A.K. Menon, V.L. Stevens, Phosphatidylethanolamine is the donor of the ethanolamine residue linking a glycosylphosphatidylinositol anchor to protein, J. Biol. Chem. 267 Ž1992. 15277–15280. w32x D.R. Voelker, Phosphatidylserine decarboxylase, Biochim. Biophys. Acta 1348 Ž1997. Chapter XVIIB. w33x M. Houweling, L.B.M. Tijburg, W.J. Vaartjes, J.J. Batenburg, G.B. Kalmar, R.B. Cornell, L.M.G. van Golde, Evidence that CTP:choline-phosphate cytidylyltransferase is regulated at a pretranslational level in rat liver after partial hepatectomy, Eur. J. Biochem. 214 Ž1993. 927–933.
Chapter VII
CTP:phosphoethanolamine cytidylyltransferase Bellinda A. Bladergroen, Lambert M.G. van Golde
)
Department of Veterinary Basic Sciences, DiÕision of Biochemistry, Graduate School of Animal Health and Institute of Biomembranes, UniÕersity of Utrecht, P.O. Box 80176, 3508 TD Utrecht, The Netherlands Received 9 April 1997; accepted 20 May 1997
Abstract CTP:phosphoethanolamine cytidylyltransferase ŽET. catalyzes the conversion of phosphoethanolamine into CDP-ethanolamine. Immunogold electron microscopy studies have demonstrated that, in hepatocytes, ET is localized predominantly in areas of the cytoplasm that are rich in rough endoplasmic reticulum ŽRER.. Within these areas the enzyme shows a bimodal distribution between the cisternae of the RER and the cytosolic space. Studies on the substrate specificity of ET have shown that it can utilize both CTP and dCTP as substrates, but not other trinucleotides. In addition, the enzyme shows a very pronounced specificity for phosphoethanolamine. Under most conditions ET contributes significantly to the overall regulation of the CDP-ethanolamine pathway. Reversible binding of the enzyme to the endoplasmic reticulum could potentially play a key-role in metabolic channeling of phosphatidylethanolamine synthesis. ET has been purified from rat liver. Convincing evidence has been provided that ET and CTP:phosphocholine cytidylyltransferase ŽCT., the analogous enzyme in the CDP-choline pathway, are separate activities that reside on different proteins. The gene coding for yeast ET has been cloned. The deduced amino acid sequence contained a region in the N-terminal half with significant similarities to the conserved catalytic domain of both yeast and rat CT. The human cDNA for ET was also cloned recently. The predicted amino acid sequence of human ET shows a high degree of similarity Ž36% identity. to that of yeast ET, but the human protein is longer than the yeast protein, especially at the C-terminal region. Interestingly, both yeast and human ET have a large repetitive sequence in their N-terminal and C-terminal half. q 1997 Elsevier Science B.V. Keywords: CTP:phosphoethanolamine cytidylyltransferase; Phosphatidylethanolamine; CDP-ethanolamine; CTP:phosphocholine cytidylyltransferase; Phosphatidylcholine; Hepatocyte
1. Discovery, occurrence and cellular localization 1.1. DiscoÕery and occurrence As is the case for most enzymes involved in the biogenesis of phospholipids, the credit for the discovAbbreviations: ET, CTP:phosphoethanolamine cytidylyltransferase; RER, rough endoplasmic reticulum; CT, CTP:phosphocholine cytidylyltransferase; PE, phosphatidylethanolamine; PC, phosphatidylcholine ) Corresponding author. Fax: q31 30 2535492.
ery of CTP: phosphoethanolamine cytidylyltransferase ŽET. goes to Eugene Kennedy and his associates. It was about four decades ago that Kennedy and Weiss elucidated the CDP-ethanolamine pathway for the synthesis de novo of phosphatidylethanolamine ŽPE.. In their famous 1956 paper w1x they reported that particulate fractions of rat liver catalyzed the incorporation of phosphoethanolamine into PE and that this conversion specifically required the addition of CTP. They proved that CDP-ethanolamine was the immediate precursor of PE and that this novel intermediate was formed from phosphoethanolamine and
0005-2760r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 5 - 2 7 6 0 Ž 9 7 . 0 0 1 1 3 - 6
92
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
CTP in a reaction catalyzed by a cytidylyltransferase activity: phosphoethanolamineq CTP ° CDP y ethanolamineq PPi The hydrolysis of pyrophosphate, catalyzed by ubiquitous intracellular pyrophosphatases, drives this reaction, which would otherwise be freely reversible, in the direction of synthesis of CDP-ethanolamine w1–3x. ET has been found in all eukaryotic cells investigated so far, including lower eukaryotes such as yeast, where the CDP-ethanolamine pathway plays only an auxiliary role w4x. However, in most cell types the activity of ET is substantially less than that of CTP:phosphocholine cytidylyltransferase Ž CT. , the analogous enzyme in the biosynthesis of phosphatidylcholine ŽPC.. 1.2. Subcellular organization Until recently it was generally accepted that ET of mammalian cells was a soluble enzyme that, in contrast to CT, did not associate with cellular membranes. This conclusion was based on classical differential centrifugation studies as well as on enzyme-release measurements from digitonin-permeabilized
hepatocytes Žsee Ref. w5x and references cited therein.. However, it is not always certain that enzymes, classified as soluble from such studies, are indeed soluble in situ. The possibility cannot be excluded that in the cell they are weakly bound to membranes or to other structural elements, from which they could be released during the mechanical fractionation of the tissue or during cell permeabilization employing cholesterol-sequestering agents such as digitonin. Van Hellemond et al. w6x performed immunogold electron microscopy studies on the cellular localization of ET in rat liver using a polyclonal affinity-purified antibody against the enzyme. Their results showed that ET was not randomly distributed in hepatocytes. As illustrated in Fig. 1 the ET label was concentrated particularly in areas that contained cisternae of the rough endoplasmic reticulum ŽRER.. Other cellular organelles, including nuclei, mitochondria and plasma membranes, were only marginally labeled. Doublelabel experiments for ET and established markers for either soluble or integral endoplasmic reticulum proteins suggested a bimodal distribution of ET between the cisternae of the RER and the cytosolic space w6x. These findings suggested that ET might reversibly interact with membranes of the endoplasmic reticulum, which could bring the enzyme into close prox-
Fig. 1. Immunogold electron microscopy study on the localization of ET in rat hepatocytes. ET label is strongest in RER-rich areas of the cytoplasma Žsolid star.. Less labeling is observed in other cytoplasm regions Žopen star.. Almost no gold markers are present in the nucleus Žn., mitochondria Žm., and plasma membranes Žarrows.. Reprinted with permission ŽRef. w6x..
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
imity of ethanolaminephosphotransferase, an established integral microsomal protein w7x that catalyzes the final step in the CDP-ethanolamine pathway. The virtual absence of ET in rat liver nuclei is interesting in view of recent studies on the subcellular localization of CT. Until a few years ago it was unanimously accepted that CT was localized in both the cytosolic and endoplasmic reticulum compartment of all animal cells Žfor review see Ref. w8x.. However, intriguing immunofluorescence studies of Kent and her colleagues have shown that CT is predominantly present in the nucleus of CHO cells and in several other cell lines w4,9,10x. In addition, these investigators showed that activation of PC synthesis in HeLa cells treated with oleate resulted in translocation of inactive soluble nuclear CT to the nuclear membrane Žreviewed in Ref. w4x.. Houweling et al. w11x recently employed a combined approach of indirect immunofluorescence microscopy, immunogold electron microscopy, and biochemical techniques to localize CT in CHO cells and in hepatocytes. Although they confirmed the abundant nuclear localization of CT in CHO cells, they provided strong evidence that CT is not predominantly a nuclear enzyme in all cells, as had been suggested by Kent and colleagues. In rat hepatocytes CT appeared to be —like ET—primarily a cytoplasmic enzyme w11x, despite the fact that rat liver CT possesses the nuclear localization sequences w4x. The significance of the fact that at least in some cells a substantial portion of CT is located in the nuclei w4,8–11x is still unclear.
2. Substrate specificity and regulation 2.1. Substrate specificity The original studies of Kennedy and Weiss had already shown that CTP could not be replaced as a substrate for ET and CT by ATP, ITP, UTP or GTP w1x. Further investigations by Kennedy and his associates w2x showed that ET and CT did not only catalyze the formation of CDP-ethanolamine and CDP-choline, respectively, but also that of the corresponding deoxyribonucleotides dCDP-ethanolamine and dCDP-choline. Later investigations with a mutant cell line derived from CHO cells, in which the specific activity of CT was reduced to 1–10% of wild type,
93
provided convincing evidence that the formation of CDP-choline and dCDP-choline was catalyzed by a single enzyme as the pool sizes of both CDP-choline and dCDP-choline were both strongly depressed w12x. Likewise it is highly likely that one single enzyme ŽET. is responsible for the formation of both CDPethanolamine and its deoxy analog. The biological function of dCDP-choline and dCDP-ethanolamine remains an enigma w3x. The presence of relatively high levels of these deoxyribonucleotides in rapidly metabolizing cells, may simply reflect the high levels of dCTP in the cytoplasma of such cells. On the other hand, it cannot be excluded that dCDP-ethanolamine and dCDP-choline are selectively utilized, e.g. for the formation of distinct pools or distinct molecular species of PE and PC, respectively w3x. Studies on the specificity of purified ET for phosphorylated bases with a varying degree of N-methylation, which will be discussed in more detail in Section 3.2, showed that ET displays a high specificity for phosphoethanolamine w13x. 2.2. Role of ET in the regulation of the CDP-ethanolamine pathway Whereas the overall regulation of PC synthesis via the CDP-choline pathway has been subject of many intense studies w4,8x, there is much less information on the control of the CDP-ethanolamine pathway. The evidence available so far strongly suggests, however, that these two pathways are under independent metabolic and hormonal control Ž for review see Ref. w14x.. Several studies have addressed the question which of the individual steps in the CDP-ethanolamine pathway could provide an important contribution to the overall metabolic regulation of the process. On the basis of theoretical considerations, Infante w15x suggested that both the reaction catalyzed by ethanolamine kinase Žcholinerethanolamine kinase. and that catalyzed by ET could be important regulatory steps in the CDP-ethanolamine pathway in rat liver. Sundler ˚ w16x demonstrated that exposure of hepand Akesson atocytes to increasing concentrations of ethanolamine resulted in an enhanced entry of labeled glycerol into PE that was accompanied by a considerable increase of the pool size of phosphoethanolamine. As the pool size of CDP-ethanolamine remained constant, these
94
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
observations strongly suggested that the reaction catalyzed by ET should indeed be considered as a potentially important regulatory site in PE biosynthesis via the CDP-ethanolamine route. Further evidence for such a role of ET came from studies by Tijburg et al. w17x who showed that exposure of hepatocytes to phorbol ester led to an increased rate of incorporation of labeled ethanolamine into PE, increased activity of ET, and a reduction in the pool size of phosphoethanolamine. Interestingly, the phorbol ester treatment also caused an enhanced conversion of CDPethanolamine into PE. The last enzyme in the CDPethanolamine pathway, ethanolaminephosphotransferase, is thought to be at equilibrium in vivo and as such is in itself not an important regulatory enzyme. The effect was probably due to the increased amount of cellular diacylglycerol that was generated by the treatment of the cells with phorbol ester. Further evidence for a potentially important role of diacylglycerol supply to the regulation of PE synthesis came from a subsequent study on the effect of glucagon on PE synthesis in hepatocytes w18x. Exposure of hepatocytes to glucagon or cyclic AMP analogs decreased the rate of incorporation of labeled ethanolamine into PE, but did not affect the activity of ET or the pool size of phosphoethanolamine. The inhibitory effect of glucagon on PE synthesis could, however, be attributed to a diminished supply of diacylglycerol resulting in a decreased conversion of CDP-ethanolamine into PE. At least in liver there is very little experimental evidence suggesting important control of PE synthesis at the site of ethanolamine kinase. Experiments of ˚ w16x and Houweling et al. w19x Sundler and Akesson with hepatocytes showed that at low ethanolamine concentrations in the medium the rate of PE synthesis was controlled by the supply of ethanolamine, whereas at higher ethanolamine concentrations Ž ) 30–50 mM. phosphoethanolamine started to accumulate indicating that ET had become rate-limiting. It is, however, important to emphasize that the contribution of each step to overall metabolic regulation of the pathway may differ among tissues. For example, there is evidence from studies with isolated perfused hamster heart that ethanolamine kinase could contribute significantly to the overall control of the CDP-ethanolamine pathway, particularly at high ethanolamine concentrations in the medium w20x.
Summarizing, it can be stated that regulation of PE synthesis can take place at multiple sites in the CDP-ethanolamine pathway, although under most conditions the supply of CDP-ethanolamine and diacylglycerol appear to be principal factors governing the rate of this process. 2.3. Potential key-role of ET in metabolic substrate channeling It has been well established that the subcellular organization of CT is a major determinant in regulating the activity of this enzyme and that of the CDPcholine pathway as a whole w4,8x. The soluble form of CT appears to represent an inactive reservoir. From this reservoir the enzyme can rapidly and reversibly translocate to cellular membranes Žendoplasmic reticulum membranes w8x andror the nuclear envelope w4x. where it becomes activated by interactions with lipids. A comparable translocation mechanism for regulation of the activity of ET, the putative key-regulatory enzyme in the CDP-ethanolamine pathway, seemed less likely as it was until recently generally accepted that ET of mammalian cells was a fully soluble enzyme that did not bind significantly to membranes. In addition, the activity of ET is not stimulated by interactions with lipids w5x. However, as mentioned in a preceding section, the immunogold electron microscopy studies by van Hellemond et al. w6x demonstrated that ET was present predominantly in RER-rich areas of hepatocytes, and that within these areas the enzyme partitioned between ER membranes and the cytosol. Although difficult to prove experimentally, reversible binding of ET to the ER could be an essential step in metabolic channeling of PE synthesis as proposed schematically in Fig. 2. The early observation by Sundler w21x that phosphoethanolamine formed from exogenous ethanolamine did not freely equilibrate with the endogenous liver phosphoethanolamine pool already suggested that the product of the ethanolamine kinase reaction may be specifically channeled towards ET. As ethanolaminephosphotransferase, the next and final enzyme in the pathway, is an integral microsomal enzyme, reversible binding of ET to the membrane could indeed play a key-role in the topodynamic regulation w22x of the CDPethanolamine pathway as a whole Ž Fig. 2. . The con-
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
cept of compartmentalization of the aqueous precursors in the CDP-ethanolamine pathway finds also support in a recent study of Shiao and Vance w23x. These investigators presented evidence for the operation of an ethanolamine-PE cycle in CHO cells. Via this cycle the ethanolamine moiety of PE synthesized from labeled ethanolamine was continuously released from PE and recycled back to PE via the CDPethanolamine route. Importantly, the labeled ethanolamine released from PE was re-incorporated into PE without dilution if the medium did not contain unlabeled ethanolamine. If, however, the cells were incubated with unlabeled ethanolamine in the medium, the specific radioactivity of the intracellular ethanol-
95
amine pool and that of PE decreased as a result of dilution by the exogenous ethanolamine. Earlier studies by George et al. w24x had provided evidence for the compartmentalization of the enzymes of the CDP-choline pathway in cultured glioma cells. Only choline that was taken up by the high affinity uptake system was converted into PC, whereas phosphocholine and CDP-choline that were administered to the cells by electropermeabilization did not enter into PC. It would be of great interest if this compartmentalization could be demonstrated for other cell types and by other techniques of introducing phosphocholine, CDP-choline, and their ethanolamine analogs into the cells.
Fig. 2. Potential key-role of ET in compartmentalization of the CDP-ethanolamine pathway. The bimodal distribution of ET between RER membranes and the cytosolic compartment in RER-rich areas could provide the possibility of substrate channeling by topographic linking of ethanolamine kinase and ethanolaminephosphotransferase. Abbreviations: Eth, ethanolamine; P-Eth, phosphoethanolamine; PE, phosphatidylethanolamine; DAG, diacylglycerol; Eth-K, cholinerethanolamine kinase; ET, phosphoethanolamine cytidylyltransferase; EPT, ethanolaminephosphotransferase. The dashed line encircles the proposed metabolon of the CDP-ethanolamine pathway.
96
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
3. Purification Already in 1975 ET had been purified about 1000-fold from a post-microsomal fraction of rat liver w25x. More recently a new procedure was developed to purify the rat-liver enzyme to apparent homogeneity w5x. This procedure included the following steps: ammonium sulphate precipitation, basic anion exchange chromatography on DEAE cellulose, dye– ligand pseudo affinity on Matrex gel red A, hydrophobic interaction chromatography on octyl-Sepharose CL-4B, and dye–ligand pseudo affinity chromatography on Matrix gel blue A. The procedure resulted in a 1400-fold purification of ET with a final specific activity Ž6.5 mmolrmin per mg protein. that was a factor of five higher than that reported by Sundler w25x. Polyacrylamide gel electrophoresis under denaturing and reducing conditions showed the presence of a single protein with an estimated molecular mass of 49.6 kDa. The purified protein was identified as the ET enzyme based on the following observations: Ž i. it also showed one single protein band on gel under native polyacrylamide gel conditions Žthis protein carried ET enzyme activity and showed one single 49.6-kDa band on SDS gel. ; Ž ii. an affinity-purified polyclonal antibody raised against ET was strongly immunoreactive against the 49.6-kDa protein band on SDS gel; and Žiii. this antibody had also the ability to precipitate ET activity from rat liver cytosol. 3.1. Properties of the purified enzyme Superose 12 gel filtration revealed a molecular weight of the native enzyme of 100 000–120 000 which agreed with the earlier estimate by Sundler w25x. This may indicate that the enzyme occurs as a dimer. ET has one sharp pH optimum at pH 7.8 and a broader one, with lower maximal activity, around pH 6–6.5 w13,25x. Interestingly, two apparent K m values of 70 and 185 mM were found for phosphoethanolamine w13x. This could imply that ET contains a second binding site for phosphoethanolamine. Kinetic studies by Sundler had shown that the reaction catalyzed by ET probably proceeds via an ordered sequential reaction with CTP being the first substrate to bind to the enzyme and CDP-ethanolamine the last product to be released w25x.
The requirement of ET for a divalent cation is absolute: Mg 2q Ž5–10 mM. was most efficient, and the activity of ET was reduced to 50% if Mg 2q was replaced by optimal concentrations Ž 2–4 mM. of Mn2q. Other divalent cations were much less effective ŽCa2q . or completely inactive w25x. 3.2. Substrate specificity of purified ET The observations of Vermeulen et al. w5x provided convincing evidence that, unlike ethanolamine and choline kinase w26x, ET and CT of rat liver are separate activities residing on different proteins: Ž i. CT activity could not be detected in purified ET, whereas pure CT did not exhibit any ET activity; Ž ii. antibodies raised against ET did not cross-react with CT Žand vice versa.; Žiii. both cytosolic and purified ET were not markedly affected by the presence of a variety of exogenous lipids Žthis is in contrast to the well-known high dependence of the activities of cytosolic and purified CT on the presence of exogenous lipids w8x.; and Živ. the subcellular localization of ET and CT are distinct, as discussed in a previous section. These conclusions are obviously in line with recent studies showing that ET and CT are coded by different genes w27,28x Ž see Section 4. . As already predicted by the early experiments of Kennedy et al. with crude fractions from rat liver w2x, purified ET also utilized dCTP as a substrate w13x. Substitution of CTP by dCTP hardly affected the K m value for the nucleotide Ž 50 mM for CTP vs. 54 mM for dCTP.. However, the Vmax value decreased by a factor of 4.4 when CTP was substituted by dCTP. Vermeulen et al. w13x also studied the specificity of pure ET for phosphorylated bases with a varying degree of N-methylation. Increasing the number of N-methyl groups on phosphoethanolamine led to a strong increase in the apparent K m value and to an even more pronounced decrease of the Vmax of ET. Phosphomonomethylethanolamine, phosphodimethylethanolamine and phosphocholine appeared to be weak competitive inhibitors of the reaction catalyzed by ET when phosphoethanolamine was used as a substrate, with Ki values of 7.0, 6.8 and 52.9 mM, respectively. Vermeulen et al. calculated the specificity constants k catrK m of ET for these various phosphobases and compared the values with the corresponding data published for CT w29x. The data
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
demonstrated that increasing the number of N-methyl groups on phosphoethanolamine drastically decreased the specificity constants of ET by factors of about 10, 350 and 2.10 5, respectively. Decreasing the number of N-methyl groups on phosphocholine had a similar but smaller effect on the specificity constants of CT Žby factors of about 8, 40 and 570, respectively. . These data indicate that both ET and CT can tolerate a difference of one N-methyl group when compared to their regular substrates phosphoethanolamine and phosphocholine, respectively. While ET strongly favors phosphomonomethylethanolamine above phosphodimethylethanolamine, the reverse is true for CT. However, it is clear that both ET and CT show a very pronounced specificity for their regular substrates phosphoethanolamine and phosphocholine, respectively.
4. Cloning of the gene Min-Seok et al. w27x recently reported the isolation and partial characterization of the gene coding for ET of Saccharomyces cereÕisiae. These workers isolated several mutants of S. cereÕisiae that were unable to utilize extracellular ethanolamine for PE synthesis. Two of these mutants carried recessive chromosomal mutations in the same gene and they were both deficient in ET. By screening a genomic library, they were able to isolate three overlapping clones of different sizes that complemented the mutation. These shared a 2.8-kb DNA region comprising an open reading frame of 969 bp in length that appeared to code for ET. This was proven by expression studies in E. coli. It is now clear that mammalian CT contains at least four structural domains Žfor review see Refs. w4,8x.. Near the N-terminal is a nuclear localization signal, followed by a catalytic domain Ž residues 75– 235. in which CTs from yeast and rat are 65% identical. This is followed by the lipid-binding amphipathic helical domain Žresidues 239–298. and the phosphorylation domain of the enzyme Ž residues 315–367.. These two domains are involved in the regulation of CT. The predicted amino acid sequence of yeast ET contained one region Ž residues 7–160. with significant similarities to the conserved catalytic domain of both yeast and rat CT Ž55 amino acids
97
identical with rat CT and 52 with yeast CT. . As suggested by Min-Seok et al. w27x it is likely that this region of the yeast ET sequence also encompasses the catalytic domain of the enzyme. Another rather conserved region in yeast ET appeared to be a repeat of the first half of the conserved region in the Nterminal part of the enzyme. This repetition had not been observed for yeast and rat CT. The amino acid sequence of yeast ET lacks a hydrophobic domain that is long enough to span a bilayer and, in contrast to CT, no amphipathic helical domains could be predicted for yeast ET w27x. Short stretches of hydrophobic amino acids occur after residue 229 of yeast ET, but it is uncertain whether these could be involved in possible binding of ET to membranes. Interestingly, the expression studies of Min-Seok et al. w27x also provided preliminary evidence that yeast ET, like rat liver ET w6x, may associate with the endoplasmic reticulum. Very recently Nakashima et al. w28x reported the cloning of a human cDNA for ET by complementation in vivo of a yeast mutant of which the ECT1 gene was disrupted. The deduced protein encoded by the cDNA consists of 389 amino acids with a calculated molecular mass of approx. 43.8 kDa. The predicted amino acid sequence shows a high degree of similarity Ž36% identity. to that of yeast ET w27x, but the human protein is longer than the yeast protein in both the N- and C-terminal regions. Interestingly, as in yeast ET, there is a large repetitive sequence in the N- and C-terminal halves of human ET. As suggested by Nakashima et al., this could imply the presence of two catalytic domains in human and yeast ET. Whether this is related to the earlier mentioned suggestion from kinetic studies with ET purified from rat liver that the enzyme might contain two binding sites for phosphoethanolamine w13x, remains to be investigated.
5. Future developments The progress in our understanding of ET since it was first reported in 1956 has been much less impressive than that of the simultaneously discovered CT. The major reason is obviously the fact that the CDPcholine pathway has traditionally received much more attention from researchers than the analogous CDP-
98
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
ethanolamine route. However, evidence is accruing for specific metabolic functions of PE in mammalian cells, e.g. as a substrate for the production of second messengers in signal transduction w30x and its role as donor of the ethanolamine residue to the phosphoethanolamine bridge that links the glycosylphosphatidylinositol anchor to surface glycoproteins of eukaryotes w31x. This will undoubtedly intensify interest in the study of overall PE metabolism and its regulation in eukaryotic cells. As there is evidence from a number of studies that the CDP-ethanolamine pathway can play an important role in the biosynthesis of PE and its plasmalogen analog in several mammalian tissues Žfor review see Ref. w14x. , it is to be expected that ET and its regulation will attract increasing interest. However, in this context it is important to emphasize that the CDP-ethanolamine pathway is not the only route leading to PE synthesis. As discussed in another contribution to this special issue w32x, there is convincing evidence that decarboxylation of phosphatidylserine is another major contributor to PE biogenesis in mammalian cells. The specific roles of each of these two biosynthetic pathways and their relative contribution to overall PE synthesis are not yet precisely known but might be different among different mammalian tissues. The availability of the cDNA w28x, complemented with that of purified ET and specific antibodies against the enzyme, will undoubtedly provide impetus to further studies on the properties of ET and the mechanisms by which the enzyme is regulated in mammalian cells. It is obvious that studies will be performed to pinpoint various domains in the structure of the protein, with identification of the catalytic domain, the domain via which the enzyme may bind to the endoplasmic reticulum, and domains that could otherwise play a role in regulation of ET activity. It would be of interest to investigate whether ET, like CT, can be reversibly phosphorylated in the mammalian cell and whether such phosphorylationrdephosphorylation mechanism could contribute to regulating the activity of the enzyme. Another interesting question is whether regulation of ET can also occur at a pretranslational level, as it has been shown that the expression of CT mRNA can be modulated under certain conditions, e.g. in the regenerating liver w33x. It would also be desirable to perform ET knockout experiments in mice, as such an approach might
provide useful information on the role and importance of ET and the CDP-ethanolamine pathway in cellular phospholipid metabolism and function.
Acknowledgements It is an enormous pleasure to dedicate this review to Professor Eugene Kennedy. After opening up the field of research on phospholipid metabolism in the early fifties, Gene Kennedy has remained a global leader in this field for a period of more than four decades. For many investigators he has been a highly appreciated and stimulating teacher and an experienced guide in this intriguing but complex area of research. One of the authors of this review Ž Bert van Golde. had the excellent opportunity of spending one year as a research associate in Gene Kennedy’s laboratory in Boston, where he performed, under Gene’s stimulating guidance, the initial studies on the membrane-derived oligosaccharides in Escherichia coli. He would like to take this opportunity of formally thanking Gene again for his warm hospitality and for providing a friendly and scientifically highly stimulating atmosphere in the laboratory. In such an atmosphere, in which each of Gene’s co-workers had almost unlimited access to him to discuss ups and downs in their project, it was almost impossible not to perform good research.
References w1x E.P. Kennedy, S.B. Weiss, The function of cytidine coenzymes in the biosynthesis of phospholipids, J. Biol. Chem. 222 Ž1956. 193–214. w2x E.P. Kennedy, L.F. Borkenhagen, S.W. Smith, Possible metabolic functions of deoxycytidine diphosphate choline and deoxycytidine diphosphate ethanolamine, J. Biol. Chem. 234 Ž1959. 1998–2000. w3x E.P. Kennedy, The biosynthesis of phospholipids, in: J.A.F. Op den Kamp, B. Roelofsen, K.W.A. Wirtz ŽEds.., Lipids and Membranes. Past, Present and Future, Elsevier Science Publishers, Amsterdam, 1986, pp. 171–206. w4x C. Kent, Eukaryotic phospholipid biosynthesis, Annu. Rev. Biochem. 64 Ž1995. 315–343. w5x P.S. Vermeulen, L.B.M. Tijburg, M.J.H. Geelen, L.M.G. Van Golde, Immunological characterization, lipid dependence, and subcellular localization of CTP:phosphoethanolamine cytidylyltransferase purified from rat liver.
B.A. Bladergroen, L.M.G. Õan Golder Biochimica et Biophysica Acta 1348 (1997) 91–99
w6x
w7x
w8x
w9x
w10x
w11x
w12x
w13x
w14x
w15x
w16x
w17x
w18x
w19x
Comparison with CTP:phoshocholine cytidylyltransferase, J. Biol. Chem. 268 Ž1993. 7458–7464. J.J. Van Hellemond, J.W. Slot, M.J.H. Geelen, L.M.G. van Golde, P.S. Vermeulen, Ultrastructural localization of CTP:phosphoethanolamine cytidylyltransferase in rat liver, J. Biol. Chem. 269 Ž1994. 15415–15418. J.E. Vance, D.E. Vance, Does rat liver Golgi have the capacity to synthesize phospholipids for lipoprotein secretion?, J. Biol. Chem. 263 Ž1988. 5898–5909. D.E. Vance, Glycerolipid biosynthesis in eukaryotes, in: D.E. Vance, J.E. Vance ŽEds.., Biochemistry of Lipids, Lipoproteins and Membranes, Elsevier Science Publishers, Amsterdam, 1996, pp. 153–182. J.D. Watkins, C. Kent, Immunolocalization of membraneassociated CTP:phosphocholine cytidylyltransferase in phosphatidylcholine-deficient Chinese Hamster Ovary cells, J. Biol. Chem. 267 Ž1992. 5686–5692. Y. Wang, T.D. Sweitzer, P.A. Weinhold, C. Kent, Nuclear localization of soluble CTP:phosphocholine cytidylyltransferase, J. Biol. Chem. 268 Ž1993. 5899–5904. M. Houweling, Z. Cui, C.D. Anfuso, M. Bussiere, ` M.H. Chen, D.E. Vance, CTP:phosphocholine cytidylyltransferase is both a nuclear and cytoplasmic protein in primary hepatocytes, Eur. J. Cell Biol. 69 Ž1996. 55–63. J.D. Esko, M.M. Wermuth, C.R.M. Raetz, Thermolabile CDP-choline synthetase in an animal cell mutant defective in lecithin formation, J. Biol. Chem. 256 Ž1981. 7388–7393. P.S. Vermeulen, M.J.H. Geelen, L.M.G. van Golde, Substrate specificity of CTP:phosphoethanolamine cytidylyltransferase purified from rat liver, Biochim. Biophys. Acta 1211 Ž1994. 343–349. P.S. Vermeulen, M.J.H. Geelen, L.B.M. Tijburg, L.M.G. van Golde, The CDP-ethanolamine pathway in mammalian cells, in: R.W. Gross ŽEd.., Advances in Lipobiology, vol. 2, JAI Press, Greenwich, CT, 1997, pp. 287–322. J.P. Infante, Rate-limiting steps in the cytidine pathway for the synthesis of phosphatidylcholine and phosphatidylethanolamine, Biochem. J. 167 Ž1977. 847–849. ˚ R. Sundler, B. Akesson, Regulation of phospholipid biosynthesis in isolated rat hepatocytes. Effect of different substrates, J. Biol. Chem. 250 Ž1975. 3359–3367. L.B.M. Tijburg, M. Houweling, M.J.H. Geelen, L.M.G. van Golde, Stimulation of phosphatidylethanolamine synthesis in isolated rat hepatocytes by phorbol 12-myristate 13-acetate, Biochim. Biophys. Acta 922 Ž1987. 184–190. L.B.M. Tijburg, M. Houweling, M.J.H. Geelen, L.M.G. van Golde, Inhibition of phosphatidylethanolamine synthesis by glucagon in isolated rat hepatocytes, Biochem. J. 257 Ž1989. 645–650. M. Houweling, L.B.M. Tijburg, W.J. Vaartjes, L.M.G. van Golde, Phosphatidylethanolamine metabolism in rat liver after partial hepatectomy. Control of biosynthesis of phosphatidylethanolamine by the availability of ethanolamine, Biochem. J. 283 Ž1992. 55–61.
99
w20x C.R. McMaster, P.C. Choy, Newly imported ethanolamine is preferentially utilized for phosphatidylethanolamine biosynthesis in the hamster heart, Biochim. Biophys. Acta 1124 Ž1992. 13–16. w 21 x R . Sundler, B iosynthesis of rat liver phos phatidylethanolamines from intraportally injected ethanolamine, Biochim. Biophys. Acta 306 Ž1973. 218–226. w22x B.N. Kholodenko, H.V. Westerhoff, Metabolic channeling and control of the flux, FEBS Lett. 320 Ž1993. 71–74. w23x Y.-J. Shiao, J.E. Vance, Evidence for an ethanolamine cycle: differential recycling of the ethanolamine moiety of phosphatidylethanolamine derived from phosphatidylserine and ethanolamine, Biochem. J. 310 Ž1995. 673–679. w24x T.P. George, S.C. Morash, H.W. Cook, D.M. Byers, F.B.St.C. Palmer, M.W. Spence, Phosphatidylcholine biosynthesis in cultured glioma cells: evidence for channeling of intermediates, Biochim. Biophys. Acta 1004 Ž1989. 283–291. w25x R. Sundler, Ethanolaminephosphate cytidylyltransferase. Purification and characterization of the enzyme from rat liver, J. Biol. Chem. 250 Ž1975. 8585–8590. w26x T.J. Porter, C. Kent, Purification and characterization of cholinerethanolamine kinase from rat liver, J. Biol. Chem. 265 Ž1990. 414–422. w27x R. Min-Seok, Y. Kawamata, H. Nakamura, A. Ohta, M. Takagi, Isolation and characterization of ECT1 gene encoding CTP:phosphoethanolamine cytidylyltransferase of Saccharomyces cereÕisiae, J. Biochem. 120 Ž1996. 1040–1047. w28x A. Nakashima, K. Hosaka, J. Nikawa, Cloning of a human cDNA for CTP-phosphoethanolamine cytidylyltransferase by complementation in vivo of a yeast mutant, J. Biol. Chem. 272 Ž1997. 9567–9572. w29x H. Jamil, D.E. Vance, Substrate specificity of CTP:phosphocholine cytidylyltransferase, Biochim. Biophys. Acta 1086 Ž1991. 335–339. w30x J.J. Mukherjee, T. Chung, D.K. Ways, Z. Kiss, Protein kinase C a is a major mediator of the stimulatory effect of phorbol ester on phospholipase D-mediated hydrolysis of phosphatidylethanolamine, J. Biol. Chem. 271 Ž1996. 28912–28917. w31x A.K. Menon, V.L. Stevens, Phosphatidylethanolamine is the donor of the ethanolamine residue linking a glycosylphosphatidylinositol anchor to protein, J. Biol. Chem. 267 Ž1992. 15277–15280. w32x D.R. Voelker, Phosphatidylserine decarboxylase, Biochim. Biophys. Acta 1348 Ž1997. Chapter XVIIB. w33x M. Houweling, L.B.M. Tijburg, W.J. Vaartjes, J.J. Batenburg, G.B. Kalmar, R.B. Cornell, L.M.G. van Golde, Evidence that CTP:choline-phosphate cytidylyltransferase is regulated at a pretranslational level in rat liver after partial hepatectomy, Eur. J. Biochem. 214 Ž1993. 927–933.