Myristoyl-CoA:Protein N-Myristoyltransferase, an Essential Enzyme and Potential Drug Target in Kinetoplastid Parasites

Eur. J. Biochem. 214, 853-867 (1993) 0 FEBS 1993

Myristoyl-CoA :protein N-myristoyltransferase activity in cancer cells Purification and characterization of a cytosolic isoform from the murine leukemia cell line L1210 Jean A. BOUTIN', Gilles FERRY ', Anne-Pascale ERNOULD', Pierrette MAES*, Georges REMOND3 and Michel VINCENT3 ' Dipartement de Cancirologie Expirimentale, Institut de Recherches Servier, Suresnes, France Chimie des Biomolicules URA CNRS 1309, Institut Pasteur, Lille, France Division de Chimie D, Institut de Recherches Servier, Suresnes, France (Received February 22March 22,1993) - EJB 930273/3

Myristoylation is a co-translational maturation process of proteins. It is extremely specific for the cosubstrate (myristoyl-CoA) and for the substrate protein that should bear a glycine at the Nterminus of the protein to be myristoylated. This acylation is catalyzed by the myristoyl-CoA:protein N-myristoyltransferase. Most of the molecular biochemistry and biology concerning this enzyme has been done on Saccharomyces cerevisiae. Because of the major importance of this pathway in several types of pathology, it is essential to study intensively the enzyme(s) isolated from mammalian tissue(s) to confirm that the enormous amount of work done on the yeast enzyme can be transposed to mammalian tissues. In earlier studies, we demonstrated the existence of a microsomal N-myristoyltransferase from the murine leukemia cell line L1210 [Boutin, J. A., Clarenc, J.-P., Ferry, G., Ernould, A. P., Remond, G., Vincent, M. & Atassi, G. (1991) E m J. Biochem. 201,2572631, a feature which is not shared by yeast, and examined the N-myristoyltransferase activities associated with L1210 cytosol. In the present work, we purified to homogeneity one of the isoforms (A) of the transferase from L1210 cytosol. The purified enzyme showed on SDSPAGE an apparent molecular mass of 67.5 kDa, distinct from the 53-kDa yeast cytosolic enzyme. The purified enzyme from L1210 cytosol could be labeled with ['4C]myristoyl-CoA.Rabbit antibodies were raised against the A isoform and used to immunoprecipitate the enzyme and immunoinhibit the activity from the same source. A survey of the specificity of the partially and completely purified isoforms was performed using peptides derived from the NH,-terminus of 42 proteins which are potential substrates for myristoylation, including oncogene products and virus structural proteins. We synthesized a series of compounds capable of inhibiting the cytosol activities of the enzyme. For example, a myristoyltetrahydroquinoleinderivative showed an IC,, of about 0.1 pM.Based on both biophysical and biochemical evidence, the N-myristoyltransferases extracted from mammalian cell cytosols seem to be different from the extensively studied yeast enzyme.

the retroviral structural protein gag or the equivalent in other types of viruses [6-91. In both cases, using molecular biology tools, it has been demonstrated that this maturation step is absolutely dependent on the presence of the glycine residue at the N-terminus. Furthermore, when the myristoylation signal (usually from p6Ov-src or from gag) is added to known oncogene products, the fusion proteins present enhanced transforming properties (see among others BrooksWilson et al. forfps [lo], Bruskin et al. for v-erbB [ l l ] and Buss et al. for ras [12]). It has also been demonstrated that the important protein kinase C substrate 'MARCKS' was myristoylated [13]. Some GTP-binding proteins are also myristoylated [14] and, for Go-a, this acylation increases its Correspondence fo J. A. Boutin, Dipartement de Chimie G, In- affinity for the G-PIG-), complex [15], suggesting again an stitut de Recherches Servier, 11, rue des Moulineaux, F-92150 important functional role for myristoylation since alteration Suresnes, France of this acylation of Go-a affects the cellular distribution of Far: +33 147287766. this protein [16]. More recent evidence shows that the myrisAbbreviations. NMT, myristoyl-CoA: protein N-myristoyltranstoylation of the fusion protein (gag-fos) from the murine sarmedian inhibitory ferase; NaCI/Pi, phosphate-buffered saline ; coma virus plays a key role, together with other factors, in concentration. Enzymes. Myristoyl-CoA:protein N-myristoyltransferase (EC its lack of transrepression effect of the serum-response element [17, 181. Incidently, this observation suggests a patho2.3.1.97); acyl-CoA synthetase (EC 6.2.1.3).

Myristoyl-CoA:protein N-myristoyltransferase (NMT) is the enzyme responsible for a co-translational modification involving the rare fatty acid, myristic acid, and having as a binding site a glycine residue at the N-terminal of target proteins [l]. Numerous oncogene products from mammalian transforming retroviruses are either myristoylated or expressed as gag-onc fusion proteins which are myristoylated on the N-terminal glycine of the gag protein [2]. This maturation process has been proved to be essential in several types of pathology either by addressing the implicated oncogene product (like src) to a specific site at the plasma membrane level [3-51 or by permitting the regular assembly of

854 logical role for a virus non-structwal myristoylated protein such as p27neffrom HIV [19, 201. Finally, the level of unidentified proteins that are myristoylated varies with differentiation of HL60 cells due to 12myristoyl-phorbol 13-acetate [21] and with transformation [22], while the myristoylated protein level is also sensitive to pretreatment with interferon and tumor necrosis factor-a [23]. This suggests a possible role of myristoylation in the cellular response to these factors. Furthermore, we found that microsomal NMT activities are inducible by 3-methylcholanthrene in rat liver (Ferry, Burbridge, Boutin, unpublished observations) suggesting a regulation of NMT activity by carcinogens. In other words, myristoylation does not only play a role in addressing some oncogene products to a specific receptor site at the plasma membrane level, accounting for the transforming capacity of these oncogene products; it also plays a role in the maturation process and the infectivity capacity of some types of virus. Because point mutations at the N-terminal glycine level (or in its vicinity) impair the transformational potency of an oncogene product [24] or the budding and spreading of infectious virus particles [7, 251, the enzyme responsible for myristoylation should be studied in relevant tissues (that is, those susceptible to becoming transformed, or to being infected by viruses). An enormous amount of work has been published on the purification, characterization and cloning of NMT from a strain of Saccharomyces cerevisiae (Towler et al. [26], Gordon et al. [27] and references therein) while only a few data have been published on NMT from mammalian tissues: rat brain [28], rat tissues [29], bovine and rat brain [30, 311, rat liver [32], and BC3H1, a murine cell line [33]. The recent work by Duronio et al. [34] sheds some light on the relationship between the yeast and the human enzymes. Choosing the indirect rescuing method [35], these authors cloned the human NMT gene from the hepatocarcinoma cell line HepG2 and found it to be highly conserved when compared to the yeast enzyme (44% similarity). Since we found that the NMT activity associated with the microsomal fraction of L1210, a murine leukemia cell line, showed dramatic differences in specificity compared to the yeast enzyme [36], the purification and characterization of the L1210 cytosolic enzyme was completed. These studies on L1210 myristoylation capacity laid the background of our molecular tools for the synthesis and the screening of specific inhibitors of these NMTs (Boutin et al., unpublished). The present paper demonstrates that at least one of the purified cytosolic NMT isoenzymes is a 67.5-kDa protein, substantially different from that already described from the cytosol of eucaryotic microorganisms.

MATERIALS AND METHODS

oyl-CoA (0.01 M sodium acetate pH 6.0 solution 2.0 GBq/ mmol; CFA 726) were from Amersham.

Inhibitors 12-Methoxydodecanoic acid (13-oxomyristic acid) has been obtained as described by Heuckeroth et al. [37]. NMyristoyl and N-(13-oxomyristoyl) amino acids were prepared by a general synthesis method for N-acylated amino acids previously described by Lapidot et al. [38] and in a French patent (FR 9101502; purity >99%). Since these compounds were myristoylated amino acids, we named this series of compounds 'myraas' . Biological sources L1210, a murine leukemia cell line, was obtained from the American Tissue Culture Collection. Cells were grown in 750-ml rollers in RPMI 1640 supplemented with penicillin (50 Ulml), streptomycin (50 pg/ml), glutamine (2 mM), Hepes (1OmM) and 10% fetal calf serum (all from Gibco, France). The rollers were gased with 5% C02/95% air at 37 "C. Batches of about 10" cells were obtained weekly. Cells were collected by low-speed centrifugation (600 X g) in a Jouan GR-411 centrifuge, washed three times with phosphate-buffered saline (NaCW,), counted and checked for viability by the trypan blue exclusion method. The NaC1/Pi solution we used had the following composition: CaCI?, 0.1 gA; KCl, 0.2 g/l; KHzPo4, 0.2 g/l; MgC12, 0.047 g / l ; NaCl, 8 g / l ; NaZHPO4,1.15 g/l. The final cell pellet was suspended in 3 vol. buffer A (50 mM HepesNaOH pH 7.4, 2 mM EGTA, 1 mM dithiothreitol, 1 mh4 phenylmethylsulfonyl fluoride, 5 pglml aprotinin, 10 pg/ml trypsin inhibitor, 10 pg/ml leupeptin) and submitted to a series of 10-s sonication bursts until all cells were lysed, as checked under the microscope. The homogenate was centrifuged (15 000 X g for 10 min) and the supernatant further centrifuged at 105000 X g for 60 min. The final supernatant was used as the starting material for the purifications ; it is designated hereafter as 'cytosol' . HL-60, a human promyelocytic cell line, and CCRFCEM cells, a human acute lymphoblastic leukemia cell line (ATCC CCL 119), were grown using the same system of culture. Lysed yeasts were obtained from Springler Fould (Maison-Alfort, France) and were submitted to clarification and to a DEAE-Sepharose chromatographic step as described by Towler et al. [33]. Frozen rat brains and mouse livers were obtained from Euromedex (Strasbourg, France). They were carefully thawed at 4°C in 10 vol. buffer A, minced with scissors, and homogenized with a glasdglass dounce homogenizer. The homogenate was submitted to the same differential centrifugation preparation as described above.

Peptides and chemicals

NMTassay

Peptide GNAAAAK was synthesized by Novabiochem (Switzerland), GNAAAARR, GLAAAARR, GQTVTTPL and GSSKSKPKDP by NeoSystem Laboratories (Strasbourg, France). All the other peptides were synthesized by the solidphase method on a Milligen 9050 peptide synthesizer using the fluoren-9-ylmethoxycarbonyl (Fmoc) strategy and purified by reverse-phase HPLC. Only peptides at least 95% pure, as checked by HPLC, were used. [3H]Myristic acid (ethanol sulution; 2.0 TBq/mmol, TRK 907) and [14C]myrist-

The assay used in this work is the assay described by Towler and Glaser [39] with minor modifications. Briefly, for about 30 assays, 80 pl [3H]myristicacid (0.9 pCi/pl, Amersham) in ethanol was dissolved in 800 p1 buffer B (20 mM Hepes pH7.4, 1OmM MgClz, 0.2mM EGTA, 2 mM dithiothreitol) to which were added 50 pl 20 mM lithium salt of CoA and 100 pl 50 mM ATP (both from Sigma, St Louis MO). The reaction was started by adding 200 p1 acyl-CoA synthetase (Sigma, St Louis, MO) at 1 mU/p1 in 50mM

855 #

9

0

E

c

0

co

N

Fraction number Fig. 1. Second affinity chromatography step on CoA-agarose of cytosolic NMT isoform A. The affinity column (step 7, Table 3) was developed and submitted to a 0-0.5 M KCl gradient (A), as described under Materials and Methods. NMT activities ( 0 )were measured with GSSKSKPKDP peptide as substrate and are expressed in arbitrary units. Protein (0) was measured as absorbance at 280 nm. Aliquots of some of the fractions eluted from the column were submitted to SDSPAGE on a 10-20% acrylamide gel (right). Examples are given for fraction 58 (lane a), the main contaminant (aldolase A, left arrow), and fraction 70 (lance c, right arrow) N-myristoyltransferase with a molecular mass of 67.5 kDa; lane B, molecular mass standards (same as in Fig. 2).

Hepes pH 7.3. After 30 min at 37"C, aliquots of 45 p1 were added to conical plastic tubes with 50 pl buffer C (10 mh4 Hepes pH 7.4, 5 mM MgCI2, 0.1 mM EGTA, 1 mM dithiothreitol), 20 p1 enzyme source and 20 pl peptide substrate (1 mg/ml in water). The reaction ran for 30 min at 37OC and was stopped by adding 200 pl methanol/trichloroacetic acid (100/10, by vol.). Tubes were kept in ice for 30min and centrifuged at 13000Xg for 5 min; 7O-pl aliquots of the supernatants were injected into a Waters HPLC station equipped with a column (0.46X 10 cm) of Nucleosyl C,, (7 pm pore size). The radioactivity was monitored on-line with a Berthold HPLC monitor LB 507 A in the presence of HPLC scintillator fluid (4 d m i n , Zinsser 303). The analysis took 30 rnin and was run at 1.5 mumin using a gradient from 0.05% triethylamine and 0.1% trifluoroacetic acid in water/ 0.1 % trifluoroacetic acid in acetonitrile (60/40; mobile phase A) to 100%0.1 % trifluoroacetic acid in acetonitrile (mobile phase B). The exact timing of the gradient was as follows: 0-5 min 100% A, 5-20 rnin 0-100% B, 20-22 rnin 100% B, 22-25 min 0-100% A and 25-30 min 100% A. In inhibition experiments, the same conditions as above were used. Inhibitors were solubilized in methanol as stock solutions (10 mM) and sequentially diluted in methanol (up to 100 pM) and then in water. Controls were systematically run with the solvent alone. Methanol, at these final concentrations, has no effect on the cytosolic NMT activity. For metal requirement experiments, metal salts (i.e. MgC12,MnCI,, CoCl,, CaCl,, CdCI,) were added as aqueous solutions (15 pl) to the following reaction medium: 90 p1 buffer C, 5 pl ['4C]myristoyl-CoA, 20 pl pure enzyme and 20 pl peptide substrate (1 mg/ml in water). The reaction proceeded then as described above. Precipitation Cytosol was obtained by pooling four batches of grown cells and led to 2 1 clear material in which the NMT activity

could be easily measured. The ammonium sulfate fractionation of this material was at 0-20%, 20-40%, 40-60% and 60 -80% saturation. The activities towards two peptides were systematically measured: GSSKSKPKDP and GQTVTTPL. The myristoylation activity of the former peptide was precipitated, together with a significant amount of myristoylation activity on the latter peptide at 60% ammonium sulfate. A residual activity for the latter peptide was precipitated by 80% ammonium sulfate, and was apparently free of GSSKSKPKDP activity. We purified the activities from both these pellets. The yield of precipitation for GSSKSKPKDP myristoylating activity was about 50%. Purification of isoform A The 60% saturation ammonium sulfate pellet was suspended in 370 ml buffer A (40 mM Hepes/NaOH pH 7.4, 2 mM EGTA, 1 mh4 dithiothreitol) supplemented with a cocktail of protease inhibitors and applied to a column (1.5 X 30 cm) filled with Phenyl-Sepharose previously equilibrated with the same buffer containing 1 M ammonium sulfate. The column was developed with a linear gradient of 1-0 M ammonium sulfate (2 X 100 ml) with a simultaneous gradient of 0-15% glycerol. The activity eluted at low ( 4 0 0 mM) salt concentration. The fractions were pooled and dialyzed twice against 4.5 1 buffer A containing 1 mM phenylmethylsulfonyl fluoride and 15% (by vol.) glycerol (buffer B). This dialyzed material was then applied at 3 d m i n on a column (2.5 X 20 cm) of DEAE-Sepharose CL6B equilibrated in the same buffer. The column was developed by a linear gradient of 01 M NaCl in the equilibration buffer. More than 80% of the activity was excluded from the column under these conditions, while more that 70% of the protein was retained in the column. The excluded material (335 ml) was pooled and applied at the same flow rate on a CM-Sepharose CL6B (Pharmacia) column of the same size (2.5 X 20 cm) equili-

856 Table 1. Comparative specificities between N-myristoyltransferases from various cytosolic or microsomal origins. All biological sources were prepared identically by differential centrifugation as described in Materials and methods. Microsomal fraction activities were measured without addition of detergent [36]. Activity from yeast was partially purified as described in [36]. All activities are expressed as rate of myristoylation of peptide/mass protein. Peptide sequences

Origin

NMT specific activity yeast

GNEASYPL GVRMESHT GAQLSTLG GNHLQISV GNA?iAARR

calcineurin immunoglobulin precursor NADH: cytochrome b, reductase lipase precursor protein kinase A

CCRF-CEM

Cytosol

Microsomes

rat brain

L1210

rat liver

rat brain

L1210

pmol . min-'

mg protein-'

105.8 39.0

19.6